diff --git a/evalstudent/utils.py b/evalstudent/utils.py
index ecdb44d..9188713 100644
--- a/evalstudent/utils.py
+++ b/evalstudent/utils.py
@@ -1,3 +1,4 @@
+import os
 import pandas as pd
 pd.options.mode.chained_assignment = None # Disabling pandas warnings triggered by display_classes
 from IPython.core.display import HTML
@@ -20,15 +21,15 @@ def display_classes(essay_id, train_df):
     Prints the essay text (keeping its exact original formatting) using colors to highlight discourse elements and their classes.
     Uses only `predictionstring`, which is useful to display models predictions.
     '''
-    
+
     # Handling submission format :
     discourse_type = "class" if "discourse_type" not in train_df.columns else "discourse_type"
-    
+
     elements_df = train_df[train_df["id"] == essay_id]
     essay_text = open(f'../../raw_data/train/{essay_id}.txt').read()
     essay_words = essay_text.split()
     formatted_essay = ""
-    
+
     # First we make sure discourse elements are in the text order
     elements_df["prediction_list"] = elements_df["predictionstring"].map(lambda x : x.split())
     elements_df["start_word_index"] = elements_df["prediction_list"].map(lambda x : int(x[0]))
@@ -38,7 +39,7 @@ def display_classes(essay_id, train_df):
     # and then we highlight the exact part of the essay corresponding to the discourse class.
     end_char = 0
     for i, element in elements_df.iterrows():
-        start_word = essay_words[element["start_word_index"]] 
+        start_word = essay_words[element["start_word_index"]]
         start_char = essay_text[end_char:].find(start_word) + len(essay_text[:end_char])
         formatted_essay += essay_text[end_char:start_char]
         for word_index in element["prediction_list"]:
@@ -68,4 +69,116 @@ def generate_predictionstring(discourse_start, discourse_end, essay_text):
     word_end = word_start + len(essay_text[char_start:char_end].split())
     word_end = min( word_end, len(essay_text.split()) )
     predictionstring = " ".join( [str(x) for x in range(word_start,word_end)] )
-    return predictionstring
\ No newline at end of file
+    return predictionstring
+
+## ADD ARTHUR ##
+
+def get_essay(id,mode='train'):
+    """Function to get the full text of an essay from the .txt file.
+
+    Args:
+        id_ (str): id of the essay
+        mode (str, optional): determines whether to access *train* or *test* texts. \
+            Defaults to 'train'.
+
+    Returns:
+        str: Returns the full text of the id
+    """
+    with open(os.path.join(os.path.dirname(os.path.dirname(__file__)),
+                           'raw_data',
+                           mode,
+                           f'{id}.txt'),'r') as file:
+        txt = file.read()
+    return txt
+
+def slicering(ps,txt):
+    """
+    Allow for predictionstring to match with corresponding words of an essay.
+    Given a predictionstring of a portion of a text and the full text, the
+    function returns the portion of the text corresponding to the predictionstring.
+
+    Args:
+        ps (str): predictionstring of a discourse
+        txt (str): full text of an essay
+
+    Returns:
+        str: portion of the text corresponding to the predictionstring
+    """
+    ps_l = ps.split()
+    txt = txt.split()
+
+    return ' '.join(txt[int(ps_l[0]):int(ps_l[-1])+1])
+
+
+def css():
+    """
+    Apply custom.css into the notebook
+
+    Returns:
+        str: HTML style tag
+    """
+    styles = open("./styles/custom.css", "r").read()
+    return HTML('<style>'+styles+'</style>')
+
+
+
+def render_html(df):
+    """
+    Transforms each discourse into a html string with appropriates tags for
+    visualization.
+
+    Args:
+        df (DataFrame): dataframe containing discourse_type and discourse_text
+
+    Returns:
+        str: html string
+    """
+    if 'class' in df.keys():
+        class_='class'
+    else:
+        class_='discourse_type'
+
+    html = "<{0} style='padding: 2px'>{1} <strong> [{0}] </strong></{0}>"\
+            .format(df[class_],df['discourse_text'])
+
+    return html
+
+
+def comparison_text(prediction, ground_truth):
+    """
+    Allow for visual comparison of an essay with predicted classes vs the essay
+    with the true classes
+
+    Args:
+        prediction (str): essay with predicted classes in html formatting
+        ground_truth (str): essay with true classes in html formatting
+
+    Returns:
+        html: visual table
+    """
+
+
+    html = f"""
+    <div class="content">
+     <span style="font-size:16px">Legend --></span>
+      <lead>Lead</lead>
+      <Position>Position</Position>
+      <Claim>Claim</Claim>
+      <Counterclaim>Counterclaim</Counterclaim>
+      <Rebuttal>Rebuttal</Rebuttal>
+      <Evidence>Evidence</Evidence>
+      <Concluding_Statement>Concluding_Statement</Concluding_Statement>
+    </div>
+
+    <div class="row">
+      <div class="column">
+        <h2 class="title">Prediction</h2>
+        <p style="text-align:justify">{prediction}</p>
+      </div>
+      <div class="column">
+        <h2 class="title">Ground Truth</h2>
+        <p style="text-align:justify">{ground_truth}</p>
+      </div>
+    </div>
+    """
+    return HTML(html)
diff --git a/notebooks/arthur/findings.ipynb b/notebooks/arthur/findings.ipynb
new file mode 100644
index 0000000..b21b496
--- /dev/null
+++ b/notebooks/arthur/findings.ipynb
@@ -0,0 +1,1488 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "ace5385f",
+   "metadata": {},
+   "source": [
+    "<font size=30px><center>**Key learnings**</center></font>\n",
+    "<br><br>\n",
+    "This notebooks aims to regroup all the key learnings found during the project from EDA to deep dives on the predictions."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "6f786b60",
+   "metadata": {
+    "toc": true
+   },
+   "source": [
+    "<h1>Table of Contents<span class=\"tocSkip\"></span></h1>\n",
+    "<div class=\"toc\"><ul class=\"toc-item\"><li><span><a href=\"#Imports-and-data-loading\" data-toc-modified-id=\"Imports-and-data-loading-1\"><span class=\"toc-item-num\">1&nbsp;&nbsp;</span>Imports and data loading</a></span></li><li><span><a href=\"#Exploratory-data-analysis\" data-toc-modified-id=\"Exploratory-data-analysis-2\"><span class=\"toc-item-num\">2&nbsp;&nbsp;</span>Exploratory data analysis</a></span><ul class=\"toc-item\"><li><span><a href=\"#Is-there-a-lot-of-sentences-that-overlap-2-different-discourses-?\" data-toc-modified-id=\"Is-there-a-lot-of-sentences-that-overlap-2-different-discourses-?-2.1\"><span class=\"toc-item-num\">2.1&nbsp;&nbsp;</span>Is there a lot of sentences that overlap 2 different discourses ?</a></span></li><li><span><a href=\"#Distribution-length-of-discourses\" data-toc-modified-id=\"Distribution-length-of-discourses-2.2\"><span class=\"toc-item-num\">2.2&nbsp;&nbsp;</span>Distribution length of discourses</a></span><ul class=\"toc-item\"><li><span><a href=\"#Length-distribution-across-discourse-types\" data-toc-modified-id=\"Length-distribution-across-discourse-types-2.2.1\"><span class=\"toc-item-num\">2.2.1&nbsp;&nbsp;</span>Length distribution across discourse types</a></span></li></ul></li><li><span><a href=\"#Distribution-length-of-essays\" data-toc-modified-id=\"Distribution-length-of-essays-2.3\"><span class=\"toc-item-num\">2.3&nbsp;&nbsp;</span>Distribution length of essays</a></span></li><li><span><a href=\"#Distribution-of-discourse-types\" data-toc-modified-id=\"Distribution-of-discourse-types-2.4\"><span class=\"toc-item-num\">2.4&nbsp;&nbsp;</span>Distribution of discourse types</a></span><ul class=\"toc-item\"><li><span><a href=\"#Distribution-discourse-types-across-essays-on-average\" data-toc-modified-id=\"Distribution-discourse-types-across-essays-on-average-2.4.1\"><span class=\"toc-item-num\">2.4.1&nbsp;&nbsp;</span>Distribution discourse types across essays on average</a></span></li></ul></li></ul></li><li><span><a href=\"#First-iteration-on-predictions\" data-toc-modified-id=\"First-iteration-on-predictions-3\"><span class=\"toc-item-num\">3&nbsp;&nbsp;</span>First iteration on predictions</a></span><ul class=\"toc-item\"><li><span><a href=\"#Load-data-from-dataset\" data-toc-modified-id=\"Load-data-from-dataset-3.1\"><span class=\"toc-item-num\">3.1&nbsp;&nbsp;</span>Load data from dataset</a></span></li><li><span><a href=\"#Get-preds\" data-toc-modified-id=\"Get-preds-3.2\"><span class=\"toc-item-num\">3.2&nbsp;&nbsp;</span>Get preds</a></span></li><li><span><a href=\"#Compare-lengths-per-class\" data-toc-modified-id=\"Compare-lengths-per-class-3.3\"><span class=\"toc-item-num\">3.3&nbsp;&nbsp;</span>Compare lengths per class</a></span></li><li><span><a href=\"#Compare-distribution-of-lengths-per-class\" data-toc-modified-id=\"Compare-distribution-of-lengths-per-class-3.4\"><span class=\"toc-item-num\">3.4&nbsp;&nbsp;</span>Compare distribution of lengths per class</a></span></li><li><span><a href=\"#Number-of-discourses-under-a-certain-length\" data-toc-modified-id=\"Number-of-discourses-under-a-certain-length-3.5\"><span class=\"toc-item-num\">3.5&nbsp;&nbsp;</span>Number of discourses under a certain length</a></span></li><li><span><a href=\"#Stammering-issue\" data-toc-modified-id=\"Stammering-issue-3.6\"><span class=\"toc-item-num\">3.6&nbsp;&nbsp;</span>Stammering issue</a></span><ul class=\"toc-item\"><li><span><a href=\"#Visual-comparison\" data-toc-modified-id=\"Visual-comparison-3.6.1\"><span class=\"toc-item-num\">3.6.1&nbsp;&nbsp;</span>Visual comparison</a></span><ul class=\"toc-item\"><li><span><a href=\"#Data-preparation\" data-toc-modified-id=\"Data-preparation-3.6.1.1\"><span class=\"toc-item-num\">3.6.1.1&nbsp;&nbsp;</span>Data preparation</a></span></li></ul></li></ul></li></ul></li></ul></div>"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "f925070b",
+   "metadata": {},
+   "source": [
+    "# Imports and data loading"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 46,
+   "id": "6571aa17",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-15T16:33:56.995010Z",
+     "start_time": "2022-02-15T16:33:56.964549Z"
+    }
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "The autoreload extension is already loaded. To reload it, use:\n",
+      "  %reload_ext autoreload\n"
+     ]
+    }
+   ],
+   "source": [
+    "# imports\n",
+    "%load_ext autoreload\n",
+    "%autoreload 2 \n",
+    "\n",
+    "import pickle\n",
+    "import os\n",
+    "import random\n",
+    "from tqdm.notebook import tqdm\n",
+    "\n",
+    "import pandas as pd \n",
+    "import numpy as np\n",
+    "import matplotlib.pyplot as plt \n",
+    "import seaborn as sns\n",
+    "\n",
+    "from IPython.display import HTML \n",
+    "from termcolor import colored"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "e01ed419",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-15T16:30:05.554782Z",
+     "start_time": "2022-02-15T16:30:05.533905Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "# utils var\n",
+    "sns.set_style('whitegrid')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "155795d2",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-15T16:30:06.263710Z",
+     "start_time": "2022-02-15T16:30:05.556927Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "# data loading\n",
+    "df = pd.read_csv('../raw_data/train.csv')\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "69d9f13b",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-15T16:30:07.610686Z",
+     "start_time": "2022-02-15T16:30:06.267018Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "#Max len of essay \n",
+    "SEQ_LEN = 1024 ## THIS SHOULD NOT BE CHANGED without appropriate changes in the preprocessing \n",
+    "\n",
+    "#Train, val, test split proportion\n",
+    "VAL_SPLIT = 0.8\n",
+    "TEST_SPLIT = 0.9\n",
+    "\n",
+    "#only to get length at this stage\n",
+    "df_essays = pd.read_csv('../raw_data/preprocessed_v3.csv')\n",
+    "LEN=len(df_essays)\n",
+    "del df_essays\n",
+    "\n",
+    "idx_val=int(LEN*VAL_SPLIT)\n",
+    "idx_test=int(LEN*TEST_SPLIT)\n",
+    "\n",
+    "idx_train=list(range(0,idx_val))\n",
+    "idx_val=list(range(idx_val,idx_test))\n",
+    "idx_test=list(range(idx_test,LEN))\n",
+    "\n",
+    "assert(len(idx_test)+len(idx_train)+len(idx_val)==LEN)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "c57dcd1e",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-15T16:30:07.638833Z",
+     "start_time": "2022-02-15T16:30:07.612877Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "#labels map\n",
+    "labels_mapping = {'B-Lead' : 0,\n",
+    "                  'B-Position' : 1,\n",
+    "                  'B-Evidence' : 2,\n",
+    "                  'B-Claim' : 3,\n",
+    "                  'B-Concluding_Statement' : 4,\n",
+    "                  'B-Counterclaim' : 5,\n",
+    "                  'B-Rebuttal' : 6,\n",
+    "                  'I-Lead' : 7,\n",
+    "                  'I-Position' : 8,\n",
+    "                  'I-Evidence' : 9,\n",
+    "                  'I-Claim' : 10,\n",
+    "                  'I-Concluding_Statement' : 11,\n",
+    "                  'I-Counterclaim' : 12,\n",
+    "                  'I-Rebuttal': 13,\n",
+    "                 'O':14,\n",
+    "                 'PAD':15}\n",
+    "\n",
+    "reversed_mapping = {v:(k[2:] if v<14 else k) for k,v in labels_mapping.items()}"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "0454f0cf",
+   "metadata": {
+    "heading_collapsed": true
+   },
+   "source": [
+    "# Exploratory data analysis"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "c5be3194",
+   "metadata": {
+    "hidden": true
+   },
+   "source": [
+    "## Is there a lot of sentences that overlap 2 different discourses ? "
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "id": "3f56b957",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-15T16:30:08.773477Z",
+     "start_time": "2022-02-15T16:30:07.640570Z"
+    },
+    "hidden": true
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Proportion of sentences overlapping 2 different discourses : 33.05%\n"
+     ]
+    }
+   ],
+   "source": [
+    "# Check whether there is a lot of type change in the middle of a sentence\n",
+    "count=0\n",
+    "for i in range(len(df)-1):\n",
+    "    if '.' in df.loc[i,'discourse_text'][-3:]:\n",
+    "        count+=1        \n",
+    "        \n",
+    "print(f'Proportion of sentences overlapping 2 different discourses : {(1-count/len(df))*100:.2f}%')"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "b13d9207",
+   "metadata": {
+    "hidden": true
+   },
+   "source": [
+    "## Distribution length of discourses"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "id": "07d102e2",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-15T16:30:09.260800Z",
+     "start_time": "2022-02-15T16:30:08.775627Z"
+    },
+    "hidden": true
+   },
+   "outputs": [],
+   "source": [
+    "#create discourse_len feature\n",
+    "df['discourse_w_len']=df['discourse_text'].apply(lambda txt : len(txt.split()))\n",
+    "df['discourse_c_len']=df['discourse_text'].apply(len)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "id": "8c47696c",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-15T16:30:12.028809Z",
+     "start_time": "2022-02-15T16:30:09.263079Z"
+    },
+    "hidden": true
+   },
+   "outputs": [
+    {
+     "data": {
+      "image/png": "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\n",
+      "text/plain": [
+       "<Figure size 1080x360 with 2 Axes>"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "# plotting lenght both in term of words and characters\n",
+    "\n",
+    "fig, axs = plt.subplots(1,2,figsize=(15,5),tight_layout=True)\n",
+    "\n",
+    "sns.histplot(x=df['discourse_w_len'],kde=True,ax=axs[0],color='lightblue')\n",
+    "sns.histplot(x=df['discourse_c_len'],kde=True,ax=axs[1],color='orange')\n",
+    "\n",
+    "axs[0].set_title('Length in term of words',size=14)\n",
+    "axs[1].set_title('Length in term of characters',size=14)\n",
+    "\n",
+    "axs[0].set_xlim((0,300))\n",
+    "axs[1].set_xlim((0,1500))\n",
+    "\n",
+    "plt.suptitle('Distribution of lengths',size=18,weight='bold');"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "b6eebfa1",
+   "metadata": {
+    "hidden": true
+   },
+   "source": [
+    "### Length distribution across discourse types"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "id": "444f6d11",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-15T16:30:12.342088Z",
+     "start_time": "2022-02-15T16:30:12.031135Z"
+    },
+    "hidden": true
+   },
+   "outputs": [
+    {
+     "data": {
+      "image/png": "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\n",
+      "text/plain": [
+       "<Figure size 1152x576 with 1 Axes>"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "plt.figure(figsize=(16,8),tight_layout=True)\n",
+    "sns.boxplot(y=df['discourse_w_len'],x=df['discourse_type'],showfliers=False)\n",
+    "\n",
+    "plt.xlabel('Discourse type',size=16,labelpad=20)\n",
+    "plt.ylabel('# words',size=16,labelpad=20)\n",
+    "\n",
+    "plt.xticks(size=14)\n",
+    "plt.yticks(size=14)\n",
+    "\n",
+    "plt.title('Length distribution across discourse types',size=18,pad=20,weight='bold');\n",
+    "\n"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "19c3bc9a",
+   "metadata": {
+    "hidden": true
+   },
+   "source": [
+    "## Distribution length of essays"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "id": "ae5dc133",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-15T16:30:13.224269Z",
+     "start_time": "2022-02-15T16:30:12.347075Z"
+    },
+    "hidden": true,
+    "scrolled": true
+   },
+   "outputs": [
+    {
+     "data": {
+      "image/png": "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\n",
+      "text/plain": [
+       "<Figure size 1080x360 with 2 Axes>"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "## agg by essay id\n",
+    "df_essays = df.groupby('id').agg({'discourse_c_len':sum,'discourse_w_len':sum})\n",
+    "\n",
+    "#plot\n",
+    "fig, axs = plt.subplots(1,2,figsize=(15,5),tight_layout=True)\n",
+    "\n",
+    "sns.histplot(x=df_essays['discourse_w_len'],kde=True,ax=axs[0],color='lightblue')\n",
+    "sns.histplot(x=df_essays['discourse_c_len'],kde=True,ax=axs[1],color='orange')\n",
+    "\n",
+    "axs[0].set_title('Length in term of words',size=14)\n",
+    "axs[1].set_title('Length in term of characters',size=14)\n",
+    "\n",
+    "#sns.lineplot(x=[1024,1024],y=[1,740],color='r',ax=axs[0])\n",
+    "axs[0].axvline(1024,ls='--',lw=1.5,color='r')\n",
+    "axs[0].arrow(x=1200,y=600,dx=-160,dy=-100,width=10,length_includes_head=True,)\n",
+    "axs[0].annotate('SEQ_LEN',(1200,600),size=12)\n",
+    "axs[0].annotate('1024',(1030,760),size=12)\n",
+    "\n",
+    "plt.suptitle('Distribution of lengths',size=18, weight='bold');"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "dbb613a2",
+   "metadata": {
+    "hidden": true
+   },
+   "source": [
+    "## Distribution of discourse types"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "id": "95058dec",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-15T16:30:13.453224Z",
+     "start_time": "2022-02-15T16:30:13.226426Z"
+    },
+    "hidden": true
+   },
+   "outputs": [
+    {
+     "data": {
+      "image/png": "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\n",
+      "text/plain": [
+       "<Figure size 936x504 with 1 Axes>"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "plt.figure(figsize=(13,7))\n",
+    "\n",
+    "sns.countplot(x=df['discourse_type'],palette='Set3',\n",
+    "              order=df['discourse_type'].value_counts().index);\n",
+    "\n",
+    "plt.xticks(rotation=90,size=14,);\n",
+    "plt.xlabel(None)\n",
+    "plt.ylabel(None)\n",
+    "\n",
+    "plt.title('Distribution of discourse types in the training set',size=16,pad=20,weight='bold');"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "e4834267",
+   "metadata": {
+    "hidden": true
+   },
+   "source": [
+    "<font size=\"4\"> __Learning__ </font> \n",
+    "<br><br>\n",
+    "<font size=\"3\">The classes are __imbalanced__. Plan of action :\n",
+    "    <ol>\n",
+    "        <li>Evaluate impact of training an imbalanced dataset on the result. One hypothesis is that it might not have a negative impact as essays naturally have classes longer than some others</li>\n",
+    "        <li>Create a balanced dataset and iterate. A way to do so would be to create length per discourse for each essay, get the std and train on the essays with this lower std.</li>\n",
+    "    \n",
+    "</font>"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "140623ae",
+   "metadata": {
+    "hidden": true
+   },
+   "source": [
+    "### Distribution discourse types across essays on average"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 12,
+   "id": "bb9e98a4",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-15T16:30:14.444030Z",
+     "start_time": "2022-02-15T16:30:13.456024Z"
+    },
+    "hidden": true,
+    "scrolled": false
+   },
+   "outputs": [
+    {
+     "data": {
+      "image/png": "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\n",
+      "text/plain": [
+       "<Figure size 720x504 with 1 Axes>"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "#agg\n",
+    "df_dist = df.groupby(['id','discourse_type']).agg({'discourse_text':'count'}).reset_index()\n",
+    "\n",
+    "#plot\n",
+    "plt.figure(figsize=(10,7),tight_layout=True)\n",
+    "sns.barplot(x='discourse_type',y='discourse_text',\n",
+    "            data=df_dist,estimator=np.mean,palette='Set3',\n",
+    "            order=df['discourse_type'].value_counts().index)\n",
+    "\n",
+    "plt.xticks(rotation=90,size=14);\n",
+    "plt.yticks(size=14)\n",
+    "plt.xlabel(None)\n",
+    "plt.ylabel('# of discourse',size=16);\n",
+    "\n",
+    "plt.title('Average # of discourse type per essay',size=16,weight='bold',pad=20);"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "1de81c03",
+   "metadata": {},
+   "source": [
+    "# First iteration on predictions\n",
+    "\n",
+    "These learnings are derived from model_v3 ; analysis on the test split "
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "7cd82c86",
+   "metadata": {},
+   "source": [
+    "__Key take away__\n",
+    "\n",
+    "- Length of each class is not fully captured by the model when looking at average length\n",
+    "- This is most likely driven by discourses predicted to be of very small length, kind of stammering from the model \n",
+    "    - could we set threshold per class to avoid this ?"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "ca34385c",
+   "metadata": {},
+   "source": [
+    "## Load data from dataset \n",
+    "\n",
+    "Let's also take the original train.csv file instead of our re-processed dataset for comparison. We will use it for **high level** comparisons. "
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 13,
+   "id": "0ab81d64",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-15T16:30:31.919643Z",
+     "start_time": "2022-02-15T16:30:14.446380Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "# load data from pickles and csv files\n",
+    "with open('../raw_data/dataset_v3.pickle','rb') as file:\n",
+    "    dataset = pickle.load(file)\n",
+    "    \n",
+    "with open('../raw_data/preds_on_testsplit.pickle','rb') as file:\n",
+    "    y_pred_proba = pickle.load(file)\n",
+    "\n",
+    "    \n",
+    "df_essays = pd.read_csv('../raw_data/preprocessed_v3.csv',converters={'predictionstring':eval,\n",
+    "                                                                     'label':eval})\n",
+    "\n",
+    "df_raw = pd.read_csv('../raw_data/train.csv')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 14,
+   "id": "f30d457f",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-15T16:30:32.142023Z",
+     "start_time": "2022-02-15T16:30:31.924577Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "# get X_test,y_test\n",
+    "X_test = {\n",
+    "    'input_ids' : dataset['inputs']['input_ids'][idx_test],\n",
+    "    'attention_mask' : dataset['inputs']['attention_mask'][idx_test]\n",
+    "}\n",
+    "\n",
+    "y_test = dataset['labels'][idx_test]\n",
+    "ps_test = dataset['predictionstrings'][idx_test]"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "34735cd3",
+   "metadata": {},
+   "source": [
+    "## Get preds"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 15,
+   "id": "d1bbaebf",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-15T16:30:32.172066Z",
+     "start_time": "2022-02-15T16:30:32.143482Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "def get_preds(y_pred,ps):\n",
+    "    \"\"\"\n",
+    "    Generate readable predictions from the output of the model.\n",
+    "\n",
+    "    Args:\n",
+    "        y_pred (ndarray): output of the model\n",
+    "        ps (ndarray): predictionstring referring to the token predicted\n",
+    "\n",
+    "    Returns:\n",
+    "        DataFrame : DataFrame with class and predictionstrings\n",
+    "    \"\"\"\n",
+    "\n",
+    "\n",
+    "    labels = []\n",
+    "    predictionstrings = []\n",
+    "    counts = []\n",
+    "    \n",
+    "    counter=dict()\n",
+    "    \n",
+    "    for tok,pos in zip(y_pred,ps):\n",
+    "        \n",
+    "        if tok <= 13:\n",
+    "            lab = reversed_mapping[tok]\n",
+    "            labels.append(lab)\n",
+    "            predictionstrings.append(pos)\n",
+    "            if len(labels)<2:\n",
+    "                counts.append(str(1))\n",
+    "                counter.setdefault(lab,1)\n",
+    "                continue\n",
+    "            if lab == labels[-2]:\n",
+    "                counts.append(str(counter[lab]))\n",
+    "            else: \n",
+    "                try:\n",
+    "                    counter[lab]+=1\n",
+    "                except KeyError:\n",
+    "                    counter.setdefault(lab,1)\n",
+    "                counts.append(str(counter[lab]))\n",
+    "    \n",
+    "    preds = pd.DataFrame([labels,counts,predictionstrings],index=['class','count','predictionstring']).T\n",
+    "    preds['class'] += ' ' + preds['count'].astype(str)\n",
+    "    preds = preds.groupby('class',sort=False).agg({'predictionstring':list}).reset_index()\n",
+    "    preds['class']=preds['class'].apply(lambda txt : txt.split()[0])\n",
+    "    preds['predictionstring']=preds['predictionstring'].apply(lambda l_ : [str(l) for l in l_])\n",
+    "    preds['predictionstring']=preds['predictionstring'].apply(lambda l_ : ' '.join(l_))\n",
+    "    \n",
+    "    return preds"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 16,
+   "id": "6df9a277",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-15T16:30:57.249410Z",
+     "start_time": "2022-02-15T16:30:32.174062Z"
+    },
+    "scrolled": true
+   },
+   "outputs": [
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "61c288c575dd46f3952b50263c899653",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "  0%|          | 0/1560 [00:00<?, ?it/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "##creating pred_df\n",
+    "y_pred = np.argmax(y_pred_proba,axis=-1)\n",
+    "pred_df=pd.DataFrame()\n",
+    "for i,idx in tqdm(enumerate(idx_test),total=len(idx_test)):\n",
+    "    \n",
+    "    pred_ = get_preds(y_pred[i],ps_test[i])\n",
+    "    \n",
+    "    pred_['id']=df_essays.iloc[idx]['id']\n",
+    "    \n",
+    "    pred_df = pred_df.append(pred_)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 17,
+   "id": "3fb90245",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-15T16:30:57.281470Z",
+     "start_time": "2022-02-15T16:30:57.251720Z"
+    }
+   },
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>class</th>\n",
+       "      <th>predictionstring</th>\n",
+       "      <th>id</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>Lead</td>\n",
+       "      <td>0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18...</td>\n",
+       "      <td>E6870101D8EE</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>Evidence</td>\n",
+       "      <td>31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 4...</td>\n",
+       "      <td>E6870101D8EE</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>Lead</td>\n",
+       "      <td>65 66 67 68 69 70 71 72 73 74 75 76 77 78</td>\n",
+       "      <td>E6870101D8EE</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>Claim</td>\n",
+       "      <td>79 80 81 82</td>\n",
+       "      <td>E6870101D8EE</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>Lead</td>\n",
+       "      <td>83 84</td>\n",
+       "      <td>E6870101D8EE</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "      class                                   predictionstring            id\n",
+       "0      Lead  0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18...  E6870101D8EE\n",
+       "1  Evidence  31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 4...  E6870101D8EE\n",
+       "2      Lead          65 66 67 68 69 70 71 72 73 74 75 76 77 78  E6870101D8EE\n",
+       "3     Claim                                        79 80 81 82  E6870101D8EE\n",
+       "4      Lead                                              83 84  E6870101D8EE"
+      ]
+     },
+     "execution_count": 17,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# let's check\n",
+    "pred_df.head()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 18,
+   "id": "d009da2c",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-15T16:30:57.305823Z",
+     "start_time": "2022-02-15T16:30:57.283092Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "#create list of ids in the test split for later use\n",
+    "ids_test_split = pred_df.id.unique()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "bd835677",
+   "metadata": {},
+   "source": [
+    "## Compare lengths per class "
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 19,
+   "id": "990a540a",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-15T16:30:57.760835Z",
+     "start_time": "2022-02-15T16:30:57.307891Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "#compute lengths\n",
+    "df_raw['discourse_length'] = df_raw['predictionstring'].apply(lambda txt : len(txt.split()))\n",
+    "pred_df['discourse_length'] = pred_df['predictionstring'].apply(lambda txt : len(txt.split()))\n",
+    "\n",
+    "#create df \n",
+    "a = df_raw.groupby('discourse_type').agg({'discourse_length':'mean'}).reset_index() #take mean of length\n",
+    "a['class'] = a['discourse_type'].replace('Concluding Statement','Concluding_Statement')\n",
+    "a.drop('discourse_type',axis=1,inplace=True)\n",
+    "\n",
+    "b = pred_df.groupby('class').agg({'discourse_length':'mean'}).reset_index() #take mean of length\n",
+    "\n",
+    "comparison_length = a.merge(b,on='class',suffixes=('_true','_pred')) \n",
+    "\n",
+    "#house cleaning for plotting purposes\n",
+    "comparison_length.rename({'discourse_length_pred':'prediction','discourse_length_true':'ground_truth'},\n",
+    "                         inplace=True, axis=1)\n",
+    "comparison_length.set_index(\"class\",inplace=True)\n",
+    "comparison_length = comparison_length.unstack().reset_index()\n",
+    "comparison_length.rename({'level_0':'type',0:'value'},axis=1,inplace=True)\n",
+    "comparison_length.sort_values('value',inplace=True,ascending=False)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 20,
+   "id": "f73eed57",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-15T16:30:58.018528Z",
+     "start_time": "2022-02-15T16:30:57.763685Z"
+    }
+   },
+   "outputs": [
+    {
+     "data": {
+      "image/png": "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\n",
+      "text/plain": [
+       "<Figure size 720x360 with 1 Axes>"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "#plot\n",
+    "plt.figure(figsize=(10,5))\n",
+    "\n",
+    "sns.barplot(data=comparison_length, y='value', x='class',hue='type')\n",
+    "\n",
+    "#pretty plot\n",
+    "plt.xticks(rotation=90,size=12);\n",
+    "plt.yticks(size=12)\n",
+    "\n",
+    "plt.xlabel(None)\n",
+    "plt.ylabel(None)\n",
+    "\n",
+    "plt.title('Comparison average of the length per class',size=16,weight='bold',pad=20)\n",
+    "plt.legend();"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "68d9af1b",
+   "metadata": {},
+   "source": [
+    "## Compare distribution of lengths per class"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 21,
+   "id": "5c8c1673",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-15T16:30:58.523313Z",
+     "start_time": "2022-02-15T16:30:58.020486Z"
+    }
+   },
+   "outputs": [
+    {
+     "data": {
+      "image/png": "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\n",
+      "text/plain": [
+       "<Figure size 1080x432 with 2 Axes>"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "#df, replace, order\n",
+    "a = df_raw.sort_values('discourse_length',ascending=False)\n",
+    "b = pred_df.sort_values('discourse_length',ascending=False)\n",
+    "\n",
+    "a.replace('Concluding Statement','Concluding_Statement',inplace=True)\n",
+    "\n",
+    "order = a.groupby('discourse_type',sort=False).mean()\\\n",
+    "            .sort_values('discourse_length',ascending=False).index\n",
+    "\n",
+    "\n",
+    "#plot\n",
+    "fig,axs = plt.subplots(1,2,figsize=(15,6), sharey= True)\n",
+    "\n",
+    "sns.boxplot(data = a,x='discourse_type',y='discourse_length',ax=axs[0],showfliers=False,order=order)\n",
+    "sns.boxplot(data = b,x='class',y='discourse_length',ax=axs[1],showfliers=False,order=order)\n",
+    "\n",
+    "#pretty plot\n",
+    "axs[0].set_xticklabels(labels = order, rotation=90,size=12)\n",
+    "axs[1].set_xticklabels(labels = order, rotation=90,size=12)\n",
+    "\n",
+    "axs[0].set_xlabel(None)\n",
+    "axs[0].set_ylabel('# words',size=14,labelpad=20)\n",
+    "axs[1].set_xlabel(None)\n",
+    "axs[1].set_ylabel(None)\n",
+    "\n",
+    "axs[0].set_title('Ground Truth',size=14,pad=10)\n",
+    "axs[1].set_title('Predictions',size=14,pad=10);\n",
+    "\n",
+    "plt.suptitle('Distribution of discourse length per class',size=16,weight='bold');"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "f7626833",
+   "metadata": {},
+   "source": [
+    "## Number of discourses under a certain length"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 22,
+   "id": "5cfb3f16",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-15T16:30:58.551971Z",
+     "start_time": "2022-02-15T16:30:58.525018Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "def show_values_on_bars(axs,counts):\n",
+    "    \"\"\"\n",
+    "    This function adds value on top of bar. Should be customed for each use case.\n",
+    "    \"\"\"\n",
+    "    def _show_on_single_plot(ax,counts):\n",
+    "        for p,v in zip(ax.patches,counts):\n",
+    "            _x = p.get_x() + p.get_width() / 2\n",
+    "            _y = p.get_y() + p.get_height() + 1\n",
+    "            value = f'{p.get_height()/v*100:.1f}%'\n",
+    "            ax.text(_x, _y+1.5, value, ha=\"center\",size=12)\n",
+    "\n",
+    "    if isinstance(axs, np.ndarray):\n",
+    "        for idx, ax in np.ndenumerate(axs):\n",
+    "            _show_on_single_plot(ax,counts)\n",
+    "    else:\n",
+    "        _show_on_single_plot(axs,counts)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 23,
+   "id": "83b314e5",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-15T16:30:59.131731Z",
+     "start_time": "2022-02-15T16:30:58.554103Z"
+    },
+    "scrolled": false
+   },
+   "outputs": [
+    {
+     "data": {
+      "image/png": "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\n",
+      "text/plain": [
+       "<Figure size 1152x504 with 2 Axes>"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "## define the threshold \n",
+    "LIMIT = 3\n",
+    "\n",
+    "#df, order, replace\n",
+    "a = df_raw.sort_values('discourse_type')\n",
+    "b = pred_df.sort_values('class')\n",
+    "\n",
+    "a.replace('Concluding Statement','Concluding_Statement',inplace=True)\n",
+    "\n",
+    "\n",
+    "order = a.groupby('discourse_type',sort=False).mean()\\\n",
+    "            .sort_values('discourse_length',ascending=False).index\n",
+    "\n",
+    "## computing total count per class to get %\n",
+    "v_a=a.groupby('discourse_type').count().loc[order]['id']\n",
+    "v_b=b.groupby('class').count().loc[order]['id']\n",
+    "\n",
+    "##filtering according to limot\n",
+    "a = a[a['discourse_length']<LIMIT]\n",
+    "b = b[b['discourse_length']<LIMIT]\n",
+    "\n",
+    "#plot\n",
+    "fig,axs = plt.subplots(1,2,figsize=(16,7),tight_layout=False,sharey=True)\n",
+    "\n",
+    "sns.countplot(data = a,x='discourse_type',ax=axs[0],order=order)\n",
+    "sns.countplot(data = b,x='class',ax=axs[1],order=order)\n",
+    "\n",
+    "#pretty plot\n",
+    "axs[0].set_xticklabels(labels = order, rotation=90,size=12)\n",
+    "axs[1].set_xticklabels(labels = order, rotation=90,size=12)\n",
+    "\n",
+    "axs[0].set_xlabel(None)\n",
+    "axs[0].set_ylabel('# discourses',size=14,labelpad=20)\n",
+    "axs[1].set_xlabel(None)\n",
+    "axs[1].set_ylabel(None)\n",
+    "\n",
+    "show_values_on_bars(axs[0],v_a)\n",
+    "show_values_on_bars(axs[1],v_b)\n",
+    "\n",
+    "axs[0].set_title('Ground Truth',size=14,pad=10)\n",
+    "axs[1].set_title('Predictions',size=14,pad=10);\n",
+    "\n",
+    "fig.text(.05,.95,'% is computed regarding to the # of discourses \\n of the predictions or the gt')\n",
+    "plt.suptitle(f'Count of discourse with word_length < {LIMIT}',size=16,weight='bold');\n",
+    "\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 24,
+   "id": "00beb4d0",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-15T16:30:59.159838Z",
+     "start_time": "2022-02-15T16:30:59.134450Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "## TODO : identify patterns ; maybe there are some proba thresholds that could be set up (idea is that sometimes the model 'hesitate' between two classes thus they both have high probability in the output; could be cleared setting up a minimal probability for a token to be classified)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "0b3a60d4",
+   "metadata": {},
+   "source": [
+    "## Stammering issue"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "c9305ee5",
+   "metadata": {},
+   "source": [
+    "### Visual comparison"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 25,
+   "id": "f53d8a81",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-15T16:30:59.186554Z",
+     "start_time": "2022-02-15T16:30:59.161320Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "from feedback_prize import utils"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 26,
+   "id": "735eb1eb",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-15T16:30:59.212357Z",
+     "start_time": "2022-02-15T16:30:59.188746Z"
+    }
+   },
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<style>\n",
+       ".column {\n",
+       "  float: left;\n",
+       "  width: 50%;\n",
+       "  /*margin-top:10px;*/\n",
+       "  border-right: 2px dotted black;\n",
+       "  text-align:justify;\n",
+       "  }\n",
+       ".column:first-child {\n",
+       "  padding-right:16px;\n",
+       "  padding-left:0px;\n",
+       "  }\n",
+       "\n",
+       ".column:last-child {\n",
+       "  padding-right:0px;\n",
+       "  padding-left:15px;\n",
+       "  border: none;\n",
+       "  }\n",
+       "\n",
+       ".row:after {\n",
+       "  content: \"\";\n",
+       "  display: table;\n",
+       "  clear: both;\n",
+       "  text-align:left\n",
+       "  }\n",
+       "\n",
+       ".title {\n",
+       "  text-align:center;\n",
+       "  padding-bottom: 15px;\n",
+       "  padding-top: 0;\n",
+       "  margin-top: 0;\n",
+       "  border-bottom: 2px dotted black\n",
+       " }\n",
+       "\n",
+       "\n",
+       "div.content > * {\n",
+       "  text-align: center;\n",
+       "  font-weight: bold;\n",
+       "  padding-right: 10px;\n",
+       "  padding-left: 10px;\n",
+       "  padding-top: 5px ;\n",
+       "  padding-bottom:5px;\n",
+       " }\n",
+       "\n",
+       ".content {\n",
+       "  display:flex;\n",
+       "  justify-content:space-evenly;\n",
+       "  margin: 0px auto;\n",
+       "  margin-bottom: 5px;\n",
+       "  background: #FFFFFF;\n",
+       "  padding-block: 10px;\n",
+       "  max-width:auto;\n",
+       "  border: 2px solid black;\n",
+       "  }\n",
+       "\n",
+       "/*styles for each discourse type*/\n",
+       "\n",
+       "Lead {\n",
+       "    background-color:#ff8585;\n",
+       "  }\n",
+       "Position {\n",
+       "  background-color: #7cf0ff;\n",
+       "  }\n",
+       "Evidence {\n",
+       "  background-color: #badcfc;\n",
+       "  }\n",
+       "Claim {\n",
+       "  background-color: #3a3dff;\n",
+       "  }\n",
+       "Concluding_Statement {\n",
+       "  background-color: #ff7df9;\n",
+       "  }\n",
+       "Counterclaim {\n",
+       "  background-color: #4d92e0;\n",
+       "  }\n",
+       "Rebuttal {\n",
+       "  background-color: #ffd57c;\n",
+       "  }\n",
+       "</style>"
+      ],
+      "text/plain": [
+       "<IPython.core.display.HTML object>"
+      ]
+     },
+     "execution_count": 26,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "utils.css()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "51d39727",
+   "metadata": {},
+   "source": [
+    "#### Data preparation"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 65,
+   "id": "645519e8",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-15T17:55:55.101288Z",
+     "start_time": "2022-02-15T17:55:54.667684Z"
+    }
+   },
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>class</th>\n",
+       "      <th>predictionstring</th>\n",
+       "      <th>id</th>\n",
+       "      <th>discourse_length</th>\n",
+       "      <th>flag</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>Lead</td>\n",
+       "      <td>0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18...</td>\n",
+       "      <td>E6C2FD3578B3</td>\n",
+       "      <td>119</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>Position</td>\n",
+       "      <td>119 120 121 122 123 124 125 126 127 128 129 13...</td>\n",
+       "      <td>E6C2FD3578B3</td>\n",
+       "      <td>28</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>Claim</td>\n",
+       "      <td>147</td>\n",
+       "      <td>E6C2FD3578B3</td>\n",
+       "      <td>1</td>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "      class                                   predictionstring            id  \\\n",
+       "0      Lead  0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18...  E6C2FD3578B3   \n",
+       "1  Position  119 120 121 122 123 124 125 126 127 128 129 13...  E6C2FD3578B3   \n",
+       "2     Claim                                                147  E6C2FD3578B3   \n",
+       "\n",
+       "   discourse_length  flag  \n",
+       "0               119     0  \n",
+       "1                28     0  \n",
+       "2                 1     1  "
+      ]
+     },
+     "execution_count": 65,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# flag essay with a length issue\n",
+    "pred_df['flag'] = np.where(pred_df['discourse_length']<3,1,0)\n",
+    "# list essays with > 3 stammers for exploration purporses\n",
+    "stammerers = pred_df.groupby('id').sum()\n",
+    "stammerers = stammerers[stammerers['flag']>3].index\n",
+    "\n",
+    "stam_df = pred_df[pred_df['id'].isin(stammerers)].copy()\n",
+    "stam_df.head(3)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 66,
+   "id": "c35d8930",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-15T17:55:57.374675Z",
+     "start_time": "2022-02-15T17:55:57.058351Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "#link with the essays\n",
+    "stam_df['text'] = stam_df['id'].apply(utils.get_essay)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 67,
+   "id": "2aa03cb4",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-15T17:55:58.683652Z",
+     "start_time": "2022-02-15T17:55:58.109043Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "# connect the predictions to text\n",
+    "slicer_vect = np.vectorize(utils.slicering)\n",
+    "stam_df['discourse_text'] = slicer_vect(stam_df['predictionstring'],stam_df['text'])\n",
+    "\n",
+    "#transform into html string for later visualization\n",
+    "stam_df['html'] = stam_df.apply(utils.render_html,axis=1)\n",
+    "stam_df = stam_df.groupby('id').agg({'html':' '.join,'flag':sum})\n",
+    "#display(stam_df.head(3))\n",
+    "\n",
+    "#same for truth df\n",
+    "df_true = df_raw[df_raw['id'].isin(stammerers)].copy() #selecting only essays of interest\n",
+    "df_true['discourse_type'].replace('Concluding Statement','Concluding_Statement',inplace=True)\n",
+    "df_true['html'] = df_true.apply(utils.render_html,axis=1)\n",
+    "df_true = df_true.groupby('id').agg({'html':' '.join})\n",
+    "#display(df_true.head(3))\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 92,
+   "id": "0dc7a92c",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-15T18:14:14.180797Z",
+     "start_time": "2022-02-15T18:14:14.152519Z"
+    },
+    "scrolled": false
+   },
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "\n",
+       "    <div class=\"content\">\n",
+       "     <span style=\"font-size:16px\">Legend --></span>\n",
+       "      <lead>Lead</lead>\n",
+       "      <Position>Position</Position>\n",
+       "      <Claim>Claim</Claim>\n",
+       "      <Counterclaim>Counterclaim</Counterclaim>\n",
+       "      <Rebuttal>Rebuttal</Rebuttal>\n",
+       "      <Evidence>Evidence</Evidence>\n",
+       "      <Concluding_Statement>Concluding_Statement</Concluding_Statement>\n",
+       "    </div>\n",
+       "\n",
+       "    <div class=\"row\">\n",
+       "      <div class=\"column\">\n",
+       "        <h2 class=\"title\">Prediction</h2>\n",
+       "        <p style=\"text-align:justify\"><Position style='padding: 2px'>The used of this technology in the classroom would benifit the teachers. <strong> [Position] </strong></Position> <Claim style='padding: 2px'>The teachers could teach and use the technology to see if the students were bored or confused. Another way this technology would help with is teachers could make a video teaching the lesson and use the technology to see how the students reacted. The use of this technology would be great for teachers to know if their students are confused <strong> [Claim] </strong></Claim> <Evidence style='padding: 2px'>or bored. The teacher could teach a different way each day and see how the students reacted. Then which ever way kept the students more interested and made them pay attention more would be the way the teacher could start teaching so that the teacher gets the best results from their students. If the students are confused and don't want to ask a question the teacher could see that the student is confused and then the teacher can talk to the student one on one to help the student best understand the topic that he or <strong> [Evidence] </strong></Evidence> <Claim style='padding: 2px'>she is confused with. The technology could be used if the teacher was gone. <strong> [Claim] </strong></Claim> <Evidence style='padding: 2px'>The teacher could make a video teaching the lesson and the facial action coding system could see the students face and change the way the lesson is being taught to each student making the video lesson more personal towards the student. This would help the student learn the lesson how they like to learn. This would help them absorb all of the information in hope that they can retain the information for <strong> [Evidence] </strong></Evidence> <Concluding_Statement style='padding: 2px'>the test. The technology would be great for classrooms. It helps the students learn the lesson the way they like to learn. It also helps the teachers teach the students to the best of their abilities. It will help the teachers <strong> [Concluding_Statement] </strong></Concluding_Statement> <Evidence style='padding: 2px'>teach <strong> [Evidence] </strong></Evidence> <Concluding_Statement style='padding: 2px'>the <strong> [Concluding_Statement] </strong></Concluding_Statement> <Evidence style='padding: 2px'>students the best way for the students. The teachers will know if their students are bored and that the teacher need to change up the way that they are presenting the lesson to the students. It will also help the teachers know if a student needs one on one but the student won't ask the teacher for <strong> [Evidence] </strong></Evidence> <Concluding_Statement style='padding: 2px'>help. With <strong> [Concluding_Statement] </strong></Concluding_Statement> <Evidence style='padding: 2px'>this technology <strong> [Evidence] </strong></Evidence> <Concluding_Statement style='padding: 2px'>the teacher will be able to go to the student and help the student. These reasons are why the facial action coding <strong> [Concluding_Statement] </strong></Concluding_Statement></p>\n",
+       "      </div>\n",
+       "      <div class=\"column\">\n",
+       "        <h2 class=\"title\">Ground Truth</h2>\n",
+       "        <p style=\"text-align:justify\"><Position style='padding: 2px'>The used of this technology in the classroom would benifit the teachers. <strong> [Position] </strong></Position> <Claim style='padding: 2px'>The teachers could teach and use the technology to see if the students were bored or confused  <strong> [Claim] </strong></Claim> <Claim style='padding: 2px'>Another way this technology would help with is teachers could make a video teaching the lesson and use the technology to see how the students reacted.  <strong> [Claim] </strong></Claim> <Claim style='padding: 2px'>The use of this technology would be great for teachers to know if their students are confused or bored. <strong> [Claim] </strong></Claim> <Evidence style='padding: 2px'>The teacher could teach a different way each day and see how the students reacted. Then which ever way kept the students more interested and made them pay attention more would be the way the teacher could start teaching so that the teacher gets the best results from their students. If the students are confused and don't want to ask a question the teacher could see that the student is confused and then the teacher can talk to the student one on one to help the student best understand the topic that he or she is confused with.  <strong> [Evidence] </strong></Evidence> <Claim style='padding: 2px'>The technology could be used if the teacher was gone  <strong> [Claim] </strong></Claim> <Evidence style='padding: 2px'>The teacher could make a video teaching the lesson and the facial action coding system could see the students face and change the way the lesson is being taught to each student making the video lesson more personal towards the student. This would help the student learn the lesson how they like to learn. This would help them absorb all of the information in hope that they can retain the information for the test.  <strong> [Evidence] </strong></Evidence> <Concluding_Statement style='padding: 2px'>The technology would be great for classrooms. It helps the students learn the lesson the way they like to learn. It also helps the teachers teach the students to the best of their abilities. It will help the teachers teach the students the best way for the students. The teachers will know if their students are bored and that the teacher need to change up the way that they are presenting the lesson to the students. It will also help the teachers know if a student needs one on one but the student won't ask the teacher for help. With this technology the teacher will be able to go to the student and help the student. These reasons are why the facial action coding system should be implimented into the classroom. <strong> [Concluding_Statement] </strong></Concluding_Statement></p>\n",
+       "      </div>\n",
+       "    </div>\n",
+       "    "
+      ],
+      "text/plain": [
+       "<IPython.core.display.HTML object>"
+      ]
+     },
+     "execution_count": 92,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "## showing a comparison choosing a random essay \n",
+    "random_idx = random.choice(stam_df.index)\n",
+    "utils.comparison_text(stam_df.loc[random_idx,'html'],df_true.loc[random_idx,'html'])\n",
+    "\n",
+    "##NB : difference comes from my shitty model which apparently tends to classify a lot of tokens as O -__-"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "b9c44900",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.8.12"
+  },
+  "toc": {
+   "base_numbering": 1,
+   "nav_menu": {},
+   "number_sections": true,
+   "sideBar": true,
+   "skip_h1_title": false,
+   "title_cell": "Table of Contents",
+   "title_sidebar": "Contents",
+   "toc_cell": true,
+   "toc_position": {
+    "height": "calc(100% - 180px)",
+    "left": "10px",
+    "top": "150px",
+    "width": "220px"
+   },
+   "toc_section_display": true,
+   "toc_window_display": true
+  },
+  "varInspector": {
+   "cols": {
+    "lenName": 16,
+    "lenType": 16,
+    "lenVar": 40
+   },
+   "kernels_config": {
+    "python": {
+     "delete_cmd_postfix": "",
+     "delete_cmd_prefix": "del ",
+     "library": "var_list.py",
+     "varRefreshCmd": "print(var_dic_list())"
+    },
+    "r": {
+     "delete_cmd_postfix": ") ",
+     "delete_cmd_prefix": "rm(",
+     "library": "var_list.r",
+     "varRefreshCmd": "cat(var_dic_list()) "
+    }
+   },
+   "types_to_exclude": [
+    "module",
+    "function",
+    "builtin_function_or_method",
+    "instance",
+    "_Feature"
+   ],
+   "window_display": false
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}
diff --git a/notebooks/arthur/inference_kaggle.ipynb b/notebooks/arthur/inference_kaggle.ipynb
new file mode 100644
index 0000000..8f446b9
--- /dev/null
+++ b/notebooks/arthur/inference_kaggle.ipynb
@@ -0,0 +1,1228 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "59cb5ea2",
+   "metadata": {
+    "papermill": {
+     "duration": 0.022357,
+     "end_time": "2022-02-11T10:32:08.844251",
+     "exception": false,
+     "start_time": "2022-02-11T10:32:08.821894",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "source": [
+    "# Imports and Variables "
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "9db7b815",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T13:16:56.018595Z",
+     "start_time": "2022-02-09T13:16:56.01273Z"
+    },
+    "execution": {
+     "iopub.execute_input": "2022-02-11T10:32:08.891267Z",
+     "iopub.status.busy": "2022-02-11T10:32:08.889764Z",
+     "iopub.status.idle": "2022-02-11T10:32:08.901047Z",
+     "shell.execute_reply": "2022-02-11T10:32:08.900526Z",
+     "shell.execute_reply.started": "2022-02-09T13:48:24.698621Z"
+    },
+    "papermill": {
+     "duration": 0.03576,
+     "end_time": "2022-02-11T10:32:08.901179",
+     "exception": false,
+     "start_time": "2022-02-11T10:32:08.865419",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "#variables \n",
+    "import os \n",
+    "\n",
+    "SEQ_LEN = 1024 \n",
+    "\n",
+    "BATCH_SIZE = 8\n",
+    "\n",
+    "\n",
+    "#PATHS \n",
+    "\n",
+    "LOAD_BACKBONE_FROM = '../input/backbone/'\n",
+    "LOAD_MODEL_WEIGHTS_FROM = '../input/mymodel/mymodel'\n",
+    "LOAD_TXT_FROM = '../input/feedback-prize-2021/test/'\n",
+    "\n",
+    "#GPU and info message for tf\n",
+    "os.environ[\"CUDA_VISIBLE_DEVICES\"]='0' \n",
+    "os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "89d5ea39",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T13:16:14.992257Z",
+     "start_time": "2022-02-09T13:16:14.981509Z"
+    },
+    "execution": {
+     "iopub.execute_input": "2022-02-11T10:32:08.946557Z",
+     "iopub.status.busy": "2022-02-11T10:32:08.946069Z",
+     "iopub.status.idle": "2022-02-11T10:32:15.810602Z",
+     "shell.execute_reply": "2022-02-11T10:32:15.810107Z",
+     "shell.execute_reply.started": "2022-02-09T13:48:32.513703Z"
+    },
+    "papermill": {
+     "duration": 6.888819,
+     "end_time": "2022-02-11T10:32:15.810763",
+     "exception": false,
+     "start_time": "2022-02-11T10:32:08.921944",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "# canonicals \n",
+    "import pandas as pd \n",
+    "import numpy as np \n",
+    "\n",
+    "#deep\n",
+    "import tensorflow as tf\n",
+    "from transformers import TFAutoModel, AutoConfig, AutoTokenizer"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "621e9a0e",
+   "metadata": {
+    "papermill": {
+     "duration": 0.020572,
+     "end_time": "2022-02-11T10:32:15.852491",
+     "exception": false,
+     "start_time": "2022-02-11T10:32:15.831919",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "source": [
+    "# Loading and preprocessing data"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "263e8156",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T13:18:16.81691Z",
+     "start_time": "2022-02-09T13:18:16.811361Z"
+    },
+    "execution": {
+     "iopub.execute_input": "2022-02-11T10:32:15.898409Z",
+     "iopub.status.busy": "2022-02-11T10:32:15.897831Z",
+     "iopub.status.idle": "2022-02-11T10:32:15.917097Z",
+     "shell.execute_reply": "2022-02-11T10:32:15.917471Z",
+     "shell.execute_reply.started": "2022-02-09T13:48:43.995593Z"
+    },
+    "papermill": {
+     "duration": 0.044566,
+     "end_time": "2022-02-11T10:32:15.917606",
+     "exception": false,
+     "start_time": "2022-02-11T10:32:15.873040",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>id</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>0FB0700DAF44</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>D72CB1C11673</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>18409261F5C2</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>DF920E0A7337</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>D46BCB48440A</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "             id\n",
+       "0  0FB0700DAF44\n",
+       "1  D72CB1C11673\n",
+       "2  18409261F5C2\n",
+       "3  DF920E0A7337\n",
+       "4  D46BCB48440A"
+      ]
+     },
+     "execution_count": 3,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "txt_ids = [f.split('.')[0] for f in os.listdir(LOAD_TXT_FROM)]\n",
+    "df_test = pd.DataFrame(txt_ids,columns=['id'])\n",
+    "df_test"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "1b3ca4ec",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T13:23:48.85536Z",
+     "start_time": "2022-02-09T13:23:48.85116Z"
+    },
+    "execution": {
+     "iopub.execute_input": "2022-02-11T10:32:15.965307Z",
+     "iopub.status.busy": "2022-02-11T10:32:15.964505Z",
+     "iopub.status.idle": "2022-02-11T10:32:15.966316Z",
+     "shell.execute_reply": "2022-02-11T10:32:15.966917Z",
+     "shell.execute_reply.started": "2022-02-09T13:48:47.851725Z"
+    },
+    "papermill": {
+     "duration": 0.02812,
+     "end_time": "2022-02-11T10:32:15.967056",
+     "exception": false,
+     "start_time": "2022-02-11T10:32:15.938936",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "def get_essay(id_):\n",
+    "    \"\"\"Function to get the full text of an essay from the .txt file.\n",
+    "\n",
+    "    Args:\n",
+    "        id_ (str): id of the essay\n",
+    "        mode (str, optional): determines whether to access *train* or *test* texts. \\\n",
+    "        Defaults to 'train'.\n",
+    "\n",
+    "    Returns:\n",
+    "        str: Returns the full text of the id\n",
+    "    \"\"\"\n",
+    "    with open(os.path.join(LOAD_TXT_FROM,f'{id_}.txt'),'r') as file:\n",
+    "        txt = file.read()\n",
+    "        return txt.strip()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "e3eecc9f",
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2022-02-11T10:32:16.016664Z",
+     "iopub.status.busy": "2022-02-11T10:32:16.016099Z",
+     "iopub.status.idle": "2022-02-11T10:32:16.036234Z",
+     "shell.execute_reply": "2022-02-11T10:32:16.035802Z",
+     "shell.execute_reply.started": "2022-02-09T13:48:49.192967Z"
+    },
+    "papermill": {
+     "duration": 0.04807,
+     "end_time": "2022-02-11T10:32:16.036346",
+     "exception": false,
+     "start_time": "2022-02-11T10:32:15.988276",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "# store text in df\n",
+    "df_test['essays'] = df_test['id'].apply(get_essay)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "89373b90",
+   "metadata": {
+    "papermill": {
+     "duration": 0.0211,
+     "end_time": "2022-02-11T10:32:16.078692",
+     "exception": false,
+     "start_time": "2022-02-11T10:32:16.057592",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "source": [
+    "# Tokenize texts"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "9386c3ca",
+   "metadata": {
+    "papermill": {
+     "duration": 0.020633,
+     "end_time": "2022-02-11T10:32:16.120312",
+     "exception": false,
+     "start_time": "2022-02-11T10:32:16.099679",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "source": [
+    "## Tokenizer"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "id": "511ceffc",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T13:22:17.698739Z",
+     "start_time": "2022-02-09T13:22:17.669997Z"
+    },
+    "execution": {
+     "iopub.execute_input": "2022-02-11T10:32:16.171123Z",
+     "iopub.status.busy": "2022-02-11T10:32:16.170332Z",
+     "iopub.status.idle": "2022-02-11T10:32:16.177500Z",
+     "shell.execute_reply": "2022-02-11T10:32:16.177081Z",
+     "shell.execute_reply.started": "2022-02-09T13:48:53.532126Z"
+    },
+    "papermill": {
+     "duration": 0.03649,
+     "end_time": "2022-02-11T10:32:16.177612",
+     "exception": false,
+     "start_time": "2022-02-11T10:32:16.141122",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "def tokenize_labelize(essay,tokenizer,predictionstring=None,labels=None,max_len=SEQ_LEN):\n",
+    "    \"\"\"Tokenize an essay and match each token with the corresponding label.\n",
+    "\n",
+    "    Args:\n",
+    "        essay (str): Text to tokenize\n",
+    "        tokenizer (tokenizer): Tokenizer from HF.\n",
+    "        predictionstring (pandas.Series | numpy.array, optional): As a unique string, list of index position of words with a label. Must be provided with labels. Defaults to None.\n",
+    "        labels (pandas.Series | numpy.array, optional): As a unique string, list of labels of each word. Must be provided with labels. Defaults to None.\n",
+    "        max_len (int): Maximum sequence length for padding/truncating.\n",
+    "        \n",
+    "\n",
+    "    Returns:\n",
+    "        dict : Returns a dictionnary with input_ids,attention_mask and labels if passed.\n",
+    "    \"\"\"\n",
+    "    \n",
+    "    tokens = tokenizer(essay,\n",
+    "                       return_attention_mask = True,\n",
+    "                       return_token_type_ids = False,\n",
+    "                       padding = 'max_length',\n",
+    "                       max_length = SEQ_LEN,\n",
+    "                       truncation = True,\n",
+    "                       return_tensors='np'\n",
+    "                      )\n",
+    "    \n",
+    "    word_ids=tokens.word_ids()\n",
+    "    \n",
+    "    labels_mapping = {\n",
+    "                      'B-Lead' : 0,\n",
+    "                      'B-Position' : 1,\n",
+    "                      'B-Evidence' : 2,\n",
+    "                      'B-Claim' : 3,\n",
+    "                      'B-Concluding_Statement' : 4,\n",
+    "                      'B-Counterclaim' : 5,\n",
+    "                      'B-Rebuttal' : 6,\n",
+    "                      'I-Lead' : 7,\n",
+    "                      'I-Position' : 8,\n",
+    "                      'I-Evidence' : 9,\n",
+    "                      'I-Claim' : 10,\n",
+    "                      'I-Concluding_Statement' : 11,\n",
+    "                      'I-Counterclaim' : 12,\n",
+    "                      'I-Rebuttal': 13\n",
+    "                        }\n",
+    "    \n",
+    "    if labels:\n",
+    "        match = {p:labels_mapping[l] for p,l in zip(predictionstring,labels)}\n",
+    "        labels_matched = [15 if (w==None or w==word_ids[i-1]) \\\n",
+    "                            else match.get(str(w),14) \\\n",
+    "                            for i,w in enumerate(word_ids)]\n",
+    "                            \n",
+    "        \n",
+    "        return {\n",
+    "        'input_ids' : tokens['input_ids'][0],\n",
+    "        'attention_mask' : tokens['attention_mask'][0],\n",
+    "        'labels': np.array(labels_matched), \n",
+    "        'predictionstring':np.array(word_ids)\n",
+    "        }\n",
+    "    \n",
+    "    return {\n",
+    "        'input_ids' : tokens['input_ids'][0],\n",
+    "        'attention_mask' : tokens['attention_mask'][0],\n",
+    "        'predictionstring':np.array(word_ids)\n",
+    "        }"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "id": "8b54766a",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T13:22:43.130645Z",
+     "start_time": "2022-02-09T13:22:43.113541Z"
+    },
+    "execution": {
+     "iopub.execute_input": "2022-02-11T10:32:16.222294Z",
+     "iopub.status.busy": "2022-02-11T10:32:16.221468Z",
+     "iopub.status.idle": "2022-02-11T10:32:16.225883Z",
+     "shell.execute_reply": "2022-02-11T10:32:16.225454Z",
+     "shell.execute_reply.started": "2022-02-09T13:48:55.068076Z"
+    },
+    "papermill": {
+     "duration": 0.027485,
+     "end_time": "2022-02-11T10:32:16.225992",
+     "exception": false,
+     "start_time": "2022-02-11T10:32:16.198507",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "## vectorize the function tokenizer above\n",
+    "tokenize_labelize_vect = np.vectorize(tokenize_labelize,excluded=['SEQ_LEN'],otypes=['object'])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "id": "5b4ded61",
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2022-02-11T10:32:16.271298Z",
+     "iopub.status.busy": "2022-02-11T10:32:16.270792Z",
+     "iopub.status.idle": "2022-02-11T10:32:16.416115Z",
+     "shell.execute_reply": "2022-02-11T10:32:16.416531Z",
+     "shell.execute_reply.started": "2022-02-09T13:48:57.546827Z"
+    },
+    "papermill": {
+     "duration": 0.169916,
+     "end_time": "2022-02-11T10:32:16.416717",
+     "exception": false,
+     "start_time": "2022-02-11T10:32:16.246801",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "## load tokenizer\n",
+    "tokenizer = AutoTokenizer.from_pretrained(LOAD_BACKBONE_FROM)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "id": "a5fb1846",
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2022-02-11T10:32:16.464389Z",
+     "iopub.status.busy": "2022-02-11T10:32:16.463619Z",
+     "iopub.status.idle": "2022-02-11T10:32:16.486745Z",
+     "shell.execute_reply": "2022-02-11T10:32:16.486301Z",
+     "shell.execute_reply.started": "2022-02-09T13:49:00.259629Z"
+    },
+    "papermill": {
+     "duration": 0.048676,
+     "end_time": "2022-02-11T10:32:16.486863",
+     "exception": false,
+     "start_time": "2022-02-11T10:32:16.438187",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "#create tokens\n",
+    "tokens_test = tokenize_labelize_vect(df_test.essays,tokenizer,max_len=SEQ_LEN)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "e7ed71b2",
+   "metadata": {
+    "papermill": {
+     "duration": 0.020753,
+     "end_time": "2022-02-11T10:32:16.528989",
+     "exception": false,
+     "start_time": "2022-02-11T10:32:16.508236",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "source": [
+    "## Dataset creation"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "id": "54940360",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T13:29:39.724058Z",
+     "start_time": "2022-02-09T13:29:39.716081Z"
+    },
+    "execution": {
+     "iopub.execute_input": "2022-02-11T10:32:16.580729Z",
+     "iopub.status.busy": "2022-02-11T10:32:16.578379Z",
+     "iopub.status.idle": "2022-02-11T10:32:16.582622Z",
+     "shell.execute_reply": "2022-02-11T10:32:16.583062Z",
+     "shell.execute_reply.started": "2022-02-09T13:49:03.531009Z"
+    },
+    "papermill": {
+     "duration": 0.033186,
+     "end_time": "2022-02-11T10:32:16.583187",
+     "exception": false,
+     "start_time": "2022-02-11T10:32:16.550001",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "def dataset_creator(tokens):\n",
+    "    \"\"\"\n",
+    "    Creates a dictionnary with tokens attributes as a numpy array.\n",
+    "\n",
+    "    Args:\n",
+    "        tokens (list): list of dictionnaries, outputs from tokeniner\n",
+    "\n",
+    "    Returns:\n",
+    "        dict: dict with list of size BATCH_SIZE of inputs_id, attention mask, predictionstring and labels if provided.\n",
+    "    \"\"\"\n",
+    "    \n",
+    "    keys = tokens[0].keys()\n",
+    "\n",
+    "    inputs = {\n",
+    "        'input_ids':[],\n",
+    "        'attention_mask':[]\n",
+    "        }\n",
+    "    predictionstring = []\n",
+    "    labels = []\n",
+    "       \n",
+    "    for t in tokens:\n",
+    "        inputs['input_ids'].append(t['input_ids'])\n",
+    "        inputs['attention_mask'].append(t['attention_mask'])\n",
+    "        predictionstring.append(t['predictionstring'])\n",
+    "        if 'labels' in keys:\n",
+    "                labels.append(t['labels'])\n",
+    "\n",
+    "        \n",
+    "    inputs['input_ids'] = np.array(inputs['input_ids'])\n",
+    "    inputs['attention_mask'] = np.array(inputs['attention_mask'])\n",
+    "    predictionstring = np.array(predictionstring)\n",
+    "    labels = np.array(labels)\n",
+    "    \n",
+    "    if 'labels' in tokens[0].keys():\n",
+    "        \n",
+    "        #OHE labels\n",
+    "        labels_ohe = np.zeros((len(labels),SEQ_LEN,16))\n",
+    "        \n",
+    "        dim1 = np.arange(len(labels))\n",
+    "        dim2 = np.arange(SEQ_LEN)\n",
+    "        \n",
+    "        labels_ohe[dim1[:,None,None],dim2[None,:,None],labels[:,:,None]] = 1\n",
+    "        \n",
+    "        return inputs, labels_ohe, predictionstring\n",
+    "    \n",
+    "    return inputs, predictionstring"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "id": "ceb5d2ba",
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2022-02-11T10:32:16.629303Z",
+     "iopub.status.busy": "2022-02-11T10:32:16.628544Z",
+     "iopub.status.idle": "2022-02-11T10:32:16.630525Z",
+     "shell.execute_reply": "2022-02-11T10:32:16.630961Z",
+     "shell.execute_reply.started": "2022-02-09T13:49:04.545139Z"
+    },
+    "papermill": {
+     "duration": 0.026963,
+     "end_time": "2022-02-11T10:32:16.631085",
+     "exception": false,
+     "start_time": "2022-02-11T10:32:16.604122",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "## creating test dataset\n",
+    "X_test,ps_test = dataset_creator(tokens_test)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "e59a69a9",
+   "metadata": {
+    "papermill": {
+     "duration": 0.021075,
+     "end_time": "2022-02-11T10:32:16.673192",
+     "exception": false,
+     "start_time": "2022-02-11T10:32:16.652117",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "source": [
+    "# Model prediction"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "7206db1a",
+   "metadata": {
+    "papermill": {
+     "duration": 0.020733,
+     "end_time": "2022-02-11T10:32:16.715129",
+     "exception": false,
+     "start_time": "2022-02-11T10:32:16.694396",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "source": [
+    "## Model architecture"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 12,
+   "id": "7db24263",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T13:37:01.648522Z",
+     "start_time": "2022-02-09T13:37:01.158808Z"
+    },
+    "execution": {
+     "iopub.execute_input": "2022-02-11T10:32:16.761456Z",
+     "iopub.status.busy": "2022-02-11T10:32:16.760922Z",
+     "iopub.status.idle": "2022-02-11T10:32:30.602549Z",
+     "shell.execute_reply": "2022-02-11T10:32:30.603198Z",
+     "shell.execute_reply.started": "2022-02-09T13:49:07.057087Z"
+    },
+    "papermill": {
+     "duration": 13.867367,
+     "end_time": "2022-02-11T10:32:30.603410",
+     "exception": false,
+     "start_time": "2022-02-11T10:32:16.736043",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "All model checkpoint layers were used when initializing TFLongformerModel.\n",
+      "\n",
+      "All the layers of TFLongformerModel were initialized from the model checkpoint at ../input/backbone/tf_model.h5.\n",
+      "If your task is similar to the task the model of the checkpoint was trained on, you can already use TFLongformerModel for predictions without further training.\n"
+     ]
+    }
+   ],
+   "source": [
+    "## Instantiate model Longformer to be used as backbone\n",
+    "config = AutoConfig.from_pretrained(os.path.join(LOAD_BACKBONE_FROM,'config.json'))\n",
+    "backbone = TFAutoModel.from_pretrained(os.path.join(LOAD_BACKBONE_FROM,'tf_model.h5'),config=config)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 13,
+   "id": "de774ab7",
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2022-02-11T10:32:30.657658Z",
+     "iopub.status.busy": "2022-02-11T10:32:30.656163Z",
+     "iopub.status.idle": "2022-02-11T10:32:30.659590Z",
+     "shell.execute_reply": "2022-02-11T10:32:30.660033Z",
+     "shell.execute_reply.started": "2022-02-09T13:49:23.360148Z"
+    },
+    "papermill": {
+     "duration": 0.033231,
+     "end_time": "2022-02-11T10:32:30.660175",
+     "exception": false,
+     "start_time": "2022-02-11T10:32:30.626944",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "def init_model():\n",
+    "    input_ids = tf.keras.layers.Input(shape=(SEQ_LEN,),dtype='int32')\n",
+    "    attention_mask = tf.keras.layers.Input(shape=(SEQ_LEN,),dtype='int32')\n",
+    "    \n",
+    "    x = backbone({'input_ids':input_ids,\n",
+    "                 'attention_mask':attention_mask})[0]\n",
+    "\n",
+    "    backbone.trainable = False\n",
+    "\n",
+    "\n",
+    "    x = tf.keras.layers.Bidirectional(tf.keras.layers.LSTM(units = 512,\n",
+    "                                                           activation = 'tanh',\n",
+    "                                                           dropout=.2,\n",
+    "                                                           return_sequences=True))(x)\n",
+    "    x_res = tf.keras.layers.Bidirectional(tf.keras.layers.LSTM(units = 512,\n",
+    "                                                           activation = 'tanh',\n",
+    "                                                           dropout=.2,\n",
+    "                                                           return_sequences=True))(x)\n",
+    "    \n",
+    "\n",
+    "    x = tf.keras.layers.add([x,x_res])\n",
+    "    output = tf.keras.layers.TimeDistributed(tf.keras.layers.Dense(16,activation = 'softmax'))(x)\n",
+    "\n",
+    "    model = tf.keras.models.Model(inputs={'input_ids':input_ids,\n",
+    "                                          'attention_mask':attention_mask},outputs=output)\n",
+    "    \n",
+    "    \n",
+    "    return model"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 14,
+   "id": "59651021",
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2022-02-11T10:32:30.711176Z",
+     "iopub.status.busy": "2022-02-11T10:32:30.710516Z",
+     "iopub.status.idle": "2022-02-11T10:32:49.698672Z",
+     "shell.execute_reply": "2022-02-11T10:32:49.699094Z",
+     "shell.execute_reply.started": "2022-02-09T13:49:24.262996Z"
+    },
+    "papermill": {
+     "duration": 19.017529,
+     "end_time": "2022-02-11T10:32:49.699240",
+     "exception": false,
+     "start_time": "2022-02-11T10:32:30.681711",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...\n",
+      "To disable this warning, you can either:\n",
+      "\t- Avoid using `tokenizers` before the fork if possible\n",
+      "\t- Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)\n",
+      "huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...\n",
+      "To disable this warning, you can either:\n",
+      "\t- Avoid using `tokenizers` before the fork if possible\n",
+      "\t- Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)\n"
+     ]
+    },
+    {
+     "data": {
+      "image/png": "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\n",
+      "text/plain": [
+       "<IPython.core.display.Image object>"
+      ]
+     },
+     "execution_count": 14,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "## instantiate the model, plot the graph\n",
+    "model = init_model()\n",
+    "\n",
+    "tf.keras.utils.plot_model(model,show_shapes=True)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "41995070",
+   "metadata": {
+    "papermill": {
+     "duration": 0.02412,
+     "end_time": "2022-02-11T10:32:49.747839",
+     "exception": false,
+     "start_time": "2022-02-11T10:32:49.723719",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "source": [
+    "## Model compilation"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 15,
+   "id": "0d637ae6",
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2022-02-11T10:32:49.804197Z",
+     "iopub.status.busy": "2022-02-11T10:32:49.803431Z",
+     "iopub.status.idle": "2022-02-11T10:32:49.805448Z",
+     "shell.execute_reply": "2022-02-11T10:32:49.805879Z",
+     "shell.execute_reply.started": "2022-02-09T13:49:49.416398Z"
+    },
+    "papermill": {
+     "duration": 0.034061,
+     "end_time": "2022-02-11T10:32:49.806006",
+     "exception": false,
+     "start_time": "2022-02-11T10:32:49.771945",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "# creating homemade metric\n",
+    "\n",
+    "def accuracy_masked_func(y_true,y_pred):\n",
+    "    y_pred = tf.cast(tf.argmax(y_pred,axis=-1),'int32')\n",
+    "    y_true = tf.cast(y_true,'int32')\n",
+    "    y_true = tf.cast(tf.argmax(y_true,axis=-1),'int32') #for y_pred and y_true to match\n",
+    "    mask = tf.cast(y_true != 15,'int32') #create a mask\n",
+    "    matches = tf.cast(tf.equal(y_true,y_pred),'int32')*mask #calculate the matches ignoring the masking\n",
+    "    accuracy = tf.math.reduce_sum(matches,axis=-1)/tf.maximum(tf.math.reduce_sum(mask,axis=-1),1)\n",
+    "    \n",
+    "    return accuracy"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 16,
+   "id": "fdb1dcd8",
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2022-02-11T10:32:49.864353Z",
+     "iopub.status.busy": "2022-02-11T10:32:49.863561Z",
+     "iopub.status.idle": "2022-02-11T10:32:49.883890Z",
+     "shell.execute_reply": "2022-02-11T10:32:49.883455Z",
+     "shell.execute_reply.started": "2022-02-09T13:49:50.243674Z"
+    },
+    "papermill": {
+     "duration": 0.053655,
+     "end_time": "2022-02-11T10:32:49.884015",
+     "exception": false,
+     "start_time": "2022-02-11T10:32:49.830360",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "# define loss and metrics \n",
+    "loss = tf.keras.losses.CategoricalCrossentropy(name='categorical_crossentropy')\n",
+    "cat_accuracy = tf.keras.metrics.CategoricalAccuracy()\n",
+    "masked_accuracy = tf.keras.metrics.MeanMetricWrapper(fn=accuracy_masked_func)\n",
+    "\n",
+    "# RMSProp optimizer with clip value and small lr to avoid exploiding gradient \n",
+    "opt = tf.keras.optimizers.RMSprop(clipvalue=.5,learning_rate=0.0001)\n",
+    "\n",
+    "#compile\n",
+    "model.compile(optimizer=opt,loss=loss,metrics=[cat_accuracy,masked_accuracy])"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "9c0a32e9",
+   "metadata": {
+    "papermill": {
+     "duration": 0.024517,
+     "end_time": "2022-02-11T10:32:49.937271",
+     "exception": false,
+     "start_time": "2022-02-11T10:32:49.912754",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "source": [
+    "## Load model pretrained weights"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 17,
+   "id": "534a0fbb",
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2022-02-11T10:32:50.025175Z",
+     "iopub.status.busy": "2022-02-11T10:32:50.024371Z",
+     "iopub.status.idle": "2022-02-11T10:32:57.777162Z",
+     "shell.execute_reply": "2022-02-11T10:32:57.776594Z",
+     "shell.execute_reply.started": "2022-02-09T13:49:56.848183Z"
+    },
+    "papermill": {
+     "duration": 7.81584,
+     "end_time": "2022-02-11T10:32:57.777304",
+     "exception": false,
+     "start_time": "2022-02-11T10:32:49.961464",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "<tensorflow.python.training.tracking.util.CheckpointLoadStatus at 0x7fa9e0120110>"
+      ]
+     },
+     "execution_count": 17,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "model.load_weights(os.path.join(LOAD_MODEL_WEIGHTS_FROM))"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "a25fa713",
+   "metadata": {
+    "papermill": {
+     "duration": 0.024352,
+     "end_time": "2022-02-11T10:32:57.826677",
+     "exception": false,
+     "start_time": "2022-02-11T10:32:57.802325",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "source": [
+    "# Get predictions"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 18,
+   "id": "a068cbca",
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2022-02-11T10:32:57.880940Z",
+     "iopub.status.busy": "2022-02-11T10:32:57.880180Z",
+     "iopub.status.idle": "2022-02-11T10:33:16.530032Z",
+     "shell.execute_reply": "2022-02-11T10:33:16.530749Z",
+     "shell.execute_reply.started": "2022-02-09T13:50:05.619127Z"
+    },
+    "papermill": {
+     "duration": 18.679681,
+     "end_time": "2022-02-11T10:33:16.530952",
+     "exception": false,
+     "start_time": "2022-02-11T10:32:57.851271",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "y_pred = model.predict(X_test,batch_size=BATCH_SIZE)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 19,
+   "id": "a73309fd",
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2022-02-11T10:33:16.619754Z",
+     "iopub.status.busy": "2022-02-11T10:33:16.618610Z",
+     "iopub.status.idle": "2022-02-11T10:33:16.622436Z",
+     "shell.execute_reply": "2022-02-11T10:33:16.623209Z",
+     "shell.execute_reply.started": "2022-02-09T13:52:49.634319Z"
+    },
+    "papermill": {
+     "duration": 0.05607,
+     "end_time": "2022-02-11T10:33:16.623391",
+     "exception": false,
+     "start_time": "2022-02-11T10:33:16.567321",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "labels_mapping = {'B-Lead' : 0,\n",
+    "                  'B-Position' : 1,\n",
+    "                  'B-Evidence' : 2,\n",
+    "                  'B-Claim' : 3,\n",
+    "                  'B-Concluding_Statement' : 4,\n",
+    "                  'B-Counterclaim' : 5,\n",
+    "                  'B-Rebuttal' : 6,\n",
+    "                  'I-Lead' : 7,\n",
+    "                  'I-Position' : 8,\n",
+    "                  'I-Evidence' : 9,\n",
+    "                  'I-Claim' : 10,\n",
+    "                  'I-Concluding_Statement' : 11,\n",
+    "                  'I-Counterclaim' : 12,\n",
+    "                  'I-Rebuttal': 13,\n",
+    "                 'O':14,\n",
+    "                 'PAD':15}\n",
+    "\n",
+    "reversed_mapping = {v:(k[2:] if v<14 else k) for k,v in labels_mapping.items()}"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 20,
+   "id": "022ae676",
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2022-02-11T10:33:16.718680Z",
+     "iopub.status.busy": "2022-02-11T10:33:16.712589Z",
+     "iopub.status.idle": "2022-02-11T10:33:16.731130Z",
+     "shell.execute_reply": "2022-02-11T10:33:16.731891Z",
+     "shell.execute_reply.started": "2022-02-09T13:52:51.914194Z"
+    },
+    "papermill": {
+     "duration": 0.066045,
+     "end_time": "2022-02-11T10:33:16.732035",
+     "exception": false,
+     "start_time": "2022-02-11T10:33:16.665990",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "def get_preds(y_pred,ps):\n",
+    "    \"\"\"\n",
+    "    Generate readable predictions from the output of the model.\n",
+    "\n",
+    "    Args:\n",
+    "        y_pred (ndarray): output of the model\n",
+    "        ps (ndarray): predictionstring referring to the token predicted\n",
+    "\n",
+    "    Returns:\n",
+    "        DataFrame : DataFrame with class and predictionstrings\n",
+    "    \"\"\"\n",
+    "    \n",
+    "    labels = []\n",
+    "    predictionstrings = []\n",
+    "    counts = []\n",
+    "    \n",
+    "    counter=dict()\n",
+    "    \n",
+    "    for tok,pos in zip(y_pred,ps):\n",
+    "        \n",
+    "        if tok <= 13:\n",
+    "            lab = reversed_mapping[tok]\n",
+    "            labels.append(lab)\n",
+    "            predictionstrings.append(pos)\n",
+    "            if len(labels)<2:\n",
+    "                counts.append(str(1))\n",
+    "                counter.setdefault(lab,1)\n",
+    "                continue\n",
+    "            if lab == labels[-2]:\n",
+    "                counts.append(str(counter[lab]))\n",
+    "            else: \n",
+    "                try:\n",
+    "                    counter[lab]+=1\n",
+    "                except KeyError:\n",
+    "                    counter.setdefault(lab,1)\n",
+    "                counts.append(str(counter[lab]))\n",
+    "    \n",
+    "    preds = pd.DataFrame([labels,counts,predictionstrings],index=['class','count','predictionstring']).T\n",
+    "    preds['class'] += ' ' + preds['count'].astype(str)\n",
+    "    preds = preds.groupby('class',sort=False).agg({'predictionstring':list}).reset_index()\n",
+    "    preds['class']=preds['class'].apply(lambda txt : txt.split()[0])\n",
+    "    preds['predictionstring']=preds['predictionstring'].apply(lambda l_ : [str(l) for l in l_])\n",
+    "    preds['predictionstring']=preds['predictionstring'].apply(lambda l_ : ' '.join(l_))\n",
+    "    \n",
+    "    return preds"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 21,
+   "id": "05015116",
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2022-02-11T10:33:16.819694Z",
+     "iopub.status.busy": "2022-02-11T10:33:16.818466Z",
+     "iopub.status.idle": "2022-02-11T10:33:17.084510Z",
+     "shell.execute_reply": "2022-02-11T10:33:17.083921Z",
+     "shell.execute_reply.started": "2022-02-09T13:52:53.90828Z"
+    },
+    "papermill": {
+     "duration": 0.313023,
+     "end_time": "2022-02-11T10:33:17.084673",
+     "exception": false,
+     "start_time": "2022-02-11T10:33:16.771650",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "preds = np.argmax(y_pred,axis=-1)\n",
+    "preds_df = pd.DataFrame()\n",
+    "\n",
+    "for i,idx in enumerate(df_test.index): \n",
+    "    \n",
+    "    pred_ = get_preds(preds[i],ps_test[i])\n",
+    "    \n",
+    "    pred_['id']=df_test.iloc[idx]['id']\n",
+    "    \n",
+    "    preds_df = preds_df.append(pred_)\n"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "be6138d5",
+   "metadata": {
+    "papermill": {
+     "duration": 0.026636,
+     "end_time": "2022-02-11T10:33:17.139195",
+     "exception": false,
+     "start_time": "2022-02-11T10:33:17.112559",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "source": [
+    "# Submission"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 22,
+   "id": "3d5d8852",
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2022-02-11T10:33:17.200959Z",
+     "iopub.status.busy": "2022-02-11T10:33:17.200125Z",
+     "iopub.status.idle": "2022-02-11T10:33:17.205973Z",
+     "shell.execute_reply": "2022-02-11T10:33:17.206518Z",
+     "shell.execute_reply.started": "2022-02-09T13:53:19.555159Z"
+    },
+    "papermill": {
+     "duration": 0.040808,
+     "end_time": "2022-02-11T10:33:17.206691",
+     "exception": false,
+     "start_time": "2022-02-11T10:33:17.165883",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "sub = preds_df[['id','class','predictionstring']]\n",
+    "sub['class'] = sub['class'].replace('Concluding_Statement','Concluding Statement')\n",
+    "sub.reset_index(inplace=True,drop=True)\n",
+    "sub.to_csv('submission.csv',index=False)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "89609b2c",
+   "metadata": {
+    "papermill": {
+     "duration": 0.024892,
+     "end_time": "2022-02-11T10:33:17.258700",
+     "exception": false,
+     "start_time": "2022-02-11T10:33:17.233808",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.7.12"
+  },
+  "papermill": {
+   "default_parameters": {},
+   "duration": 80.332934,
+   "end_time": "2022-02-11T10:33:21.050153",
+   "environment_variables": {},
+   "exception": null,
+   "input_path": "__notebook__.ipynb",
+   "output_path": "__notebook__.ipynb",
+   "parameters": {},
+   "start_time": "2022-02-11T10:32:00.717219",
+   "version": "2.3.3"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}
diff --git a/notebooks/arthur/metrics.ipynb b/notebooks/arthur/metrics.ipynb
new file mode 100644
index 0000000..bc2a2aa
--- /dev/null
+++ b/notebooks/arthur/metrics.ipynb
@@ -0,0 +1,1284 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "13115bc0",
+   "metadata": {},
+   "source": [
+    "<center><font size=18><strong> Metrics</strong></font></center>\n",
+    "<br><br>\n",
+    "This notebook wraps up the metrics used during the training of the models and the evaluation metric for the kaggle competition. It is intended to be used for evaluating the results on the test split."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "165075d8",
+   "metadata": {},
+   "source": [
+    "## Imports"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "id": "c422e5a0",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-12T11:27:19.667516Z",
+     "start_time": "2022-02-12T11:27:19.664254Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "import os \n",
+    "import pickle\n",
+    "\n",
+    "from tqdm.notebook import tqdm"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "id": "ab1e01c2",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-12T11:27:22.723904Z",
+     "start_time": "2022-02-12T11:27:22.720647Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "import pandas as pd \n",
+    "import numpy as np \n",
+    "import matplotlib.pyplot as plt\n",
+    "import seaborn as sns\n",
+    "import tensorflow as tf\n",
+    "\n",
+    "#import seqeval.metrics #not used\n",
+    "\n",
+    "from transformers import AutoTokenizer"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "bedf28c7",
+   "metadata": {
+    "heading_collapsed": true
+   },
+   "source": [
+    "## Categorical Accuracy with masking "
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "1a9ad0bf",
+   "metadata": {
+    "hidden": true
+   },
+   "source": [
+    "### Subclassing -- WIP"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "id": "c7ea652e",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-12T11:27:23.696135Z",
+     "start_time": "2022-02-12T11:27:23.690834Z"
+    },
+    "hidden": true
+   },
+   "outputs": [],
+   "source": [
+    "#Create dummy data\n",
+    "\n",
+    "labels = {\n",
+    "    'B-First':0,\n",
+    "    'I-First':1,\n",
+    "    'B-Second':2,\n",
+    "    'I-Second':3,\n",
+    "    'O':4,\n",
+    "    'PAD':-100\n",
+    "}\n",
+    "\n",
+    "reversed_labels={v:k for k,v in labels.items()}\n",
+    "\n",
+    "y_true = np.array([[0,1,1,1,4,2,3,3,3,-100,-100,-100],[-100,0,1,4,4,2,3,3,3,0,1,-100]])\n",
+    "y_pred = np.array([[0,1,1,1,4,2,3,3,4,4,4,4],[-100,0,1,1,2,2,3,3,0,0,0,4]])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 204,
+   "id": "5cdcc9ff",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-04T12:00:26.638087Z",
+     "start_time": "2022-02-04T12:00:26.630017Z"
+    },
+    "hidden": true
+   },
+   "outputs": [],
+   "source": [
+    "class AccuracyMasked(tf.keras.metrics.Metric):\n",
+    "    \n",
+    "    def __init__(self,class_to_ignore,name='accuracy_masked',**kwargs):\n",
+    "        super().__init__(name=name,**kwargs),\n",
+    "        self.class_to_ignore = class_to_ignore\n",
+    "        self.accs = self.add_weight(name = 'accs',initializer = 'zeros',dtype='float64')        \n",
+    "    \n",
+    "    def update_state(self,y_true,y_pred):\n",
+    "        #y_pred = tf.argmax(y_pred,axis=-1)\n",
+    "        #y_true = tf.argmax(y_true,axis=-1)#for y_pred and y_true to match\n",
+    "        #mask = tf.cast(tf.not_equal(y_true,self.class_to_ignore),'int32') #create a mask\n",
+    "        #matches = tf.cast(tf.equal(y_true,y_pred),'int32')*mask #calculate the matches ignoring the masking\n",
+    "        #accuracy = tf.math.reduce_sum(matches,axis=-1)/tf.maximum(tf.math.reduce_sum(mask,axis=-1),1)\n",
+    "        accuracy=tf.math.reduce_sum(tf.cast(y_true == y_pred,'int32')*tf.cast(y_true != -100,'int32'))/tf.math.reduce_sum(tf.cast(y_true != -100,'int32'))\n",
+    "        self.accs.assign_add(tf.math.reduce_mean(accuracy))\n",
+    "        \n",
+    "    \n",
+    "    def result(self):\n",
+    "        return self.accs #tf.math.reduce_mean(accuracy)\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 205,
+   "id": "963ca44b",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-04T12:00:27.157765Z",
+     "start_time": "2022-02-04T12:00:27.146240Z"
+    },
+    "hidden": true
+   },
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "<tf.Tensor: shape=(), dtype=float64, numpy=0.6086956521739131>"
+      ]
+     },
+     "execution_count": 205,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "accuracymasked = AccuracyMasked(-100)\n",
+    "\n",
+    "accuracymasked(y_pred,y_true)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 181,
+   "id": "d2b72fec",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-04T11:57:45.527829Z",
+     "start_time": "2022-02-04T11:57:45.517677Z"
+    },
+    "hidden": true
+   },
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "<tf.Tensor: shape=(), dtype=float64, numpy=0.7368421052631579>"
+      ]
+     },
+     "execution_count": 181,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "tf.math.reduce_sum(tf.cast(y_true == y_pred,'int32')*tf.cast(y_true != -100,'int32'))/tf.math.reduce_sum(tf.cast(y_true != -100,'int32'))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 99,
+   "id": "02560a48",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-04T10:13:51.823339Z",
+     "start_time": "2022-02-04T10:13:51.817445Z"
+    },
+    "hidden": true
+   },
+   "outputs": [],
+   "source": [
+    "mask = tf.cast(tf.not_equal(y_true,-100),'int32') #create a mask\n",
+    "matches = tf.cast(tf.equal(y_true,y_pred),'int32')*mask #calculate the matches ignoring the masking\n",
+    "accuracy = tf.math.reduce_sum(matches,axis=-1)/tf.maximum(tf.math.reduce_sum(mask,axis=-1),1)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 103,
+   "id": "40ee6f96",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-04T10:13:56.068300Z",
+     "start_time": "2022-02-04T10:13:56.062516Z"
+    },
+    "hidden": true
+   },
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "<tf.Tensor: shape=(), dtype=float64, numpy=0.7444444444444445>"
+      ]
+     },
+     "execution_count": 103,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "tf.math.reduce_mean(accuracy)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "f16d9a53",
+   "metadata": {
+    "heading_collapsed": true,
+    "hidden": true
+   },
+   "source": [
+    "### Using simple function"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 208,
+   "id": "0f26c8fb",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-04T13:35:38.446847Z",
+     "start_time": "2022-02-04T13:35:38.439726Z"
+    },
+    "hidden": true
+   },
+   "outputs": [],
+   "source": [
+    "def accuracy_masked_func(y_true,y_pred):\n",
+    "    \"\"\"\n",
+    "    Compute the accuracy ignoring the class 15 (PAD).\n",
+    "\n",
+    "    Args:\n",
+    "        y_true (tf.Tensor): target of shape (None, 1024, 16)\n",
+    "        y_pred (tf.Tensor): targets predicted of shape (None, 1024, 16)\n",
+    "\n",
+    "    Returns:\n",
+    "        float: accuracy \n",
+    "    \"\"\"\n",
+    "    y_pred = tf.cast(tf.argmax(y_pred,axis=-1),'int32')\n",
+    "    y_true = tf.cast(tf.argmax(y_true,axis=-1),'int32') #for y_pred and y_true to match\n",
+    "    mask = tf.cast(y_true != 15,'int32') #create a mask for 15 = PAD\n",
+    "    matches = tf.cast(tf.equal(y_true,y_pred),'int32')*mask #calculate the matches ignoring the masking\n",
+    "    accuracy = tf.math.reduce_sum(matches,axis=-1)/tf.maximum(tf.math.reduce_sum(mask,axis=-1),1)\n",
+    "    \n",
+    "    return accuracy"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 207,
+   "id": "e812f1eb",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-04T13:30:47.443487Z",
+     "start_time": "2022-02-04T13:30:47.384912Z"
+    },
+    "hidden": true
+   },
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "<tf.Tensor: shape=(2,), dtype=float64, numpy=array([0.88888889, 0.6       ])>"
+      ]
+     },
+     "execution_count": 207,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "accuracy_masked_func(y_true,y_pred)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 210,
+   "id": "1d71adf6",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-04T13:35:53.801253Z",
+     "start_time": "2022-02-04T13:35:53.682910Z"
+    },
+    "hidden": true
+   },
+   "outputs": [],
+   "source": [
+    "#simpler than subclassing\n",
+    "accuracy_masked = tf.keras.metrics.MeanMetricWrapper(fn=accuracy_masked_func)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 211,
+   "id": "b5dec6c5",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-04T13:36:03.220012Z",
+     "start_time": "2022-02-04T13:36:02.983247Z"
+    },
+    "hidden": true
+   },
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "<tf.Tensor: shape=(), dtype=float32, numpy=0.7444445>"
+      ]
+     },
+     "execution_count": 211,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "accuracy_masked(y_true,y_pred)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "20541fd0",
+   "metadata": {},
+   "source": [
+    "## Evaluation metric\n",
+    "\n",
+    "Create a metric to evaluate model performance according to the Kaggle\n",
+    "competition rules.\n",
+    "\n",
+    "- For each sample, all ground truths and predictions for a given class are \n",
+    "    compared.\n",
+    "- If the overlap between the ground truth and prediction is >= 0.5, \n",
+    "  and the overlap between the prediction and the ground truth >= 0.5, \n",
+    "  the prediction is a match and considered a true positive. If multiple \n",
+    "  matches exist, the match with the highest pair of overlaps is taken.\n",
+    "- Any unmatched ground truths are false negatives and any unmatched \n",
+    "  predictions are false positives.\n",
+    "    "
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "0acd78d9",
+   "metadata": {},
+   "source": [
+    "### Get the data\n",
+    "Only for testing purposes"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "id": "04b7374c",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-12T11:27:45.796473Z",
+     "start_time": "2022-02-12T11:27:28.711604Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "#get the data\n",
+    "df_essays = pd.read_csv('../raw_data/preprocessed_v3.csv',converters={'predictionstring':eval,\n",
+    "                                                                     'label':eval})\n",
+    "with open('../raw_data/preds_on_testsplit.pickle','rb') as file:\n",
+    "    y_pred = pickle.load(file)\n",
+    "    \n",
+    "with open('../raw_data/dataset_v3.pickle','rb') as file:\n",
+    "    dataset = pickle.load(file)\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 12,
+   "id": "812217ff",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-12T11:27:45.934390Z",
+     "start_time": "2022-02-12T11:27:45.928556Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "#Max len of essay \n",
+    "SEQ_LEN = 1024 ## THIS SHOULD NOT BE CHANGED without appropriate changes in the preprocessing \n",
+    "\n",
+    "#Train, val, test split proportion\n",
+    "VAL_SPLIT = 0.8\n",
+    "TEST_SPLIT = 0.9\n",
+    "\n",
+    "LEN=len(dataset['labels'])\n",
+    "\n",
+    "idx_val=int(LEN*VAL_SPLIT)\n",
+    "idx_test=int(LEN*TEST_SPLIT)\n",
+    "\n",
+    "idx_train=list(range(0,idx_val))\n",
+    "idx_val=list(range(idx_val,idx_test))\n",
+    "idx_test=list(range(idx_test,LEN))\n",
+    "\n",
+    "assert(len(idx_test)+len(idx_train)+len(idx_val)==LEN)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 13,
+   "id": "fc7a9007",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-12T11:27:46.285310Z",
+     "start_time": "2022-02-12T11:27:46.072217Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "#creating X_test, y_test, ps_test\n",
+    "\n",
+    "X_test = {\n",
+    "    'input_ids' : dataset['inputs']['input_ids'][idx_test],\n",
+    "    'attention_mask' : dataset['inputs']['attention_mask'][idx_test]\n",
+    "}\n",
+    "\n",
+    "y_test = dataset['labels'][idx_test]\n",
+    "ps_test = dataset['predictionstrings'][idx_test]"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 14,
+   "id": "f9dad965",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-12T11:27:46.398761Z",
+     "start_time": "2022-02-12T11:27:46.394171Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "#Labels mapping\n",
+    "\n",
+    "labels_mapping = {'B-Lead' : 0,\n",
+    "                  'B-Position' : 1,\n",
+    "                  'B-Evidence' : 2,\n",
+    "                  'B-Claim' : 3,\n",
+    "                  'B-Concluding_Statement' : 4,\n",
+    "                  'B-Counterclaim' : 5,\n",
+    "                  'B-Rebuttal' : 6,\n",
+    "                  'I-Lead' : 7,\n",
+    "                  'I-Position' : 8,\n",
+    "                  'I-Evidence' : 9,\n",
+    "                  'I-Claim' : 10,\n",
+    "                  'I-Concluding_Statement' : 11,\n",
+    "                  'I-Counterclaim' : 12,\n",
+    "                  'I-Rebuttal': 13,\n",
+    "                 'O':14,\n",
+    "                 'PAD':15}\n",
+    "\n",
+    "reversed_mapping = {v:(k[2:] if v<14 else k) for k,v in labels_mapping.items()}"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "d23afa93",
+   "metadata": {},
+   "source": [
+    "### Creating predictions df"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 15,
+   "id": "da5e8108",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-12T11:27:46.511743Z",
+     "start_time": "2022-02-12T11:27:46.504988Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "def get_preds(y_pred,ps):\n",
+    "    \"\"\"\n",
+    "    Generate readable predictions from the output of the model.\n",
+    "\n",
+    "    Args:\n",
+    "        y_pred (ndarray): output of the model\n",
+    "        ps (ndarray): predictionstring referring to the token predicted\n",
+    "\n",
+    "    Returns:\n",
+    "        DataFrame : DataFrame with class and predictionstrings\n",
+    "    \"\"\"\n",
+    "\n",
+    "\n",
+    "    labels = []\n",
+    "    predictionstrings = []\n",
+    "    counts = []\n",
+    "    \n",
+    "    counter=dict()\n",
+    "    \n",
+    "    for tok,pos in zip(y_pred,ps):\n",
+    "        \n",
+    "        if tok <= 13:\n",
+    "            lab = reversed_mapping[tok]\n",
+    "            labels.append(lab)\n",
+    "            predictionstrings.append(pos)\n",
+    "            if len(labels)<2:\n",
+    "                counts.append(str(1))\n",
+    "                counter.setdefault(lab,1)\n",
+    "                continue\n",
+    "            if lab == labels[-2]:\n",
+    "                counts.append(str(counter[lab]))\n",
+    "            else: \n",
+    "                try:\n",
+    "                    counter[lab]+=1\n",
+    "                except KeyError:\n",
+    "                    counter.setdefault(lab,1)\n",
+    "                counts.append(str(counter[lab]))\n",
+    "    \n",
+    "    preds = pd.DataFrame([labels,counts,predictionstrings],index=['class','count','predictionstring']).T\n",
+    "    preds['class'] += ' ' + preds['count'].astype(str)\n",
+    "    preds = preds.groupby('class',sort=False).agg({'predictionstring':list}).reset_index()\n",
+    "    preds['class']=preds['class'].apply(lambda txt : txt.split()[0])\n",
+    "    preds['predictionstring']=preds['predictionstring'].apply(lambda l_ : [str(l) for l in l_])\n",
+    "    preds['predictionstring']=preds['predictionstring'].apply(lambda l_ : ' '.join(l_))\n",
+    "    \n",
+    "    return preds"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 16,
+   "id": "bb83207b",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-12T11:28:11.536940Z",
+     "start_time": "2022-02-12T11:27:46.654887Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "## Working essay per essay, building pred_df\n",
+    "\n",
+    "preds = np.argmax(y_pred,axis=-1)\n",
+    "pred_df=pd.DataFrame()\n",
+    "for i,idx in tdqm(enumerate(idx_test),total=len(idx_test)): ## CHANGE idx_test\n",
+    "    \n",
+    "    pred_ = get_preds(preds[i],ps_test[i])\n",
+    "    \n",
+    "    pred_['id']=df_essays.iloc[idx]['id']\n",
+    "    \n",
+    "    pred_df = pred_df.append(pred_)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 18,
+   "id": "1f1c075b",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-12T11:28:57.763276Z",
+     "start_time": "2022-02-12T11:28:32.577541Z"
+    }
+   },
+   "outputs": [
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "fde323963f1f43819251f0134373b26e",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "  0%|          | 0/1560 [00:00<?, ?it/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "# Working essay per essay, building true_df\n",
+    "\n",
+    "## NB : this does not yield exactly the same result as if \n",
+    "## we processed train.csv. \n",
+    "## TODO : First glance shows delta to be negligible. Yet to further investigate.\n",
+    "\n",
+    "true = np.argmax(y_test,axis=-1)\n",
+    "true_df=pd.DataFrame()\n",
+    "for i,idx in tqdm(enumerate(idx_test),total=len(idx_test)): ## CHANGE idx_test\n",
+    "    \n",
+    "    true_ = get_preds(true[i],ps_test[i])\n",
+    "    \n",
+    "    true_['id']=df_essays.iloc[idx]['id']\n",
+    "    \n",
+    "    true_df = true_df.append(true_)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "94cdc11d",
+   "metadata": {},
+   "source": [
+    "### Scoring functions"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 19,
+   "id": "b2f51aee",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-12T11:29:00.900607Z",
+     "start_time": "2022-02-12T11:29:00.894481Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "def scoring(predictionstring_true,predictionstring_pred):\n",
+    "    \"\"\"\n",
+    "    Identify each prediction to be a True Positive, a False Positive or a False \n",
+    "    Negative according to the competition rules : \n",
+    "    - For each sample, all ground truths and predictions for a given class are \n",
+    "    compared.\n",
+    "    - If the overlap between the ground truth and prediction is >= 0.5, \n",
+    "    and the overlap between the prediction and the ground truth >= 0.5, \n",
+    "    the prediction is a match and considered a true positive. If multiple \n",
+    "    matches exist, the match with the highest pair of overlaps is taken.\n",
+    "    - Any unmatched ground truths are false negatives and any unmatched \n",
+    "    predictions are false positives.\n",
+    "    \n",
+    "    predictionstring_true and predictionstring_pred are a possible match from an\n",
+    "    outer join of all predictions.\n",
+    "\n",
+    "    Args:\n",
+    "        predictionstring_true (str): predictionstring of a true discourse\n",
+    "        predictionstring_pred (str): predictionstring of a predicted discourse\n",
+    "\n",
+    "    Returns:\n",
+    "        str: TP, FP, or FP \n",
+    "    \"\"\"\n",
+    "    \n",
+    "    if predictionstring_true is np.nan :\n",
+    "        return 'FP'\n",
+    "    \n",
+    "    elif predictionstring_pred is np.nan :\n",
+    "        return 'FN'\n",
+    "    \n",
+    "    else:\n",
+    "        \n",
+    "        ps_true = set(predictionstring_true.split(' '))\n",
+    "        ps_pred = set(predictionstring_pred.split(' '))\n",
+    "\n",
+    "        inter = ps_pred.intersection(ps_true)\n",
+    "        overlap_1 = len(inter)/len(ps_true)\n",
+    "        overlap_2 = len(inter)/len(ps_pred)\n",
+    "\n",
+    "        if overlap_1 >= .5 and overlap_2 >= .5:\n",
+    "            return 'TP'\n",
+    "        else:\n",
+    "            return 'FP'\n",
+    "\n",
+    "## vectorize the funct\n",
+    "scoring_vect = np.vectorize(scoring)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 20,
+   "id": "ce1ca8c1",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-12T11:29:02.131685Z",
+     "start_time": "2022-02-12T11:29:02.128171Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "def f1_score(fp,fn,tp):\n",
+    "    return tp/(tp+.5*(fp+fn))*100"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 65,
+   "id": "e9933025",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-12T11:52:47.478241Z",
+     "start_time": "2022-02-12T11:52:47.464874Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "def scores_df(merged_df):\n",
+    "    \"\"\"\n",
+    "    Computes f1-score summary.\n",
+    "\n",
+    "    Args:\n",
+    "        merged_df (DataFrame): DF with all correspondances of ground truth and \n",
+    "        predictions.\n",
+    "\n",
+    "    Returns:\n",
+    "        df, plot: Returns a dataframe summarizing f1-score per class and a\n",
+    "        barplot showing f1 levels in comparison to the Macro f1-score\n",
+    "    \"\"\"\n",
+    "    \n",
+    "    merged_df['FP'] = np.where(merged_df['score']=='FP',1,0)\n",
+    "    merged_df['FN'] = np.where(merged_df['score']=='FN',1,0)\n",
+    "    merged_df['TP'] = np.where(merged_df['score']=='TP',1,0)\n",
+    "\n",
+    "    merged_df['key'] = merged_df['id']+merged_df['class']+merged_df['predictionstring_pred']+\\\n",
+    "                        merged_df['predictionstring_true']\n",
+    "    \n",
+    "    idx_potential_duplicates = merged_df[(merged_df['key'].duplicated()) & (merged_df['TP']==1)].index\n",
+    "    \n",
+    "    merged_df.drop(idx_potential_duplicates,inplace=True,axis=0)\n",
+    "    \n",
+    "    merged_df.drop('key',axis=1,inplace=True)\n",
+    "    \n",
+    "    score_df = merged_df.groupby('class').agg({'TP':sum,'FP':sum,'FN':sum})\n",
+    "    score_df['F1'] = f1_score(score_df.FP,score_df.FN,score_df.TP)\n",
+    "    \n",
+    "    #Create a total row\n",
+    "    score_df.loc['Total']=score_df.mean()\n",
+    "\n",
+    "    ## weighted average\n",
+    "    score_df['Support'] = true_df.groupby('class').count()['id']\n",
+    "    score_df['F1_Weighted']=score_df['F1']*(score_df['Support']/score_df['Support'].sum())\n",
+    "    score_df.loc['Total','Support']=score_df.Support.sum()\n",
+    "    score_df.loc['Total','F1_Weighted']=score_df.F1_Weighted.sum()\n",
+    "\n",
+    "    # impute correct values for the Total row for TP FN FP \n",
+    "    score_df.loc['Total','FP'] = score_df.loc[:'Rebuttal','FP'].sum()\n",
+    "    score_df.loc['Total','FN'] = score_df.loc[:'Rebuttal','FN'].sum()\n",
+    "    score_df.loc['Total','TP'] = score_df.loc[:'Rebuttal','TP'].sum()    \n",
+    "    \n",
+    "    #Print aggregated scores\n",
+    "    print(f\"F1 Macro Score = {score_df.loc['Total','F1']:.2f}%\")\n",
+    "    print(f\"F1 Micro Score = {f1_score(score_df.loc['Total','FP'],score_df.loc['Total','FN'],score_df.loc['Total','TP']):.2f}%\")\n",
+    "    print(f\"F1 Weighted Score = {score_df.loc['Total','F1_Weighted']:.2f}%\")\n",
+    "    \n",
+    "    ## Creating figure\n",
+    "    sns.barplot(y=score_df.reset_index().loc[:6,'F1'],x=score_df.index[:-1],palette='Set2')\n",
+    "    plt.xticks(rotation=90)\n",
+    "    locs,_=plt.xticks()\n",
+    "    plt.plot([locs[0]-.5,locs[-1]+.5],[score_df.loc['Total','F1'],score_df.loc['Total','F1']],c='r');\n",
+    "    plt.title('Macro F1',size= 16);\n",
+    "\n",
+    "    ## FORMATING\n",
+    "    score_df[['TP','FN','FP','Support']] = score_df[['TP','FN','FP','Support']].applymap('{:.0f}'.format)\n",
+    "    score_df[['F1','F1_Weighted']] = score_df[['F1','F1_Weighted']].applymap('{:.2f}%'.format)\n",
+    "    \n",
+    "    \n",
+    "    return score_df"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "69ece2c1",
+   "metadata": {},
+   "source": [
+    "### Building F1 report"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 66,
+   "id": "74444e77",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-12T11:52:49.103916Z",
+     "start_time": "2022-02-12T11:52:48.691932Z"
+    },
+    "scrolled": true
+   },
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>class</th>\n",
+       "      <th>predictionstring_pred</th>\n",
+       "      <th>id</th>\n",
+       "      <th>predictionstring_true</th>\n",
+       "      <th>score</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>Lead</td>\n",
+       "      <td>0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18...</td>\n",
+       "      <td>E6870101D8EE</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>FP</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>Lead</td>\n",
+       "      <td>65 66 67 68 69 70 71 72 73 74 75 76 77 78</td>\n",
+       "      <td>E6870101D8EE</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>FP</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>Lead</td>\n",
+       "      <td>83 84</td>\n",
+       "      <td>E6870101D8EE</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>FP</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>Evidence</td>\n",
+       "      <td>31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 4...</td>\n",
+       "      <td>E6870101D8EE</td>\n",
+       "      <td>0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18...</td>\n",
+       "      <td>FP</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>Evidence</td>\n",
+       "      <td>31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 4...</td>\n",
+       "      <td>E6870101D8EE</td>\n",
+       "      <td>110 111 112 113 114 115 116 117 118 119 120 12...</td>\n",
+       "      <td>FP</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "      class                              predictionstring_pred            id  \\\n",
+       "0      Lead  0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18...  E6870101D8EE   \n",
+       "1      Lead          65 66 67 68 69 70 71 72 73 74 75 76 77 78  E6870101D8EE   \n",
+       "2      Lead                                              83 84  E6870101D8EE   \n",
+       "3  Evidence  31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 4...  E6870101D8EE   \n",
+       "4  Evidence  31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 4...  E6870101D8EE   \n",
+       "\n",
+       "                               predictionstring_true score  \n",
+       "0                                                NaN    FP  \n",
+       "1                                                NaN    FP  \n",
+       "2                                                NaN    FP  \n",
+       "3  0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18...    FP  \n",
+       "4  110 111 112 113 114 115 116 117 118 119 120 12...    FP  "
+      ]
+     },
+     "execution_count": 66,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "#Merge true_df and pred_df to get all possible matches\n",
+    "merged_df = pred_df.merge(true_df,how = 'outer',on=['id','class'],suffixes=('_pred','_true'))\n",
+    "merged_df['score'] = scoring_vect(merged_df.predictionstring_true,\n",
+    "                                  merged_df.predictionstring_pred) # apply scoring func to determine for each\n",
+    "                                                                    # pred if it FP,FN,TP\n",
+    "merged_df.head()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 67,
+   "id": "88eed805",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-12T11:52:49.416950Z",
+     "start_time": "2022-02-12T11:52:49.106575Z"
+    }
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "F1 Macro Score = 44.92%\n",
+      "F1 Micro Score = 39.31%\n",
+      "F1 Weighted Score = 42.86%\n"
+     ]
+    },
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>TP</th>\n",
+       "      <th>FP</th>\n",
+       "      <th>FN</th>\n",
+       "      <th>F1</th>\n",
+       "      <th>Support</th>\n",
+       "      <th>F1_Weighted</th>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>class</th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>Claim</th>\n",
+       "      <td>1570</td>\n",
+       "      <td>7897</td>\n",
+       "      <td>49</td>\n",
+       "      <td>28.32%</td>\n",
+       "      <td>3412</td>\n",
+       "      <td>8.06%</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Concluding_Statement</th>\n",
+       "      <td>1106</td>\n",
+       "      <td>605</td>\n",
+       "      <td>27</td>\n",
+       "      <td>77.78%</td>\n",
+       "      <td>1346</td>\n",
+       "      <td>8.73%</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Counterclaim</th>\n",
+       "      <td>141</td>\n",
+       "      <td>573</td>\n",
+       "      <td>202</td>\n",
+       "      <td>26.68%</td>\n",
+       "      <td>542</td>\n",
+       "      <td>1.21%</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Evidence</th>\n",
+       "      <td>2400</td>\n",
+       "      <td>9860</td>\n",
+       "      <td>0</td>\n",
+       "      <td>32.74%</td>\n",
+       "      <td>3792</td>\n",
+       "      <td>10.35%</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Lead</th>\n",
+       "      <td>774</td>\n",
+       "      <td>618</td>\n",
+       "      <td>50</td>\n",
+       "      <td>69.86%</td>\n",
+       "      <td>963</td>\n",
+       "      <td>5.61%</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Position</th>\n",
+       "      <td>955</td>\n",
+       "      <td>909</td>\n",
+       "      <td>68</td>\n",
+       "      <td>66.16%</td>\n",
+       "      <td>1535</td>\n",
+       "      <td>8.47%</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Rebuttal</th>\n",
+       "      <td>57</td>\n",
+       "      <td>571</td>\n",
+       "      <td>199</td>\n",
+       "      <td>12.90%</td>\n",
+       "      <td>402</td>\n",
+       "      <td>0.43%</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Total</th>\n",
+       "      <td>7003</td>\n",
+       "      <td>21033</td>\n",
+       "      <td>595</td>\n",
+       "      <td>44.92%</td>\n",
+       "      <td>11992</td>\n",
+       "      <td>42.86%</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                        TP     FP   FN      F1 Support F1_Weighted\n",
+       "class                                                             \n",
+       "Claim                 1570   7897   49  28.32%    3412       8.06%\n",
+       "Concluding_Statement  1106    605   27  77.78%    1346       8.73%\n",
+       "Counterclaim           141    573  202  26.68%     542       1.21%\n",
+       "Evidence              2400   9860    0  32.74%    3792      10.35%\n",
+       "Lead                   774    618   50  69.86%     963       5.61%\n",
+       "Position               955    909   68  66.16%    1535       8.47%\n",
+       "Rebuttal                57    571  199  12.90%     402       0.43%\n",
+       "Total                 7003  21033  595  44.92%   11992      42.86%"
+      ]
+     },
+     "execution_count": 67,
+     "metadata": {},
+     "output_type": "execute_result"
+    },
+    {
+     "data": {
+      "image/png": "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",
+      "text/plain": [
+       "<Figure size 432x288 with 1 Axes>"
+      ]
+     },
+     "metadata": {
+      "needs_background": "light"
+     },
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "f1_report = scores_df(merged_df)\n",
+    "f1_report"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "2fe56a67",
+   "metadata": {},
+   "source": [
+    "## Confusion matrix"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 284,
+   "id": "d8ddb586",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-08T09:42:21.575452Z",
+     "start_time": "2022-02-08T09:42:21.571645Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "from sklearn.metrics import confusion_matrix"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 420,
+   "id": "0edffa7e",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-08T10:57:41.466079Z",
+     "start_time": "2022-02-08T10:57:41.456484Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "def show_confusion_matrix(y_true,y_pred):\n",
+    "    \"\"\"\n",
+    "    Plots a confusion matrix.\n",
+    "\n",
+    "    Args:\n",
+    "        y_true (np.array): targets of shape (None, 1024)\n",
+    "        y_pred (np.array): targets predicted of shape (None, 1024)\n",
+    "    \"\"\"\n",
+    "\n",
+    "    reversed_mapping = {\n",
+    "                        0: 'Lead',\n",
+    "                        1: 'Position',\n",
+    "                        2: 'Evidence',\n",
+    "                        3: 'Claim',\n",
+    "                        4: 'Concluding_Statement',\n",
+    "                        5: 'Counterclaim',\n",
+    "                        6: 'Rebuttal',\n",
+    "                        7: 'Lead',\n",
+    "                        8: 'Position',\n",
+    "                        9: 'Evidence',\n",
+    "                        10: 'Claim',\n",
+    "                        11: 'Concluding_Statement',\n",
+    "                        12: 'Counterclaim',\n",
+    "                        13: 'Rebuttal',\n",
+    "                        14: 'O',\n",
+    "                        15: 'PAD'}\n",
+    "    \n",
+    "    y_true_flat = [reversed_mapping[y] for y in y_true.flatten()]\n",
+    "    y_pred_flat = [reversed_mapping[y] for y in y_pred.flatten()]\n",
+    "    \n",
+    "    LABELS = ['Lead','Position','Claim','Counterclaim','Rebuttal','Evidence','Concluding_Statement','O','PAD']\n",
+    "\n",
+    "    cfn = confusion_matrix(y_true_flat,y_pred_flat,labels=LABELS)\n",
+    "    \n",
+    "    fig,ax = plt.subplots(1,1,figsize=(10,10))\n",
+    "    plt.title('Confusion Matrix',size=18,pad=20)\n",
+    "    sns.heatmap(cfn/np.sum(cfn,axis=0)*100,cmap='Blues',annot = True,fmt='.2f',annot_kws={'size':10},ax=ax);\n",
+    "    plt.xticks(np.arange(len(LABELS))+.5,LABELS,rotation = 90,size=12);\n",
+    "    plt.yticks(np.arange(len(LABELS))+.5,LABELS,rotation = 0,size=12);\n",
+    "    plt.xlabel('PREDICTED',size=16);\n",
+    "    plt.ylabel('ACTUAL',size=16);\n",
+    "    for t in ax.texts: t.set_text(t.get_text() + \" %\")\n",
+    "    \n",
+    "    plt.show()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 421,
+   "id": "55ecfde3",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-08T10:57:50.337916Z",
+     "start_time": "2022-02-08T10:57:44.304107Z"
+    }
+   },
+   "outputs": [
+    {
+     "data": {
+      "image/png": "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",
+      "text/plain": [
+       "<Figure size 720x720 with 2 Axes>"
+      ]
+     },
+     "metadata": {
+      "needs_background": "light"
+     },
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "show_confusion_matrix(y_true,y_pred)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "b3e3e48c",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.8.12"
+  },
+  "toc": {
+   "base_numbering": "1",
+   "nav_menu": {},
+   "number_sections": true,
+   "sideBar": true,
+   "skip_h1_title": true,
+   "title_cell": "Table of Contents",
+   "title_sidebar": "Contents",
+   "toc_cell": false,
+   "toc_position": {
+    "height": "calc(100% - 180px)",
+    "left": "10px",
+    "top": "150px",
+    "width": "268px"
+   },
+   "toc_section_display": true,
+   "toc_window_display": true
+  },
+  "varInspector": {
+   "cols": {
+    "lenName": 16,
+    "lenType": 16,
+    "lenVar": 40
+   },
+   "kernels_config": {
+    "python": {
+     "delete_cmd_postfix": "",
+     "delete_cmd_prefix": "del ",
+     "library": "var_list.py",
+     "varRefreshCmd": "print(var_dic_list())"
+    },
+    "r": {
+     "delete_cmd_postfix": ") ",
+     "delete_cmd_prefix": "rm(",
+     "library": "var_list.r",
+     "varRefreshCmd": "cat(var_dic_list()) "
+    }
+   },
+   "position": {
+    "height": "505.844px",
+    "left": "1063px",
+    "right": "20px",
+    "top": "120px",
+    "width": "357px"
+   },
+   "types_to_exclude": [
+    "module",
+    "function",
+    "builtin_function_or_method",
+    "instance",
+    "_Feature"
+   ],
+   "window_display": false
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}
diff --git a/notebooks/arthur/preprocessing.ipynb b/notebooks/arthur/preprocessing.ipynb
new file mode 100644
index 0000000..0391fb0
--- /dev/null
+++ b/notebooks/arthur/preprocessing.ipynb
@@ -0,0 +1,1071 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "ff304fae",
+   "metadata": {},
+   "source": [
+    "# Imports"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 25,
+   "id": "774a69c4",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T14:47:57.059503Z",
+     "start_time": "2022-02-09T14:47:56.945646Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "import os \n",
+    "import csv\n",
+    "import pickle\n",
+    "\n",
+    "from tqdm.notebook import tqdm\n",
+    "\n",
+    "import pandas as pd \n",
+    "import numpy as np \n",
+    "import matplotlib.pyplot as plt \n",
+    "\n",
+    "from transformers import AutoTokenizer"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 26,
+   "id": "e834accd",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T14:47:57.806090Z",
+     "start_time": "2022-02-09T14:47:57.795868Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "## Utilities variables\n",
+    "\n",
+    "#Sample mode \n",
+    "SAMPLE_MODE = None\n",
+    "\n",
+    "#Max len of essay\n",
+    "SEQ_LEN = 1024\n",
+    "\n",
+    "#path\n",
+    "PATH_RAW_DATA='/Users/arthurcollard/code/arthurcol/feedback_prize/raw_data/'\n",
+    "\n",
+    "VERSION = 3\n",
+    "NAME_OUTPUT_FILE = f'preprocessed_v{VERSION}.csv'\n",
+    "NAME_TEST_FILE = f'test_preprocessed_v{VERSION}.csv'"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "a17de51a",
+   "metadata": {},
+   "source": [
+    "# Data loading and preparation"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "91aae15f",
+   "metadata": {},
+   "source": [
+    "## Loading training set"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 27,
+   "id": "76f1bd5e",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T14:48:01.136989Z",
+     "start_time": "2022-02-09T14:47:59.717075Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "#load data from csv file \n",
+    "df = pd.read_csv(PATH_RAW_DATA+'train.csv',nrows=SAMPLE_MODE)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "5cc7bdb6",
+   "metadata": {},
+   "source": [
+    "## Preparation of the training data"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 28,
+   "id": "4867ca17",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T14:48:01.758156Z",
+     "start_time": "2022-02-09T14:48:01.745441Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "## Def a function for labelling discourses per word\n",
+    "\n",
+    "def labelizer(label,len_,flag):\n",
+    "    \"\"\"Repeat the label according to the length of the sentence. Makes use of B/I notation according to the position of the word within the sentence and the sentence within the essay.\n",
+    "\n",
+    "    Args:\n",
+    "        label (str): NER label of the sentence.\n",
+    "        len_ (int): Length of the sentence (n° of words).\n",
+    "        flag (int): 1 if the sentence follows a sentence with the same label. 0 otherwise.\n",
+    "\n",
+    "    Returns:\n",
+    "        str: Returns a string of length (n° of words) len_ with B/I-label repeated len_ times.\n",
+    "    \"\"\"\n",
+    "    if flag==0:\n",
+    "        label_first = f'B-{label} '\n",
+    "    else:\n",
+    "        label_first = f'I-{label} '\n",
+    "        \n",
+    "    return (label_first + f'I-{label} '*(len_-1)).strip()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 29,
+   "id": "b65d4e94",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T14:48:02.820733Z",
+     "start_time": "2022-02-09T14:48:02.343098Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "## Creating features for labeling needs : \n",
+    "\n",
+    "    #Flag if the discourse is the same as the previous one\n",
+    "df['previous_discourse_flag']=np.where(df['discourse_type'].shift(1)==df['discourse_type'],1,0)\n",
+    "\n",
+    "    #Get length of predictionstring\n",
+    "df['predictionstring_len'] = df['predictionstring'].apply(lambda txt:len(txt.split()))\n",
+    "\n",
+    "    # Remove spaces in labels\n",
+    "df['discourse_type']=df['discourse_type'].str.replace('Concluding Statement','Concluding_Statement')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 30,
+   "id": "73c807ed",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T14:48:23.387677Z",
+     "start_time": "2022-02-09T14:48:03.517512Z"
+    }
+   },
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>id</th>\n",
+       "      <th>discourse_id</th>\n",
+       "      <th>discourse_start</th>\n",
+       "      <th>discourse_end</th>\n",
+       "      <th>discourse_text</th>\n",
+       "      <th>discourse_type</th>\n",
+       "      <th>discourse_type_num</th>\n",
+       "      <th>predictionstring</th>\n",
+       "      <th>previous_discourse_flag</th>\n",
+       "      <th>predictionstring_len</th>\n",
+       "      <th>label</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>423A1CA112E2</td>\n",
+       "      <td>1.622628e+12</td>\n",
+       "      <td>8.0</td>\n",
+       "      <td>229.0</td>\n",
+       "      <td>Modern humans today are always on their phone....</td>\n",
+       "      <td>Lead</td>\n",
+       "      <td>Lead 1</td>\n",
+       "      <td>1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 1...</td>\n",
+       "      <td>0</td>\n",
+       "      <td>44</td>\n",
+       "      <td>B-Lead I-Lead I-Lead I-Lead I-Lead I-Lead I-Le...</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "             id  discourse_id  discourse_start  discourse_end  \\\n",
+       "0  423A1CA112E2  1.622628e+12              8.0          229.0   \n",
+       "\n",
+       "                                      discourse_text discourse_type  \\\n",
+       "0  Modern humans today are always on their phone....           Lead   \n",
+       "\n",
+       "  discourse_type_num                                   predictionstring  \\\n",
+       "0             Lead 1  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 1...   \n",
+       "\n",
+       "   previous_discourse_flag  predictionstring_len  \\\n",
+       "0                        0                    44   \n",
+       "\n",
+       "                                               label  \n",
+       "0  B-Lead I-Lead I-Lead I-Lead I-Lead I-Lead I-Le...  "
+      ]
+     },
+     "execution_count": 30,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# vectorize labelizer func and apply to our df \n",
+    "labelizer_vect = np.vectorize(labelizer)\n",
+    "df['label']=labelizer_vect(df['discourse_type'],df['predictionstring_len'],df['previous_discourse_flag'])\n",
+    "df.head(1)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 31,
+   "id": "58b4fc6d",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T14:48:26.232046Z",
+     "start_time": "2022-02-09T14:48:24.060013Z"
+    }
+   },
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>id</th>\n",
+       "      <th>discourse_id</th>\n",
+       "      <th>discourse_start</th>\n",
+       "      <th>discourse_end</th>\n",
+       "      <th>discourse_text</th>\n",
+       "      <th>discourse_type</th>\n",
+       "      <th>discourse_type_num</th>\n",
+       "      <th>predictionstring</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>423A1CA112E2</td>\n",
+       "      <td>1.622628e+12</td>\n",
+       "      <td>8.0</td>\n",
+       "      <td>229.0</td>\n",
+       "      <td>Modern humans today are always on their phone....</td>\n",
+       "      <td>Lead</td>\n",
+       "      <td>Lead 1</td>\n",
+       "      <td>1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 1...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>423A1CA112E2</td>\n",
+       "      <td>1.622628e+12</td>\n",
+       "      <td>230.0</td>\n",
+       "      <td>312.0</td>\n",
+       "      <td>They are some really bad consequences when stu...</td>\n",
+       "      <td>Position</td>\n",
+       "      <td>Position 1</td>\n",
+       "      <td>45 46 47 48 49 50 51 52 53 54 55 56 57 58 59</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>423A1CA112E2</td>\n",
+       "      <td>1.622628e+12</td>\n",
+       "      <td>313.0</td>\n",
+       "      <td>401.0</td>\n",
+       "      <td>Some certain areas in the United States ban ph...</td>\n",
+       "      <td>Evidence</td>\n",
+       "      <td>Evidence 1</td>\n",
+       "      <td>60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "             id  discourse_id  discourse_start  discourse_end  \\\n",
+       "0  423A1CA112E2  1.622628e+12              8.0          229.0   \n",
+       "1  423A1CA112E2  1.622628e+12            230.0          312.0   \n",
+       "2  423A1CA112E2  1.622628e+12            313.0          401.0   \n",
+       "\n",
+       "                                      discourse_text discourse_type  \\\n",
+       "0  Modern humans today are always on their phone....           Lead   \n",
+       "1  They are some really bad consequences when stu...       Position   \n",
+       "2  Some certain areas in the United States ban ph...       Evidence   \n",
+       "\n",
+       "  discourse_type_num                                   predictionstring  \n",
+       "0             Lead 1  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 1...  \n",
+       "1         Position 1       45 46 47 48 49 50 51 52 53 54 55 56 57 58 59  \n",
+       "2         Evidence 1    60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75  "
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>id</th>\n",
+       "      <th>predictionstring</th>\n",
+       "      <th>label</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>0000D23A521A</td>\n",
+       "      <td>[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13,...</td>\n",
+       "      <td>[B-Position, I-Position, I-Position, I-Positio...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>00066EA9880D</td>\n",
+       "      <td>[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13,...</td>\n",
+       "      <td>[B-Lead, I-Lead, I-Lead, I-Lead, I-Lead, I-Lea...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>000E6DE9E817</td>\n",
+       "      <td>[2, 3, 4, 5, 6, 7, 8, 10, 11, 12, 13, 14, 15, ...</td>\n",
+       "      <td>[B-Position, I-Position, I-Position, I-Positio...</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "             id                                   predictionstring  \\\n",
+       "0  0000D23A521A  [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13,...   \n",
+       "1  00066EA9880D  [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13,...   \n",
+       "2  000E6DE9E817  [2, 3, 4, 5, 6, 7, 8, 10, 11, 12, 13, 14, 15, ...   \n",
+       "\n",
+       "                                               label  \n",
+       "0  [B-Position, I-Position, I-Position, I-Positio...  \n",
+       "1  [B-Lead, I-Lead, I-Lead, I-Lead, I-Lead, I-Lea...  \n",
+       "2  [B-Position, I-Position, I-Position, I-Positio...  "
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "## Groupby ID to get predictionstrings and labels as a unique string\n",
+    "\n",
+    "df_essays = df.groupby('id').agg({'predictionstring':' '.join,'label':' '.join})\n",
+    "\n",
+    "## Transform into lists\n",
+    "\n",
+    "df_essays['label'] = df_essays['label'].apply(lambda txt : txt.split())\n",
+    "df_essays['predictionstring'] = df_essays['predictionstring'].apply(lambda txt : txt.split())\n",
+    "\n",
+    "#remove utilities columns created in the original df\n",
+    "df.drop(['previous_discourse_flag','predictionstring_len','label'],axis=1,inplace=True)\n",
+    "\n",
+    "#reset index\n",
+    "df_essays.reset_index(inplace=True)\n",
+    "\n",
+    "display(df.head(3),df_essays.head(3))"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "592ef22e",
+   "metadata": {},
+   "source": [
+    "## Create dataframe for the test set"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 32,
+   "id": "7a898c54",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T14:48:28.767641Z",
+     "start_time": "2022-02-09T14:48:28.753338Z"
+    }
+   },
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>id</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>DF920E0A7337</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>0FB0700DAF44</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>D46BCB48440A</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>18409261F5C2</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>D72CB1C11673</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "             id\n",
+       "0  DF920E0A7337\n",
+       "1  0FB0700DAF44\n",
+       "2  D46BCB48440A\n",
+       "3  18409261F5C2\n",
+       "4  D72CB1C11673"
+      ]
+     },
+     "execution_count": 32,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "ids = [t.split('.')[0] for t in os.listdir(os.path.join(PATH_RAW_DATA,'test'))]\n",
+    "df_test = pd.DataFrame(ids,columns=['id'])\n",
+    "df_test"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "ac4a7ee9",
+   "metadata": {},
+   "source": [
+    "# Retrieve full text properly"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 33,
+   "id": "fbed858f",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T14:48:29.456781Z",
+     "start_time": "2022-02-09T14:48:29.447066Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "#Function\n",
+    "\n",
+    "def get_essay(id_,mode='train'):\n",
+    "    \"\"\"Function to get the full text of an essay from the .txt file.\n",
+    "\n",
+    "    Args:\n",
+    "        id_ (str): id of the essay\n",
+    "        mode (str, optional): determines whether to access *train* or *test* texts. \\\n",
+    "        Defaults to 'train'.\n",
+    "\n",
+    "    Returns:\n",
+    "        str: Returns the full text of the id\n",
+    "    \"\"\"\n",
+    "    with open(os.path.join(PATH_RAW_DATA,mode,f'{id_}.txt'),'r') as file:\n",
+    "        txt = file.read()\n",
+    "        return txt.strip()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "c47b29dc",
+   "metadata": {},
+   "source": [
+    "# Tokenizer"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 34,
+   "id": "8d0c7d8a",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T14:48:30.490257Z",
+     "start_time": "2022-02-09T14:48:30.472338Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "#Function\n",
+    "\n",
+    "def tokenize_labelize(essay,tokenizer,predictionstring=None,labels=None,max_len=SEQ_LEN):\n",
+    "    \"\"\"Tokenize an essay and match each token with the corresponding label.\n",
+    "\n",
+    "    Args:\n",
+    "        essay (str): Text to tokenize\n",
+    "        tokenizer (tokenizer): Tokenizer from HF.\n",
+    "        predictionstring (pandas.Series | numpy.array, optional): As a unique string, list of index position of words with a label. Must be provided with labels. Defaults to None.\n",
+    "        labels (pandas.Series | numpy.array, optional): As a unique string, list of labels of each word. Must be provided with labels. Defaults to None.\n",
+    "        max_len (int): Maximum sequence length for padding/truncating.\n",
+    "        \n",
+    "\n",
+    "    Returns:\n",
+    "        dict : Returns a dictionnary with input_ids,attention_mask and labels if passed.\n",
+    "    \"\"\"\n",
+    "    \n",
+    "    tokens = tokenizer(essay,\n",
+    "                       return_attention_mask = True,\n",
+    "                       return_token_type_ids = False,\n",
+    "                       padding = 'max_length',\n",
+    "                       max_length = SEQ_LEN,\n",
+    "                       truncation = True,\n",
+    "                       return_tensors='np'\n",
+    "                      )\n",
+    "    \n",
+    "    word_ids=tokens.word_ids()\n",
+    "    \n",
+    "    labels_mapping = {\n",
+    "                      'B-Lead' : 0,\n",
+    "                      'B-Position' : 1,\n",
+    "                      'B-Evidence' : 2,\n",
+    "                      'B-Claim' : 3,\n",
+    "                      'B-Concluding_Statement' : 4,\n",
+    "                      'B-Counterclaim' : 5,\n",
+    "                      'B-Rebuttal' : 6,\n",
+    "                      'I-Lead' : 7,\n",
+    "                      'I-Position' : 8,\n",
+    "                      'I-Evidence' : 9,\n",
+    "                      'I-Claim' : 10,\n",
+    "                      'I-Concluding_Statement' : 11,\n",
+    "                      'I-Counterclaim' : 12,\n",
+    "                      'I-Rebuttal': 13\n",
+    "                        }\n",
+    "    \n",
+    "    if labels:\n",
+    "        match = {p:labels_mapping[l] for p,l in zip(predictionstring,labels)}\n",
+    "        labels_matched = [15 if (w==None or w==word_ids[i-1]) \\\n",
+    "                            else match.get(str(w),14) \\\n",
+    "                            for i,w in enumerate(word_ids)]\n",
+    "                            \n",
+    "        \n",
+    "        return {\n",
+    "        'input_ids' : tokens['input_ids'][0],\n",
+    "        'attention_mask' : tokens['attention_mask'][0],\n",
+    "        'labels': np.array(labels_matched), \n",
+    "        'predictionstring':np.array(word_ids)\n",
+    "        }\n",
+    "    \n",
+    "    return {\n",
+    "        'input_ids' : tokens['input_ids'][0],\n",
+    "        'attention_mask' : tokens['attention_mask'][0],\n",
+    "        'predictionstring':np.array(word_ids)\n",
+    "        }"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "7d43342b",
+   "metadata": {},
+   "source": [
+    "# Create preprocessed data\n",
+    "\n",
+    "Working per batch for RAM issues"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 43,
+   "id": "bd001f31",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T14:49:41.762812Z",
+     "start_time": "2022-02-09T14:49:41.609070Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "## Instantiate tokenizer from HF\n",
+    "tokenizer = AutoTokenizer.from_pretrained('backbone')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 44,
+   "id": "6f321835",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T14:49:44.576833Z",
+     "start_time": "2022-02-09T14:49:44.573401Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "## vectorize the function tokenizer above\n",
+    "tokenize_labelize_vect = np.vectorize(tokenize_labelize,excluded=['SEQ_LEN'],otypes=['object'])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 45,
+   "id": "05a661ad",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T14:50:37.492415Z",
+     "start_time": "2022-02-09T14:49:48.077341Z"
+    }
+   },
+   "outputs": [
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "2b6550187b734dab98348745e3db5030",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Processing...:   0%|          | 0/313 [00:00<?, ?it/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "# Create CSV file with tokens (input_ids, attention_mask, predictionstring, labels)\n",
+    "# also stored in an array tokens\n",
+    "\n",
+    "batch_size = 50\n",
+    "nbatch = int(len(df_essays)/batch_size)+1\n",
+    "\n",
+    "fieldnames = ['id','predictionstring','label','essays']\n",
+    "tokens = np.array([])\n",
+    "\n",
+    "with open(PATH_RAW_DATA+NAME_OUTPUT_FILE,'w') as file :\n",
+    "    writer = csv.DictWriter(file,fieldnames = fieldnames)\n",
+    "    writer.writeheader()\n",
+    "\n",
+    "for i in tqdm(range(nbatch+1),desc='Processing...'):\n",
+    "    df_ = df_essays.loc[i*batch_size:(i+1)*batch_size-1].copy()\n",
+    "    df_['essays'] = df_['id'].apply(get_essay)\n",
+    "    tokens = np.append(tokens,tokenize_labelize_vect(df_.essays,tokenizer,\n",
+    "                                           df_.predictionstring, df_.label ,max_len=SEQ_LEN))\n",
+    "    df_.to_csv(PATH_RAW_DATA+NAME_OUTPUT_FILE,mode='a',header=False)\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 46,
+   "id": "bfeff648",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T14:51:15.947759Z",
+     "start_time": "2022-02-09T14:51:13.383076Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "## sanity check\n",
+    "result = pd.read_csv(PATH_RAW_DATA+NAME_OUTPUT_FILE)\n",
+    "assert(tokens.shape[0]==result.shape[0])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 47,
+   "id": "4455a2dc",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T14:51:16.676568Z",
+     "start_time": "2022-02-09T14:51:16.639286Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "##create tokens_test array\n",
+    "\n",
+    "df_test['essays'] = df_test['id'].apply(get_essay,mode='test')\n",
+    "tokens_test = tokenize_labelize_vect(df_test.essays,tokenizer,max_len=SEQ_LEN)\n",
+    "\n",
+    "## saving as csv\n",
+    "df_test.to_csv(PATH_RAW_DATA+f'preprocessed_inf_v{VERSION}.csv')"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "170b3d19",
+   "metadata": {},
+   "source": [
+    "# Build dataset"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 48,
+   "id": "3de728d9",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T14:51:17.818262Z",
+     "start_time": "2022-02-09T14:51:17.804759Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "def dataset_creator(tokens):\n",
+    "    \"\"\"\n",
+    "    Creates a dictionnary with tokens attributes as a numpy array.\n",
+    "\n",
+    "    Args:\n",
+    "        tokens (list): list of dictionnaries, outputs from tokeniner\n",
+    "\n",
+    "    Returns:\n",
+    "        dict: dict with list of size BATCH_SIZE of inputs_id, attention mask, predictionstring and labels if provided.\n",
+    "    \"\"\"\n",
+    "    \n",
+    "    keys = tokens[0].keys()\n",
+    "\n",
+    "    inputs = {\n",
+    "        'input_ids':[],\n",
+    "        'attention_mask':[]\n",
+    "        }\n",
+    "    predictionstring = []\n",
+    "    labels = []\n",
+    "       \n",
+    "    for t in tqdm(tokens,desc='Aggregating dataset'):\n",
+    "        inputs['input_ids'].append(t['input_ids'])\n",
+    "        inputs['attention_mask'].append(t['attention_mask'])\n",
+    "        predictionstring.append(t['predictionstring'])\n",
+    "        if 'labels' in keys:\n",
+    "                labels.append(t['labels'])\n",
+    "\n",
+    "        \n",
+    "    inputs['input_ids'] = np.array(inputs['input_ids'])\n",
+    "    inputs['attention_mask'] = np.array(inputs['attention_mask'])\n",
+    "    predictionstring = np.array(predictionstring)\n",
+    "    labels = np.array(labels)\n",
+    "    \n",
+    "    if 'labels' in tokens[0].keys():\n",
+    "        \n",
+    "        #OHE labels\n",
+    "        labels_ohe = np.zeros((len(labels),SEQ_LEN,16))\n",
+    "        \n",
+    "        dim1 = np.arange(len(labels))\n",
+    "        dim2 = np.arange(SEQ_LEN)\n",
+    "        \n",
+    "        labels_ohe[dim1[:,None,None],dim2[None,:,None],labels[:,:,None]] = 1\n",
+    "        \n",
+    "        return inputs, labels_ohe, predictionstring\n",
+    "    \n",
+    "    return inputs, predictionstring"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 49,
+   "id": "9465810d",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T14:51:22.975169Z",
+     "start_time": "2022-02-09T14:51:19.285701Z"
+    }
+   },
+   "outputs": [
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "6e57bccdfaab41979d186cab4d8881dd",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Aggregating dataset:   0%|          | 0/15594 [00:00<?, ?it/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "#training dataset\n",
+    "if 'labels' in tokens[0].keys():\n",
+    "    inputs,labels,predictionstrings = dataset_creator(tokens)\n",
+    "else:\n",
+    "    inputs,predictionstrings = dataset_creator(tokens)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 50,
+   "id": "5a6d73d2",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T14:51:23.707341Z",
+     "start_time": "2022-02-09T14:51:23.675222Z"
+    }
+   },
+   "outputs": [
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "1837e1233a0049df9578b850c79ab579",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Aggregating dataset:   0%|          | 0/5 [00:00<?, ?it/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "## creating test dataset\n",
+    "inputs_test,ps_test = dataset_creator(tokens_test)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "5d7f188d",
+   "metadata": {},
+   "source": [
+    "# Save datasets"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 51,
+   "id": "4d6e94b9",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T14:51:25.873361Z",
+     "start_time": "2022-02-09T14:51:24.901074Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "## Store all objects in a single dictionnary for training\n",
+    "\n",
+    "if 'labels' in tokens[0].keys():\n",
+    "    dataset = {\n",
+    "        'inputs':inputs,\n",
+    "        'labels':labels,\n",
+    "        'predictionstrings':predictionstrings\n",
+    "    }\n",
+    "else:\n",
+    "    dataset = {\n",
+    "        'inputs':inputs,\n",
+    "        'predictionstrings':predictionstrings\n",
+    "    }"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 52,
+   "id": "d09bf01c",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T14:51:26.522434Z",
+     "start_time": "2022-02-09T14:51:26.519539Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "## store test objects stored in a dict\n",
+    "\n",
+    "dataset_test = {\n",
+    "        'inputs':inputs_test,\n",
+    "        'predictionstrings':ps_test\n",
+    "    }"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 53,
+   "id": "321206c8",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T14:51:35.630267Z",
+     "start_time": "2022-02-09T14:51:27.083694Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "## dump dataset dictionnary with as a pickle file\n",
+    "\n",
+    "with open(f'../raw_data/dataset_v{VERSION}.pickle','wb') as file : \n",
+    "    pickle.dump(dataset,file)\n",
+    "    \n",
+    "with open(f'../raw_data/dataset_test_v{VERSION}.pickle','wb') as file : \n",
+    "    pickle.dump(dataset_test,file)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 22,
+   "id": "7fefe1cc",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-08T13:23:34.140576Z",
+     "start_time": "2022-02-08T13:23:34.137608Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "### the end ###"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.8.12"
+  },
+  "toc": {
+   "base_numbering": 1,
+   "nav_menu": {},
+   "number_sections": true,
+   "sideBar": true,
+   "skip_h1_title": false,
+   "title_cell": "Table of Contents",
+   "title_sidebar": "Contents",
+   "toc_cell": false,
+   "toc_position": {
+    "height": "calc(100% - 180px)",
+    "left": "10px",
+    "top": "150px",
+    "width": "288px"
+   },
+   "toc_section_display": true,
+   "toc_window_display": true
+  },
+  "varInspector": {
+   "cols": {
+    "lenName": 16,
+    "lenType": 16,
+    "lenVar": 40
+   },
+   "kernels_config": {
+    "python": {
+     "delete_cmd_postfix": "",
+     "delete_cmd_prefix": "del ",
+     "library": "var_list.py",
+     "varRefreshCmd": "print(var_dic_list())"
+    },
+    "r": {
+     "delete_cmd_postfix": ") ",
+     "delete_cmd_prefix": "rm(",
+     "library": "var_list.r",
+     "varRefreshCmd": "cat(var_dic_list()) "
+    }
+   },
+   "types_to_exclude": [
+    "module",
+    "function",
+    "builtin_function_or_method",
+    "instance",
+    "_Feature"
+   ],
+   "window_display": false
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}
diff --git a/notebooks/arthur/training_v2.ipynb b/notebooks/arthur/training_v2.ipynb
new file mode 100644
index 0000000..7879a47
--- /dev/null
+++ b/notebooks/arthur/training_v2.ipynb
@@ -0,0 +1,1563 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "5a81d29d",
+   "metadata": {
+    "id": "5a81d29d",
+    "toc": true
+   },
+   "source": [
+    "<h1>Table of Contents<span class=\"tocSkip\"></span></h1>\n",
+    "<div class=\"toc\"><ul class=\"toc-item\"><li><span><a href=\"#Imports-&amp;-Variables\" data-toc-modified-id=\"Imports-&amp;-Variables-1\"><span class=\"toc-item-num\">1&nbsp;&nbsp;</span>Imports &amp; Variables</a></span><ul class=\"toc-item\"><li><span><a href=\"#Colab\" data-toc-modified-id=\"Colab-1.1\"><span class=\"toc-item-num\">1.1&nbsp;&nbsp;</span>Colab</a></span></li><li><span><a href=\"#Imports\" data-toc-modified-id=\"Imports-1.2\"><span class=\"toc-item-num\">1.2&nbsp;&nbsp;</span>Imports</a></span></li><li><span><a href=\"#Variables\" data-toc-modified-id=\"Variables-1.3\"><span class=\"toc-item-num\">1.3&nbsp;&nbsp;</span>Variables</a></span></li></ul></li><li><span><a href=\"#Load-data\" data-toc-modified-id=\"Load-data-2\"><span class=\"toc-item-num\">2&nbsp;&nbsp;</span>Load data</a></span></li><li><span><a href=\"#Modeling\" data-toc-modified-id=\"Modeling-3\"><span class=\"toc-item-num\">3&nbsp;&nbsp;</span>Modeling</a></span><ul class=\"toc-item\"><li><span><a href=\"#Model-architecture\" data-toc-modified-id=\"Model-architecture-3.1\"><span class=\"toc-item-num\">3.1&nbsp;&nbsp;</span>Model architecture</a></span></li><li><span><a href=\"#Loss,-Optimizer,-Metrics\" data-toc-modified-id=\"Loss,-Optimizer,-Metrics-3.2\"><span class=\"toc-item-num\">3.2&nbsp;&nbsp;</span>Loss, Optimizer, Metrics</a></span></li><li><span><a href=\"#Model-training\" data-toc-modified-id=\"Model-training-3.3\"><span class=\"toc-item-num\">3.3&nbsp;&nbsp;</span>Model training</a></span></li><li><span><a href=\"#Model-evaluation\" data-toc-modified-id=\"Model-evaluation-3.4\"><span class=\"toc-item-num\">3.4&nbsp;&nbsp;</span>Model evaluation</a></span></li><li><span><a href=\"#Model-evaluation\" data-toc-modified-id=\"Model-evaluation-3.5\"><span class=\"toc-item-num\">3.5&nbsp;&nbsp;</span>Model evaluation</a></span><ul class=\"toc-item\"><li><span><a href=\"#Get-predictions\" data-toc-modified-id=\"Get-predictions-3.5.1\"><span class=\"toc-item-num\">3.5.1&nbsp;&nbsp;</span>Get predictions</a></span></li><li><span><a href=\"#Evaluate-test-split\" data-toc-modified-id=\"Evaluate-test-split-3.5.2\"><span class=\"toc-item-num\">3.5.2&nbsp;&nbsp;</span>Evaluate test split</a></span><ul class=\"toc-item\"><li><span><a href=\"#F1-Report\" data-toc-modified-id=\"F1-Report-3.5.2.1\"><span class=\"toc-item-num\">3.5.2.1&nbsp;&nbsp;</span>F1 Report</a></span></li><li><span><a href=\"#Confusion-Matrix\" data-toc-modified-id=\"Confusion-Matrix-3.5.2.2\"><span class=\"toc-item-num\">3.5.2.2&nbsp;&nbsp;</span>Confusion Matrix</a></span></li></ul></li></ul></li></ul></li></ul></div>"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "6e66f60a",
+   "metadata": {
+    "id": "6e66f60a"
+   },
+   "source": [
+    "This notebook is used for the training of our model and its evaluation on the test split. "
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "7af21fa2",
+   "metadata": {
+    "id": "7af21fa2"
+   },
+   "source": [
+    "# Imports & Variables"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "aaafcdc1",
+   "metadata": {
+    "id": "aaafcdc1"
+   },
+   "source": [
+    "## Colab"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "9eb34847",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T15:19:54.049278Z",
+     "start_time": "2022-02-09T15:19:54.041429Z"
+    },
+    "colab": {
+     "base_uri": "https://localhost:8080/"
+    },
+    "id": "9eb34847",
+    "outputId": "98e6dcb8-22ba-454f-c5e4-5b82a5226033"
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Running the notebook on \u001b[34myour machine\u001b[0m\n"
+     ]
+    }
+   ],
+   "source": [
+    "from termcolor import colored\n",
+    "\n",
+    "try:\n",
+    "    from google.colab import drive\n",
+    "    COLAB = True\n",
+    "    print('Running the notebook on',colored('Colab','yellow'))\n",
+    "except:\n",
+    "    COLAB = False\n",
+    "    print('Running the notebook on',colored('your machine','blue'))"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "6443d60f",
+   "metadata": {
+    "id": "6443d60f"
+   },
+   "source": [
+    "## Imports"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "4c0d1bee",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T15:20:05.161147Z",
+     "start_time": "2022-02-09T15:19:55.526591Z"
+    },
+    "colab": {
+     "base_uri": "https://localhost:8080/"
+    },
+    "id": "4c0d1bee",
+    "outputId": "b87ae6e0-1133-427e-c9a5-70eb095ffd25"
+   },
+   "outputs": [],
+   "source": [
+    "## utilities\n",
+    "import os \n",
+    "import pickle\n",
+    "from datetime import datetime\n",
+    "\n",
+    "## classics \n",
+    "import numpy as np \n",
+    "import pandas as pd \n",
+    "import matplotlib.pyplot as plt\n",
+    "import seaborn as sns \n",
+    "\n",
+    "## deep\n",
+    "import tensorflow as tf\n",
+    "\n",
+    "if COLAB:\n",
+    "    !pip install --quiet transformers\n",
+    "\n",
+    "from transformers import AutoTokenizer, TFAutoModel, AutoConfig"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "460a7dec",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T15:20:05.320773Z",
+     "start_time": "2022-02-09T15:20:05.175602Z"
+    },
+    "id": "460a7dec"
+   },
+   "outputs": [],
+   "source": [
+    "#evaluation metrics\n",
+    "from sklearn.metrics import confusion_matrix"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "387e4f08",
+   "metadata": {
+    "id": "387e4f08"
+   },
+   "source": [
+    "## Variables"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "4acaafab",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T15:20:05.337097Z",
+     "start_time": "2022-02-09T15:20:05.333617Z"
+    },
+    "id": "4acaafab"
+   },
+   "outputs": [],
+   "source": [
+    "#Max len of essay \n",
+    "SEQ_LEN = 1024 ## THIS SHOULD NOT BE CHANGED without appropriate changes in the preprocessing \n",
+    "\n",
+    "#Train, val, test split proportion\n",
+    "VAL_SPLIT = 0.8\n",
+    "TEST_SPLIT = 0.9\n",
+    "\n",
+    "#Batch size\n",
+    "BATCH_SIZE = 16\n",
+    "\n",
+    "#Data version\n",
+    "VERSION=3\n",
+    "\n",
+    "# Load weights of trained model\n",
+    "MODEL_NAME = 'mymodel'\n",
+    "LOAD_MODEL = True\n",
+    "LOAD_BACKBONE_FROM = '/content/drive/MyDrive/feedback-prize/backbone'\n",
+    "LOAD_MODEL_WEIGHTS_FROM = '/content/drive/MyDrive/feedback-prize/mymodel/mymodel'"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "7efa402a",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T15:20:11.117941Z",
+     "start_time": "2022-02-09T15:20:11.114592Z"
+    },
+    "colab": {
+     "base_uri": "https://localhost:8080/"
+    },
+    "id": "7efa402a",
+    "outputId": "d4aeaada-14d5-4089-a0fb-a959ffc80b02"
+   },
+   "outputs": [],
+   "source": [
+    "## Paths\n",
+    "\n",
+    "## if running in colab, mount drive\n",
+    "if COLAB:\n",
+    "    drive.mount('/content/drive')\n",
+    "    PATH='/content/drive/MyDrive/feedback-prize/'\n",
+    "else:\n",
+    "    PATH='../'"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "178d6319",
+   "metadata": {
+    "id": "178d6319"
+   },
+   "source": [
+    "# Load data\n",
+    "\n",
+    "The data is already preprocessed in another notebook.\n",
+    "\n",
+    "The preprocessed data is loaded and splitted in `train`, `val`, `test`"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "id": "98e9bc4d",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T15:20:21.289653Z",
+     "start_time": "2022-02-09T15:20:18.512222Z"
+    },
+    "id": "98e9bc4d"
+   },
+   "outputs": [],
+   "source": [
+    "#Load train and test data\n",
+    "with open(os.path.join(PATH,'raw_data',f'dataset_v{VERSION}.pickle'),'rb') as file:\n",
+    "    dataset = pickle.load(file)\n",
+    "    \n",
+    "with open(os.path.join(PATH,'raw_data',f'dataset_test_v{VERSION}.pickle'),'rb') as file:\n",
+    "    dataset_inf = pickle.load(file)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "id": "c01c00ca",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T15:20:24.039041Z",
+     "start_time": "2022-02-09T15:20:22.704665Z"
+    },
+    "id": "c01c00ca"
+   },
+   "outputs": [],
+   "source": [
+    "#Load preprocessed.csv file as it will be needed to retrieve predictions\n",
+    "df_essays = pd.read_csv(os.path.join(PATH,'raw_data',f'preprocessed_v{VERSION}.csv'))\n",
+    "df_inf = pd.read_csv(os.path.join(PATH,'raw_data',f'preprocessed_inf_v{VERSION}.csv'),index_col=0)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "id": "06633885",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T15:20:24.658232Z",
+     "start_time": "2022-02-09T15:20:24.653276Z"
+    },
+    "id": "06633885"
+   },
+   "outputs": [],
+   "source": [
+    "## Creating splits indexes\n",
+    "\n",
+    "LEN=len(dataset['labels'])\n",
+    "\n",
+    "idx_val=int(LEN*VAL_SPLIT)\n",
+    "idx_test=int(LEN*TEST_SPLIT)\n",
+    "\n",
+    "idx_train=list(range(0,idx_val))\n",
+    "idx_val=list(range(idx_val,idx_test))\n",
+    "idx_test=list(range(idx_test,LEN))\n",
+    "\n",
+    "assert(len(idx_test)+len(idx_train)+len(idx_val)==LEN)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "id": "d5694515",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T15:20:29.000487Z",
+     "start_time": "2022-02-09T15:20:25.580342Z"
+    },
+    "id": "d5694515"
+   },
+   "outputs": [],
+   "source": [
+    "## Splitting dataset\n",
+    "\n",
+    "#train\n",
+    "X_train = {\n",
+    "    'input_ids' : dataset['inputs']['input_ids'][idx_train],\n",
+    "    'attention_mask' : dataset['inputs']['attention_mask'][idx_train]\n",
+    "}\n",
+    "\n",
+    "y_train = dataset['labels'][idx_train]\n",
+    "ps_train = dataset['predictionstrings'][idx_train]\n",
+    "\n",
+    "#val\n",
+    "X_val = {\n",
+    "    'input_ids' : dataset['inputs']['input_ids'][idx_val],\n",
+    "    'attention_mask' : dataset['inputs']['attention_mask'][idx_val]\n",
+    "}\n",
+    "\n",
+    "y_val = dataset['labels'][idx_val]\n",
+    "ps_val = dataset['predictionstrings'][idx_val]\n",
+    "\n",
+    "\n",
+    "#test\n",
+    "X_test = {\n",
+    "    'input_ids' : dataset['inputs']['input_ids'][idx_test],\n",
+    "    'attention_mask' : dataset['inputs']['attention_mask'][idx_test]\n",
+    "}\n",
+    "\n",
+    "y_test = dataset['labels'][idx_test]\n",
+    "ps_test = dataset['predictionstrings'][idx_test]\n",
+    "\n",
+    "#inference\n",
+    "X_inf = dataset_inf['inputs']\n",
+    "ps_inf = dataset_inf['predictionstrings']"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "06895f0e",
+   "metadata": {
+    "id": "06895f0e"
+   },
+   "source": [
+    "# Modeling"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "ce61fc80",
+   "metadata": {
+    "id": "ce61fc80"
+   },
+   "source": [
+    "## Model architecture"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 14,
+   "id": "dbf3d6ae",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-08T15:05:09.982796Z",
+     "start_time": "2022-02-08T15:04:51.874202Z"
+    },
+    "colab": {
+     "base_uri": "https://localhost:8080/"
+    },
+    "id": "dbf3d6ae",
+    "outputId": "0ba8c2c1-1edd-4c6c-d10a-e17735c0811d",
+    "scrolled": true
+   },
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "All model checkpoint layers were used when initializing TFLongformerModel.\n",
+      "\n",
+      "All the layers of TFLongformerModel were initialized from the model checkpoint at /content/drive/MyDrive/feedback-prize/backbone/tf_model.h5.\n",
+      "If your task is similar to the task the model of the checkpoint was trained on, you can already use TFLongformerModel for predictions without further training.\n"
+     ]
+    }
+   ],
+   "source": [
+    "## Instantiate model Longformer to be used as backbone\n",
+    "config = AutoConfig.from_pretrained(os.path.join(LOAD_BACKBONE_FROM,'config.json'))\n",
+    "backbone = TFAutoModel.from_pretrained(os.path.join(LOAD_BACKBONE_FROM,'tf_model.h5'),config=config)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 15,
+   "id": "4b31d749",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-08T15:05:19.028832Z",
+     "start_time": "2022-02-08T15:05:19.014958Z"
+    },
+    "id": "4b31d749"
+   },
+   "outputs": [],
+   "source": [
+    "## TODO : retrieve the attention mask from backbone and pass it to the two LSTM ; currently highly possible the attention mask got lost\n",
+    "\n",
+    "# init model\n",
+    "\n",
+    "def init_model():\n",
+    "    input_ids = tf.keras.layers.Input(shape=(SEQ_LEN,),dtype='int32')\n",
+    "    attention_mask = tf.keras.layers.Input(shape=(SEQ_LEN,),dtype='int32')\n",
+    "    \n",
+    "    x = backbone({'input_ids':input_ids,\n",
+    "                 'attention_mask':attention_mask})[0]\n",
+    "\n",
+    "    backbone.trainable = False\n",
+    "\n",
+    "\n",
+    "    x = tf.keras.layers.Bidirectional(tf.keras.layers.LSTM(units = 512,\n",
+    "                                                           activation = 'tanh',\n",
+    "                                                           dropout=.2,\n",
+    "                                                           return_sequences=True))(x)\n",
+    "    x_res = tf.keras.layers.Bidirectional(tf.keras.layers.LSTM(units = 512,\n",
+    "                                                           activation = 'tanh',\n",
+    "                                                           dropout=.2,\n",
+    "                                                           return_sequences=True))(x)\n",
+    "    \n",
+    "\n",
+    "    x = tf.keras.layers.add([x,x_res])\n",
+    "    output = tf.keras.layers.TimeDistributed(tf.keras.layers.Dense(16,activation = 'softmax'))(x)\n",
+    "\n",
+    "    model = tf.keras.models.Model(inputs={'input_ids':input_ids,\n",
+    "                                          'attention_mask':attention_mask},outputs=output)\n",
+    "    \n",
+    "    \n",
+    "    return model"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 16,
+   "id": "805fd880",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-08T15:08:40.730243Z",
+     "start_time": "2022-02-08T15:08:32.179279Z"
+    },
+    "colab": {
+     "base_uri": "https://localhost:8080/",
+     "height": 360
+    },
+    "id": "805fd880",
+    "outputId": "c4bd4814-003e-4232-876f-94e18ef4a459"
+   },
+   "outputs": [
+    {
+     "data": {
+      "image/png": "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\n",
+      "text/plain": [
+       "<IPython.core.display.Image object>"
+      ]
+     },
+     "execution_count": 16,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "## instantiate the model, plot the graph\n",
+    "model = init_model()\n",
+    "\n",
+    "tf.keras.utils.plot_model(model,show_shapes=True)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "ba134f3f",
+   "metadata": {
+    "id": "ba134f3f"
+   },
+   "source": [
+    "## Loss, Optimizer, Metrics"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 17,
+   "id": "2e741083",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-08T15:19:20.793041Z",
+     "start_time": "2022-02-08T15:19:20.726503Z"
+    },
+    "id": "2e741083"
+   },
+   "outputs": [],
+   "source": [
+    "# creating homemade metric\n",
+    "\n",
+    "def accuracy_masked_func(y_true,y_pred):\n",
+    "    y_pred = tf.cast(tf.argmax(y_pred,axis=-1),'int32')\n",
+    "    y_true = tf.cast(y_true,'int32')\n",
+    "    y_true = tf.cast(tf.argmax(y_true,axis=-1),'int32') #for y_pred and y_true to match\n",
+    "    mask = tf.cast(y_true != 15,'int32') #create a mask\n",
+    "    matches = tf.cast(tf.equal(y_true,y_pred),'int32')*mask #calculate the matches ignoring the masking\n",
+    "    accuracy = tf.math.reduce_sum(matches,axis=-1)/tf.maximum(tf.math.reduce_sum(mask,axis=-1),1)\n",
+    "    \n",
+    "    return accuracy"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 18,
+   "id": "39f32227",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-08T15:20:32.714566Z",
+     "start_time": "2022-02-08T15:20:32.434987Z"
+    },
+    "id": "39f32227"
+   },
+   "outputs": [],
+   "source": [
+    "# define loss and metrics \n",
+    "loss = tf.keras.losses.CategoricalCrossentropy(name='categorical_crossentropy')\n",
+    "cat_accuracy = tf.keras.metrics.CategoricalAccuracy()\n",
+    "masked_accuracy = tf.keras.metrics.MeanMetricWrapper(fn=accuracy_masked_func)\n",
+    "\n",
+    "# RMSProp optimizer with clip value and small lr to avoid exploiding gradient \n",
+    "opt = tf.keras.optimizers.RMSprop(clipvalue=.5,learning_rate=0.0001)\n",
+    "\n",
+    "#compile\n",
+    "model.compile(optimizer=opt,loss=loss,metrics=[cat_accuracy,masked_accuracy])"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "af59f074",
+   "metadata": {
+    "id": "af59f074"
+   },
+   "source": [
+    "## Model training"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 21,
+   "id": "8de8c166",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-08T15:35:00.705395Z",
+     "start_time": "2022-02-08T15:35:00.630108Z"
+    },
+    "id": "8de8c166"
+   },
+   "outputs": [],
+   "source": [
+    "#### CALLBACKS\n",
+    "\n",
+    "timestamp = datetime.today().__format__('%d%m_%Hh%M')\n",
+    "\n",
+    "checkpoints_path = f'{PATH}{MODEL_NAME}/{MODEL_NAME}_{timestamp}.ckpt'\n",
+    "logdir = '/content/drive/MyDrive/feedback-prize/logs/'\n",
+    "\n",
+    "\n",
+    "#early stopping\n",
+    "es = tf.keras.callbacks.EarlyStopping(patience=1,restore_best_weights=True)\n",
+    "\n",
+    "#save weights at every epoch\n",
+    "checkpoint_saver = tf.keras.callbacks.ModelCheckpoint(filepath=checkpoints_path,\n",
+    "                                                      save_weights_only=True,\n",
+    "                                                      save_best_only = True,\n",
+    "                                                      monitor = 'val_categorical_accuracy',\n",
+    "                                                      mode = 'max',\n",
+    "                                                      verbose = 1)\n",
+    "\n",
+    "#logs for tensorboard\n",
+    "tensorboard = tf.keras.callbacks.TensorBoard(log_dir=logdir)\n",
+    "\n",
+    "# list callbacks\n",
+    "\n",
+    "callbacks=[es,checkpoint_saver,tensorboard]"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 22,
+   "id": "d14a118f",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-08T15:36:39.914302Z",
+     "start_time": "2022-02-08T15:36:39.910221Z"
+    },
+    "colab": {
+     "base_uri": "https://localhost:8080/"
+    },
+    "id": "d14a118f",
+    "outputId": "85f537a3-c0a1-45cd-9946-b0f60a430f3d"
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Loading model weights below\n"
+     ]
+    }
+   ],
+   "source": [
+    "## TRAINING\n",
+    "if not LOAD_MODEL:\n",
+    "    history = model.fit(X_train,y_train,\n",
+    "                        validation_data= (X_val,y_val),\n",
+    "                        epochs=30,callbacks=callbacks,batch_size=BATCH_SIZE)\n",
+    "else:\n",
+    "    print('Loading model weights below')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "da39c25d",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "## saving\n",
+    "if not LOAD_MODEL:\n",
+    "    os.mkdir(MODEL_NAME)\n",
+    "    model.save_weights(MODEL_NAME)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "5c9b34e7",
+   "metadata": {
+    "id": "5c9b34e7"
+   },
+   "source": [
+    "## Model evaluation"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 27,
+   "id": "e4b0466a",
+   "metadata": {
+    "colab": {
+     "base_uri": "https://localhost:8080/"
+    },
+    "id": "e4b0466a",
+    "outputId": "ee35de22-350b-4c42-b124-69e1246ed7d4",
+    "scrolled": true
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "98/98 [==============================] - 193s 2s/step - loss: 0.3133 - categorical_accuracy: 0.8977 - mean_metric_wrapper: 0.7678\n"
+     ]
+    },
+    {
+     "data": {
+      "text/plain": [
+       "[0.3132975399494171, 0.8977440595626831, 0.7678290605545044]"
+      ]
+     },
+     "execution_count": 27,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "## model eval on val set \n",
+    "model.evaluate(X_val,y_val,batch_size=BATCH_SIZE)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "9153f4d8",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "#predict on test split\n",
+    "y_pred = model.predict(X_test,batch_size=BATCH_SIZE)\n",
+    "\n",
+    "#dump pickle\n",
+    "with open(os.path.join(PATH,'raw_data','preds_on_testsplit.pickle'),'wb') as file:\n",
+    "    pickle.dump(y_pred,file)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "b2bb7321",
+   "metadata": {},
+   "source": [
+    "> Evaluation on a separate notebook"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "9a0d1923",
+   "metadata": {
+    "heading_collapsed": true,
+    "id": "9a0d1923"
+   },
+   "source": [
+    "## Model evaluation\n",
+    "\n",
+    "<font size=6 color='red'>No longer accurate</font>"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "fb0aa2f6",
+   "metadata": {
+    "hidden": true,
+    "id": "fb0aa2f6"
+   },
+   "source": [
+    "### Get predictions"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "id": "9LPEEQCzC3LI",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T15:20:45.557610Z",
+     "start_time": "2022-02-09T15:20:45.532672Z"
+    },
+    "hidden": true,
+    "id": "9LPEEQCzC3LI"
+   },
+   "outputs": [],
+   "source": [
+    "labels_mapping = {'B-Lead' : 0,\n",
+    "                  'B-Position' : 1,\n",
+    "                  'B-Evidence' : 2,\n",
+    "                  'B-Claim' : 3,\n",
+    "                  'B-Concluding_Statement' : 4,\n",
+    "                  'B-Counterclaim' : 5,\n",
+    "                  'B-Rebuttal' : 6,\n",
+    "                  'I-Lead' : 7,\n",
+    "                  'I-Position' : 8,\n",
+    "                  'I-Evidence' : 9,\n",
+    "                  'I-Claim' : 10,\n",
+    "                  'I-Concluding_Statement' : 11,\n",
+    "                  'I-Counterclaim' : 12,\n",
+    "                  'I-Rebuttal': 13,\n",
+    "                 'O':14,\n",
+    "                 'PAD':15}\n",
+    "\n",
+    "reversed_mapping = {v:(k[2:] if v<14 else k) for k,v in labels_mapping.items()}"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "id": "4586e676",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T15:20:45.632079Z",
+     "start_time": "2022-02-09T15:20:45.600034Z"
+    },
+    "hidden": true,
+    "id": "4586e676"
+   },
+   "outputs": [],
+   "source": [
+    "def get_preds(y_pred,ps):\n",
+    "    \"\"\"\n",
+    "    Generate readable predictions from the output of the model.\n",
+    "\n",
+    "    Args:\n",
+    "        y_pred (ndarray): output of the model\n",
+    "        ps (ndarray): predictionstring referring to the token predicted\n",
+    "\n",
+    "    Returns:\n",
+    "        DataFrame : DataFrame with class and predictionstrings\n",
+    "    \"\"\"\n",
+    "\n",
+    "\n",
+    "    labels = []\n",
+    "    predictionstrings = []\n",
+    "    counts = []\n",
+    "    \n",
+    "    counter=dict()\n",
+    "    \n",
+    "    for tok,pos in zip(y_pred,ps):\n",
+    "        \n",
+    "        if tok <= 13:\n",
+    "            lab = reversed_mapping[tok]\n",
+    "            labels.append(lab)\n",
+    "            predictionstrings.append(pos)\n",
+    "            if len(labels)<2:\n",
+    "                counts.append(str(1))\n",
+    "                counter.setdefault(lab,1)\n",
+    "                continue\n",
+    "            if lab == labels[-2]:\n",
+    "                counts.append(str(counter[lab]))\n",
+    "            else: \n",
+    "                try:\n",
+    "                    counter[lab]+=1\n",
+    "                except KeyError:\n",
+    "                    counter.setdefault(lab,1)\n",
+    "                counts.append(str(counter[lab]))\n",
+    "    \n",
+    "    preds = pd.DataFrame([labels,counts,predictionstrings],index=['class','count','predictionstring']).T\n",
+    "    preds['class'] += ' ' + preds['count'].astype(str)\n",
+    "    preds = preds.groupby('class',sort=False).agg({'predictionstring':list}).reset_index()\n",
+    "    preds['class']=preds['class'].apply(lambda txt : txt.split()[0])\n",
+    "    preds['predictionstring']=preds['predictionstring'].apply(lambda l_ : [str(l) for l in l_])\n",
+    "    preds['predictionstring']=preds['predictionstring'].apply(lambda l_ : ' '.join(l_))\n",
+    "    \n",
+    "    return preds"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "dbfbd22c",
+   "metadata": {
+    "hidden": true,
+    "id": "dbfbd22c"
+   },
+   "source": [
+    "### Evaluate test split"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 30,
+   "id": "74d63b23",
+   "metadata": {
+    "hidden": true,
+    "id": "74d63b23"
+   },
+   "outputs": [],
+   "source": [
+    "#predict on test split\n",
+    "y_pred = model.predict(X_test,batch_size=BATCH_SIZE)\n",
+    "\n",
+    "#dump pickle\n",
+    "with open(os.path.join(PATH,'raw_data','preds_on_testsplit.pickle'),'wb') as file:\n",
+    "    pickle.dump(y_pred,file)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "234d1115",
+   "metadata": {
+    "hidden": true,
+    "id": "234d1115"
+   },
+   "source": [
+    "#### F1 Report"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 16,
+   "id": "87454dba",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T15:23:57.539068Z",
+     "start_time": "2022-02-09T15:23:08.212220Z"
+    },
+    "colab": {
+     "base_uri": "https://localhost:8080/",
+     "height": 381
+    },
+    "hidden": true,
+    "id": "87454dba",
+    "outputId": "e76feccd-2868-47ab-eb6b-a2936fcaa7c5"
+   },
+   "outputs": [],
+   "source": [
+    "## Create two DF with preds and ground truth\n",
+    "\n",
+    "y_true = np.argmax(y_test,axis=-1)\n",
+    "y_pred = np.argmax(y_pred_,axis=-1)\n",
+    "\n",
+    "ps = ps_test\n",
+    "\n",
+    "true_df = pd.DataFrame()\n",
+    "pred_df = pd.DataFrame()\n",
+    "\n",
+    "for i,idx in enumerate(idx_test): ## CHANGE idx_test\n",
+    "    \n",
+    "    true_ = get_preds(y_true[i],ps[i])\n",
+    "    pred_ = get_preds(y_pred[i],ps[i])\n",
+    "    \n",
+    "    true_['id']=df_essays.iloc[idx]['id']\n",
+    "    pred_['id']=df_essays.iloc[idx]['id']\n",
+    "    \n",
+    "    true_df = true_df.append(true_)\n",
+    "    pred_df = pred_df.append(pred_)\n",
+    "    \n",
+    "true_df['unique_id'] = pd.util.hash_pandas_object(true_df,hash_key='1234567890123456')\n",
+    "pred_df['unique_id'] = pd.util.hash_pandas_object(pred_df,hash_key='azerty1234567890')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 18,
+   "id": "d6d45514",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T15:24:11.107286Z",
+     "start_time": "2022-02-09T15:24:11.101012Z"
+    },
+    "hidden": true,
+    "id": "d6d45514"
+   },
+   "outputs": [],
+   "source": [
+    "## Determine whether a prediction is a true positive or not\n",
+    "\n",
+    "def true_positive(predictionstring_true,predictionstring_pred):\n",
+    "    ps_true = set(predictionstring_true.split(' '))\n",
+    "    ps_pred = set(predictionstring_pred.split(' '))\n",
+    "    \n",
+    "    inter = ps_pred.intersection(ps_true)\n",
+    "    overlap_1 = len(inter)/len(ps_true)\n",
+    "    overlap_2 = len(inter)/len(ps_pred)\n",
+    "    \n",
+    "    if overlap_1 >= .5 and overlap_2 >= .5:\n",
+    "        return 1\n",
+    "    return 0\n",
+    "\n",
+    "## vectorize the funct\n",
+    "true_positive_vect = np.vectorize(true_positive)\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 19,
+   "id": "e811bd53",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T15:24:13.910117Z",
+     "start_time": "2022-02-09T15:24:13.487097Z"
+    },
+    "hidden": true,
+    "id": "e811bd53"
+   },
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>class</th>\n",
+       "      <th>predictionstring_pred</th>\n",
+       "      <th>id</th>\n",
+       "      <th>unique_id_pred</th>\n",
+       "      <th>predictionstring_true</th>\n",
+       "      <th>unique_id_true</th>\n",
+       "      <th>FP</th>\n",
+       "      <th>FN</th>\n",
+       "      <th>TP</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>Lead</td>\n",
+       "      <td>0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18...</td>\n",
+       "      <td>E6870101D8EE</td>\n",
+       "      <td>1.172241e+19</td>\n",
+       "      <td></td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>1</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>Lead</td>\n",
+       "      <td>65 66 67 68 69 70 71 72 73 74 75 76 77 78</td>\n",
+       "      <td>E6870101D8EE</td>\n",
+       "      <td>9.178975e+18</td>\n",
+       "      <td></td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>1</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>Lead</td>\n",
+       "      <td>83 84</td>\n",
+       "      <td>E6870101D8EE</td>\n",
+       "      <td>2.577067e+18</td>\n",
+       "      <td></td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>1</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>Evidence</td>\n",
+       "      <td>31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 4...</td>\n",
+       "      <td>E6870101D8EE</td>\n",
+       "      <td>5.320668e+18</td>\n",
+       "      <td>0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18...</td>\n",
+       "      <td>1.447722e+19</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>Evidence</td>\n",
+       "      <td>31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 4...</td>\n",
+       "      <td>E6870101D8EE</td>\n",
+       "      <td>5.320668e+18</td>\n",
+       "      <td>110 111 112 113 114 115 116 117 118 119 120 12...</td>\n",
+       "      <td>1.562738e+19</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "      class                              predictionstring_pred            id  \\\n",
+       "0      Lead  0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18...  E6870101D8EE   \n",
+       "1      Lead          65 66 67 68 69 70 71 72 73 74 75 76 77 78  E6870101D8EE   \n",
+       "2      Lead                                              83 84  E6870101D8EE   \n",
+       "3  Evidence  31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 4...  E6870101D8EE   \n",
+       "4  Evidence  31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 4...  E6870101D8EE   \n",
+       "\n",
+       "   unique_id_pred                              predictionstring_true  \\\n",
+       "0    1.172241e+19                                                      \n",
+       "1    9.178975e+18                                                      \n",
+       "2    2.577067e+18                                                      \n",
+       "3    5.320668e+18  0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18...   \n",
+       "4    5.320668e+18  110 111 112 113 114 115 116 117 118 119 120 12...   \n",
+       "\n",
+       "   unique_id_true  FP  FN  TP  \n",
+       "0             NaN   1   0   0  \n",
+       "1             NaN   1   0   0  \n",
+       "2             NaN   1   0   0  \n",
+       "3    1.447722e+19   0   0   0  \n",
+       "4    1.562738e+19   0   0   0  "
+      ]
+     },
+     "execution_count": 19,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "## Merge the 2 DF to compute F1 \n",
+    "\n",
+    "merged_df = pred_df.merge(true_df,how = 'outer',on=['id','class'],suffixes=('_pred','_true'))\n",
+    "\n",
+    "## if a pred is not matched it is a FN ; if the truth is not matched it is a FP\n",
+    "\n",
+    "#creating separate columns \n",
+    "merged_df['FP'] = np.where(merged_df.predictionstring_true.isna(), 1, 0)\n",
+    "merged_df['FN'] = np.where(merged_df.predictionstring_pred.isna(), 1, 0)\n",
+    "\n",
+    "\n",
+    "#cleaning nan for the true positive function\n",
+    "merged_df['predictionstring_pred'].fillna('',inplace=True)\n",
+    "merged_df['predictionstring_true'].fillna('',inplace=True)\n",
+    "\n",
+    "\n",
+    "merged_df['TP'] = true_positive_vect(merged_df['predictionstring_true'],merged_df['predictionstring_pred'])\n",
+    "\n",
+    "merged_df.head()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 20,
+   "id": "b518be6c",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T15:24:16.473582Z",
+     "start_time": "2022-02-09T15:24:16.470122Z"
+    },
+    "hidden": true,
+    "id": "b518be6c"
+   },
+   "outputs": [],
+   "source": [
+    "## creating f1 function\n",
+    "def f1_score(fp,fn,tp):\n",
+    "    return tp/(tp+.5*(fp+fn))*100"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 40,
+   "id": "62ad9adb",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T15:26:19.380276Z",
+     "start_time": "2022-02-09T15:26:19.352593Z"
+    },
+    "hidden": true,
+    "id": "62ad9adb"
+   },
+   "outputs": [],
+   "source": [
+    "## group by class for F1 score calculation\n",
+    "f1_df = merged_df.groupby('class').sum()\n",
+    "f1_df.drop(['unique_id_pred','unique_id_true'],axis=1,inplace=True)\n",
+    "\n",
+    "#apply to the df\n",
+    "f1_df['f1']=f1_score(f1_df.FP,f1_df.FN,f1_df.TP)\n",
+    "\n",
+    "#Create a total row\n",
+    "f1_df.loc['Total']=f1_df.mean()\n",
+    "\n",
+    "## weighted average\n",
+    "\n",
+    "f1_df['support'] = true_df.groupby('class').count()['id']\n",
+    "\n",
+    "f1_df['f1_weighted']=f1_df['f1']*(f1_df['support']/f1_df['support'].sum())\n",
+    "f1_df.loc['Total','support']=f1_df.support.sum()\n",
+    "f1_df.loc['Total','f1_weighted']=f1_df.f1_weighted.sum()\n",
+    "\n",
+    "# impute correct values for the Total row for TP FN FP \n",
+    "\n",
+    "f1_df.loc['Total','FP'] = f1_df.loc[:'Rebuttal','FP'].sum()\n",
+    "f1_df.loc['Total','FN'] = f1_df.loc[:'Rebuttal','FN'].sum()\n",
+    "f1_df.loc['Total','TP'] = f1_df.loc[:'Rebuttal','TP'].sum()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 41,
+   "id": "d296aa49",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T15:26:19.534016Z",
+     "start_time": "2022-02-09T15:26:19.530003Z"
+    },
+    "hidden": true,
+    "id": "d296aa49"
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "F1 Macro Score = 76.36%\n",
+      "F1 Micro Score = 90.47%\n",
+      "F1 Weighted Score = 90.70%\n"
+     ]
+    }
+   ],
+   "source": [
+    "## PRINT REPORT\n",
+    "\n",
+    "print(f\"F1 Macro Score = {f1_df.loc['Total','f1']:.2f}%\")\n",
+    "\n",
+    "\n",
+    "print(f\"F1 Micro Score = {f1_score(f1_df.loc['Total','FP'],f1_df.loc['Total','FN'],f1_df.loc['Total','TP']):.2f}%\")\n",
+    "\n",
+    "print(f\"F1 Weighted Score = {f1_df.loc['Total','f1_weighted']:.2f}%\")\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 42,
+   "id": "e3cd3b9d",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T15:26:19.827938Z",
+     "start_time": "2022-02-09T15:26:19.681316Z"
+    },
+    "hidden": true,
+    "id": "e3cd3b9d"
+   },
+   "outputs": [
+    {
+     "data": {
+      "image/png": "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\n",
+      "text/plain": [
+       "<Figure size 432x288 with 1 Axes>"
+      ]
+     },
+     "metadata": {
+      "needs_background": "light"
+     },
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "sns.barplot(y=f1_df.reset_index().loc[:6,'f1'],x=f1_df.index[:-1],palette='Set2')\n",
+    "plt.xticks(rotation=90)\n",
+    "locs,_=plt.xticks()\n",
+    "plt.plot([locs[0]-.5,locs[-1]+.5],[f1_df.loc['Total','f1'],f1_df.loc['Total','f1']],c='r');"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 43,
+   "id": "ee4dd298",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T15:26:20.595497Z",
+     "start_time": "2022-02-09T15:26:20.579822Z"
+    },
+    "hidden": true,
+    "id": "ee4dd298"
+   },
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>FP</th>\n",
+       "      <th>FN</th>\n",
+       "      <th>TP</th>\n",
+       "      <th>f1</th>\n",
+       "      <th>support</th>\n",
+       "      <th>f1_weighted</th>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>class</th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>Claim</th>\n",
+       "      <td>79</td>\n",
+       "      <td>49</td>\n",
+       "      <td>1570</td>\n",
+       "      <td>96.08%</td>\n",
+       "      <td>3412</td>\n",
+       "      <td>27.34%</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Concluding_Statement</th>\n",
+       "      <td>214</td>\n",
+       "      <td>27</td>\n",
+       "      <td>1106</td>\n",
+       "      <td>90.18%</td>\n",
+       "      <td>1346</td>\n",
+       "      <td>10.12%</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Counterclaim</th>\n",
+       "      <td>162</td>\n",
+       "      <td>202</td>\n",
+       "      <td>141</td>\n",
+       "      <td>43.65%</td>\n",
+       "      <td>542</td>\n",
+       "      <td>1.97%</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Evidence</th>\n",
+       "      <td>25</td>\n",
+       "      <td>0</td>\n",
+       "      <td>2400</td>\n",
+       "      <td>99.48%</td>\n",
+       "      <td>3792</td>\n",
+       "      <td>31.46%</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Lead</th>\n",
+       "      <td>249</td>\n",
+       "      <td>50</td>\n",
+       "      <td>774</td>\n",
+       "      <td>83.81%</td>\n",
+       "      <td>963</td>\n",
+       "      <td>6.73%</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Position</th>\n",
+       "      <td>23</td>\n",
+       "      <td>68</td>\n",
+       "      <td>955</td>\n",
+       "      <td>95.45%</td>\n",
+       "      <td>1535</td>\n",
+       "      <td>12.22%</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Rebuttal</th>\n",
+       "      <td>128</td>\n",
+       "      <td>199</td>\n",
+       "      <td>57</td>\n",
+       "      <td>25.85%</td>\n",
+       "      <td>402</td>\n",
+       "      <td>0.87%</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Total</th>\n",
+       "      <td>880</td>\n",
+       "      <td>595</td>\n",
+       "      <td>7003</td>\n",
+       "      <td>76.36%</td>\n",
+       "      <td>11992</td>\n",
+       "      <td>90.70%</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                       FP   FN    TP      f1 support f1_weighted\n",
+       "class                                                           \n",
+       "Claim                  79   49  1570  96.08%    3412      27.34%\n",
+       "Concluding_Statement  214   27  1106  90.18%    1346      10.12%\n",
+       "Counterclaim          162  202   141  43.65%     542       1.97%\n",
+       "Evidence               25    0  2400  99.48%    3792      31.46%\n",
+       "Lead                  249   50   774  83.81%     963       6.73%\n",
+       "Position               23   68   955  95.45%    1535      12.22%\n",
+       "Rebuttal              128  199    57  25.85%     402       0.87%\n",
+       "Total                 880  595  7003  76.36%   11992      90.70%"
+      ]
+     },
+     "execution_count": 43,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# Full report\n",
+    "f1_df[['FP','FN','TP','support']]=f1_df[['FP','FN','TP','support']].applymap('{:.0f}'.format)\n",
+    "f1_df['f1_weighted']=f1_df['f1_weighted'].map('{:.2f}%'.format)\n",
+    "f1_df['f1']=f1_df['f1'].map('{:.2f}%'.format)\n",
+    "\n",
+    "f1_df"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "f0514f5e",
+   "metadata": {
+    "hidden": true,
+    "id": "f0514f5e"
+   },
+   "source": [
+    "#### Confusion Matrix"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 44,
+   "id": "c3a81d7f",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T15:26:26.334535Z",
+     "start_time": "2022-02-09T15:26:26.322643Z"
+    },
+    "hidden": true,
+    "id": "c3a81d7f"
+   },
+   "outputs": [],
+   "source": [
+    "def show_confusion_matrix(y_true,y_pred):\n",
+    "\n",
+    "    reversed_mapping = {\n",
+    "                        0: 'Lead',\n",
+    "                        1: 'Position',\n",
+    "                        2: 'Evidence',\n",
+    "                        3: 'Claim',\n",
+    "                        4: 'Concluding_Statement',\n",
+    "                        5: 'Counterclaim',\n",
+    "                        6: 'Rebuttal',\n",
+    "                        7: 'Lead',\n",
+    "                        8: 'Position',\n",
+    "                        9: 'Evidence',\n",
+    "                        10: 'Claim',\n",
+    "                        11: 'Concluding_Statement',\n",
+    "                        12: 'Counterclaim',\n",
+    "                        13: 'Rebuttal',\n",
+    "                        14: 'O',\n",
+    "                        15: 'PAD'}\n",
+    "    \n",
+    "    y_true_flat = [reversed_mapping[y] for y in y_true.flatten()]\n",
+    "    y_pred_flat = [reversed_mapping[y] for y in y_pred.flatten()]\n",
+    "    \n",
+    "    LABELS = ['Lead','Position','Claim','Counterclaim','Rebuttal','Evidence','Concluding_Statement','O','PAD']\n",
+    "\n",
+    "    cfn = confusion_matrix(y_true_flat,y_pred_flat,labels=LABELS)\n",
+    "    \n",
+    "    fig,ax = plt.subplots(1,1,figsize=(10,10))\n",
+    "    plt.title('Confusion Matrix',size=18,pad=20)\n",
+    "    sns.heatmap(cfn/np.sum(cfn,axis=0)*100,cmap='Blues',annot = True,fmt='.2f',annot_kws={'size':10},ax=ax);\n",
+    "    plt.xticks(np.arange(len(LABELS))+.5,LABELS,rotation = 90,size=12);\n",
+    "    plt.yticks(np.arange(len(LABELS))+.5,LABELS,rotation = 0,size=12);\n",
+    "    plt.xlabel('PREDICTED',size=16);\n",
+    "    plt.ylabel('ACTUAL',size=16);\n",
+    "    for t in ax.texts: t.set_text(t.get_text() + \" %\")\n",
+    "    \n",
+    "    plt.show()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 45,
+   "id": "80189c2f",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-09T15:26:32.922434Z",
+     "start_time": "2022-02-09T15:26:27.324606Z"
+    },
+    "hidden": true,
+    "id": "80189c2f"
+   },
+   "outputs": [
+    {
+     "data": {
+      "image/png": "iVBORw0KGgoAAAANSUhEUgAAArkAAALyCAYAAADewcwbAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/YYfK9AAAACXBIWXMAAAsTAAALEwEAmpwYAAEAAElEQVR4nOzdd1wURxvA8d/AgaJgFzSR2LtGTTSx964I2Hsvsaa8iS3GFhONJWrsvTfsvfcWTWJi7xVUQGNDpR3M+8cdBLxDQRAVnq8fPsLuzO48t+Xmnp3dU1prhBBCCCGESEps3nYDhBBCCCGESGjSyRVCCCGEEEmOdHKFEEIIIUSSI51cIYQQQgiR5EgnVwghhBBCJDnSyRVCCCGEEEmOdHKFEOItU0oVV0rtVko9VEpppdTQN7Se9ublV34Ty09KzK/T/LfdDiHE65NOrhAi2VJKpVJKfaWUOqiUeqCUClVK+Smltpg7hIZEaIMBWA3kBX4A2gBr3vR63xalVA5zB1IrpTbFUMZOKXXPXOZGPNbl8aY+MAgh3n1KvgxCCJEcKaXyAJuBfMAuYAdwH3AGqpt/xmit+77hduQDLgL/01r/+obXZQvYASFa6/A3ua6XtCEHcB0IMrfFVWt994UyjYBV5jJ+Wuscr7mu+UA7rbV6jbopgTCtdejrrFsI8fa98SyFEEK8a5RSDsAmIBfQSGv9Yub0F6VUKaBUIjQni/n/B296RVrrMCDsTa8nljYBHpgy16NfmNcROAXYAo6J1SDzfhGqtTZqrYMSa71CiDdDhisIIZKjzkB+YJyVDi4AWus/tNZTo04zX/4+rJR6ppR6av7d/cW6SqkbSql9SqkCSqnNSqkApdRjpdQqpVSWKOX2AfvNf86Lchk/x8vGz5qXfeOFaWWVUluVUr5KqSCl1G3zsIvSUcpYXaZSKpNSaopSylspFWL+f4pSKuML5SLqV1VKfauUuqqUClZKXVJKtbP2Or6EH7AF6PDCOrICtYB51ioppT5TSs03r/O5+bU9rJTyfPE1AtqZf9dRftqbp803/51ZKTVXKeUHPAOyRakzP8ryepin/fDCej4wD604r5RKHcfXQAjxBkkmVwiRHDU2/z8zthWUUj2AKcAFYLh5cntgnVKqm9b6xWV9COwD1gLfAcWAbkAaoKa5zE/AYWCguS0HzdPvxT4UUErlB3YCvsBETB1IF6C8eb2/v6RuWuAIkAeYC5wASgDdgapKqc+01gEvVPsZcABmAMHmsvOVUle01ofj0PS5mF6/Mlrro+Zp7TBlmxdj+jDyIk+gAOAF3AQymuusUUq10lovNZf7CVMipwKmbHGEIy8sL+J1+xFIDTy11lCt9VSlVDVgiFJqr9b6kFLKBlgCOAHVtdbPYh+6EOJNk06uECI5KgI80Vpfi01hpVR6TJfUrwKfa62fmKdPA/4GximlvLTWj6JUywM001p7RVlOONBDKZVfa31Ra71TKRWKqZN7VGu9OErZuMRTC0gFtNBaH49LRaAvppveekbNXCul/gEmm+f/8EKdFEAprXWIuewq4BrQC1OnPbY2Y+qQdwAiOrkdgI1a6/sxvAYjtNYDok5QSv2GaTsMApYCmF/bVkCFqK+rFWe01q1j2d7OwKfAUqVUMUzxVgZ6a61PxnIZQohEIsMVhBDJURrgxezky9TAlOX7LaKDC2D+/TdM40arv1DnTtQOrtke8/9549bcV3ps/t/dfMNUXHhiyhy/mImeYZ7uaVEDpkZ0cAG01reBS8QxLq21EVgENFNKOSilymG6EXDuS+pEZkuV6ekYGTF18PcABZVSaeLSBmBsHNr7EGgJZAW2AkOADVrryXFcpxAiEUgnVwiRHD3BdIk5tnKa/z9rZV7EtFwvTLeWJf7X/H9GK/PiYzmmJ0QMBB4opfYopfoppbLHom5O4KK5wxnJ/PclLOOCmGN7nbjmYfrQ0QjTDWd3gO0xFVZKOSulZkYZQ3sfU2f8C3ORdHFc/6W4FNZaHwF+AT43r7djHNcnhEgk0skVQiRHZ4A0SilrHbiE8rKnGMRmLMLLnu8YbaiZ1jpYa10DU8drpHndw4ELL96QlUBiii3Oj+rSWp8DjgE9gabAQvNTICwXbhq/sAPTGNwFQDOgNqZMe8RY3Di9r2mtn8elvFLKHtPwEIAMwEdxqS+ESDzSyRVCJEerzf9bu7HJmojMZWEr8wq9UCahRDxSLIOVeTmtTENrfVxr/aO5w5sHU6ZzxCvWcw3Ir1744gvz3/lI+LismQuUxjTsI8ahCsDHmG6kG6W17qu19tJab9da78L0uLEXvYkHwY8ESmIaq/wEWC5PVRDi3SSdXCFEcjQb0xcwfGvtEWAASqlPzU9UANMd+M+A3koppyhlnIDemO7I35nAbYy4jB5trK9SqgXwwQvTMlmp74Ppcrq1TnJU64DMWHb4u5inr41dc+NlOTAM+FJrffkl5SIyvNEyxkqpIlgfO/zUPP9Vr0GsKKXqAF8DC7TWYzDdJJcP0w16Qoh3jDxdQQiR7Gitnyul6mO6u3+dUmoHpk7qv5g6dlUwXZIebS7/SCnVF9MjxI5FeX5qe0wZ025a68ckIK31RaXULqCb+TL9P0BxTJ25K5i+LSzCIKVUTUxfsHAdUyfQDdOjtl78ooUXjQaaAFOUUp9gekpBCaATpg8Cr6ofb+Yb+IbGouh5TGOg+yqlUmFqXz5Mj2Y7jenJB1H9jukJCFOVUpuBUOCY1vp6XNtofn7vAuCyeZlorTcppSYCXyqltmutl8d1uUKIN0c6uUKIZElrfUUpVQJTB6kR8D2my+UPgD8xjftcGqX8VKXUXUzPvB1innwS8NRar3tDzWwDTAJamX8/iKkDPg3IEaXcOkx3/DfF9HzcQEydsS7AnJetQGv92PxUg2FAA0zZST9gOjDEyjNy3xqtdZhSqh6mJyK0w/TEizPm34th2cldhqnD3hxTR94GU3xx6uSan4e7CPMzjrXWUZ+l2xeoCMxQSr1WB1oI8WYord/EkCUhhBBCCCHeHhmTK4QQQgghkhzp5AohhBBCiCRHOrlCCCGEECLJkU6uEEIIIYRIcqSTK4QQQgghkhzp5AohhBBCiCRHOrlCCCGEECLJkU6uEEIIIYRIcqSTK4QQQgghkhzp5AohhBBCiCRHOrlCCCGEECLJkU6uEEIIIYRIcqSTK4QQQgghkhzp5AohhBBCiCRHOrlCCCGEECLJkU6uEEIIIYRIcqSTK4QQQgghkhzp5AohhBBCiCRHOrlCCCGEECLJMbztBggRG1m7rtZvuw3xdXRkvbfdhATh7JTibTch3mxs1NtugohCv/dHNyjZpUQCS2kgUfcqhxK9Eu1IDPx7cqLEJplcIYQQQgiR5EgmVwghhBAiuVNJL++Z9CISQgghhBDJnmRyhRBCCCGSuyQ4sFwyuUIIIYQQIsmRTK4QQgghRHInY3KFEEIIIYR490kmVwghhBAiuZMxuUIIIYQQQrz7JJMrhBBCCJHcyZhcIYQQQggh3n3SyRVCCCGEEEmODFcQQgghhEju5MYzIYQQQggh3n2SyRVCCCGESO7kxjMhhBBCCCHefZLJFUIIIYRI7mRMrhBCCCGEEO8+yeQKIYQQQiR3MiZXCCGEEEKId59kcoUQQgghkjsZkyuEEEIIIcS7TzK5QgghhBDJnYzJFSLhKKXaK6UOve12CCGEECLpkU6uiJFS6oZSqvrbboc1uV0c2flDtcifSxMb0KVaHgAKZUvLxn6V2TOkOgt6lsUxZcwXLGwU7BhUjYW9ykab3t+jMId+rMmBYTXoVDW3Rb1SuTOye3B1tg2sSk5nRwDSONix/KvycRrWNO6nwTStW5murRpGmz5r8q90au7OF20aM6z/VzwNeGK1/h+/H6ZT8wa0b1KfFQvnRE7XWjNv+iQ6NnOjcwsP1nktsah79tTffNGmMb06tuC2900AngY8YcCX3QgPD499EC8Y+sNAqlYqS2NPt1eWPXvmNCWLF2bnjm0AXLxwnratmtHIoz5NGzZg+7YtVutN/HUsTRs2YNDAfpHTNm/cwJJFC1673S86fPAADerVon7tGsyZNdNqGa8Vy2jk4UbThu60a92Cq1euAHD7tg+fffIxTRu607ShOz8OG2y1/oC+/6Oxpxu/Tfg1ctrM6VPZs3tXosYREhLCd//7ivq1a9CqeRNu3/YBIDQ0lEED+tHIww0PtzrMmTXDat3uXTvR0L0+K5b9t58NH/ID58+dTZgYDh3AvX4t3OrUYO5s6zGsX7eGKhVK07SRO00bubNm1crIeZ98XDBy+pe9vrBaf8KvY2ji6cagAX0jp23euJ7Fi+YnSAwQv20BcOniBdq0bIZng3o08nAjODjYov6b3qfic1wAzJk1g/q1a9CgXi0OHzpoUTcx9qfYxvGybfGuxJGglEq8n0QinVzxXrrq95QaP+6mxo+7qTViN4EhYWz9+w4A49p+ws9rz1B12C62/nObHjXzxbicLtXycvlu9A5ks7LZ+SC9AxUG76DikJ2s+8PHot4XNfLS+rfDDPY6SduKOQH4ql4BJm65gNaxj6NmXXd+Gj/NYvonpUozc/Fqpi9axYeu2VkepQMbISwsjCljf2bEuKnMWrqWvbu2cfP6VQB2bF7PPX9fZi9bz+xl66hcvbZF/dXLFvLjuCl88eV3bFpr6hAsnT+L5u06Y2Pz+qcGN3dPpkyb9cpyYWFhTBw/ltJlykVOS5kyJT/+/Aur121i8vRZjP1lJAFPom+fgIAAzp8/h9eaDdjZ2XH50kWCgoLYsH4NTZu3fO12v9i2n38aztTps1m7YTPbtmyK9kYdoW49N1av24jXmvV06NiZsaNHRs7L5voRXmvW47VmPT8MGW5R99LFC6RImZJVazdy9sxpAgICuHfPn9OnTlG1WsJ8toxtHGtXryRNmjRs2raT1m3bM+HXsQDs3L6NkNAQVq/byDKvNazyWhHtjR7gyKGDlPjkU1at3cCmjRsAuHjhAmHhYRQsVDhBYhg5YjhTps1mTUQMVy1jAKhZuy5eq9fjtXo9DRs3iZyeIkXKyOkTJ0+3qBcQEMD5c+dYuXZjtH1q/bo1NGveKt4xRMQRn21hNBoZ2P87Bg0extoNm5kzfyEGQ/QP8G96n4rvcXH1yhW2bdnMmg2bmTpjNj+PGEZYWFi0um96f4pLHDFti3clDvFq0skVcaKUslFK9VdKXVVK/auU8lJKZYgyf6VSylcp9VgpdUApVTjKvIxKqQ1KqSdKqeOAZYr0NVQo6MyNe0/xefAcgFwuThy9dB+AA+f8qffJh1brZU3nQLWiWVh66Ea06e0q5eLXTecjO6v/BlhmS0LDwnGwt8XB3pbQME32zKn5IL1D5Hpjq2iJT3FKk8Zi+qefl8XW/AZWsMjH3L/nb1Hm4rkzfJDNlawfZsPOzo7K1Wtz9OA+ADat9aJVx26RndV0GTJa1Lc1GAgOCiQ4KAiDwcAdH2/u+ftS7JNScYrBou0lS5E2bdpXllu+dDHVqtckQ4bI3YfsOXKSPXsOAJydXUifIQMPHj6IVs/GRmE0hqK1JigwEIPBjoXz59K8RWvs7Ozi1fYIZ06fwtU1O9lcXbGzt6d23Xrs27vbopyjo2Pk74GBgag4ZCgMBjuCg4IIDw/HaDRia2PD1Em/0aNX7wSJAWIfx949e2jg7glAjZq1OP77UbTWKKUIfB6I0WgkODgIg50djqkdo9U12BkICgrCaDSizQfNlEkT6Nn7y4SL4SNzDHb21KpTj317LGOID9M+ZWp/oPl4WDh/Ds1btkn0fSqmbXH0yGHy5stP/gIFAEiXLj22trbR6r7pfSq+x8W+vbupXbce9vb2ZMvmiqtrds6cPhU9hje8P8Uljpi2xbsSR4JTNon3k0ikkyviqjfgAVQCPgAeAlOizN8K5AWcgRNA1OvkU4AgICvQ0fwTb+6lXKNlWy/eeULt4h8A4PZpNj7I4GC13vBmHzNi9WnCX0i9Zs+cGvdS2dg2sCpL+pSLHI4Q1aStF/mtY0l618nPvL1X6e9RmF/Wn0uIcCxs37SOUqXLWUz/954/mV2yRP6dKbMz9+/5AXD3tg/7d22nV8cWfP9Nj8jhCFE1b9OJMT8OYvmiOTRo3IL5MybRvmuvNxLDi/z9/NizeydNmrWIscyZ06cwhobi6vpRtOmpUztSvkIlmjfxJFPmzDg6OXLm9EmqJFD2M6J9WbL+99o6u7jg5+dntezypUuoV7s6438dQ7+BgyKn377tQ9NGHnRs15oTf/1pUS9X7tykT5+B5o09qVi5Crdu3SJchydolie2cfj7+5ElS1YADAYDjk5OPHr0kOo1a+GQyoHqlctTq3oV2rXvSNp06aLVLV2mHHdu36Z1i6a0bNWGfXt2U7BQYZydXRImBn8/smT5LwYXFxf8/a1vi907d9DE041vv+6D7927kdNDQoJp2bQhbVo2tXrZPnVqR8pXrEizxh5kzpwZRyenBM2oQ/y3xc0b11FK8UWXTjRr7Mm8OZZXS970PhXf48LPzw+XqNsyiwv+L9R/0/tTXOKIaVu8K3GIV5OnK4i4+gLopbX2AVBKDQVuKaXaaK2NWuu5EQXN8x4qpdICT4FGQFGt9TPgjFJqAVAxPo2xs1XUKpaVn9eciZz2zYK/GNG8GF/XK8D2k3cJMVqOL61eNAv3A4I5desRZfJlijYvhcGWoNBwav+8h7olPmB8u0/xGLM/WpmzPo+pP2ofAKXzZsL/cRAKmN7lM4xhmqErT3HfSgY4rpbOn4WtrS1Va9WLU73Q0BDs7e2ZPHcZh/btYtzPQ/h12vxoZXLnK8DEWYsBOP33X2TIlBmtNT/98B0GWwNd+3xLeisZ4IQw5pef+fLrb2McFnHvnj+DBvZl+IhRVsu079iZ9h07AzBsyCC69+zDmtUr+d2c7erSrfsbabc1zVu2onnLVmzZtJFZ06cxYuQvZM7szPZde0mXLj3nzp7hqz49WbN+c7QMF0DfAd9H/t67xxf8MHQYs2ZM49LFC5QuU45GTZomWhzWnDl9ClsbG3buPciTJ0/o0LYlpcuUJZura2QZg8HAqDHjANMY3u5dOzFx8lTG/DIS37t3cWvgTuWq1d54WytVrkKduvWxt7dnlddyfvi+H7PmLgRgy469uLi44OPtTZdO7cibNx+uH0X/8NShYxc6dOwCwLDB39OjVx/WrFrJ0aOHyJcvP1269XjjMbxMWFgYf5/4i6UrVpEypQNdO7WnUOEifF66TLRy78o+Ze24iI13ZX+Kr6QSx/tOMrkirrIDa5VSj5RSj4DzQBjgopSyVUqNMg9leALcMNfJBGTG9KHKO8qyLNOLUSiluiql/lRK/fn8/E6rZaoWycLpW4+idSiv+AbQfMIhav20h3XHvbl575lFvc/yZKRmsawc/7k207t8TvkCmZnc0XSZ/u6jQLacuA3Alr/vUDDbyy+9f1WvAOM3ned/bgUZsfoMSw5dp7P5Jrj42LF5PccPH6Df0JFWL4NnzOzMPT/fyL/v3/MnU2ZTliBTZhfKVzadQMtVqsb1K5djXI/WmqXzZ9KyQ1cWz51B5x5fU8e9Eeu8lsY7hpicO3eG/n2/oW6tquzauYORPw1nrznD9vTpU/r0/IKevb/i42LFX7qcC+fPobUmR46c7NqxjdHjJuDjfYubN2/Eq33OLi743v3vtfX388PF5eUZmNp167F3jykGe3t70qVLD0ChwkVwdf2Imzeux1h3755dFCpcmOfPn+PtfYsxv05k547tBAYGJkoczs4u+PqaMp9Go5GnAQGkS5eerZs3UbZ8Bezs7MiYMSPFS3zC2bOnY1yf1/KluDXw4NTJkzg5OTF63HgWLpgXvxicXfD1/S8GPz8/q9mwdOnSY29vD4BnoybRbu6JiDmbqyslS33GhQsxX3WJuk/t3LGNMeMm4u3tnWj7VEzbwtklC59+Wor06TPg4OBA+QoVX3oD05vYp+J7XLi4uOAXdVv6+uH8kvpvYn+C+G+LdyWOBCfDFYTAG6ijtU4X5Sel1vo20BJwB6oDaYEc5joKuAcYAdcoy4qeSnmB1nqm1rqk1rpkqoI1rJbx+MyVtce9o03L6JTCtFJl6oAuPHDNot7Pa8/yab+tfDZwG1/MOsahC/foNfcPALb+fYdyBTIDUCZfJq75BcTYxiZlPmL3aV8ePQ/Fwd6WcK0JD9c42NvGWCc2/vj9MCuXzGfo6ImkTGl9uEX+goW57XML3zs+hIaGsm/XNkqXrwRA2YpVOHnCFM+pv/8km2v2GNe1a+tGSpWtQJo0aQkOCkTZ2KCUDcHBQfGK4WU2b9vNlu172LJ9D9Vr1GTA94OpUq06oaEh/O+rXtR3c6dGTcub5V40dfJEevTqg9FoJCzMlLFXNjYEBcWv7YWLFOXWrRv4+HgTGhLCti2bqVSlqkW5qB2fA/v38VF20+v84MGDyBtRfMwdpGzZXC3qgynLs3jhAtp37ExwUHDkB5rw8DBCQ0MTJY7KVaqyYf1aAHbu2M5nn5dGKUWWrFk5fuwYAM+fP+f0yZPkzJnL6rqePH7Mgf37cHP3ICjINA5TKZVg2+K2jzehoSFs32o9hntRxq3v37uHnLlyR7YrJCQEgIcPH/DP3yfIlTvmD6FTJk2kR+8vCTUaCQ83bUMbpQgKTJx9KqZtUa5ceS5fvkRgoGmM9F9//hFjHG9qn4rvcVGpSlW2bdlMSEgIPj7e3Lp1gyJFP7a6rje1P8Uljpi2xbsSh3g1Ga4gXsVOKZUyyt+zgZ+UUu201jeVUpmBslrr9YATEAz8C6QCfo6opLUOU0qtAYYqpTpi6gC3479sb5w52NtSsaAzfRefiDbds5Qr7auY3oi3nLjD8sOmhLFL2pSMa/sprScdfulyJ2+7yJTOpehaPS/Pgoz8b+EJq+Uc7G1pVjY7zSeYHvU7Y+dlFvcpR6gxnB6zj8cqhpGD+3Hq7z95/OgRrdxr0KZzd2q7NWTKuJGEhoYw4CvT444KFC7Kl31/4N97/owfNYwR46ZgazDQ85sBDPy6O+Fh4dSs70GOXKY3vWZtOvLL0IGsWb4YB4dUfDVgiNX1BwUFsmPLekZOMN1x3rB5W374X08Mdnb0HzoqVjG8qH/fb/jrjz949OghtapV4ouevfFs2JiVXssBaNK0eYx1d2zbxom//uTRo0eRby7DR4wkf4GCFmX37t5FocJFIrN6+QsUoImnm+nmnPwFXqvtEQwGAwO+H0z3rp0JDw/Dw7MRefLkBUydoMKFi1C5ajWWL13M70ePYmcw4JQmDT/+bLoke+LPP5gy+TfsDAaUjQ2DBg+zGMsaYcWyJTRw98TBwYF8+fMTFBhEIw83yleoSBorNyW+iTg8GzXm+/7fUb92DdKkTcvoseMBaN6iFYMHDcCzQT3QGnfPhuSL4bWdMW0Knbt+gY2NDWXLVWD5sqU08nCjSbOYt3dsY+g/cDDdu3UmPCwM9ygxTJ08kUKFi1C5SjWWLV7Evn17MNjakiZtWoaPMN3Rf+3aVUYMH4KNUoRrTcdOXcgdQ+dwz4v7VP6CNPZ0I2++fJE3fMUnjvhsizRp09KmXXtaNmuMUooKFSpSsVJlq+t6U/tUfI+LPHnyUrN2HTwb1MXW1paBgwZb3DwX4U3tT3GJI6Zt8a7EkeBskt7X+iodl+cdiWRFKXUD0/CEqEYC94FumG488wdWaK0HKqUcMd1oVhV4APwALADyaq2vmDvE8zCNw70AbAeqaK3Lv6otWbuufu931KMj4zau9l3lbM6Uv89skuDJ/H2WFN6GEvHRnyKZSGkgUfcqhyo/JtqRGLj3h0SJTTK5IkZa6xwvmf3rixO01k8xDVeIamGU+feA+gnSOCGEEEIkHPlaXyGEEEIIId59kskVQgghhEjukuCYG8nkCiGEEEKIJEcyuUIIIYQQyZ2MyRVCCCGEEOLdJ5lcIYQQQojkTsbkCiGEEEII8e6TTK4QQgghRHInY3KFEEIIIYR490kmVwghhBAiuZMxuUIIIYQQQrz7pJMrhBBCCCGSHBmuIIQQQgiR3MmNZ0IIIYQQQrz7JJMrhBBCCJHcyY1nQgghhBBCvPskkyuEEEIIkdzJmFwhhBBCCCHefZLJFUIIIYRI7mRMrhBCCCGEEO8+yeQKIYQQQiR3MiZXCCGEEEKId59kcoUQQgghkjvJ5AohhBBCCPHuk0yuEEIIIURylwSfriCdXPFe+Htsg7fdhHhrMe+Pt92EBLGua+m33YR4S2GT9E7m4u0yhum33YQEYZsEjo0k2FcTr0k6uUIIIYQQyZ2MyRVCCCGEEOLdJ51cIYQQQgiR5MhwBSGEEEKI5C4JDmaWTK4QQgghhEhyJJMrhBBCCJHcyY1nQgghhBBCvPskkyuEEEIIkdzJmFwhhBBCCCHefZLJFUIIIYRI5pRkcoUQQgghhHj3SSZXCCGEECKZk0yuEEIIIYQQ7wHJ5AohhBBCJHdJL5ErmVwhhBBCCJH0SCZXCCGEECKZkzG5QgghhBBCvAckkyuEEEIIkcxJJlcIIYQQQoj3gHRyhRBCCCFEkiPDFYQQQgghkjkZriCEEEIIIcR7QDK5QgghhBDJnGRyRbKjlBqolJr9kvmtlFI7ErNNQgghhBCvIpncJEgpdQNwAcKAZ8BWoJfW+mlcl6W1/jnKcnMA1wE7rbXRPH8JsCT+rRZCCCHEW5P0ErmSyU3C3LTWjsAnQElg0FtuzxvltXQh7Zq6076ZB8O+/47g4GCLMiEhIQwd8D9aetbhi/YtuHvnNgB379ymRvlP6dSyEZ1aNmLcyGFW1/HjoH50aOHJzCkTIqctnDODg/t2v3a7U9vbMrRufha0Kc78NsUplMURgA6lXZndqhizWhZjtEchMqa2s6ibO1MqJjctyrzWxZndqhhV8maMnDexcRFmtTTVX9mpJD/Wz29Rv0hWJ2a3Ksb05h/zYbqUke0Z7VHotc91wcHBtG/VlJZNPWjWsD4zp06yWs737h26d25H62YNadnEncMH9wNw7Ohh2rZoRIvGDWjbohF/HP/dav0fBnxHyybuTP1tfOS0ObOmsW/PrtdsuaXDBw/QoF4t6teuwZxZM62WWTh/Hp5udWns6UaXju24Y96nAMaPG0ND9/o0dK/Ptq1brNYfP24MjT3d+H5A38hpmzauZ/HC+YkaR0hICN/97yvq165Bq+ZNuH3bB4DQ0FAGDehHIw83PNzqMGfWDKt1u3ftREP3+qxY9t/n3eFDfuD8ubMJE8OhA7jXr4VbnRrMnW09hgi7dm6neJH8nD1z2hxDCIMHDaCxpxtNGzbgj+PHrNYb0O9/NPF047cJv0ZOmzVjKnt2J8w+5et7l66d2tLYox5NPOuzdPFCq+W01oweNQL3ejVp1qhB5Gt48cJ52rduRhPP+jRr1IAd26zvU7+NH0uzRg0YPLBf5LQtmzawdNGCBIkjKWwLiN9xATBn1gzq165Bg3q1OHzooNW6b/q4EK8mndwkTmt9G1Mmt4hSqoFS6qxS6pFSap9SqmBEOaVUP6XUbaVUgFLqolKqmnn6UKXUYnOxA+b/Hymlniqlyiil2iulDkVZTlml1B9Kqcfm/8tGmbdPKfWjUuqweT07lFKZ4hvjPX8/Vq9YwsyFK5i/Yh3h4eHs2bHVotzm9WtwSpOGpWu30qRlG2ZM+u8E+uGHrsxZupo5S1fzvwFDLOpevXyRFClTMG/ZWi6cO8PTpwH8e/8e586cokLlaq/d9t6VcnL85kPaLfqHzktOcvNBIAArTtyh85KTdFl6kt+vP6Dt564WdYON4YzccZkOi/+h37pz9KyUk9T2tgB8ueoMXZaa6p/zDeDglQcW9Zt88gH9159n8oHrNCiaBYA2n2VjyR8+6NeMx97enqmz5rHUax1LVqzl6JFDnD71j0W5ubOmU61mbRavWMOIUeMY/fNwANKlT8+4idNYtmoDQ34cydDv+1nUvXzJtC2WrlzPubOneRoQwP17/pw9fYrKVau/ZsujCwsL4+efhjN1+mzWbtjMti2buHrlikW5AgULstRrNavWbqRGzVqMHzcGgAP793Hh/Dm8Vq9j8TIvFs6bw9On0S+kBAQEcOH8OVat3YidnR2XL10kKCiI9WvX0KxFq0SNY+3qlaRJk4ZN23bSum17Jvw6FoCd27cREhrC6nUbWea1hlVeK6K90QMcOXSQEp98yqq1G9i0cQMAFy9cICw8jIKFCidIDCNHDGfKtNmsiYjhqmUMAM+ePWXp4oUU/bhY5LTVq1YCsGrtRqbPmsevY38hPDw8Wr1LFy+QMkVKVq7dyNkzpwkICODePX9OnzpF1WoJs0/Z2try9f/6sWrdZuYvXs7KFUu4ZiWOw4cO4H3zJus2bWfQ4OGMHGH60J0yZUqG//QLK9duYvK0WYwdPZKAJ0+i1Y3Yp1as3oAhyj61Yd0amjRvGe8Yksq2iO9xcfXKFbZt2cyaDZuZOmM2P48YRlhYWLS6b/q4eBOUUon2k1ikk5vEKaVcgbpAALAM+ArIDGwBNiql7JVS+YFeQCmttRNQC7hhZXEVzf+n01o7aq2PvrCuDMBm4DcgI/ArsFkplTFKsZZAB8AZsAe+TYAwCTMaCQ4Oxmg0EhwUSKbMmS3KHD6wh1r13AGoVLUmJ/44htax687ZGgwEBwUTHh5OmNGIjY0tc6ZPpmO3nq/d5tT2tnz8YRq2nPUHwBiueRZiOlE+D/nvhJnSzhZrzfR5FMTtR0EA/PsslEfPQ0mXKnrGN5W9LSWypeXQNctObli4JqXBhpQGG4zh4XyQNgXOTik4efuJRdnYUkqRKlVqUzxGI0ZjqNUTmlKKZ89Mnb6nTwPIlNkZgPwFCpHZ2fR7rtx5CQ4OJiQkJFpdQ5RtYTQasbG1YcbUSXTt3uu12/2iM6dP4eqanWyurtjZ21O7bj327bXM2H/2eWkcHBwAKFqsOP6+vgBcu3qFTz4ticFgIFWqVOTNn5/Dhw5Eq2tjozAajWitCQoMwmAwsGDeHFq0aoOdnWXm/k3GsXfPHhq4ewJQo2Ytjv9+FK01SikCnweajqvgIAx2djimdoxW12BnICgoKDIWgCmTJtCz95cJF8NH5hjs7KlVpx779li/ejJl0kTad+yCvX2KyGnXrl7hs88+ByBDxow4OTlx9uyZ6DEY7AgKDorcp2xtbZg6+Te69+ydIDEAZM7sHNm5SZ3akZw5c+Pv72dRbv/e3dRzc0cpRdFixXka8IR79/zJniMnH2XPYVqWswsZMmTg4cPox7Vpnwo17VNBgRjs7Fi0YC7NWrZOkH0qqWyL+B4X+/bupnbdetjb25Mtmyuurtk5c/pU9Dje8HEhYkc6uUnXOqXUI+AQsB84B2zWWu/UWocCYwEHoCymsbspgEJKKTut9Q2t9dXXWGc94LLWepHW2qi1XgZcANyilJmntb6ktQ4EvIDirxlfpMzOLjRv3Z6mbtVpWKcKqVM7Uap0OYty9/39cXYxZSwNBgOpHR15/PgRYBqy0KlVY/p0bc/Jv/+yqJsjZ27SpU9Pl9ZNKFuhMre9b6F1OPkKFHrtdmdJk4JHgaH0q5GHmS0+5ttquUlp+O+Q7FTmI1Z0/JTq+TMz7/dbL11WARdHDLaKO+ZOb4TyuTJwwvtxtE5zhCV/+DCgZl5alszG2pO+dCqTnTlHX76e2AgLC6NVU09qVS3PZ6XLUqRoMYsyXb7oybbNG6lfszJf9/qCb/tbjqbZs2sH+QsWxN7ePtr0nLlykz59eto0b0SFSlXwuXWLcB1OgYIJlx3x9/MjS9YskX87u7jg52fZIYlq7epVlKtg+hyYL38Bjhw6SGBgIA8fPuCP48fwNXeAI6RO7Uj5ChVp1siDTJkz4+jkxOnTCZetiksc/v5+ZMmSFTAdG45OTjx69JDqNWvhkMqB6pXLU6t6Fdq170jadOmi1S1dphx3bt+mdYumtGzVhn17dlOwUGGcnV0SJgZ/P7Jk+S8GFxcXq53D8+fO4ufrS8VKlaNNz5e/APv27cFoNHLbx5tz587i53s3WplcuXOTPn0GmjfxpFLlKty6dQsdHv7GMm53bvtw4cJ5q8eGv78fLuZtAeDskoV7L8R75vQpQkNDyeb6UbTpqVM7Uq58JVo29TTtU46OnDl9kioJdIUjqWyL+B4Xfn5+uER9HbK44P9C/Td9XLwJSTGTKzeeJV0eWuvIAUxKqWnAzYi/tdbhSilv4EOt9T6l1FfAUKCwUmo78I3W+k4c1/lB1HWY3QQ+jPJ31Hf650D0tFAUSqmuQFeA0ROm0qZDZ6vlAp485tCBvSxfvx1HJyeG9P8fO7ZspGZdN6vlX5QxU2a8Nu4kbbp0XDx/lu+/7cOCFetJ7Ri9ab3/1z/y9/5f9+TbgUNYNHcGVy5fouRnZXDzbByr9UWwtVHkc3Zk0r7rnPd7Sq+KOWhR8kPm/e4NwJyjt5hz9BYtS36IZ7GszDdPf1GGVHYMqJWXUTsuWwwzqJo/E1vOWu+cXb3/nJ5eprFyH3+Qhn+fh6CAwXXyYQzXTDt4g4fPQ+MUE5guyy7xWkvAkyf0/aY3V69cIneefNHKbN+2hfoNPGnVtgOnTv7N0EH9WLZqAzY2pk7+1SuXmTxxHJOmWX+wxzd9B/73e5/uDBg0jLmzpnP50kU+L10Gj0ZN49zu+Ni0cT3nzp5h7gLTyJ6y5cpz9sxp2rVqTvoMGShWrDi2NpY5hQ6dutChUxcAhg7+np69+rBm1UqOHjlE3nz56fpFj0SN40VnTp/C1saGnXsP8uTJEzq0bUnpMmXJ5vrf8BmDwcCoMeMA0xje7l07MXHyVMb8MhLfu3dxa+BO5aqvP6QnNsLDwxk7ehTDfxppMc/DsxHXr12lZbNGfPDBBxQrXgIbG1uLcn37fx/5e5+eXzBoyDBmzZjGpUsXKF2mHI0aJ8w+9fz5M777pg/f9h2Ao2OMp78Y3bvnz+CBfRk2YlTk8RJVu46dadfRdK4cPmQQX/Tow9rVK/n96GHy5stP567d4x3Dy7xP2+JNeheOCyGZ3OTkDpA94g9l+ijlCtwG0Fov1VqXN5fRwC9WlvGqa/vR1mH2UcQ64kprPVNrXVJrXTKmDi7An8d/J+sHH5IufQYMBjsqVKnGGSvjQDM5O+PvZ+pjG41Gnj19Stq06bC3t4/MTuUvWJgPs7nifetGjOs7tH8P+QsWIvD5c277eDNs5Dj279lBUFBgnOK79zSEe0+DOe9numy//8q/5HO2fNPbdfEeFXNntJgOpuEII90LMufILc77Rh/zmSalgQIujhy9/vCVbWn9WTYWHfOh3eeuzDh0k81n/GhYLOsr672MU5o0fFrqM44ePmQxb8PaVVSvWRuAj4uVIDg4mEePTO308/Ol7ze9GfrjKItM1Yv2791NgYKFeR74jNs+3owcM57du3YQFBi3bfEiZxcXfO/+93nM388PFxfrGZjfjx5h9szpTJw8LVrWuUu37nitWc+M2fPQGrLnyBnj+s6fP4fWmuw5crJj+zbG/DoRb29vbt68kShxODu74GvOqBmNRp4GBJAuXXq2bt5E2fIVsLOzI2PGjBQv8Qlnz56OcX1ey5fi1sCDUydP4uTkxOhx41m4YF78YnB2iZYF9/Pzs8iGPXv2jKtXLtG5Q1vq1KzK6VP/8FXv7pw9cxqDwcB3/QbitXo9EyZNI+BJANlz5IhxfXv37KJgocIEPn+Oj/ctxoybyK4d2wmM5z4Fps7Od9/0oU49N6pWrxljvFGzm/5+vmQ2x/v06VO+7PkFPXp/RdFixV+6rgvnzwGaHDlysmvnNn4ZOwEf71vcisc+lVS2RXyPCxcXF/yivg6+fjjHcH6AN3NcvAlJMZMrndzkwwuop5SqppSyA/4HBANHlFL5lVJVlVIpgCAgEAi3sox75um5YljHFiCfUqqlUsqglGoGFAI2JXQwUblkycq506cICgpEa82JP46RPadlE8tVqML2zesB2L9nByVKfY5SikcPH0TeNHDHxxsf71t88KHljV4ARmMoK5ctokXbjgQHB0UerOFh4YSGxi3r+fB5KP4BIbian2zwiWs6bjx4DhD5tAOAcrkycOuh5UndYKP4sX5+dpy/x4Er/1rMr5Q3I79ff0ho2Ms/m9QqmJljNx4SEGwkhcGGcK0J15oUdnE/PTx88CDyZpigoCCO/X6U7DktO3dZsn7AH8dMT064fu0qISHBpE+fgYAnT/i69xf0+vIbipX45KXrMoaGsnzJQtq270RwUHDk42/Cw8PivC1eVLhIUW7duoGPjzehISFs27KZSlWqWpQ7f/4cPw4bzMTJ08iY8b8PImFhYZGd9ksXL3Dp0kXKlLUcQhNhyqSJ9Oz9JUajkfBw075oY6MICgyKsU5CxlG5SlU2rF8LwM4d2/ns89IopciSNSvHj5nugH/+/DmnT54kp5VjC+DJ48cc2L8PN3cPgoICI9/MgoISJobbPt6EhoawfatlDE5OTuw7dIytO/awdccein5cnAmTplG4SFECAwMJfG46ro4eOYzBYEvu3Hmsris0NJQlixbQvmNngoKC/zu+E2Cf0lrz45BB5MyZm9ZtO8RYrmLlqmzeuB6tNadP/oOjkxOZMzsTGhrCt1/1or6be+QHxJeZNmUi3Xv2Me1TYabTuY2NTby2R1LZFvE9LipVqcq2LZsJCQnBx8ebW7duUKTox1bX9aaOCxE7MlwhmdBaX1RKtQYmYRo+8A+mx4yFmDu3o4CCQChwBPMwgReW8Vwp9RNw2NxRrv3C/H+VUvWBicA04ApQX2t9/81FBoWKfEylajXo0roptra25MlfADfPJgDMmT6ZAgULU65SFeq6N+SnIQNo6VkHpzRpGfKT6U74k3//xdzpkzEYDCgbG77pP5g0adNaXddar+XUrudOypQO5M6bn6CgINo396R0uQo4OaWJc9t/23eN72vnw2CruPs4iF92mu7w7VouO67pHAhH4/ckmPF7rgGQzzk1DYpmYezuq1TOm5GPP0hDmpR21C5kullr1I7LXL1vehOpmi8TS/98eRI9hcGGWgWd+W7dOQBW/n2HUe4FMYZpRmy7FOd47t+/x7AfBhAeHkZ4eDjVa9amQsUqAMyY+hsFCxWhYuWqfPlNX34ePpilSxagUAweNhKlFF4rluBz6xazZ0xj9oxpAEyaPpsMGSwz2StXLKWemwcpHRzImy8/wUFBtGjcgLLlK+KUJu7bIiqDwcCA7wfTvWtnwsPD8PBsRJ48eQFTh7Rw4SJUrlqN8WNH8/z5c7772nQzSZasWfltynSMRiMd2piekJDa0ZGfR43BYLB+ut2zexeFCxeJzIjlL1CQRh5u5MuXj/wFCiRKHJ6NGvN9/++oX7sGadKmZfRY06PZmrdoxeBBA/BsUA+0xt2zIfnyW2/TjGlT6Nz1C2xsbChbrgLLly2lkYcbTZo1j3cM/QcOpnu3zoSHheEeJYapkydSqHARKleJ+bLvgwf/0qNbJ2yUDc4uLowYOTrGsiuWL8HN3RMHBwfy5Tcd34093ShfoSJp4rlP/fP3CTZvWk+evPlo0cQDgJ59vqZ8hUqs8loOQOOmzSlfoRKHDx7AvV5NUqZMydAfTY8q37l9GydO/Mnjx4/YuMHU8Rr640jyFyhosa69e3ZRqFCRyAxwvvwFaNrQjbz58se4/WIjqWyL+B4XefLkpWbtOng2qIutrS0DBw3G1tZy2AW8uePiTUjMDGtiUbG9u1yIt8n3Seh7v6O2mPfH225CgljXtfTbbkK8vU6WWrw5SeFtKCw8CQSB6V6B911S6aulNCTu1zNkbLss0Xbifxe2SJTYJJMrhBBCCJHcJZEPB1FJOkMIIYQQQiQ50skVQgghhBBJjgxXEEIIIYRI5pLijWeSyRVCCCGEEEmOZHKFEEIIIZI5yeQKIYQQQgjxHpBMrhBCCCFEMieZXCGEEEIIId4DkskVQgghhEjukl4iVzK5QgghhBDi3aKU+lopdVYpdUYptUwplVIplVMpdUwpdUUptUIpZf+yZUgnVwghhBAimVNKJdpPLNryIdAHKKm1LgLYAs2BX4DxWus8wEOg08uWI51cIYQQQgjxrjEADkopA5AKuAtUBVaZ5y8APF62AOnkCiGEEEIkc4mZyVVKdVVK/Rnlp2vUtmitbwNjgVuYOrePgb+AR1pro7mYD/Dhy2KSG8+EEEIIIUSi0VrPBGbGNF8plR5wB3ICj4CVQO24rkc6uUIIIYQQydw79pzc6sB1rfU9AKXUGqAckE4pZTBnc7MBt1+2EBmuIIQQQggh3iW3gNJKqVTK1PuuBpwD9gKNzWXaAetfthDJ5AohhBBCJHPvUiZXa31MKbUKOAEYgb8xDW/YDCxXSo0wT5vzsuVIJ1cIIYQQQrxTtNZDgCEvTL4GfBbbZUgnVwghhBAiuXt3ErkJRsbkCiGEEEKIJEc6uUIIIYQQIsmR4QpCCCGEEMncu3TjWUKRTK4QQgghhEhyJJMrhBBCCJHMSSZXCCGEEEKI94BkcsV7wcHO9m03Id4WtPn0bTchQbRbcuJtNyHefnUv8rabkCDSp7Z7201IEMHG8LfdhHhLkzJpvJ2GJIFtkcJO8nevQzK5QgghhBBCvAeSxkdPIYQQQgjx+pJeIlcyuUIIIYQQIumRTK4QQgghRDInY3KFEEIIIYR4D0gmVwghhBAimZNMrhBCCCGEEO8ByeQKIYQQQiRzkskVQgghhBDiPSCZXCGEEEKIZE4yuUIIIYQQQrwHpJMrhBBCCCGSHBmuIIQQQgiR3CW90QqSyRVCCCGEEEmPZHKFEEIIIZI5ufFMCCGEEEKI94BkcoUQQgghkjnJ5AohhBBCCPEekEyuEEIIIUQylwQTuZLJFUIIIYQQSY9kcoUQQgghkjkZkyuEEEIIIcR7QDK5QgghhBDJXBJM5EomV7yaUmqoUmpxLMtOV0r98KbbJIQQQgjxMpLJFZGUUi2Bb4ACQADwD/BTXJahtf4i4Vv2csHBwXTr2IaQ0BDCjEaqVa9F1x69Lcr9OmYkf/1xHICgoEAePnjAnkPHuXvnNn2/6U14uMZoDKVpi9Y0atLcov4PA77j6pVLlK9QmR59vgZgzqxp5M6dl8pVq79W28eOGMyxI/tJlz4Ds5asjZy+f/cOFs2Zxq0b15g0Zyn5Cxa2qOvv58vo4d/z8MG/KKWo696Ihs1aR85ft3IpG1Ytx8bWls/LVqBLr2+i1fe+eZ2RQ/pjNBr5qt9gChUtRpjRyICvuzN8zG+kTOkQ6zhmNitKYGgY4RrCwjXfrj8PQM4MDnxRPjv2tjaEhWtmHLnF5XvPotXNmcGBbuWyk8relnCtWfnPXQ5fexitTOcyrlTLl4kWC/62WHeZHOlo8emHPA02MnLnFQKCw8jilILWpT5k7J5rsY4B4NefB3P8yAHSpc/A9EVrIqcvnDWZo4f2YaNsSJs+Pf/7/kcyZnK2qD/om+5cOHeawh8XZ9joyZHTfe/4MGpIP548eUze/AX59oefsbOzi1b30L5dLJo9Bac0aRk8cgJp0qbjzm1vFsz4jQHDx8QpjqjCwsLo0KoJmZ1dGPfbNIv5a1YuZ7XXMmxsbHBIlZoBg4aSM3ceAC5fusgvI4by7NlTbGxsmLvYixQpUkSrP3jgd1y9cpnyFSrRvbfpuJg7azq58+ShUpXXOy6iunXjOsMGfhv59507PnTs2osmLdtEKxfw5DGjfvyBOz7e2NunoN8PP5IrT14Ajh05xKRxowgPD6OeeyNate9ssZ7pk37l2JGD5MlXgO+HjQRgx5aNPH70yGJdr2voDwM5cGAfGTJkZNXajVbL/PnHMb7u05MPPswGQNVqNejWvac5xicMGzqIq5cvo5RiyPCfKFa8RLT6y5YsYvUqL7Jkycr43yZjZ2fP3yf+YvfOHXzbb0C82h/bc+2m9Wv5bcIYMmd2AaBJ85Z4NGwCwKQJYzl8cD8Anbp2p0atuhb1J00Yy9HDB8mbvwDDRvwCwNbNG3j08CEtWreLVwwRDh88wC+jfiI8LBzPRk3o1KWrRZmQkBC+H9CX82fPkjZdOkaPG8+H5u0yZ9YM1q5ehY2tDf0GDKJc+QoWdb/s1R0/Pz+aNW9BsxatABg+5AeaNGtOwUKW5/S3TcbkiiRLKfUNMAH4GXABPgKmAu5vsVmxYm9vz9RZ81jqtY4lK9Zy9MghTp/6x6LcN98NYInXWpZ4raVpi9ZUqVYDgEyZMzNn4XKWeK1l3uIVLJw7i3v+/tHqXr50kRQpU7B05XrOnT3N04AA7t/z5+zpU6/dwQWoWa8BP4+37HjkyJ2HISN/pWjxT2Osa2trS7c+/2POsnX8NmsxG1av4Ob1qwD889dxjhzYy/RFq5i9dC2NW1q+MWxet4oeX/fjp1+nsHLpfAA2rvWiWu36cergRhi0+RJfrz0X2cEFaPdZNlacuMPXa8+x7K87tPssm0W9YGM4E/dfp8/qswzbdplOpV1JbW8bOT93plQ42sf8ebxeYRe+XXee7RfuUTF3RgBalvyAJX/ejnMMNeq6M2Kc5fZo1LI90xasYsp8Lz4vW5Gl82ZYrd+oZXu+HTTCYvrcaRPxaNaauSs24eiUhu2b1lqU2bBqGRNnL6Wue2P27twCwMKZk2nbpVec44hqxdJF5MiZO8b5terUZ8nK9SxasZbW7Toy8dfRABiNRoYO6ke/74ewbPVGps5agMEQfTtcvnSRFClSssRrHefOnjEfF/c4e+ZUgnRwAT7KkZM5S1czZ+lqZi7yImWKlFSoUs2i3OJ5s8ibrwDzlq1l4LCfmTRuFGDq5E8YPYLRE6exwGsDu3ds4ca1q9HqPn0awKUL55i3bC12dnZcvXKJ4KAgtm5ch2dTyw+8r8vN3ZMp02a9slyJTz5lxap1rFi1LrKDCzD6l58oW64CazduZcXqdeTKZbldt27ZhNfq9RQrXoIjhw+htWbWjKl0+aJ7vNsf23MtQI2adSLPtxEd3EMH9nHx/DkWrzCdaxcvmMfTp0+j1XsaEMDF8+dYunI9dnZ2XLl8iaCgIDauX0uTZi3jHQOY9omffxrO1OmzWbthM9u2bOLqlSsW5dauXkmaNGnYtG0nrdu2Z8KvYwG4euUK27ZsZs2GzUydMZufRwwjLCwsWt0jhw5S4pNPWbV2A5s2bgDg4oULhIWHvZMd3KRKOrkCpVRaYDjQU2u9Rmv9TGsdqrXeqLX+zkr5lUopX6XUY6XUAaVU4Sjz5iulRph/r6yU8lFK9VVK+Sul7iqlPJRSdZVSl5RSD5RSAxOg/aRKlRowvTEbjaGv/ES6Y+tmatY2ZRDs7Oyxt7cHTJ++w7W2KG8wGAgOCiY8PByj0YiNrQ0zpk6ia/f4dUA+LlESpzRpLaZnz5EL1+w5X1o3Y6bM5M1fCIBUqVPzUY6c3L9n6pxvXONF8zadIuNKnyGjRX1bg4GgoCCCg4IwGOx4GvCEo4f2U6OOW7xiikoDDuYOayp7Wx48C7Eoc+dJMHefBAPw8HkojwONpElp6kzZKGj/WTYWHPeJcR3hWmNnq0hha4MxXFPIxZFHz42Ry4yLosU/xSlNGovpqVM7Rv4eFBQU4+C1EiU/j9wXI2itOXniOBUqmz5UVa/TgKMH91jUtbFRhIaEEhQUhMFg4MzJE6TPmIkPXbPHOY4I/n6+HDm0nwaejWIsk9oxSmyBgZG/Hz96mDx585E3fwEA0qZLh62tbbS6BoOB4OCgaMfFzGmT6PJF/I6LmJz443c+yOZKlqwfWMy7cf0qJUp+DpiOH9+7t3nw733Onz3Nh64f8UE2V+zs7Khaow6H9kd//W2UDWFGI1rryNd/+eL5NGzWEoPBzmJdr+vTkqVIm9byeI+NgIAATvz1J54NGwOm85a1fVVrjdFoJCgoEIPBjs2bNlCufEXSpk0Xn6YDr3eujer6tauU+LQkBoMBB4dU5MmXj6OHD0Zfh40NxohtEWjaFksWzqVp81YY7BJmW5w5fQpX1+xkc3XFzt6e2nXrsW/vbotye/fsoYG7JwA1atbi+O9H0Vqzb+9uateth729PdmyueLqmp0zp09Fq2uwM51fI2IBmDJpAj17f5kgMbwJSiXeT2KRTq4AKAOkBCzTS9ZtBfICzsAJYMlLymYxL/tDYDAwC2gNfApUAH5QSr28NxcLYWFhtGrqSa2q5fmsdFmKFC0WY9m7d25z544PJT8rHTnNz/cuLZu441a7Km3bdyKzc/RL0Tlz5SZ9+vS0ad6ICpWq4HPrFuE6nAJWhhG8Db53b3Pl0gUKFC4KgI/3TU6f/IvenVryTfcOXDx3xqKOe6PmLFswm9E/DqJFu84snjeTlu06Y2MT99OCBobWycs4j4LUzJ8pcvqc371p/1k2Zjf/mPafZ2PRK7KreTOnxmCr8DV3UOsWcub4rcc8DAyNsc7qk74Mr5uPUtnTcfDqA5qWyIrXP3fiHMOrzJ8xiTYNa7J3x2badOoR63pPHj8itaMTtuYsaKbMLvx7z9+iXNM2nRj4VVeOHd5P5ep1WDp/Bi3bW15CjYvxY0bR68tvUa/YpqtWLKWRWy0mTxzHN31Nnztv3bqJUoove3ShbYtGLJo/x6Ke6bjIQLsWjShfsTI+3rfQ4eEUKFgoXu2Oye4dW6lm5fI2QO68+Tm4dxcA58+exs/3Lvf8/bh/zx9nlyyR5TK7uER+GIyQKnVqPi9Xkc6tGpMxY2YcHZ04f/YUFSpbZowTw6mT/9C0kTs9v+jC1SuXAbhz24f06TMwZNAAmjfxZNiQQQQ+f25Rt1mLVrRt1Qzfu3cpXqIEG9atoWnzhMmAQuzPtXt276BlE3f6f/slfr53AcibrwBHDx8iKDCQRw8f8tcfx/H3841WL3Xq1JQtX5HWzRqSKXNmHB0dORPPK2Yv8vfzI0vW//YJZxcX/Pz8LMv5+5ElS1bA9IHO0cmJR48e4ufnh0uW/+q7ZHHB/4X6pcuU487t27Ru0ZSWrdqwb89uChYqjLOzS4LFIV5NxuQKgIzAfa21MTaFtdZzI35XSg0FHiql0mqtH1spHgr8pLUOU0otB2YCE7XWAcBZpdQ5oBhwPT4B2NrassRrLQFPntD3m95cvXKJ3HnyWS27Y/sWqlavFS0r5ZIlK0tXrueevz/ffd2LqjVqkTFjpmj1It78Ab7p050Bg4Yxd9Z0Ll+6yOely+DRqGl8Qnhtgc+fM3zAN3T/qm9kxjE8zEjAkyf8NnsJF8+dYcSgb1m4emu0rItzlqyMm2ralLe9b3Hf34+PcuRi1LCBGENDad+1J9k+yhGrNgzYeIEHz0NJm9LA0Dr58HkcxDnfp9QumJm5v3tz9MYjyuVMT68KORiy9ZLVZaR3sOOrSjmZuP86Gkifyo6yOdMzaPPFl6775O0n/O/2EwAq58nIX96P+SBNSjw+duFpcBizj3oTEhYeqzhepn233rTv1psVi+awcc3yOHV0Y+OTUmX4pFQZAHZt3UipMhXwuXWT1csX4OSUhm5f9o3TMJJDB/aRPkMGChQqzF9/Hn9p2cbNWtK4WUu2b93E/NkzGPzjSMLCjJz8+wTzFnuRMmVKenXrSIGChSj1eZlodb/+7r9xnv/7sgf9vx/KvNnTuXLpIqVKl428VB1foaGhHDmwj649v7I6v1W7zvw2bhSdWjYiZ5685MlXABsbW6tlrWnZtiMt23YEYPSIwXTs1otN61bxx7Gj5M6Tj7aduiVEGK9UoGBhtuzYQ6pUqTl4YD9ff9mLDZu3YwwzcuH8OfoNGETRj4sxetRPzJ0zyyIzWN/NnfpuplFmM6ZNoUXLNhw+dJBNG9eRxSUr33zX77U+yEaIzbm2fKXK1KxjynSuWbWCoT8MYNqs+ZQuW45zZ0/TqV1L0qdPT9GPi1ttS9sOnWnbwTRuesSwQXTr0Zt1a1Zy7OgR8uTLR6cu8R968aYZDAZGjRkHmPbd7l07MXHyVMb8MhLfu3dxa+BO5apv50NUciKZXAHwL5BJKfXKDz1KKVul1Cil1FWl1BPghnlWphiq/Ku1jhisFHEtNOpH3kDAESuUUl2VUn8qpf6cP2fmK4MAcEqThk9LfcbRw4diLLNz21Zq1a5ndV5mZ2dy58nLPyf+irH+/r27KVCwMM8Dn3Hbx5uRY8aze9eOaJd6E4vRGMqwgd9QtVY9KlT+L9ORKbML5StXQylFgcJFUTY2PH70MMblzJsxifbderHWawl13BrSpefXLJozPdbtePDclGl9HGTk2M1H5M1suqRZJW9Gjt54BMDh6w8jp7/Iwc6GQbXysPjP21wy35iWK2MqsqZJyfSmRZnZrCgpDDZMa1IkxjbY29pQNV9Gtpy7R4tPP2Di/huc93tKpTwZYh1HbFSpUZfD+3bFunyatOl49jSAMKPpM+T9e35kzGx501qEoKBAdm1dj1vDZiyeO5Vvv/+RQh+XYO+OLXFq56l/TnBw/1486lbnh/7/488/jjHk+74vrVOjVl327zNdtnV2zkKJT0qSLn16Ujo4ULZ8RS5eOBdj3QN7d1OgYCECA59z28ebn0aPZ28CHhfHjhwkb4GCZMho/VST2tGRAUNGMGfpar4fNpLHjx7ywYfZyJTZOVq28J6fH5le8vpfungerTWu2XOwb/cOho0cx20fb3xu3UyQOF7F0dExckhAhYqVMBpDefjwIS4uWXB2caHox6bMafUatbhwPubt4e/vx9kzp6hSrTqLFszllzHjcUrjxPHfjyZIO192rk2XLn3kUCl3z8ZcOH82cl7HLl+wxGstk2fMRWvNR9lzxLiOixfOobUme46c7N65nZFjxnPb25tbN2/Eq+3OLi743v1vn/D388PFxTLD6uzsgq85C200GnkaEEC6dOlxcXHBz/e/+n6+fjhbqR/Ba/lS3Bp4cOrkSZycnBg9bjwLF8yLVwxvgo2NSrSfRIsp0dYk3mVHgWDAIxZlW2K6Ga06kBbIYZ6e4Hut1nqm1rqk1rpk+04xX7Z9+OABAU9MmbygoCCO/X6U7Dmtj4C4cf0aAU8eU7RY8chpfn6+pnGWwJMnj/nn77/InsN6fWNoKMuXLKRt+04EBwVHRh0eHkZoaMyX1N8ErTXjfhrCR9lz0rhF22jzylasyj9//QGAz60bGENDSZsuvdXlnDzxJxkzZSaba3aCg4KwsVEoGxuCza/Jq6Qw2JDSziby9+IfpuHWQ1PH5sHzUIpkdQLg4w+cuPvEcpkGG8WA6nnYd/lfjt74ryP+l/djOiw9SdcVp+m64jTBxnC6r7QcdhHB82MXNp/1J0xr7G1t0Nr0GqUwxP80d9v7vw7O0UN7yfaK8dJRKaX4uEQpDu7bCcCurRsoU75KjOVXL11Ag8amsaAhwcGgFDZKxXp7ROjR5xs2bt/Lui27+HHUOEqW+pxhP422KBe1w3D44H5czWOAPy9bjitXLhEUGIjRaOTEX3+QM1ceq+syhoayfOki2rTrRHBQEMp8YISFhRFqTJjjYvf2LVSraX2oAkBAwJPIY3DTutV8XOJTUjs6UqBQEXxu3eLubR9CQ0PZs3Mr5SrG/PrPnT6JTl/0xmg0Em6+AmBjowgKSpwPsffv34scw3nm9Cl0uCZdunRkypSZLFmycuO66Ykhx48dJVfumG8onDr5N7r37AOYnoqglEIpGwLjuB9FFdtzbdThIAf27yFnzlyAaX94ZP6wffnSRa5cvsjnZcrFuL7pU37jix5fYgw1Eh5uypUoGxV5vn5dhYsU5datG/j4eBMaEsK2LZupVKWqRbnKVaqyYb1pFN/OHdv57PPSKKWoVKUq27ZsJiQkBB8fb27dukGRoh9bXdeTx485sH8fbu4eBAUFmrdD/GMQsSPDFQRa68dKqcHAFKWUEdiBaZhBdaAKEHXglxOmDvG/QCpMT2N4q+7fv8ewHwYQHh5GeHg41WvWpoL5TWzG1N8oWKgIFSubTmA7tm2hRu260S7b37h21XRHuVKgNa3bdiRPXutDHVauWEo9Nw9SOjiQN19+goOCaNG4AWXLV7R6E8ir/DS4L6dO/MnjR49o0aA6bTv3oE6Dhhzat5spv5qyUYP+15Pc+QowasJ07t/z59eRQ/n516mcPfU3u7ZtImfuvHRra7ok3PGLPnxetgK13TwZ99NgurTyxGCw47sfRli9QURrzdL5M/n+R9Mjqup6NGbUkP6EhYXR57tBsYohnYOB/tVNnR9bG8WBqw/428f0Rjj14E06l3HFRilCw8KZetDUWcydKRW1C2ZmysGblMuVnkJZHXFKaaBqPlOW7rf917n+IPadivSp7MibOTUr/jZlXTaf82esR0GehYQxcqflXdMxGTWkH6f++ZMnjx7R2rMGbTp1p1b9hsybPhGfWzdQNjY4u2Slt/m1uXThLFvWreSr/kMB+LZHe7xv3SDo+XNae9bg6/5D+fTzcnTs/hWjhvZl4awp5M5bgJr1Pa2u/9/7/lw8f4ZWHU1P4mvQqAVfdm6Jo1Mafvh5fKzjeJWZUydRoFBhKlauyqoVS/nj2FEMBoPp8WU/mg7pNGnS0qJ1Ozq0bopSijLlK1KuQiWry1vltYy6bu6kdHAgT778BAUF0aqJO2XKV8TJKe7HxYsCA5/z5/Gj/G/gkGjT169eAYB7o2bcvH6NkcO+R6HIkSs3/X4YDpguG3/VdyDf9ulGeFgYdRt4Rj4i7UUH9+0mf8HCkZnePPny0765J7nz5CNPvgLxjqN/32/4648/ePToIbWqVeKLnr3xbNiYlV7LAWjStDm7dmxnpddybG1tSZkyJSPHjIs8dvsNGMTA/t9hDA3lw2yuDPvR+uk3IsMbcRd/nbr1adKwAVmyZKF9R8vHp8VWbM+1K5Yt5sC+PdgaDKRNk5bBw02PZDMajXTraHocW+rUqRn+02iLJ3ZE2LdnFwULFYm8PyJf/oK0aNyAPHnzky9//LaFwWBgwPeD6d61M+HhYXh4NiKP+XFzUyZNpHDhIlSuWg3PRo35vv931K9dgzRp0zJ6rOkYzJMnLzVr18GzQV1sbW0ZOGiwxU2ZEWZMm0Lnrl9gY2ND2XIVWL5sKY083GjSLOGe2pFQkuATxFDayp3kInlSSrUCvgYKYnpO7l+YnpNbE8ijtW6tlHLEdKNZVeAB8AOwAMirtb6ilJoP+GitBymlKgOLtdbZzMs3YOo859Ra3zBPOwRM11q/9MsmHgeGv/c76uOX3Dz1Pumz5vTbbkK8/eoe87CH90n61Al35//bFGyM/5jpty3iiSDvu9Cw9/5USwq7pHGROqUh4a+Qvkzh73ck2sY/+1PNRIktaRyVIkForZdg/UkJR6KUeYrls3MXRpnfPsrv+4BsUf428sKwBq11+fi0WQghhBDxJ18GIYQQQgghxHtAMrlCCCGEEMlcEkzkSiZXCCGEEEIkPZLJFUIIIYRI5mRMrhBCCCGEEO8ByeQKIYQQQiRzkskVQgghhBDiPSCZXCGEEEKIZC4JJnIlkyuEEEIIIZIeyeQKIYQQQiRzMiZXCCGEEEKI94B0coUQQgghRJIjwxWEEEIIIZK5JDhaQTK5QgghhBAi6ZFMrhBCCCFEMic3ngkhhBBCCPEekEyuEEIIIUQylwQTuZLJFUIIIYQQSY9kcoUQQgghkjkZkyuEEEIIIcR7QDK5QgghhBDJXBJM5EomVwghhBBCJD2SyRVCCCGESOZkTK4QQgghhBDvAcnkCiGEEEIkc0kwkSudXPF+sDO8/0dfZqcUb7sJCWJxm0/fdhPibez+K2+7CQmic6nsb7sJCSJ9aru33YT4e/9PUQCksJMLvCLpkE6uEEIIIUQyJ2NyhRBCCCGEeA9IJ1cIIYQQQiQ5MlxBCCGEECKZS4KjFSSTK4QQQgghkh7J5AohhBBCJHNy45kQQgghhBDvAcnkCiGEEEIkc0kwkSuZXCGEEEIIkfRIJlcIIYQQIpmTMblCCCGEEEK8BySTK4QQQgiRzEkmVwghhBBCiPeAZHKFEEIIIZK5JJjIlUyuEEIIIYRIeiSTK4QQQgiRzMmYXCGEEEIIId4DkskVQgghhEjmkmAiVzK5QgghhBAi6ZFOrhBCCCGESHJkuIIQQgghRDInN54JIYQQQgjxHpBMrhBCCCFEMpcEE7mSyRWvppSar5QaEcuyW5VS7d50m4QQQgghXkYyuW+AUqol8A1QAAgA/gF+0lofeoPr1EBerfWVN7WO2NBa13mb6xdCCCFE3NkkwVSuZHITmFLqG2AC8DPgAnwETAXc32KzXkop9d5/2Dl86CAe9WvToE5N5s6eabXMnTu36dapPU09G9C5fRv8fH0j5/Xs1pkKZUrRp0e3GNcx8dexNPVswKAB/SKnbd64gSWLFiRgHAdwr18Ltzo1YoxjzC8/07SRO00budOgXi3KlykZbf7Tp0+pWa0iI38abrX+gH7/o4mnG79N+DVy2qwZU9mze1eCxDBs8ECqVypLU0+3GMs8efKY/33Vi2aNGtC2ZROuXL4UOS/gyRP6ftOHhg3q0Mi9LqdO/m1Rf/nSRTT1dKNPj66EhoYA8PeJvxg3emS82x8eHsbWX/qwf/qwyGm/LxrPhiGd2DqqN1tH9eahzzWrda8d283G4V3YOLwL147tjpx+48/9bPm5J1tG9mLv1MEEP31sUffS/o1s+bkH+6YNIcwYCsC9q2c5sXrWa8eyesViOrX0pGMLD1YvX2S1zNOnAXz/v150ad2Iji082LZpLQB+d+/QrW1TurZpTMcWHmxc42W1/s+D+9G5VUNmT5sYOW3x3Bkc2r/bavnXMfSHgVSrVJYmL9mnFsybQ/PGHjRv7EETTzdKFivE48ePCA4Opk2LJjRr5E5jj/pMm/Kb1foTfx1L04YN+GHgmzy+X32eunv3Dl06tKV5Y0+aejbg4IH9FvPLlvqEhfPmWNQNCQmhZ7fONPZww2v50sjpPw79gfPnziZMDAcP0KBeLerXrsGcWdZjCAkJ4bv/fUX92jVo1bwJt2/7RI/hzh1KlyzBghhi6N61Ew3d67Ni2ZLI6cOHJFwMCRHHnFkzqF+7Bg3q1eLwoYNvLQ7xctLJTUBKqbTAcKCn1nqN1vqZ1jpUa71Ra/2dUiqFUmqCUuqO+WeCUiqFuW57pdShF5anlVJ5zL/PV0pNUUptVkoFKKWOKaVym+cdMFc5qZR6qpRqZp5eXyn1j1LqkVLqiFLq4yjLvqGU6qeUOgU8U0oZlFLlzeUeKaW8lVLtrcSYXim1SSl1Tyn10Px7tijz9ymlOkeJ6bBSarx5mdeUUmXN072VUv4JMbQhLCyMUSOGM3naLFZv2MS2LZu5etUyoT1+7GjqNXDHa+0GunbvyaQonby2HToxYuQvMa4jICCA8+fO4bV2A3Z2dly+dJGgoCA2rFtD0+Yt4xtCZBwjRwxnyrTZrNmwmW1bNlmN47t+A/FavR6v1etp0bI11arViDZ/yqQJfPJpKavruHTxAilTpGTl2o2cPXOagIAA7t3z5/SpU1StVj1B4nBr4MmkaS/vmM2dNYP8+QuwYvUGhv30C2N/+Tly3phffqJMuQqs2bCV5avWkTNnbov6WzdvYvnq9XxcvARHDx9Ca83smVPp3K17vNt/ad8G0rq4Wkwv7tGBOv0nUaf/JNJny2UxP/hZAGe2LqXm/36l1rfjObN1KSHPnxIeFsaJ1TOp1udn6g6YTLoPcnDpwCaL+jf+3Eed/pPJlLMgvudPoLXmzLblFK7d/LXiuH71MlvWr2bK3KXMWrSK3w/t57b3LYty61ctJ3vOXMxavJpfp85l+m9jCQ0NJUOmzEyavZiZi1YxZc5Sli2cw/17/tHqXr18EfsUKZm9ZA0Xz53h6dMA/r1/j/NnT1O+UrXXarc1bu6eTH7FPtWuQyeWr1rH8lXr6PXl13xSshRp06bD3t6eGXPms2L1epatXMvRw4c4dfKfaHUDAgK4cP4cXmteOL7XJ+zxHZvz1OwZ06hRqw7LV61l5NhfGTliWLT540aPolyFClbXceTwIYp/8ilea9azaeN6AC5euEBYWDgFCxVOkBh+/mk4U6fPZm3EOeqKZQxrV68kTZo0bNq2k9Zt2zPh17HR5o8dPYryMcVw6CAlPvmUVWs3sGnjhv9iCA9LkBgSIo6rV66wbctm1mzYzNQZs/l5xDDCwsISPY6EplTi/SQW6eQmrDJASmBtDPO/B0oDxYFiwGfAoDgsvzkwDEgPXAF+AtBaVzTPL6a1dtRar1BKlQDmAt2AjMAMYENEp9qsBVAPSAd8CGwFJgGZzW38x0obbIB5QHZMWepAYPJL2vw5cMrchqXAcqAUkAdoDUxWSjnGMn6rzpw+hetHH5HN1RU7O3tq1anLvj2WWaRrV6/y2WelASj12efs2/tfmc9LlyF1qtQxrsPGRmE0hqK1JigoEIPBjoXz59K8ZWvs7Ozi0/wX4sgeJY56VuOIauuWzdSuWz/y73Nnz/Dg338pU7ac1fIGgx1BwUGEh4djNBqxtbVh6uTf6N6zd4LEAJg7F2lfWubatauUMm+LnDlzcefObf799z4BAQH8/defeDRsDICdnT1OadJY1NdaYzQaCQo0bYstmzZQrlxF0qZNF6+2P394nztn/yBXmZpxrnv3/AmyFChBitRO2KdyJEuBEtw59xegAY0xJNjU7qDnOKTNaLkArQkPCyMsJBhla+DGH3vJWqgkKVI7vVYst25co0DhoqRM6YCtwcDHn5Tk4D7LbL1SisDnz9FaExj4HKc0abG1tcXOzg57e3sAQkJD0Drcoq7BYEeIeX8KMxqxtbFl/swptOvS47XaHJNPY7FPRbV9y2Zq16kHmOJLZT62jUYjRqPR4lFJ0Y5v8z61aP5cmrdI6OP71ecppRTPnj0F4GlAAJkzO0fO27t7Fx9+mI3cufNYXYfBYCAoMBCj0Wja7YCpkyfSo3efhIvB1XyOsrendt160c6jke3cs4cG7p4A1KhZi+O/H0VrU4P27N7Fh9k+JHeevNZjsDMQFBSE0WiMrDNl0gR69v4yQWJIiDj27d1N7br1sLe3J1s2V1xds3Pm9KlEj0O8mnRyE1ZG4L7W2hjD/FbAcK21v9b6HqYOa5s4LH+t1vq4eflLMHVEY9IVmKG1Pqa1DtNaLwCCMXWyI/ymtfbWWgcCLYFdWutl5uzzv1rrf15cqHn6aq31c611AKaOdqWXtOO61nqe1joMWAG4YnoNgrXWO4AQTB3e1+bv74dLlqyRf7u4ZOGev59FuXz587Nn104A9uzaybNnz3j06GGs1pE6tSPlK1aieWNPMmXOjKOTI2dOnaRKAmU/wRRHlixZIv92cXHB30ocEe7cuc2d2z589rlpk4aHhzNuzC98822/GOvkyp2b9Okz0LyJJ5UqV+HWrVvo8ITJ8sRFvnz52bPbtC3OnD6F7907+Pv5cue2D+kzZGDoDwNo2dST4UMGEfj8uUX9Zi1a0b51M3x971KsRAk2rFtDkwTIuJ1YM5Pi7h1RNpaphlObFrFlZC9OrJ5FWGioxfzAx/+SKl2myL9TpctI4ON/sbE1ULJpD7aM7Mm6QW157OtNrjI1LOrnrVifHb/+j2cP75E5V0Gu/b6LfBXrvXYsOXLl5fQ/J3j8+BFBQYEcO3KQe36+FuU8Grfg5o1rNK1flc6tGtLz6/7Y2JjeGvz9fOncqiEtGtSgWZuOZIrS4QLInjMXadNn4It2TSldvjK3fW4RrsPJV6DQa7c7vgIDAzly+BDVavz3QSUsLIzmjT2oXqkcn5cuS9GPi0Wrkzq1I+UqVKJFk/+O79OnE/74js15qluPXmzZtIFa1SrRu0c3+g005UGeP3/GvLmz6NajZ4zrKF2mLHfu3KZty2a0aNWafXv3ULBgIZydXRImBj8/smT97xzl7OKCn59lDKZzmSlWg8GAo5MTjx495PmzZ8ybM4svuvd6SQzluHP7Nq1bNKVlqzbs27ObgoUKJ1gMCRGHn58fLlHP1Vlc8H+hfmLEkdCUUon2k1je+7GY75h/gUxKKUMMHd0PgJtR/r5pnhZbUd+hngMvy4BmB9oppaKm6OxfWJ93lN9dgauvaoBSKhUwHqiNKaMM4KSUsjV3ZF8U9cgPBNBavzjNahxKqa6YOutMmjqdjp27vqp5L/X1t3355acRbFi/lk8+LYmziwu2Nraxrt++Y2fad+wMwLDBg+jeqw9rVq3k96OHyZsvP10S4FJ5XGzfupnqNWtha2uKwWv5UspXrBjt5GtN3/7fR/7ep+cXDBoyjFkzpnHp0gVKlylHo8ZN32i7Adp36srYX36iRRMP8uTNR/4CBbGxsSUszMiF8+f4rv8gin5cjDGjfmLe3Fn06BU9+1HPzZ16bqZh7jOnT6F5qzYcOXSQzRvX4ZIlK19/2y+yoxZbt88cJ4VjOjJ8lAe/y9GzMsUatCNlmvSEG40cXz6J87tWUaROi1gtNzzMyJVDW6jd9zccM2Xhr5XTObdjJUVeGIaQ87Oq5PysKgBnti4jfyU37pz7ixvHd5MqXWZKeHZCxSGm7Dlz0bxNR/r16UpKBwfy5C2Aja3l/v7HscPkyZefcVPmcMfHm759ulK0+CekTu2Is0sWZi9Zw/17/gzu9yUVq9QgQ8ZM0er3/Pq/D1Xf/68XX/cfzJJ5M7l65SKflipDPY/GsW5zQjiwfy/FSpSIltW3tbVl+ap1BDx5wv++6sWVy5fIkzdftHpRj+/hQwbRvWcf1q5eye9HTMd3QgyFiY1tWzbj5u5J2/YdOfnP3wwa0I9V6zYyfcpkWrdpH5mVtsZgMDBy9DgAQkND6dmtM+MnTWHs6JH43r1L/QYeVK5SNVHisGba1Mm0btuOVKlfHsOoMf/F0L1rJyZOnsqYX0wxuDVwp3LVhBsK86YklTjed5LJTVhHMWVLPWKYfwdT5zPCR+ZpAM+AVBEzlFIv76m8mjemJzqki/KTSmu9LEoZ/UJ5y8GPlv4H5Ac+11qnASKGSiT4RzOt9UytdUmtdcmXdXCdnV3w870b+befny+ZrXxadnZ2YdzESSxftZZeX34FYPVS+KtcOH8OrTU5cuRk145tjB43AR/vW9y8eSPOy3qxfb5Rbobz8/N76af+bVu3RF6SBTh58m9WLF1CnZpVGT/2FzZtWMfE8WNjrL93zy4KFipM4PPn+HjfYsy4iezasZ3AwMB4xREbjo6ODP1xJMtWrmP4T7/w8OEDPszmirNLFpxdXCIzbdVr1OLC+XMxLueevx9nz5yiStXqLF44l5FjxuPk5MTxY0fj3KZ7185x+8wxNgzpyJF5o/G7dIojC0yvn0PaDCilsLWzI1fp6vx785JFfYe0GXn+6H7k388f/YtD2oyRN6k5Zc6KUoqPPqnA/evnY2zH88f/8u/NS2QrVoYLe9ZStkM/7BxS43vpZJxjqtugIdMXeDFh+gIc06Qhm2t2izLbN62jfOXqKKX40PUjsnzwId43rkcrkymzMzlz5eH0yRMxruvwgT3kK1CIwMDn3LntzeCfxnFg706Cgt78/hTVjheOi6ic0qShZKnPOXLY8kahCFGP7507tvHLuAl4e9/iVgIc37E5T61bs5qatUwPqSlWvAQhIcE8eviQM6dPMeHXMdStWZUlixcyZ9ZMli9dHOP6Vi5fRv0G7pw+eRInRyd+GTueRQvmxi8GFxd87/53jvL388PFxfq51tccq9Fo5GlAAOnSpef0qZNMGDeWOjWqsmTRAmbPnMGyJTHH4LV8KW4NPDh18iROTk6MHjeehQvmxSuGhIjDxcUl2o3Lfr5+OFup/6bjSGg2KvF+Ei2mxFtV0qe1fgwMBqYopTyUUqmUUnZKqTpKqdHAMmCQUiqzUiqTuWzEEX4SKKyUKq6USgkMjePq/YCod8PMAr5QSn2uTFIrpeoppWIa4LcEqK6Uamq+CS2jUqq4lXJOmLKvj5RSGYAhcWxngitcpCi3bt3kto8PoaEhbN+6xWq24uHDh4SHm8YVzp01E3fPRq+1vqmTTGPcjEYjYeblKWVDUGDQ6wdBRBw3uO3jbY5jM5ViyLpcv3aVJ0+eUKx4ichpI38Zx7Zd+9i6Yw9ff9uP+g08+PLrb63WDw0NZcmiBbTv2JmgoODIy0fh4WGEWrkUn9ACnjyJfCrC2tUr+eSTUjg6OpIpU2ZcXLJy47qpY3j82FFy5Yr5s9e0Kb/xRQ/TeMPgYFMcysaGoKC4b4viDdrj8eMCGgybS9kOfXHJ9zFl25lev8DHDwDTWGCfU7+TNqtlZzFrwU/wPf83Ic+fEvL8Kb7n/yZrwU9wSJeRx77eBAWYnqjge+Fv0li5sS3C6U2LKVqvFYBpfC4KZaMICwmOc0wPH/wLgJ/vXQ7t20W1WnUtyji7ZOXvP44B8ODf+3jfukHWD7Nxz9+XYPPrGPDkMadP/o3rRzmsrsdoDGX18sU0a9OBkODgyDtLwsPCMCbC/hQhICCAv/78g8pV/suQPXzwgIAnTwAICgri99+PkCOn5Y2DEaZOnkiPXqbjOzzMdHzbvOY+FVVsz1NZsmaN/JB27epVgoODSZ8hA3MXLmHLjj1s2bGHVq3b0qlLV5q3bG11XU8eP+bA/n3Ub+BBYFAQysYGpRTBQXHfhyxjuIGPjzehISFs22L9HFW5SlU2rDfdmrJzx3Y++7w0SinmL1rK1p172LpzD63atKNz1260aPXyGNzcPQgKCoy8zB3f7ZAQcVSqUpVtWzYTEhKCj483t27doEjRjy3qv+k4xKvJcIUEprUep5TyxXRD2RJMz8n9C9PY1RNAGkw3YgGsBEaY611SSg0HdmHqRA7AdNNYbA0FFiilHICuWmsvpVQXTDeF5TUv8xBwwFplrfUtpVRdYCwwG3hsjuGfF4pOwHQD2X1MWehxxJy5ThQGg4F+A3+gR7dOhIeF4+7ZKPKmhqmTf6NQ4SJUrlKVP/84xqQJ41EKPvm0FAMGDY5cRse2rbh+/RqBz59Tq1olhgwfQdlylnf/7t29i0KFi0RmWPPnL0ATTzfy5stP/gIF4h1H/4GD6d6tM+FhYbh7NiJPZBwTzXGY3rxNWdy6rz22acXyJbi5e+Lg4EC+/PkJCgqisacb5StUJM1rZLejGtj3G/788w8ePXpIneqV6NajNx4NG7PKazkAjZs25/r1qwwZ1B+FIleevAwe9t93jfQdMIhBA74jNDSUD7O5MvTHn62uJyLDGzGeuHad+jRr1AAXlyy069A5XjG86MiCsebHfmnSfZiLUs1N4yL/vXWZK4e28nnLPqRI7UTh2s3YPuZrAIrUaR5501iROi3YPbEfytZA6gyZKd36a6vreeBtGjGUwdU0TD17ycpsGdmLVOkzUbBa3C/7Dx3wDU8eP8JgMNDn2+9xdDJt24jHgbk1bErrjt0Y/eMgOrfyRGvo0uMr0qZLz5/HjjD9t7EopdBa07RVO3LlyWd1PetXLadm3QakTOlArjz5CA4KonMrTz4rUyFynfExoO83/PWHaZ+qXa0SX/S03KcA9u7eSemy5XBIFXlRjHv37jFkUH/CwsLQWlOjZm0qVqpidT0Rx3dEhjV/gQI0NR/f+fLH//iOzXnqm+/68eOQH1i8cAFKKYaPGBnn43zm9Kl07toNGxsbypYrj9eyJTTxbEDjps3iHcOA7wfTvWtnwsPD8IhyjpoyaSKFCxehctVqeDZqzPf9v6N+7RqkSZuW0WPHx3ldM6ZNoXPXL8wxVGD5sqU08nCjSbPXe9pIQsaRJ09eataug2eDutja2jJw0ODIYWOJGUdCS8yxsrGhlEqHqT9SBNOV547ARUz39+QAbgBNtdYx3lyjIu76E+Jd9jz0/d9RVcKP6HgrwsLf+03B2P1v9TtTEkznUpYZ5fdR+tQJ8wSDt+kd6x+8tqT4hQDvq5SGxH3TqDv9eKKd3Ld88dkrY1NKLQAOaq1nK6XsMQ3pHAg80FqPUkr1B9JrrWO821qGKwghhBBCJHPv0nNylel7ByoCcwC01iFa60eYvlgr4htaFvCKK8nSyRVCCCGEEIlGKdVVKfVnlJ8X7y7PCdwD5iml/lZKzVZKpQZctNYRd3D6Yvpm2RjJmFwhhBBCCJFotNYzAevfp2xiAD4BemutjymlJgL9X1iGVkq9dIiFZHKFEEIIIZI5lYj/YsEH8NFaHzP/vQpTp9dPKZUVwPy/fwz1AenkCiGEEEKId4jW2hfwVkrlN0+qBpwDNgDtzNPaAetfthwZriCEEEIIkcwl5pc0xFJvYIn5yQrXgA6YkrNeSqlOmL419qVf0SmdXCGEEEII8U7RWv8DlLQyK9bfhyydXCGEEEKIZO5d+zKIhCBjcoUQQgghRJIjmVwhhBBCiGQuCSZyJZMrhBBCCCGSHsnkCiGEEEIkczZJMJUrmVwhhBBCCJHkSCZXCCGEECKZS4KJXMnkCiGEEEKIpEcyuUIIIYQQyZw8J1cIIYQQQoj3gGRyhRBCCCGSuSSYyJVMrhBCCCGESHqkkyuEEEIIIZIcGa4ghBBCCJHMyZdBCCGEEEII8R6QTK4QQgghRDKX9PK4kskVQgghhBBJkGRyhRBCCCGSuaT4ZRDSyRXvhaQwIF7rt92ChGFr8/5viy6fZX/bTUgQK075vO0mJIhupXO+7SbEXxI5vpPkNWuRbMW7k6uUSgvk1lqfSID2CCGEEEKIRJYE8hcWEmJMbnXgjwRYjhBCCCGEEAlChisIIYQQQiRzSXFMrjxdQQghhBBCJDmSyRVCCCGESOaSYCJXMrlCCCGEECLpiTGTq5QaHstlFEigtgghhBBCiLcgKY7JfdlwhUFxWE5SeUKgEEIIIYRIAmLs5GqtZSiDEEIIIUQyIM/JjYFSKlVCLEcIIYQQQoiEEK9OrlKqilJqHuCbQO0RQgghhBAi3uL8CDGlVF6gLdAGcAWCgTUJ3C4hhBBCCJFIktuNZ5GUUmmB5kA74HNAYbrZ7BfgF6314zfWQiGEEEIIIeIoxuEKSikbpVQ9pZQXpuEI0zBlbkcD5TB1dLdJB1cIIYQQ4v2mEvEnsbwsk3sHyAw8B1YBC4FdWmttzuwKIYQQQgjxTnpZJ9fZ/P9xYB2wT2stz8MVQgghhEhibJLgmNyXPV2hPDAL+ATwAnyVUlOVUqUTpWVCCCGEEEK8phg7uVrrI1rrbkAWoBWmjG4X4DDwD6Ybz9InQhuFEEIIIcQbpFTi/SSWVz4nV2sdrLVerrWug+nGs/7AU0xjh1crpXYppVq+4XYKIYQQQggRa3H6Mgitta/WeozWuihQEpgKfAwsehONE0IIIYQQb55SKtF+Estrf+OZ1vqE1roP8AHQMOGaJIQQQgghRPzE+HQFpdTLxuuGR/ndCKxP4HYJIYQQQohEkgQfrvDSTK4RCLX2o5QyKqXuKqWWKKUKJkI7hRBCCCGEiLWXPSd3OKYnKFhji2mYQh3ATSlVRmt9NqEbJ94spdQ+YLHWevZbWv9QII/WuvXbWL8QQgghTJLic3Jj7ORqrYe+qrJSygHTI8WGAE0TrlkiLpRSNwAXIAzTky+2Ab201k/f4DqH8kIH9W13moUQQgghIrz2jWcAWutA4FegcoK0RsSHm9baESgOlAAGvN3mJK7DBw/QoF4t6teuwZxZM62WWTh/Hp5udWns6UaXju24c+d25LzuXTtRvnRJevXoFuM6xo8bQ2NPN74f0Ddy2qaN61m8cH7CxXHoAO71a+FWpwZzZ1uPY/26NVSpUJqmjdxp2sidNatWRs775OOCkdO/7PWF1foTfh1DE083BkWJY/PG9SxelDBxxCaGCLt2bqd4kfycPXMagNDQUAYN7EdjTzc83eowZ9YMizohISH06NaJRh71WbF8SeT04UN/4Py5hLugtHr5Yjq28KRDcw9WLYv5ATL//PUHXVo3pkNzD776on20eWFhYXRt04SB3/S0Wnfm5F/p3KohI4cOjJy2c+vGl64vNsLDw1g1vCdbfxsSOe3Mng0sG9iRGV3qEBjwOMa6M7vWY9Wwnqwa1pNtk4dGTtdac3ztfJZ/35kVP3Tl9G7LWzF8r5xl5dDurB7Rh8d+puMr+PlTNo8fiA4PtygfW0N/GEi1SmVp4ukWY5k//zhGxTIlad7Yg+aNPZg5bUrkvKWLF9LE043GHvVZsmiB1frLlyyiiacbvbt3JTQ0BIC/T/zF2F9Gvna7rcVRtVJZGr8kjuvXrtG2VTM++6QoC+fPiTZv6eKFNPZ0o9FL4li2ZBGNPd3o9YbiiM25NiQkhO/+9xX1a9egVfMm3L7tA5iP7wH9aOThhsdLju/uXTvR0L0+K5ZFOb6HJOzxHZ84AObMmkH92jVoUK8Whw8dfGtxJKRk+ZzcWPAB0ibAckQC0Fr7AtsxdXZRSpVWSh1RSj1SSp1USlV+oUpupdRxpdQTpdR6pVQGc73KSimfqAWVUjeUUtWVUrWBgUAzpdRT83J/AioAk83TJpvrTFRKeZuX/5dSqkJCxxwWFsbPPw1n6vTZrN2wmW1bNnH1yhWLcgUKFmSp12pWrd1IjZq1GD9uTOS89h07M2Lk6BjXERAQwIXz51i1diN2dnZcvnSRoKAg1q9dQ7MWrRIsjpEjhjNl2mzWRMRx1TIOgJq16+K1ej1eq9fTsHGTyOkpUqSMnD5x8nSrcZw/d46VL8axbg3Nmsc/jrjE8OzZU5YuXkjRj4tFTtu5YxuhISGsWruRpV5rWLVyRbQ3FoAjhw9S4pNPWblmA5s3bgDg4oULhIeFUbBQ4XjHAHD96mU2r1/N1HlLmb14Fb8f3s9t71sW5Z4GPGHi6BGMGDuJecvXMeTncdHmr1mxmI9y5LS6jqdPA7h88Tyzl6zBzs6Oa1cuERwUxLZN6/Fo0jxe7T+zaz3ps34UbVqWPIWo/81IHDM6x1DLxNbensZDptB4yBRq9xoaOf3ikZ08fXCfZj/OpNmPM8lTqpJF3ZM71lCnz3DKNuvKuf1bADixeRkl6jZH2bz+242buyeTp816Zbnin3zK8lXrWL5qHV27mz5YXLl8ibWrV7JwqRfLV63j4P593Lp106Luli2bWLF6PcWKl+DI4UNorZk1Yypdvuj+2u22FseUV8SRNm1a+g0YRNv2HaNNv3L5EmtWr2TRUi9WrFrHgRji2LplE15vKI7YnmvXrl5JmjRp2LRtJ63btmfCr2MB2Ll9GyGhIaxet5FlXmtY5WXl+D5kOr5Xrd3ApijHd1h4wh3f8Y3j6pUrbNuymTUbNjN1xmx+HjGMsLCwRI9DvFpCdHLzAvcSYDkiASilsmEaK31FKfUhsBkYAWQAvsX0BR6Zo1RpC3QEsmK62fC3V61Da70N+BlYobV21FoX01p/DxzENEzCUWvdy1z8D0wd7gzAUmClUipl/CP9z5nTp3B1zU42V1fs7O2pXbce+/butij32eelcXBwAKBoseL4+/pGzvu8dBlSp04d4zpsbBRGoxGtNUGBQRgMBhbMm0OLVm2ws7NLuDg+MsdhZ0+tOvXYt8cyjviIGkdgkCmOhfPn0LxlwsQRlximTJpI+45dsLdPETlNKUVgYCBGo5Hg4CDs7OxwdHSMVs9gMBAYGBQZB8DUyRPo0fvLeLc/ws0b1yhYuCgpUzpgazBQrERJDu7bZVFu9/YtlK9SDZcsWQFInyFj5Lx7fr78fvggdd0bWV2HjbL5b58KCsRgMOC1ZD6eTVtgMLz+tnj64B43Tx+nQPla0aZn+igPTplcXnu55/Zt5lO3lpGdVYc06SzK2NgaMIYEYwwJxsbWlsf+d3j64D4f5P/4tdcL8GnJUqRN+3q5lOvXrlGk6Mc4ODhgMBj4tGQp9uzaaVlQa4xGI4FBgRgMdmzetIFy5SuSNm26eLU9qtjEkSFjRgoXKYrBEH00YWzj0OY4gt5AHLE91+7ds4cG7p4A1KhZi+O/H0VrbTq+n/93fBvs7HBM/cLxbWcgKCj68T1l0gR6JuDxHd849u3dTe269bC3tydbNldcXbNz5vSpRI9DvFq8OrlKqSyYMnrbE6Y5Ih7WKaUCAG/AH9M46dbAFq31Fq11uNZ6J/AnUDdKvUVa6zNa62fAD0BTpZRtQjVKa71Ya/2v1tqotR4HpADyJ9TyAfz9/MiSNUvk384uLvj5+b20ztrVqyhXoWKs15E6tSPlK1SkWSMPMmXOjKOTE6dPn6Jqteqv3e4X+fv7kSXLf3G4uLjg7289jt07d9DE041vv+6D7927kdNDQoJp2bQhbVo2Zc9uy05Z6tSOlK9YkWaNPcgcEcephIsjtjGcP3cWP19fKlaqHG169Rq1cHBwoEaV8tSuUYW27TtavDmXLlOOO3du06ZlU1q0asO+vbspULAwzs6v34F7Uc5ceTn9zwkeP35EUFAgx44cxN/P16Kc962bPH3yhK+7d6Bb26bs2LIhct6U8aPp1utrbGJ4GmOq1Kn5vGwFurZpQsZMmUnt6MT5s6cpX6lavNp+ZMUMSjfu9NqZ07DQEFaP6MPan7/i+t9HIqc/uXeXq3/sZ/WIPmyZ+EPkcISoStRpyt65Y/lnqxeFq7jxx7oFfObR9rVjiavTJ/+hWSN3en3RhatXLgOQO29e/j7xJ48ePSQwMJBDB/fj53vXom7TFq1o16oZvnfvUrxECTasW0PT5u/Ol3lai8PXShzNWrSi7RuKI7bnWtN5wPTBz2Aw4OjkxKNHD6lesxYOqRyoXrk8tapXoV37jqRNly5a3dJlynHn9m1at2hKy1Zt2LdnNwULJezxHd84/Pz8cIl6nsvigv8L9RMjjoSWFL8M4mXPyV34knoRT1coDdzH1KESb5eH1nqXUqoSpoxpJiA70EQpFXUAmB2wN8rf3lF+v2menymhGqWU+hbohGl/0UCahFz+69i0cT3nzp5h7oLFcarXoVMXOnTqAsDQwd/Ts1cf1qxaydEjh8ibLz9dv+jxJpproVLlKtSpWx97e3tWeS3nh+/7MWuu6XDdsmMvLi4u+Hh706VTO/LmzYfrR9EvW3fo2IUOHU1xDBv8PT0i4jh6iHz58tOl25uNIzw8nLGjRzH8J8sxgmdOn8LG1oYdew4S8OQJHdq1pHTpsmRzdY0sYzAYGDXaNCwgNDSUHt06MWHSVMaOHonv3bvUb+BO5Srx6yhmz5mL5m070rd3V1I6OJA7XwFsbCw/+4WFGbl04Txjp8wiJDiYXp1aU7DIx/jcukm6DBnIV7Aw//z1R4zrad6mI83bmC5Lj/1pCO279mTz+tX8eewIufLko03HmMeIW3Pz5DEc0qQjc/a83Ll46tUVrGg1agGp02fiyb27bBzXnwwf5iCt8weEGUOxtbOn0aDfuHbiMPvmj8e939hodTN9lBvPgRMAuHPpNKnSZkBrzc4ZI7GxtaVM0y6kSpP+tdr1KgUKFmbzjj2kSpWaQwf2882XvVi/eTu5cuWmfccu9OjaCQeHVOQvUBAbW8ttWd/Nnfpu7gDMnDaFFi3bcOTQQTZtXIeLS1a++a4fNvEYchFfUeNIaY7D9hVxzDDHcdgcR5a3HMeZ06ewtbFh596DPHnyhA5tW1K6jJXje8x/x3f3rp2YOHkqY34xHd9uDdypXDV+x3diSCpxvO9etqdXxDTG0trPZ4A9MA74RGvtE9NCROLSWu8H5gNjMXVgF2mt00X5Sa21HhWlimuU3z/C9Czk+8AzIFXEDHN2N+owB2uPl4s2zTz+ti+mJ2+k11qnAx4DsfoYp5TqqpT6Uyn1Z0w3BoDpU7jv3f+ybP5+fri4WP+0/PvRI8yeOZ2Jk6dhb28fm2ZYOH/+HFprsufIyY7t2xjz60S8vb25efPGay0vgrOzC75RhlD4+flZ/dSfLl36yLZ7NmoS7SaGiLizubpSstRnXLhwLsb1XTDHkSNHTnbu2MaYcfGPIzYxPHv2jKtXLtG5Q1vq1KzK6VP/8FXv7pw9c5qtWzZRrlwF7OzsyJAxI8WLf8LZs6djXJ/X8qXUb+DBqZMncXR04pex41m0YN5rtz+qug0aMmOhFxNnLMDJKQ2uH2W3KJPZ2YVSpcvi4JCKtOnS83GJT7l6+SJnTv7NkQN7aeFRix8Hfcfffx7n5yH9Y1zX5Yvn0Vrjmj0H+3fvYMjP47jj442PlTGXL+N79Rw3//mdJf3bsWvmKO5cPMnu2TGPNbcmdXrTZ9A0mbPyQb6P+df7KgCO6TORs0Q5AHKWKMuD29djXIbWmr83L+OT+i35a+MSSjfuSMEKtTlj5Wa1hOLo6EiqVKYhR+UrVsJoDOXhw4cAeDRszFKvNcxZsBinNGnInj1HjMu55+/HmTOnqFKtOosWzGXUmPE4pXHi+O9H31jbY8vTHMfcBYtJ84o4/P39OBsljl8SII7YnmtN5wFTltloNPI0IIB06dKzdfMmypY3Hd8ZM2akeIlXH99u5uPbycmJ0ePGszABju/4xuHi4oJf1POcrx/OMbznvMk4EppNIv4klpd9q1kOrXXOGH7yaq3Laa0Haa1lPO67ZwJQAziC6TnGtZRStkqplOYbyrJFKdtaKVVIKZUK07ORV2mtw4BLQEqlVD2llB0wCNNQgwh+QI4XvhnPD8gV5W8nTON87wEGpdRgTJncWNFaz9Ral9Ral+zUpWuM5QoXKcqtWzfw8fEmNCSEbVs2U6lKVYty58+f48dhg5k4eRoZM2a0sqTYmTJpIj17f4nRaCQ83HSzgY2NIigw6LWXCf/FcdvHm9DQELZvtR7HvXv+kb/v37uHnLlyA/Dk8WNCQkx3Uz98+IB//j5Brtx5XhpHj95fEho1DhW/OGITg5OTE/sOHWPrjj1s3bGHoh8XZ8KkaRQuUpSsWbNy/PgxAAKfP+f0qZPkzJnL2qp48vgxB/bvw62BB0FBgdjYmC6DBQXFbztEePjgXwD8fO9ycN8uqtWqa1GmXMWqnD75N2HmMZDnz54me45cdOn5FV6bdrNs3XZ+GDGGEiU/Y+CwURb1I8ybMZkO3XoRFm2fsiE4jrF83rADrccsptWoBVTv2p8P8hejWue+r65oFvwsgDDzHfmBAY/xvXou8ga2HMXLcOfiSQDuXjpNWucPY1zOpaO7cC1aipSpnTCGBKOUDUrZYAwJjlM8cXH//r3IsY9nTp9Ch2vSmS+FP/jXtC3v3r3D3l07qVO3fozLmTr5N7r37ANAcHAwSilslE2C7VfxETWOPa8Rh1I2BMYjjtieaytXqcqG9WsB2LljO599XhqlFFmyZuX4MdPx/fz5c06fjMXx7W46viMucyfEdohvHJWqVGXbls2EhITg4+PNrVs3KFLU+rjzNxmHeLWXDVfYA/TQWl9IxPaIBKC1vmcebtIHcAdGA8swPUf3OBD1NttFmDK/BYD9EfO01o+VUj2A2ZiGp4zG9CSNCCsxjfn9Vyl1XWv9CTARWKCU6m5e7teYntl7CVNmeDzRh0ckCIPBwIDvB9O9a2fCw8Pw8GxEnjx5AVNHrnDhIlSuWo3xY0fz/PlzvvvaNPA/S9as/DbF9ASC9m1acuP6NZ4/f06NqhUZOvwnypW3fBDEnt27KFy4SGR2Mn+BgjTycCNfvnzkL1Ag3nH0Hzj4/+zdd3gUVRfH8e9JQu8IhCq9CNgRUTo2lN47qCAIimJDsKDYQEXFhiJip6s0QVEpooL1VVGqqHQISG/p9/1jlpiQBBISstnN7+PD4+7M3NlzM7O7d8/ce4fBgwYQHxdHu0T1mPDKi9SuU5dmza9i2gfvs2zZEsJCQylcpAiPPeFd9v/777944rFHCDEj3jlu7n8LVVNp5C5Z/CW1E9ej5nl07tCG6hmsR1rrkJpuPXox6qGRdGzXCpyjbfuO1KiZcjwTX3+VAQNvJSQkhCsbNmbGtKl07tCGLl0zNjPBCY+OuJtDBw8QGhbGnfc9SMFC3u+zeR/PBKBtx65UrFyFyxo0ZECvTlhICDe07UjlqtXT9TrffLWYGufVoURJb9aDqtVr0b9nB6pUq0HVGpnXff33xXP57bNZHDu0nw9HD+Hc8y+jab9h7Nm0gTVfLaRpv2Hs37mVrz942ZvjxzkubtmVYmW9DPZF13dlyZvP8PuXcwjLk5em/Yal+DoxUZFsWPElNwx7EoALrunIpy+NIiQ0jKtuuf+MYh85/G5+/vFHDhzYT8urmnLrbUNp37EzH86cDkDnrt358vNFfDhzOqGhoeTJm5cxzz6X0Pfv3rvv4OCBA4SFhXH/g6MoVDjl39rr1npXPk6Mfm95Q2u6dmxLeOnS9Lt5wBnFntiIRPW4zlePDh07M8tXjy5du/Pvv3vo1a0zR48ewUJCmPL+e3w0dwEFCxbk3rvv4ICvHiPSUY/rb2hNl45tKV26NDdmoB5p/azt0KkzD464j9Ytr6FwkSI8M+4FALr73t8d2nrv73YdTvH+fi3p+3v6tKl0at+GLt0y/v7OaD2qVavOtS2vp0PbGwgNDeWBh0al2HXkbNcjs2VlX9msYid++SZbYRYPNHDO/ZC1IYkkFxmb6t33AkYqbzXxg31Ho/0dQqaYsSo4eooNapDyNGuBJFiaByEhwVKTwJc3LGtPqzvmrMuyb6mX2tfKkrqd6ra+IiIiIpIDBOPvG/8NFRUREREROUtOl8kdaGap92z/j3POaRoxERERkQAUjJnc0zVybz7N+hMcmitXRERERLKJ0zVyNfBMREREJMgF4+wK6pMrIiIiIkFHsyuIiIiI5HDB2CdXmVwRERERCTqnyuRWBvaZWV7nXIr3nzOzvEAu59zhsxKdiIiIiJx1Qdgl95SZ3MrAfqDRKbZpBOw3syszNSoRERERkQw4VSZ3EDDfOfdlahs45740sznAbcCKTI5NRERERLJASBCmck+VyW0EzEzDPj4EGmdOOCIiIiIiGXeqRm44sDUN+9jm21ZEREREJFs4VXeFQ0DxNOyjOHAkc8IRERERkawWjNNtnapOvwKt07CPNsAvmRKNiIiIiEgmOFUj923gRjNrn9oGZtYR6Ae8lclxiYiIiEgWMcu6f1kl1e4KzrkpvgbuR2b2CTAf2OxbXREvg9sa+Mg5N/VsByoiIiIiklanu61vN+AB4G68Rq3zLTfgADAKGHO2ghMRERGRsy8YpxA7ZSPXORcPPGFmTwP1gAq+VVuBn5xzMWbW1Mz6OeduPsuxioiIiIikyekyuQA452KAlb5/mFk14GEz64PXdeEYoEauiIiISAAKwkRu2meMMLMiZjbQzL4F1gMP4t32dzBQ9izFJyIiIiKSbqfM5JpZCNASbwaFNkBeYAfwKt6tfIc555af7SBFRERE5OwJCcJMbqqNXDN7DugJlAIigdnAu8CXQGHg9qwIUEREREQkvU6Vyb0LbzaFhcCNzrm9J1aYmUu1lIiIiIgElGCcXeFUfXInA4eBVsB6M3vFzOpnTVgiIiIiImcu1Uauc+4WoDTQC/gJGASsNLO1wP38N2euiIiIiASwHHXHMwDnXCQwDZhmZmWAPkBfYIRvk7FmNgH40LetyFlxNCrW3yFkWLBcCvpj2yF/h5BhwTLAonTB3P4OIVOce8t0f4eQYatf7uzvEDJFvlyh/g4hw/LkSvPEURLk0nwmOOd2Oueecc7VBerjzbBQHXgP2HmW4hMRERGRsyzEsu5fltXpTAo5535yzg3Fmx+3E7AsM4MSEREREcmINN3xLDW+O6HN9v0TEREREckWMtTIFREREZHAZwTJYIVE1DtbRERERIKOMrkiIiIiOVywzDqTmDK5IiIiIhJ0lMkVERERyeGUyRURERERCQDK5IqIiIjkcBYkd+VMTJlcEREREQk6yuSKiIiI5HDqkysiIiIiEgCUyRURERHJ4YKwS64yuSIiIiISfJTJFREREcnhQoIwlatMroiIiIgEHWVyRURERHI4za4gIiIiIhIA1MgVERERkWzFzELN7Bcz+8T3vLKZfW9mG81shpnlPt0+1MgVERERyeHMsu5fGt0JrE30/GngBedcNWA/0P90O1AjV0RERESyDTMrD7QC3vQ9N6AF8KFvk3eB9qfbjwaeiYiIiORwIWSrkWfjgeFAId/zc4ADzrlY3/NtQLnT7USZ3BzCzBqb2fpTrH/HzJ7IyphEREQk5zGzgWb2U6J/AxOtaw3sds79nNHXUSY3AJnZJiAciEu0+B3n3O2plXHOfQ3UPMuh+U3HVteQv0ABQkNCCA0N460pM5Nt87+ffuD+u4dStqz3469pi6u5eeAQAGZMfZ95sz8E52jboTPdevVNVn7W9CnM+WgmpUuXYezzL5ErV25+++Vnli3+gjvvHZEp9Wh/w9UUKFCAEF893pk6K9k2zjmef+YpVn67nDx58/Hw6KeodV5tAK68tC5Vq1UHILx0Wca9+Gqy8q+8+Bwrv/2aGjVq8cgTYwH4dME8Dh44QPcU6n06+/ZEMPmF0Rw6sA/DaNKyPVe37ZawfvH8mSxd8BEhISGcf9mVdLlpaJrLv/70g0Rs3wLAsaOHyV+gEI+89H6S8ru2beaNcaOIi4ul720jqFrrfOLiYhn/yDBuf2gcefLmTXM93nx+NAcP7MPMaHpde65p58UxZ8okli+aR6EiRQHo1HcwF1x2ZbJ93Hdze/Lm845fSGgoj4x/B4DXnn6QXduS1mP0y0nrsXPbZt549r96VDvPq8fzo4Zxx8Npq0dsdDRvPzaMuJgY4uPiOO/yJjTvcqNXh9eeZvPaVeTJXwCA9rcOp3Slasn28cGYEWzbuIZza9al5/CnEpa//eidREUeB+DowQOUq1aT7vc8nqTslvV/sGDyeELDctFp6IOcU6Y8kUePMOvFx+g9YiwWkva8yq3X1aRP06o451iz7SBD3/yOqJh4Xh90BRdXLk5MXDz/+3sfd7/zA7FxLknZ8ufk5707GhNiRq6wECZ9sYF3lm4E4MJKxXhlQAPy5g7ly992MHLK/5K99i1XV6df82ps23uMPi9+TUxcPJdXL0Gbyyrw0NRf0lyHJH+bTf8w+oF7E57v2LGNmwfeTpeefZJsd/jQQcY+/jA7tm0ld+483P/w41TxvacPHz7Es088wj9/bQSD+x9+nLoXXJSk/EczpjD/41mUKl2GJ8e9RK5cuVj16/9YvuQLbr/7/jOK/YSoqCgG3dyH6Jho4mJjuerq6xg4ZGiK236x6FPenOh9/lSvUYsnxo5LWHfkyBG6d2xN0+ZXcd/Ih5OVfXjkffy1cQONGjdjyB13ATB50mtUrVqdZi2uzlAdTvj26+U8PfZJ4uPi6dCpC/1vGZhsm+joaB4cOZy1q1dTpGhRnnnuBcqVK8+BA/u5Z9gdrP7jD9q278ADD41K8TVGDr+HP//cQJOmzblj2N0AvPH6BKpVr0GLqzKnHpkpK+8F4Zx7A3gjldUNgbZmdgOQFygMvAgUNbMwXza3PLD9dK+jRm7gauOc+9LfQWQnr0x8m6LFip1ymwsvupRxL01IsuyvjX8yb/aHTH5vOmG5cnH37YNo2Lgp5c+tmGS7zxd+wvszZvPuW2/w/YpvadikGW9Pep3HxjybqfV49Y13TlmPld8sZ+uWzcya+xmrf1/FM0+N5q33ZwCQJ08e3p8xO9WyRw4fZv3aNUyZOYcnRz/Mxj83UL7CuSyYN5vxr6T2eXNqIaGhdL35DipWq0XksaM8fteN1L6oPmXPrcy6VT/z6/fLeeTl98mVKzeHDuxLV/lb738yYbsZk18kf/6Cycp/9dlsetxyF+eEl2X6G88z5IGxLFv4MQ2atUxzA/dEHN36e3EcP3aUx4bdSO2L61Pu3MoAXNu+Oy079jrtfoY/9WpCY/iEwYnqMf3NF8lfIIV6fDqbHgPvokR4Waa98TzVzhvL0oUfc0XztNcjNFcu+j30HLnz5iMuNpa3H72T6hfVp3x170fQNb0GUvvypqfcx5VtuhITFcnPiz9JsvymR19MeDzzhUepeWnyRv7KBbPodf8YDuzZxU9fzue6PoNZPvsDGrfvma4Gbpli+Rh4TQ2uHLmQyJg4Jt/WkI6XV2TaN//w4cpN3DpxJQBvDL6SPk2r8vaSjUnKRxyIpOXjXxAdG0+BPGF88+T1fPbLdnYdOM64fpdx19s/8NNfe5lxT1OuuqAMi1ftTFK+8xWVaPzQp9zdpg4tzi/Nol93cG+7utzy2oo01+Fk51aqzOSpHwEQFxdH5xta0Lj5Vcm2++DtSVSvUYsnn32JzZv+ZvzTT/LCa5MBePm5sdS/oiGPPf0CMTExRPp+dCT25WcLeGvax3zw9iR+WPktVzZuynuTX2fUExn/nMqdOzcTJr1N/vwFiI2J4ZabenNFo8acf1JDe8vmTbz71iQmvTOFwoWLsG/f3iTrJ776EhddUi/F1/hzw3ry5M3D1FlzuX3QzRw5fJjIyOOs/n0V/W8ZnOE6gPf3f+rJx5g46W3Cw8Pp2a0zzZq3oGq1pD/6Zn80i8KFC/PJZ1/w6cIFjH9+HM8+N57cufNw29A72bjxTzb++WeKr7Fh/Try5M3Lh7PnM2jATRz21eP3VasYeOuQTKlHsHLOjQRGAphZM+Be51wvM5sFdAamA/2Auafbl7orBAkzy2NmB8ysbqJlJc3suJmVMrNmZrYt0bqLzex/ZnbYzGbg/VpKvL/WZvarb58rzOyCROs2mdm9ZrbKzA76pvLIm2h9O1/ZQ2b2l5m19C0vYmaTzWynmW03syfMLPSs/mHSYPM/f1On7gXkzZePsLAwLr60HsuWJP/94HDExsYSdTySsLAwPlswnysaNqbwSQ2as235V0u4oXU7zIy6F1zIkcOH+XfPnjSVtZAQYmNjcc4RFenVY+p7b9Oley/CcuU6o3iKFi9BxWq1AMibvwBlKlRi/97dACxb+DHXd+5LrlzeTC+FixZPV/kTnHP89M1i6je9Jln50LAwoqOiiI6KJDQsjGNHDvPbD99wRYsbzrge+XxxHDgpjoxyzvHjN4u5vMlp6hH6Xz2uTEc9zIzcefMBEB8XS1xcbLrTM1XqXkKefPlTXR917Cj/rP6FWvUaJq9DaBgx0ZHEREcRGhbGvogdHNq7h0q1L0pXDABhIUbe3KGEhhj5coey84DXoPsyUYP0f3/vpWyx5LHGxMUTHRsPQO6wEEJ8s9yHF8lLoby5+Okvr9E149tN3HBJ+WTlzSBXaAj5cocSE+foemUlvly1kwNHo9Ndj5T878fvKFu+AqXLlE22btM/f3FxvcsBqFipCrt2bmff3n85cuQwv/3yM63adQIgV65cFCpUOFl557zPqcjI44SFhfH5p/O5/IrGFC5SJMNxmxn5fVcCYmNjiY2NwVI4v+Z8PIvO3XpQuLD3msWLn5Owbu2a1ezb9y8Nrkh+/gCEhYURFRlFfHw8sbGxhISGMHHCywwcnOqFynT74/dVVKhQkfIVKpArd25a3tCKZUsXJ9tu6ZIltG3XAYBrrr2OH75biXOO/Pnzc8ml9ciTO0+qrxEWlouoyMiEeoSGhDDh5ZcYcnvKme/sIMSy7t8Zuh+428w24vXRnXzaOp3xS0m24pyLAj4GeiRa3BX4yjmX5JvaN7fcHOB9oDgwC+iUaP3FwFvAILwTaSIwz8wSv6O7Ai2BysAFwI2+svWB94D7gKJAE2CTr8w7QCxQDbgYuBYYcMaVTlonht12Czf17MKcj5J3VTjhj99/pW+3Dtx9+yD+/svL/lSpWo3ffvmZgwcOEHn8OCu++ZrdEbuSle3UtSe39OvBrl07Of+iS1gwbzaduvZItl1G63HHkAH069k51Xrs2b2bUqVLJzwvFR7Ont0RgHd57caeXejftztfLU3eUC9QoABXNmpC3+4dOadECQoWLMTqP1bRtHnmXDr7N2IHW/7aQJWa3m+tiB1b+HP1bzx5z808M2Iw/2xYk67yJ/y5+lcKFy1OeNlzk5VpfkNnFsx6h7deeIwbuvRj/oy3uKFrP0LSkTlMMY6/k8ax+JNZjLq9F2+Nf4KjRw6lWM7MeG7UHYy+sx/LPpuTbP2GE/Uol7weLVp1ZsHMd5j8/GO06tqPedPfolWX9NcjPj6O10cM5NlBnahy/qWUr3ZewrolM97iteED+Oy9CcTGnFmDbd1P31K5zsUJ3R4Sa9SuB7MnPM03c6dS/9r2LJkxmeZdb0r3a+zcf5xXPl3Hb8+3Zc2L7Tl0LIZlfyR9T4aFGl2vrMTi33emuI+yxfOz/InrWfVCO15asJZdB45Tplh+duw/lrDNjn3HKFMsX7Kyb375J4tGXUv5cwrww5976Nm4CpMXb0h3PVKz+PNPueq6lH+8VK1ek6997921q38nYtdO9uyOYOf27RQtWoyxox+if6/OPPPEKI4fP5asfIeuPRh8U09279rJ+RdezKfz59Cha/dMiz0uLo5eXTtwXYtG1G9wJXXPvzDZNls2b2bL5k0M6NeTm/t0Y+W3XwMQHx/Pi889zR13D091/5WrVKVYsWL06d6Jxk2bs23LFuJdPLXOq5NpddgdEUHpMkk/QyMiIpJvtzuC0qXLAF7ju2ChQhw4sD9Nr1GlalWKFStO984daNKsOVt89TivdubVIydwzi1zzrX2Pf7bOVffOVfNOdfF1+45JXVXCFxzzCw20fP7gKl4DdIHfct6+p6frAGQCxjvnHPAh2Z2d6L1A4GJzrnvfc/fNbMHfOW+8i17yTm3A8DM5gMX+Zb3B95yzn3he77dt004cANQ1Dl3HDhqZi+ceK30Vv5kr7/1PiVLhbNv316GDR5AxUpVuPjSpJfDataqzccLviB//gKs+GY5I+4eysy5n1KpSlV639ifYUNuIW++fNSoWSvFhsX1rdtyfeu2ALz1xgS69OjNyhVf8+kn8wgPL83Qu4dnqGEFMPHtDyjlq8cdt6Zcj1OZvfBLSpUKZ/u2rdw28CaqVqtB+QpJG1R9buxPnxu96QWfHP0wtwweytyPP+SH776lavWa3HzLrWcUe+TxY0wYM5Jutwwjn68BFBcXx9EjB3lg3GT++XMNE59+kDFvfpxi9iel8id8v/xz6qeQ/QQ4p1Rpho95DYCIHVvZ/+9uypSvxJvPPUpsbAztew+idAqNylPV49WnRtIjURzNb+hI2+43gxmzP5jIjDdf4uZhDyUrO/LpiRQrUYpDB/Yx7qE7KFO+IjXrXvxfPb76PMUs7ol63D82UT327qZMhUpM8tWjQxrrERISyq1j3yDy6BFmPD+K3Vv/oVSFylzVfQAFixYnLjaGTyY9z7fzptO0U/r7YP+xYgkXN0+5gVa6UjUGPP4KAJvXrqJg0XMAx4cvPk5IaCjX9r6Vgilk809WJH8ubrikPJfcO5+Dx6J5+7ZGdLmyErNWbErY5tm+9Vi5fg/fbUj5KsaOfcdo8tCnlC6aj/fvbMy8H7emuY4zV2xipu+17m1Xhze+WM/VF5SlW8NKbN93jIen/YJzp95HamJiYlixfBkDbxuW4vpe/Qbw0nNj6d+zE5WrVadajVqEhIQSFxfLn+vXcud9D1C77gW8NG4MU9+ZTP/BSTOD193Qlutu8D6n3pn0Gp269eL7b79h0cJ5lAovzZBh92Xocyo0NJQpM2dz+NAhht89lL82bqBqtRpJtomLi2Xrls28/ua7ROyOYNDNfZg2ay6fLpzPlY2aEB5eOpW9e+4e/sB/j+8YzMiHRvPWpNf5c8N6Lm9wBe07dT3j+LPS8JEPJjweOuRWHn50NJMmvsaG9etocEVDOnXJXvUIycpOuVlEmdzA1d45VzTRv0nAUiC/mV1uZpXwGp4pddAsC2z3NXBP2JzocUXgHl9XhQNmdgCo4Ct3QuK0yjHgRCfDCsBfKbxmRbyG9c5E+5wIlEqtgolHX7771qTUNgOgZKlwwLss1qT51axd/XuybQoULJhwqe3KRk2IjY3lwH7vV3mb9p14e+osXpv8HoUKFaZCxUqpvtaePbtZ88cfNG1+FdPef5fHxz5HwUKF+emH704ZY1qUSlSPpi2uYs3qVcm2KVmqFLt3/ffn3x0RkVD/E+XLla/AJfXqs2Hd2mTlT1i/bg04R8VKlVjy5SKefOYFtm/bwpbNm9Idd2xsLK+NGUmDZtdx6ZXNE5YXK1GKS65ojplRpUYdLCSEI4cOpLk8eF+Y/1u5jMsap9w4TGz2+6/Toc8gFs+fSeNr29LlptuZP+3NdNXj1aeSx1Gk2DmEhIYSEhJC0+vapZqRLlbCO50LFy3OJVc0TbLdiXqk1lhP7OP3X6dj70F8OX8mTXz1mDc17fUAyFugIJVqX8TG334EoFCxczAzwnLl5qJmLdn+17p07Q/g2KGDbP9rHTUubnDK7ZxzLJ/9AU079uarj97n6p4DuaRFK75flHp/8cSa1inN5j1H2Hs4itg4xyc/b6V+tRIJ6+9rX5cShfLy0LTkg8ZOtuvAcdZuO0iDGiXZuf9Yku4NZYvnZ+f+5P1aTyhdNB+XVDmHhf/bzpCWtej/6goOHouhae1TN9JO5fsVX1O91nkUP6dEiusLFCzIyEeeYPLUj3hw9BgOHthP2XLlKVmqNCVLhVO7rtdzrOlV17JhfepXRv7ds5t1a36ncbOrmDHlXR55ahwFCxXi5x8z/jkFUKhwYS69rD4rv/0m2bpS4aVp0rQFYblyUa5cec6tWImtWzbz+2+/MmvGVNpdfxUvvvAMCz+ZyysvPpfqa3y1dDG1zqvDseNH2b5tK2OefYHFX35O5PHUj1lalAoPZ9fOpJ+h4eHhybcrFc6uXd6VgtjYWI4cPkzRoqce95GSpUu+pHadOhw7doytW7fw7PMv8sXniziewXrI6amRG0Scc3HATLwuCz2AT5xzh1PYdCdQzpKm0xKniLYCT57UiM7vnJuWhjC2AlVTWR4FlEi0z8LOuVSv3Tjn3nDO1XPO1et38y2pvuDx48c4evRowuMfvltBlarJR43v/XcPJ9r1a/5YhXPxFClaFCBhYMSunTtYtvRLrr2+VaqvN2nCy9zi6x8WFRWJmRESYkRGRqZaJi2S1WPlCqpUrZ5su8ZNW7Dwk7k45/hj1W8ULFiIEiVLcujQQaKjvUvQB/bvZ9Wv/6NylZQOheeNCS8zcMgdxMbGEhfnTdQRYiFEpbMezjnefelJylSoxLXteyZZd3GDJqxb5c0Cs2v7FmJjYyhYuGiaywOs/fVHypSrRPESqf4eAmD97/+jaPGShJc9l+ioKCwkBLMQoqNOe0UrIY63X/TiuK5D0jgO7Ps34fH/Vn5FuYpVkpWPijzO8WNHEx6v/uWHJNut+fVHSpdPRz3K+ephId5xSUM9jh46QOTRIwDEREfx9+8/U6JsBQAO79+bUM91P35LqQqVT7u/k635fjk1Lm5AWO5T303zt+WfU/2iy8lXsDAxUZFYiGEhRkwaj8X2vceoV60E+XJ7Xfab1C7Nhh0HAejdtAot6pbmltdWpJpNLVssH3lzeWWL5M/F5TVKsHHXISIORnI4MoZ6Vb0+ot0aVuLT/21LeSfAyI7nM/Zj7wdzvtyhOBzx8S4hrjOxeNFCrro29X7Whw8fIiYmBoBP5nzEBRdfSoGCBTmnRAlKhpdmy6Z/AK9fb6XKqb+/J7/+MjcPSvo5ZWfw/k5s/759HD7kddWJjIzk++9WUrFy8vOoWfOr+PmnHwDvs2jL5k2ULV+ex8c8y/zPljD308XceddwbmjdjtvvvCfF14qNiWH6lPfoe2N/oiKjODGFa3x8XMLf50zVqXs+W7ZsYtu2rcRER/PZwgU0bd4ihXq0YN5c74fZF58vov7lDVK8CnUqMTExfPDeu9x48wCiIqMSymdGPTJbNrzjWYapu0LwmYrX33Yv/3VbONlKvL6xd5jZBKANUB8vEwwwCZhtZl8CPwD5gWbA8lQazYlNBj733Wt6KVAGKOScW2dmnwPPmdnDwBG8/rzlnXNfpb6709u3dy8j77kD8C6PX9OyFQ0aNgZg9oferAMdOndj6ZefM/vDGYSGhpInT14eGzMu4QPnwXuHcfDgAcLCwrj3/odSHNABsN6XGa3pm7Lr2pat6NO1PaXCS9Or32nvMHjaetx/94l6xHLt9a24wlePj2dNB6Bjl+5c2agJK75ZTue2LcmbNy8PPeqN3N/09988/eSjmIXgXDx9b7qFyik09gG+WvoltWrXoWQpr8FVo2YtenVpR9XqNahes1a64t645jdWLv2UcpWqMvoObzqkDn0Hc0G9K2l0dRvefukJRt3Wk7CwMG4eNgoz48DePbzz8lMMe/SFU5YH+GH5FykOOEvMOccnM99m0HBvqucmLdvx5rhHiIuPo/fg1Pv/JfanL47ylaryyFAvjhNThc16+xW2/P0nZlCiVBn63u5NGbd/7x7eeekp7hr9AgcP7OOVJ7wpmuLj47i86bWcf+kVCfv/YfkXqXZVSFyP+TPe5tb7vXo0bdmON8Y9QnxcHH2GnL4eR/bvZc5rzxAfH4dzjjoNmlLjEi+Gj195imOHD+Kco3TFqrQe4E3NtOOv9fy0eD5tB3rTW7396J38u2Mr0ZHHef62brQdeC/VLrwMgD9WLqVR21P374yJiuS35YvoPfIZABq06szUpx8gNCyMjren9pGU1M9/72Xej1tYOrolsfHx/L55P+8u8y4QPdfvMrbuPcpnD3t/y09+3sq4uau5qFJxbmxRjWFv/UCNskV4rMfFOOcwM179dB1rt3mN5Pve/YlXbrmcvLlDWbxqZ5KBbImdf66XsVu12bva8+HKzXzz5A1s33uMlxemfoXkVI4fP8ZPP6zkngceSbJ87kfe51S7Tt3Y/M/fjBn9IIZRqUpV7n/4sYTt7rz3AZ4YdT8xMTGULVeBEaOSTuF2wob1Xnw1anmfU1df14qbunegVHhpevS9+YxiB/j33z2Mfngk8fFxxMfHc/W1LWncxLviMXHCS5xXuy5NmrWgwZWN+G7lt3Tr2JqQkBDuuOvedGdAZ82YSqs27cmbLx/Va9QkKjKSHp3bcmWjJhQqnPLnc1qFhYUx8sFRDB44gPj4ONp36EQ13zRtr778InXq1KVZi6vo0KkzD464j9Ytr6FwkSI8M+6FhH1cf00Ljhw5QkxMDEuXfMnrb7yVbHYGgBnTptC2XQfy5ctHjZo1iTweSaf2bWjUuAmFM1gPOT1zZ9qxSPwmlXlyv3DOdfCt34g3oKy0cy7at6wZ8IFzrrzveT28xmw1YKFvH3865x7yrW8JPA5UB44D3wA3O+cO+15/wIkpzMzsUaCac66373kHYDReIzYCuM05t8jMigBj8RrVhYC/gaedc9NPV+e9R2MD/kQNlv5Of2xLedBVIMnA6N5sZfPho/4OIVMMffFrf4eQYatf7uzvEDJFvlx+n/Amw/LkCo6L1HnDsvYWZJN/2JJl37P965+bJXVTJjcAOecqnWZ9sp+TzrlleJMnn3j+E94MB6nt4zPgs7S8vnPu0ZOezyaFvsDOuYPAYN8/ERERkbNGjVwRERGRHC5ILjYmERw5fRERERGRRNTIFREREZGgo+4KIiIiIjlcMGY9g7FOIiIiIpLDKZMrIiIiksOl90YXgUCZXBEREREJOsrkioiIiORwwZfHVSZXRERERIKQMrkiIiIiOVyw3Ho+MWVyRURERCToKJMrIiIiksMFXx5XmVwRERERCULK5IqIiIjkcEHYJVeZXBEREREJPsrkioiIiORwuuOZiIiIiEgAUCZXREREJIcLxqxnMNZJRERERHI4NXJFREREJOiou4KIiIhIDqeBZyIiIiIiAUCZXBEREZEcLvjyuMrkioiIiEgQUiZXREREJIcLxj655pzzdwwipxUZS8CfqHHxAV8FAEKC4IMwNi7e3yFkipi44Din8uQK/IuKNYfN9XcImWL1c239HUKGBcP5BJA3LGt7EHz4284s+0DpfGGZLKmbMrkiIiIiOVxw/DRIKhjrJCIiIiI5nDK5IiIiIjlcMPbJVSZXRERERIKOMrkiIiIiOVzw5XGVyRURERGRIKRMroiIiEgOF4RdcpXJFREREZHgo0yuiIiISA4XEoS9cpXJFREREZGgo0auiIiIiAQddVcQERERyeE08ExEREREJAAokysiIiKSw5kGnomIiIiIZH/K5IqIiIjkcOqTKyIiIiISAJTJFREREcnhdDMIEREREZEAoEyuiIiISA6nPrkiIiIiIgFAmVwRERGRHE6ZXBERERGRABD0jVwzq2RmzszOKGttZpvM7Grf4wfM7M3MjVBERETEvywL/8sqfmnkmllPM/vJzI6Y2U4z+9TMGvkjlvRwzj3lnBtwNl/DzBqZ2QozO2hm+8zsWzO7zLfuRjP7Jh37ylAD/2wys2Zmts3fcYiIiEhwyvJGrpndDYwHngLCgXOBCUC7rI4luzGzwsAnwMtAcaAcMBqI8mdcgeDbr5fTttV1tG55DZMnvZHiNtHR0dx3zzBat7yGXt27sH37f23syZMm0rrlNbRtdR3ffvN1imUHD+xPx3atmTFtSsLyxx55mLVrVmdaPR59+AGuanolXTq0SXWbw4cPc+ftt9KtUzs6t2/N3NkfAbB+3Vr69epG5/at6dqxLYs+W5hi+RefH0fXjm15+IH7E5YtmD+PKe+/myl1+Pab5bRrfR1trr+Gt95M+Vic8OUXi7iobk1W//F7kuU7d+7gissu5t23JycrEx0dzZBB/enUvjUzpic6Fo9m3rHYtWsng/r3o0uH1nTt0JppU95Lcbsjhw9z19DB9OjSnq4dWjNvzscJ6156YRxdO7aha8c2fJ7KsXjphXF079yOUQ/+dywWfjKPqR9kzrHo0OpqenVtR9/uHbipV5cUt1m+bDG9u7ZP2Oa3X35OWNewXl36du9A3+4duG/YbSmWf/XF5+jdtT2jHx6RsOyzBfOYnsrf7Exk5H0B3vk0ZODNdGx7A53atWLH9uS/rzP7fVGlVEEWjWye8G/tc63o37wqABP610tYvvLxa1k0snmK++jfvApfPtSCxQ+1SCgLcF65wsy9twlfPtictwc3oGDe5HmMKqUKsnBEM754sDmXVC4GQGiIMe2OK8mbK/SM6hQVFcWNvbrSs2t7unVszRsTXk512y8WfUq3jq3p1rE1D424N2F5g0vq0KtrB3p17cA9dw5JsezL48fRs0s7Hnnov2Px6YJ5TMuk9wUEz3dGZgqxrPuXZXXKupcCMysCPAbc5pz72Dl31DkX45yb75y7z8zymNl4M9vh+zfezPL4yjYzs21mdo+Z7fZlgG9KtO98ZvacmW32ZUG/MbN8KcSQ0P3A9/xRM/sg0fM+vn3sNbMHTyqbsG2iLGk/M9tiZv8m3t4Xz7tmtt/M1prZ8DRkLmsAOOemOefinHPHnXOfO+dWmdl5wOvAFb4M+AHf67Qys1/M7JCZbTWzRxPtb7nv/wd8Za7wlbnZF9N+M1tkZhUTxe3MbIiZ/Wlmh83scTOr6ssuHzKzmWaWO9H2rc3sVzM74NvmgpP+1vea2SrfMZlhZnnNrADwKVDWF9cRMyt7mr9NquLi4njqyceY8PqbzJ63gM8WfsJfGzcm2272R7MoXLgwn3z2Bb373sj458cB8NfGjXy2cAEfz1vAhIlv8tQTo4mLi0tSdsU3X3PxJZfy4ex5fDJ/HgDr160jLj6O82rXOdPQk2nTrgOvvDbplNvMnD6FKlWqMeOjuUx66z1eGPcMMTHR5M2bl8efepoP53zCq69P4rmnx3D40KEkZQ8fPsy6tWuY+fE8cuXKxZ8b1hMZGcm8uR/TtXvPDMcfFxfHmCce49XX3uTjE8fir+THAuDo0SNM/eA9zr/gwmTrnntmLA0bN06x3IpvvWMx6+N5LEh0LOLjMu9YhIWGcte9w5k1+xPe/mAGs6ZP5e8U6jFzxlQqV6nKtFlzmDj5PcY/5x2Lb5YvY926NUydOZt3P5jBB++9zZEjR5KUPXL4MOvWrWH6h3PJlSsXG//cQGRkJPPnzqZrt4wfixNenfgO702fzdtTZqW4vl79Brw/YzbvTZ/Ng488wVOPj0pYlydPHt6b7q17dvyrycoeOXyY9evW8MHMOUnq8Mm82XTu2iPT6pCR9wXAqAfup++N/fl43kLenzaTYsXPSVL2bLwv/t59hOvGLOW6MUu5fuxSjkfH8dlvOwAYMvmnhHULf9nBp7/uSFa+ZplC9GhYidZPf8W1Ty3l6vPDqVSyAADP9r6YMXNXc/WTS/ns1x3cenX1ZOV7N67EI7NW0ffVlQnr+zapzMc/bCMyJi7Z9mmRO3duJkx6m6kz5zBlxmxWrviG31f9mmy7LZs38e5bk5j0zhRmfPwJdw8fmbAuT568TJk5mykzZ/PcixOSlT1y+DDr165h6qzk74sumfS+CKbvDDm1rM7kXgHkBWansv5BoAFwEXAhUB94KNH60kARvAxnf+BVMyvmWzcOuBS4Ei8LOhyIT09wZlYbeA3oA5QFzgHKn6ZYI6AmcBUwytcYBXgEqARUAa4BeqchhA1AnK9xfH2iuuGcWwvcCqx0zhV0zhX1rToK9AWKAq2AwWbW3reuie//RX1lVppZO+ABoCNQEvgamHZSHNfh/S0b4P0d3/DFXwGoC/QAMLOLgbeAQXh/q4nAvBM/THy6Ai2BysAFwI3OuaPA9cAOX1wFnXPJP+XT6I/fV1GhQkXKV6hArty5aXlDK5YtXZxsu6VLltC2XQcArrn2On74biXOOZYtXUzLG1qRO3duypevQIUKFfnj91VJyoblCiMyMpLY2FiccwC8+vJ4bht655mGnaJL611GkSJFTrmNmXHs2FGccxw7dozCRYoQGhpGxUqVObdiJQBKlgqnWPHi7N+/L0nZkBAjNjYG5xyRx48TFpaL9995i+49epMrV64Mx//H76uocK7vWOTKzXXXt2LZkuTHAuDVl1/kxptvIXfuPEmWL1n8JWXLlaNq1eRf3ABhYWEcP570WEx4ZTxDMvFYlChZilrneV9EBQoUoFKVquzeHZFsu9SOxd9//8Ull9QjLCyMfPnzU616DVZ+mzTbYyEhCXWIjIwkLCyMD959i249ehGWCccirfLnL4D5hlUfP348Xf3lUqrD1Pffpkv3zK1DRt4Xf/+1kbi4OBpc2RDw6psvX9L8x9l+XzSqVZLN/x5l+77jyda1ubQsc39Knv+oVroQv27aT2RMHHHxju/+3Mv1F5UBvCztd3/uBWD5uj3ccHGZZOVj4+LJlzuUvLlDiYmLp3C+XFx9fmk+/H7LGdfDzMif32tox8bGEhsbk3DuJDbn41l07taDwoW9Y1b8pB8Vp3yNxOfUce+cmvLeW3TNxHMqmL4z5NSyupF7DvCvcy42lfW9gMecc7udc3vwLtX3SbQ+xrc+xjm3EDgC1DSzEOBm4E7n3HZfFnSFcy69l/k7A58455b7yj7M6RvKo30Z19+A3/Aa5+A17p5yzu13zm0DXjrdizvnDuE1mh0wCdhjZvPMLPwUZZY55353zsU751bhNVibnuJlbgXGOOfW+o7DU8BFibO5wDPOuUPOudXAH8Dnzrm/nXMH8TKwF/u2GwhMdM597/ubv4vXtaJBon295Jzb4ZzbB8zH+wGTqXZHRFC6TOmE56XCw4mISN4g2b07gtKlvS+DsLAwChYqxIED+4mIiCC89H/lw0uHs/uk8g2uaMiO7dvp3aMrPXv1YdmSxZxXuw6lSqV6aM6abj168c/ff3FdiyZ07diW+0Y8QEhI0rfyH7+vIiYmhvIVzk2yvECBgjRs3JQeXTpQomRJChYqyO+//0bzq64mM3h/40R/y/DwFBuHa9esJmLXLpo0bZZk+bFjR3nnrUncOuT2VF+jwRUN2bFjO316dqVHrz4sW7qYWuedvWOxY/t21q9bS93zk2ecu3bvxT9//03Lq5vQvXM77h0+kpCQEGrUqMWKFd8Qefw4B/bv5+cffyBi164kZQsUKEDDRk3o1a0jJUqUpGDBgvzx+yqatcicYwFeo+TO2wZwY8/OzPloZqrbLVvyJd06tuKeO2/lwUeeSFgeHR3NTb26MKBvd75a+mWycgUKFODKhk3o16MjJUqUoGDBQqz+fRVNm2deHdIqtffF5k2bKFioEPcMG0qPLh144blnkmXdzvb7ou2l5VNsyF5e7Rz2HIrinz1Hk61bv/MQ9aueQ9ECucibK5QWdcIpWyw/ABt2HuK6C73PstYXl6VssWQXLXnnq38Yel1Nxve9lFcWbeDO62vyymcb8LW3zlhcXBy9unbguhaNqN/gyhTfF1s2b2bL5k0M6NeTm/t0S/IDLzo6ir49O3Nzn24sW5LKOdWoCb27dfSOxVl4X+S074y0CsaBZ1k9IGkvUMLMwlJp6JYFNid6vtm3LKH8SeWOAQWBEngZ4r8yGF9ZYOuJJ865o2a29zRlEn9znYgn2b5OepwqX8b2RgAzqwV8gNeHOcVrf2Z2OTAWL8OaG8gDpHxd0lMReNHMnku8G7zs+Im/feJ36/EUnp94d1cE+pnZ0ETrc5P0mJ3890lztwQzG4jXkOaVCRPpf8vAtBbNdGFhYYx91vuTxcTEMHhgf158ZQLPPj2GXTt30qZtO5q1uCpLYln57TfUqHkeEye/y9atWxgy8GYuvqQeBQt6p96ePbt5+IHhjH5ibLLGL8CNNw/gxpu98ZOPPfIQg2+7g9kfzeK7Fd9SvUZNBgwafFbjj4+PZ9wzY3nsyTHJ1r3+6iv06tMvIVuUkrCwMMY+89+xGDKoP+NfnsC4Z7xj0bptO5o1z5xjcezYUYbfcwf33Dci4e+b2MoV31CjVi1ef/Mdtm3dwm2D+nPRJfVocGVDVq/+nZv79aRosWKcf+FFhIQmPxb9bhpAv5u8Y/H4ow9x621DmfPxLL5buYJq1WswYGDGjsXrb31AqVLh7Nu3lzsHD6BipSpcfGm9ZNs1a3E1zVpczS8//8Qbr73Ey6+/BcDHC76kVKlwtm/byu2DbqJqtRrJfjj1vrE/vW/sD8BTjz3MwMFDmTf7Q77/7luqVa/JTQNuzVAd0iq190VcXCy//u9nps6cTekyZRhx313Mnzub9h07Jyl/tt4XuUKNay8ozdi5a5Kta1evPHN/2p5iuY27jjDhiz+ZOrQhx6JiWb3tIHHxXgv1nvd/4bGuF3Dn9TX5YtVOYmKTt1x37D9Ol/HeOOVKJQtQplhe/tx1mBf7XUquMOPZ+Wv5Z3fyxvXphIaGMmXmbA4fOsTwu4fy18YNVK1WI8k2cXGxbN2ymdfffJeI3REMurkP02bNpVDhwsxduJhS4d45NeSWG6lWPfk51femAfT1vS+eGP0Qg4Z474vvV66gWo0a9L/l7H5GZYbs9J2Rk2V1JnclXqavfSrrd+A1nE4417fsdP4FIoGqp9sQ7/J+/kTPSyd6vBPvkjwAZpYfL/t8JnaStKtDhdQ2TI1zbh3wDl4DFrwM78mmAvOACs65Inj9du0U228FBjnniib6l885tyK98fn29eRJ+8rvnDu5+0NKTptPcM694Zyr55yrd6oGbqnwcHbt/K8tvTsigvDw5L+WS5UKZ9eunYB3qe3I4cMULVqM8PDwJFm2iF0RlEqh/Akzp0+lTdv2rPrtNwoVKsQzz73Ae+++fbrqZJp5c2bT4uprMDPOPbciZcuVZ9M/fwNw5MgR7rztVm4bOowLLrzolPtZt3YNzjkqVarMF59/xtPPjWfr1i1s2bzpjGPz/saJ/pYREckyF0ePHuWvjRsYcFNfrr+2Bb+v+pVhQwez+o/f+f333xj//Diuv7YFUz54l8mTJjJ96gcnv0yCmdOn0tp3LAoWLMTT417g/Uw6FrExMQy/+05a3tCGFldfm+I28+d+TIurvGNR4aRj0f+WW5k6czYTJr6Fcy6hK0lKThyLihUr8+Xnixj77Ats37o1Q8cCSPjbFy9+Dk2bX8Wa1atOuf3Fl9Zjx/ZtHNi/P0n5cuUrcEm9+mxYvzbVsuvXeXU4t1IllnyxiCeffoHtW7ewdUvG6pBWqb0vSoWXpkbNWpSvUIGwsDCatbiatWuSNzhPyOz3RfM64fy+9SD/Hk56YTE0xLj+ojLM/zn1oRrTV2zmhrHL6PzCNxw8Fs3fu71+3X9FHKHXyyu4Yewy5vy0jc3/nrqxOrzteTw7by03N6/CtBWbeHL2au6+odYZ1eeEQoULc+ll9Vn5bfIJf0qFl6ZJ0xaE5cpFuXLlObdiJbZu2exbl/ScWr/u9OdUxUqVWfzFIsZk1vsih31npJVZ1v3LKlnayPVd7h6F15e2vZnlN7Ncvv6nz+Bdan/IzEqaWQnftql/w/2333i8vqHPm1lZMws1sytO6ht6wq9Ad9/r1sPronDCh0Br86bxyo03SO5M/0YzgZFmVszMygGpX3/1MbNa5g2sK+97XgEvg/udb5MIoHzigV9AIWCfcy7SzOoDiXvm78HrblEl0bLXfXHV8b1GETNLedj16U0CbjWzy81TwDcQrlAaykYA55g3GDFD6tQ9ny1bNrFt21ZioqP5bOECmjZvkWy7Zs1bMG+u1x38i88XUf/yBpgZTZu34LOFC4iOjmbbtq1s2bKJuudfkKw8wKGDB1n+1TLatGtPZORxzAwzIzIyMqPVSLPSZcrww/crAdj7779s3vQP5cpXICYmmnuG3U6rNu24+tqWp93PhFdeZMjtdxAbG0t8nNcrJyQkJEN1OXEstm/bSkxMNIs+TX4sChUqxLJvvufTz5fw6edLOP+Cixj/8mvUqXs+b783NWF5r9796H/LILr3TLk7e8KxaOsdi5CQzDsWzjkee/QhKlepQu++N6a6XenSZfjhe+/tuXevdyzKl69AXFwcBw54DcU/N6znzw3raXBFw1T38/qrLzH4tju9YxHvXUq3kIzV5fjxYxw9ejTh8fffraBKCv2ct27ZnNBncP3aNURHR1OkaFEOHTpIdLQ3cOvA/v2s+vV/VK6Seh7hjQkvM3CIdz7FJdQhY+dTeqT2vqhT93wOHz7M/n1e//Qfv/+OKlVTr0dmvy/a1SvP3B+TN2Qb1yrJXxFH2Hkg9f2eU9D7qC9bLB/XX1SWOb79nFhuBndeX5P3v/4n1X00qH4OEQci+WfPUfLlCiU+HuId5Mud/hkW9u/blzCYNTIyku+/W0nFypWTbdes+VX8/NMPgHfubNm8ibLly6f7nHr91Ze4dcidxMZk3vsCct53Rk6W5fOnOueeM7NdeAPKpgCHgZ+BJ4H/AYWBE+mGWcATKe0nBfcCY4Af8boM/IY3gOpkD+M1pvcDX+FlQov7YlttZrf5lhUAngfOdC7Xx/AalP/gZXWnADedsoT3t7gcuNvMigIH8KYUu8+3fgmwGthlZvHOuRLAEOA5M3vFV5+ZeIPQcM4dM7MngW/NLBfQ0jk328wKAtN9/XAPAl9w6i4OKXLO/WRmtwCvANXxujJ8w3+zOpyq7Dozmwb8bWahQO0zHXwWFhbGyAdHMXjgAOLj42jfoRPVqnlf5q++/CJ16tSlWYur6NCpMw+OuI/WLa+hcJEiPDPuBQCqVavOtS2vp0PbGwgNDeWBh0YRGpryF8DE115lwMBbCQkJ4cqGjZk+bSqd2rehS7fuZxJ6MiOH383PP/7IgQP7aXlVU269bSjtO3bmw5nTAejctTu3DBrMIw+NpGuHNjjgjmH3UqxYMRbMn8cvP//EwQMHmO/7YB79xBhq1jov2essXfwltevUpaQvU1ezVi26dmhD9Ro1qVHzzDM8YWFhjHhgFIMHDSA+Lo52iY7FhFdepHadupnWlWDi60mPxYxpU+ncoQ1dumb8WPz2y/9Y+Mk8qlWvQc+u3sCTIUOH0ahx0yTHYsDAITz68Ei6dWqLc46hw+6haLFiREVFcctN3nCCAgUK8PhTzxAWlvLH7bIlX3JenbqULFUKgBo1z6Nbp7YZPhb79u5lxD13AN7l42tbtuKKht6MFR9/6NWhY+fuLFvyBZ9+MpewsDDy5MnLE2Ofw8zY9M/fPP3ko4RYCPEunj433ULlKtVSfK2vln7JebXrULKkV4fqNWvRq2s7qlWvQfUaGcsYQsbeFwB33TOcQQNuBOc4r3YdOnZO+Xd9Zr8v8uUOpUmtUoyY+muydW0vLc+ck/rphhfJy7O9LqLvBO+H0xsD61OsQG5i4xwPzviNQ8djAGh/WXn6NfFyF5/+uoMZK1MfTHZHy5oMmfwjAFO+3cTLN9YjLNQYOe23dNfn33/3MPrhkcTHxxEfH8/V17akcRNv+rOJE17ivNp1adKsBQ2ubMR3K7+lW8fWhISEcMdd91K0aDFW/foLY554BAsJwcXH0/fmW6hSNeVzatmSLzmvdtL3RY/ObalWPWPvCwiu74zMlJV9ZbOKuYz2Qpc0MbPBQHfn3KkGhUkqImNP370huzvRny7QhQTBDc5j49I18Uq2FRMXHOdUnlyBf/PNmsPm+juETLH6ubb+DiHDguF8AsgblrWtzmXr92XZB0qzmsWzpG7Z7k5YwcLMyuB1E1iJl+W8By/jKSIiIpKtZOVNGrKKGrlnT268eWMr43U7mA5MMLNzgdRGPdR2zp35JIYiIiIiAqiRe9Y45zbz36wIiW3hv2nGRERERPwuGPvkBkfHFRERERGRRJTJFREREcnhgmBMcTLK5IqIiIhI0FEmV0RERCSHC8JErjK5IiIiIhJ8lMkVERERyeGC4UY/J1MmV0RERESCjhq5IiIiIhJ01F1BREREJIcLvs4KyuSKiIiISBBSJldEREQkpwvCVK4yuSIiIiISdJTJFREREcnhLAhTucrkioiIiEjQUSZXREREJIcLwntBKJMrIiIiIsFHmVwRERGRHC4IE7nK5IqIiIhI8FEmV0RERCSnC8JUrjK5IiIiIhJ0lMkVERERyeE0T66IiIiISABQJldEREQkhwvGeXLNOefvGERO60hU4J+o0bHx/g4hU8TGB3498uUK9XcImSI+4N8VweN4dJy/Q8gUlfu97e8QMmz/hwP9HUKmyBuWtf0Hft50KMs+US6tVDhL6qbuCiIiIiISdNRdQURERCSHC8LeCsrkioiIiEjwUSZXREREJKcLwlSuMrkiIiIiEnSUyRURERHJ4XQzCBERERGRAKBGroiIiEgOZ5Z1/04fi1Uws6VmtsbMVpvZnb7lxc3sCzP70/f/Yqfajxq5IiIiIpKdxAL3OOdqAw2A28ysNjACWOycqw4s9j1PlRq5IiIiIjmcZeG/03HO7XTO/c/3+DCwFigHtAPe9W32LtD+VPtRI1dEREREsiUzqwRcDHwPhDvndvpW7QLCT1VWjVwRERGRnC4LU7lmNtDMfkr0b2CKIZkVBD4ChjnnDiVe55xzgDtVlTSFmIiIiIhkGefcG8Abp9rGzHLhNXCnOOc+9i2OMLMyzrmdZlYG2H2qfSiTKyIiIpLDWRb+d9pYzAyYDKx1zj2faNU8oJ/vcT9g7qn2o0yuiIiIiGQnDYE+wO9m9qtv2QPAWGCmmfUHNgNdT7UTNXJFREREcri0zF+bVZxz35D6RAxXpXU/6q4gIiIiIkFHmVwRERGRHC4bJXIzjTK5IiIiIhJ01MgVERERkaCj7goiIiIiOV0Q9ldQJldEREREgo4yuSIiIiI5XFpu0hBolMkVERERkaCjTK5kKTO7EbgHqAocAmYDI51zB850n7t27WTUg/ezb+9ezIwOnbrSs3ffZNs553j26Sf59uvl5M2bl0cfH8N5teskrD9y5Ahd2reiWYuruP+BUcnKPzjiXjb+uYHGTZpx+513A/DmG69RtVp1mre4+kzDT6JDq6vJX6AAoSEhhIaG8faUWcm2WbRwPu+/MxmHI3/+Agx/YBTVa9QC4IlHH2TF119RrHhxpsyal+JrzJr+AXM+mkl46TI8/fzL5MqVm99++Zmli79g2L0jMhT/lk3/8MgD9yY837F9G/0H3U7Xnn2SbPf1siW8+frLhISEEBoayh33jOCCiy4BoGn9C6hSrToA4eFlGPvCK8le57WXnue7FV9TvUYtHnpsTMLf5eCBA8le60xERUVxy019iImJJi42lquuuY5BQ4Ym2y46OppHHryftWvXUKRIUcY88zxly5UjJiaapx57lDVr/iAkJIR7hj9AvcvqJyv/0Mj7Es6p2+64C/DOqWrVqtMsg+dUVFQUg27uQ/SJOlx9HQNTqMMnc2fz0vhnKVkyHIAu3XvSvmMXAHbt3MGTox8mImIXZsYLL0+kbLlyScq/PH4cK7/9muo1azH6iacB+HTBPA7s30+P3v3IqGCpB8DMqe+xYO5HmBmVq1VnxMNPkCdPniTbzP1oBrM/nE5oSAj58ufn3pGPUqlKVQD++nM948Y8xrGjR7CQECa+Mz1Z+ccfvp+//9rAFY2aMnDIMADemzyRylWr0bhZmufQT2Jom/O58ZqaOAerN+9j4MtfERUTR7MLyvJUvwaEhMDR47Hc8tIy/t51KEnZ4oXyMHX4NVxarSQfLNnAXZO+TViXKyyEFwY2pEmdMsQ7eHTKj8xZ+U+S8u2vqMzDPS5l/5Eouo75nH2Ho6hcuhCP9a5Pn3GLz6g+AN9+vZynxz5JfFw8HTp1of8tA5NtEx0dzYMjh7N29WqKFC3KM8+9QLly5QGYPGkisz/6kJDQEO4f+RANGzVOVvbO2wcTERFBt+496NajFwCPPfIwXbp1T/Ldk11kp5tBZBY1ciXLmNk9wHC8+00vBsoBE4AvzKyhcy76TPYbGhrKXffcz3m163D06BF6d+9EgyuupErVakm2+/ab5WzdvJk5nyzij1W/MeaJ0bw3dWbC+tdeeZGLL62X4mv8uWE9efLkZcZH8xgy8GYOHz5MZORx/lj1GwMGDj6TsFP16sR3KFqsWKrry5Qrz4Q336Vw4SKs/HY5Y594hMnvzQCgVZsOdOnWi8dGpd5YXbTwE96fMYd333qD71Z8S6MmzXh70uuMHvNshmM/t1Jl3p76EQBxcXF0vKEFTZon/2K9tH4DGjVtjpmx8c/1PDLiXqZ8NB+APHnyJOwjJUeOHGbDujW8O302Yx8fxV8bN1C+/LksnD+H515+PcN1AMidOzevv/k2+fMXIDYmhv439ubKRo05/4KLkmw3d/aHFCpchDmfLGLRpwt4efw4xjz7ArM/8n6czPhoHvv27uWO2wby3tRZhIT8d/HMO6fyMP3DuQwZdDNHfOfU6t9XZco5lTt3biZM+q8Ot9zUmytSqAPANddez30jH062/NGHRnDTgEFcfkVDjh07Soglvfh35PBh1q9dw9RZc3li9ENs/HMD5Sucy/y5s3np1TcyXIdgqsee3RF8NGMK782YS568eXlk5D0s+eJTrm/dPsl2V1/XinadugHw7fKlvDr+GZ59aSKxsbE88cgIHnx0DNVq1OLggQOEhSX9Cv/rz/W+989s7r59AEeOHCYqMpI1q1fRt/+gM4q7bPH8DGldh4uHziIyOo4P7ruKLo2r8sGSDbw0qBFdxnzO+m0HGHh9bUZ0vZiBL32VpHxkdByPTf2R2ucWp865xZOsu7/zxew5cJwLbpuJGRQvmLTBDjC4VR0a3TubdldUpluTary2YDWP9ryMR6f8eEb1Ae+z6aknH2PipLcJDw+nZ7fONGvegqrVkn5nzP5oFoULF+aTz77g04ULGP/8OJ59bjx/bdzIZwsX8PG8BezeHcGgATcxb8EiQkNDE8qu+OZrLr7kUgYMvJV+vb1G7vp164iLj8uWDdxgpe4KkiXMrDAwGhjqnPvMORfjnNuEd9/pSkDvM913yZKlEj40ChQoSOXKVdm9OyLZdl8tXUyrNu0wM86/8CKOHD7Enj27AVi75g/27dtLgysbpvgaYWFhREVFEh8fT2xsDKGhIbz+6sspZvfOtgsuvJjChYsAUOf8C9kd8V9dL760HoWLFDlleYcjNjaWyOPHCQsL47MF82nQsDFFihTN1Dh//vE7yparQOkyZZOty58/P+ZLG0QeP56uDEKIhRAbG4tzjqjISMLCwpj2wTt06taTsLBcmRK7mZE/fwEAYmNjiY2NSbG/2ldLl9C6bTsArrrmOn744Tucc/zz91/Uq385AMXPOYdChQqzZvUfScp651SU75yKJSQ0hNcnvMygIbefvTqk4w/9918biYuL4/IrvPdE/vwFyJsvX9LXCPnvWEQe947FlPfeomv3XoTlOovHIgDrARAXF0tUVBSxsbFERR6nRImSybYpULBgwuPjx48npNd++n4FVavVoJrvqk2RokWTNKoAQhOdU3GxsYSEhDJ54ivcPPC2DMUdFhpCvtxhhIYY+XKHsXPfUQAcUDif9/cpnD83O/cdS1b2WFQsK9ZGEBkTl2xdv6tr8uxHv3r7crD3cFSybeLjHXlyhZI/TxgxsfE0rF2aiAPH+WvnoWTbptUfv6+iQoWKlK9QgVy5c9PyhlYsW5o8K7x0yRLatusAwDXXXscP363EOceypYtpeUMrcufOTfnyFahQoSJ//L4qSdmwXGFERkYmnFcAr748ntuG3nnGcZ9tloX/sooauZJVrgTyAh8nXuicOwIsBK7JjBfZsX0b69atpe75FyZbt3t3BOGlyyQ8LxVemj27I4iPj+eFcU8z7O7hqe63cpWqFCtWnF7dOtK4aXO2btmCc/GZ/ovczLjztgHc2LMzcz6aedrt58/5iCsaNj7tdol17tqLAf26E7FrJxdcdAkL5s2mc9ceZxpyqhYv+pSrr7sh1fXLl35Jr05tGD5sCCNGPZ6wPDo6mgF9ujLoxp4sX5b8iyd/gQI0aNiEm3t15pwSJSlQsBBr/lhFkzO8FJuauLg4enbtwDXNG3F5gyupe8Gpz6mwsDAKFizEwQMHqF6jFsu/WkpsbCzbt21j7drVRETsSlLWO6eK0bt7J5o08c6p+Ph4ap2XeedUXFwcvbp24LoWjajf4MoU3xcASxZ/Ts8u7Rhx751E7NoJwJbNmyhYqBDD7x5K724deen5Z4mLS9pQKVCgAFc2akLvbh0pUbIkBQsW5I/fV2W4q0Uw1qNkqXC6976Rrm2vpuMNzSlQsBCXNUj5R/XsWdPo0aElr7/8HHfeMxKArVs2gxn3Dh3IgD5dmPreW8nKVapclaLFinFLny5c2bgZ27dtwcXHU6NW7TOOe8e+Y4yfs4oNk3ryz9u9OXQsmsW/bgdgyKvLmf3w9Wx8syc9m1VnnK/BmhZFCuQG4JGe9VjxXEem3Hc1pYrkS7bdsx/9yoLHWnHDZRWZ+fVGRnS9hDEz/3fG9QHYHRFB6TKlE56XCg8nIiJ5YmT37ghKJ35/FyrEgQP7iYiIILz0f+XDS4cnSTYANLiiITu2b6d3j6707NWHZUsWc17tOpQqFZ6h2CV91F1BskoJ4F/nXGwK63YCl2b0BY4dO8p9d9/BvcNHUjBRNuR0Zs2YSsNGTZN8aKXk3vsfSHg87PZbeXDUaCa/8TobNqzj8gZX0rFz1zOO/YTX3/qAUqXC2bdvL3cOHkDFSlVS7ULx84/fM3/Ox0x864N0vcb1rdtyfeu2AEx+YwJdevRi5Yqv+fSTeZQKL80ddw9Pcln9TMTExPDt8mUMun1Yqts0aX41TZpfza//+4k3X3+F8RPeBGDW/M8pWSqcHdu2cufg/lStVp1y5c9NUrZXv5vp1e9mAMY+Por+t97O/Dkf8uN3K6larQb9BpzZpdnEQkNDmTpzNocPHeLeu4ay8c8NVKteI01l27bvyD///EXfnl0oXaYsF1x4EaEp/E3vGf7fOXXX0ME88PBoJk96nT83rOfyBlfQoVPGzqnQ0FCm+Oow/O6h/LVxA1WrJa1Do6bNuPZ6Lyv18YczePThkbw26R3i4uL49Zef+WD6x4SXLsOD99/NJ/Nm065D5yTl+940gL43DQDgidEPMWjIUOZ8PIvvV66gWo0a9L8l410vgqEehw8d5JuvljJ9ziIKFirEIyPu4fNP53Pt9W2SbduhSw86dOnBF58t4L23JvLAo08RFxfL77/+wsR3p5M3b17uGjKAmrVqc2n9BknKDr37v65KI+6+jXtHPsL7b01k458bqHf5FbRp3/nklzulogVy07p+Rc4bNI0DR6OYOvwaujetxvSvNjK0zfl0ePxTfvxzD3e1v4Cnb76CIa8uT9N+w0KM8iUK8t26CO5/+zvuaHs+Y25qQP/xS5Nst+S37Sy5ZzYAPZtVZ9HPW6hetgjD2l/A/iNR3PvmCo5HJ88S+1tYWBhjn30O8D4PBw/sz4uvTODZp8ewa+dO2rRtR7MWmfvDPMOCsE+uMrmSVf4FSphZSj+syvjWJ2FmA83sJzP76a03T90vLiYmhvvuvoPrW7WhxdXXprhNqVLhCdkdgN0RuyhZKpxVv/3KjOlTaN2yBeOfe4YF8+fy0vjnUn2tZUu9X+THjh1j27YtPD1uPIu/XORdWsygE7/yixc/h6bNr2LN6lUpbrdxw3rGPD6KZ154hSJFi57Ra+3Zs5s1f/xO0+ZXM/X9d3h87HMUKlSIn3747kzDT/Ddt19To9Z5FD+nxGm3veiSeuzYvo0DB/YDXsYLoGz5Clx06WVsWLcu1bIb1q0F5zi3YiWWfvk5j419ju3bt3pZr0xSqHBh6l1Wn5Urvkm2LvE5FRsby5EjhylStChhYWHcc99Ips6czfMvvsqRw4c5t2KlVF9j2dLF1Kpdh2PHjrJt61bGPvsCi7/4nMhMOKdO1OHSy+qz8tvkdShatBi5c3tZtXYdOrNu7WqvbuHh1KhZi3LlKxAWFkbT5lexfu2aVF9j/bo1OOeoWKkyi79YxJhnX2D71q1s2bwpU+oQ6PX46YfvKFO2HEWLFScsLBeNm1/FH6t+PWWZq669nm++WgJ474sLL76UokWLkTdvPho0bMyG9anX45uvllCzVm2OHzvG9u1bGT3mOb5a/DmRkek7p1pcWI5Nuw/z76FIYuMcc1b+Q4Na4ZQonJfzK5/Dj3/uAeDDb/6iQa20Zyn3Ho7iaGQMc77zBpp9vOJvLqpyTqrb58sdSp8WNXh94Woe6nEpA15cxoq1u+jetHq66gPeObFr539XVnZHRBAenjz2UqXC2ZX4/X34MEWLFiM8PJyIXf+Vj9gVQakUyp8wc/pU2rRtz6rffqNQoUI889wLvPfu2+mOW9JPjVzJKiuBKKBj4oVmVhC4Hm8gWhLOuTecc/Wcc/VuHpB85Gui7Xj8kYeoXLkqvfvelOp2TZq1YMH8uTjn+P23XylYqBAlS5biybHjWPj5Uj75bAnD7hlOqzbtuGPYPSnuIyYmhqkfvEvfmwYQFRWV0DcwPi6e2JiY0/8VTuH48WMcPXo04fH3362gStXkH+C7du5gxL13MOrxsadsOJ3OGxNeYuBgr0/xibpYSEi6vwRT8uWihVx1iq4K27ZuSeintn7dGmKioylSpCiHDx0kOtobf3jgwH7++O2XhJHlKXnz9ZcZMHgosbGxxMfHAxBiRlQG67B/3z4OH/L6/EVGRvL9dyupVKlysu2aNGvOJ/PmArD4i0VcVr8BZkbk8eMcP+b1T/xu5beEhoYmGwh5QmxMDNM+eI9+N/b3HQdveXx8HDEZOKdSqkPFysnr8K+vXzrA8q+WULlyFQBq1zmfw4cPs3/fPgB++uF7Kp/iWLz+6kvcOuROYmNiiY/3MmsWYkRGRp5xHYKpHuGly7Dmj1VERh7HOcf/fvyeipWqJNtuW6IfaCu/XU75Ct5VjPoNGvL3X38SGXmc2NhYfvvfT1SqnHI9YmNjmDX9fXr0vZmoqMiE/uTx8fHpPqe27jlC/RqlyJfb6//b/IJyrN92gP1HoiicPzfVynpjAFpcVJ712w6ka98Lf9xCk7pen/1mF5Rj3dbUy9/V4UImLPiD2DhHvtxhOOeId5A/T/ovSNepez5btmxi27atxERH89nCBTRt3iLZds2at2DeXC+L/MXni6h/uff+btq8BZ8tXEB0dDTbtm1ly5ZN1D3/ghRf69DBgyz/ahlt2rUnMvK49zlrGT+fzgbLwv+yirorSJZwzh00s9HAy2Z2iKSzK2wD3j/Tff/6y/9Y8MlcqlWvQY8u7QG47Y67aNS4KR/OnA5A567dadS4Kd9+vZx2ra71TSH2VLpfa9b0qbRu2558+fJRvUZNIo8fp2vHNjRq3JRChQufaRUA2Ld3LyPuuQPwBqhc27JVQn/bjz/06tGxc3femvQahw4eZNyYxwCSTDU2auS9/O/nHzhw4ABtWzZnwK2307Z9p2SvtX6dlwGqeZ7XV+/alq3o3bUdpcLL0Ltf/wzV4/jxY/z0w0rue/CRJMvnfOjNANG+cze+WvwFny2cR1hYGHny5GX0mHGYGZv++ZtxTz2GhRgu3tGrX/9UGyTLly2m1nl1KFGyFADVa9SkX7cOVK3+3+CcM/Xvv3t45KGRxMfHER8fzzXXtqRx0+aA1wg6r05dmjZrQbsOnRn14P20b30dhQsX4alnvCsA+/bt4/bBAwgJCaFUqVI89uTTqb7WzBneOZX3xDkVGUm3Tm1p2KhJhs6pf//dw+iH/6vD1de2pHETrw4TJ7zEebXr0qRZC2ZM+4Dly5YQGhZGkcJFGOWbki00NJQ777qP2wbdhHOOWufVoX2nLim+1rIlX3Je7bqULOUdixo1z6NH57ZUq16TGjUzfiyCoR61615A06uu4ZY+XQkNDaVazVq06eDFMXniK9Q6rw4NmzTn41lT+fmH77w+oIULM/IR73OqUOEidO3Zl0H9umNmXH5lY65o1DTF15o9azotW7Ujb958VK1ek8ioSG7s0YEGVzamUKH0nVM//rmH2Sv+YeXznYiNi+e3f/YyedFa4uIdt726nGn3X0N8vOPA0SgGvezNrNDqsopcUq0Ej0/7GYB1b/SgUL5c5A4Lpc3lFWn96ELWbTvAQ+99z+RhzXm2/xX8eyiSQS8tSzGGMsXyU696KZ6a4fXFfW3BH3wzrgMHj0bTdczn6aoPeF0JRj44isEDBxAfH0f7Dp2o5pu28NWXX6ROnbo0a3EVHTp15sER99G65TUULlKEZ8a9AEC1atW5tuX1dGh7A6GhoTzw0KhkgwBPmPjaqwwYeCshISFc2bAx06dNpVP7NnTp1j3dcUv62YlsikhWMLP+wF38N0/uHGCEc27/qcodiQr8EzU6Nt7fIWSK2PjAr0e+XCl/IQWa+IB/VwSP7Ngv9ExU7hf4l9H3f5j6lb9Akjcsa3vJrt91LMs+UWqWzp8ldVMmV7KUc24yMNnfcYiIiEhwUyNXREREJIcLwskVNPBMRERERIKPGrkiIiIiEnTUXUFEREQkpwvC/grK5IqIiIhI0FEmV0RERCSHy8qbNGQVZXJFREREJOgokysiIiKSw1nwJXKVyRURERGR4KNMroiIiEgOF4SJXGVyRURERCT4KJMrIiIiktMFYSpXmVwRERERCTrK5IqIiIjkcJonV0REREQkACiTKyIiIpLDaZ5cEREREZEAoEyuiIiISA4XhIlcZXJFREREJPiokSsiIiIiQUfdFURERERyuiDsr6BMroiIiIgEHWVyRURERHI43QxCRERERCQAKJMrIiIiksMF480gzDnn7xhETisyFp2oIiJyWsUuu93fIWSK47+8kqXNzi37orLse/bc4nmypG7K5IqIiIjkcEGYyFWfXBEREREJPsrkioiIiORwwdgnV5lcEREREQk6yuSKiIiI5HjBl8pVJldEREREgo4yuSIiIiI5nPrkioiIiIgEAGVyRURERHK4IEzkKpMrIiIiIsFHjVwRERERCTrqriAiIiKSw2ngmYiIiIhIAFAmV0RERCSHsyAceqZMroiIiIgEHWVyRURERHK64EvkKpMrIiIiIsFHmVwRERGRHC4IE7nK5IqIiIhI8FEmV0RERCSH0zy5IiIiIiIBQJlcERERkRxO8+SKiIiIiAQAZXJFREREcrrgS+QqkyvpY2abzOy4mR0xswgze8fMCvrW3Whmzsy6nVSmmZnF+8ocMbNtZjbTzC7zTy1EREQk2KmRK2eijXOuIHAJUA94yLe8H7AP6JtCmR2+MoWABsA64GszuyozAvr26+W0bXUdrVtew+RJb6S4TXR0NPfdM4zWLa+hV/cubN++DYDfV62ia8d2dO3Yji4d2rL4yy9SLDt4YH86tmvNjGlTEpY/9sjDrF2zOjOqkOF6nLBzxw4a1LuYd9+e7Jd66Fh4Jk+aSOuW19C21XV8+83XfqlHRuqwcsW3dO/SkU7t29C9S0e+/25liuVHDr+Hzh3a8NL45xOWvfH6BJYs/jJT6pDRegTLOZVd6hHIn1G39WjGT7Me4OcPH+T2ns0Slp9foxzL3r2HH2c+wIfjB1GoQN50lb+gRjm+evcevps+gm+mDKdenYrJyl5xYRV+mDGSb6YMp+q5JQEoUjAf8yfchmWTaQ0sC/9lFTVy5Yw557YDnwJ1zawi0BQYCFxnZqVTKeOcc9ucc6OAN4GnMxpHXFwcTz35GBNef5PZ8xbw2cJP+GvjxmTbzf5oFoULF+aTz76gd98bGf/8OACqVa/O1JkfMfPjuUx4400eHz2K2NjYJGVXfPM1F19yKR/Onscn8+cBsH7dOuLi4zivdp2MViFT6nHCuGfG0qhx4xRf42zXQ8fCq8dfGzfy2cIFfDxvARMmvslTT4wmLi4uS+uR0ToULVaMl159jY/mzOfxp8by4MjhycpuWL+OPHnz8uHs+az+43cOHz7Mnj27+X3VKlpcdXWG65AZ9QiWcyo71COQP6NqVy3DTR2vpHGfZ6nfbQzXN6lLlQolAHhtVE8eemkul3V9inlLf+OufslzL6cq/+Sw9jz5xqc06D6Wx1/7hCeHtU9W/s4+Legw9DWGP/sht3RuBMCIW1ryzOTPcc6dcb3k1NTIlTNmZhWAG4Bf8LK3PznnPgLWAr3SsIuPgUvMrEBG4vjj91VUqFCR8hUqkCt3blre0IplSxcn227pkiW0bdcBgGuuvY4fvluJc458+fIRFuZ1T4+KikrxV3VYrjAiIyOJjY1N+EB69eXx3Db0zoyEnqn1AFiy+EvKlS9H1WrVU3yNs10PHQuvHsuWLqblDa3InTs35ctXoEKFivzx+6osrUdG63DeebUpVSocgGrVqhMVGUV0dHTSOoTlIioykvj4eGJjYwkNCWHCyy8x5PahmVKHzKhHsJxT2aEegfwZVatyaX78YxPHI2OIi4vn65830r7FRQBUO7cU3/zsNdaXfLeO9lddlK7yzkFhX/a3SMF87NxzMFn5mNg48uXNTb68uYmJjaNy+RKUDy/K1z//maF6yampkStnYo6ZHQC+Ab4CnsJr5E71rZ9Kyl0WTrYD78pF0YwEszsigtJl/ksclwoPJyIiIvl2uyMoXboMAGFhYRQsVIgDB/YDsGrVb3Ro24rO7dvy0KjRCV8mJzS4oiE7tm+nd4+u9OzVh2VLFnNe7ToJjYDMkNF6HDt6lLcnT+LWwben+hpnux46Fl49IiIiCC/9X/nw0uHsPql8IByLE778fBHn1a5N7ty5kyyvUrUqxYoVp3vnDjRp1pwtW7YQ7+IzLfuZWfUIhnMqO9QjkD+jVv+1g4YXV6N4kQLky5uLlo3qUL50MQDW/r2TNs0uAKDjNZdQPrxYusrfN+5DnhrWnj8/fZwxd3Vg1Mtzk5V/9q3Pmfx4H+67+Vpen76c0be34dEJn2SoTpnNLOv+ZRXNriBnor1zLqHDnZk1BCoD032LpgJPmtlFzrlfT7GfcoADDqS00swG4nV/4JUJE+l/y8CMR56KCy64kNnzFvD3X3/x0AP306hxE/LkyZOwPiwsjLHPPgdATEwMgwf258VXJvDs02PYtXMnbdq2o1mLTOlefMZem/AKvfv2I3+B1BPjgVCPYDgWaREo9di48U/GvzCO1994K8X1w0c+mPB46JBbefjR0Uya+Bob1q+jwRUN6dSla1aFmqpgOacCvR7+/Ixa/08Ez73zBfMn3MaxyGh+W7+NuLh4AAY9OoXnhndmxC0tWfDV70THxKWr/MAujRn+3MfMWfwrna65mNce6UWrW19JUn7Vhu007efVq+ElVdm15yCG8f7Ym4iJjWPE87PZve9wuuslp6ZMrmSGfngZ2V/NbBfwfaLlp9IB+J9z7mhKK51zbzjn6jnn6p2qgVsqPJxdO3clPN8dEUF4ePJf/aVKhbNr104AYmNjOXL4MEWLJv3FXqVqVfLnz8/GPzek+nozp0+lTdv2rPrtNwoVKsQzz73Ae+++fcqKpkVG6/H7qt8Y/9w4rr+mBVPef5c335jItCkfZGk9dCy8eoSHhxOx67/yEbsiKJVC+bNZj8w4FhG7dnHXHbfzxFNPU+Hcc0/5ekuXfEntOnU4duwYW7du4dnnX+SLzxdx/Phxv9fjhEA+p7JDPQL9M+rdOStp2OsZruk/ngOHjvHn5t0AbNgUQZshr9Kw1zPM/Oxn/tm2J13le7W+nDmLfwXgoy9+SXHgWWIjBrRkzKTPeHDQ9Tz44hzemr2CIT2anXG9Motl4X9ZRY1cyRAzywt0xcu4XpTo31Cgp5mFnbS9mVk5M3sEGAA8kNEY6tQ9ny1bNrFt21ZioqP5bOECmjZvkWy7Zs1bMG/ubAC++HwR9S9vgJmxbdvWhAEcO3ZsZ9M/f1O2XLkUX+vQwYMs/2oZbdq1JzLyOGaGmREZGZnRamS4Hu+8P5VPv1jCp18soVeffgwYOIgevXpnaT10LLx6NG3egs8WLiA6Oppt27ayZcsm6p5/QZbWI6N1OHToELcPHsidd93DxZdcesrXiomJ4YP33uXGmwcQFflff9H4+DhiYmL8Wo9gOaeyQz0C/TOqZLGCAFQoXYx2LS5kxqc/JVluZoy45TomffhNusrv3HOQxpd6fYyb1a/Bxi0pN5IBerW5nEXfrGb/oWPkz5ub+HiHi3fkz5vrjOslqVN3Bcmo9sBx4D3nXMK3mZm9BTwGtASOAGXN7AhexvcgsAJo5pz7LqMBhIWFMfLBUQweOID4+Djad+hENd+ghldffpE6derSrMVVdOjUmQdH3EfrltdQuEgRnhn3AgC//O9n3npzErnCwrCQEB54+FGKFSue4mtNfO1VBgy8lZCQEK5s2Jjp06bSqX0bunTrntFqZLge6XG26qFj4dWjWrXqXNvyejq0vYHQ0FAeeGgUoaGhWVqPjNZh+tQP2LJ1C2+89ipvvPYqAK9Neotzzjkn2WvNmDaFtu06kC9fPmrUrEnk8Ug6tW9Do8ZNKFy4sF/rESznVHaoR6B/Rk0bN4DiRQsQExvHsLEzOXjEu8rQtWU9BnVrAsDcJb/y3lzva6lMySJMGNWTDkNfO2X52x6fyrP3dSYsLISoqFhuf2Jaiq+fL28u+rS5nNZDvK4ML32whNkvDyE6JpYbH3jnjOuVWbLJTGaZyjR1hQSCyFh0ooqIyGkVuyz1gW2B5Pgvr2Rps3P/sbgs+54tlj80S+qm7goiIiIiEnTUyBURERGRoKM+uSIiIiI5XDD2yVUmV0RERESCjjK5IiIiIjlcVs5fm1WUyRURERGRoKNMroiIiEgOpz65IiIiIiIBQJlcERERkRwuCBO5yuSKiIiISPBRI1dEREREgo66K4iIiIjkdEHYX0GZXBEREREJOsrkioiIiORwuhmEiIiIiEgAUCZXREREJIfTzSBERERERAKAMrkiIiIiOVwQJnKVyRURERGR4KNMroiIiEhOF4SpXGVyRURERCToKJMrIiIiksNpnlwRERERkbPMzFqa2Xoz22hmI85kH8rkioiIiORw2WmeXDMLBV4FrgG2AT+a2Tzn3Jr07EeZXBERERHJTuoDG51zfzvnooHpQLv07kSZXAkIecPObmchMxvonHvjbL5GVlA9so9gqAMERz2CoQ6geqTV8V9eOVu7ThAsxyKxs/09m5iZDQQGJlr0xkl/z3LA1kTPtwGXp/d1lMkV8Qw8/SYBQfXIPoKhDhAc9QiGOoDqkZ0EQx38xjn3hnOuXqJ/Z+UHgxq5IiIiIpKdbAcqJHpe3rcsXdTIFREREZHs5EeguplVNrPcQHdgXnp3oj65Ip5g6VulemQfwVAHCI56BEMdQPXIToKhDtmWcy7WzG4HFgGhwFvOudXp3Y855zI9OBERERERf1J3BREREREJOmrkioiIiEjQUSNXRERERIKOBp6JiIhkQ2ZW2jm3K63LsyMzCwN6492etQTwL/Al8IFzLsafsUnw08AzEckWzKwicCFQMPFy59xU/0SUfmaWH6hG8jqs8E9EEsjM7JBzrnAKy/c554r7I6b0MLMiwBdAJWAhsBMoA1wPbAGuds4d9FuAEvTUyBURvzOzkcDDwBrgeKJVzjnXxD9RpY+Z9QVeAaJJXodz/RNVzmVmBgwAegAlnHMXmFkToLRzbqZ/o0sbMzvsnCt00rLCwN/OuRJ+CivNzGwCUBHo6pw7mmh5QWAGsNk5N8Rf8Z0JM8sHFAUOOOeOn2Zz8TM1ciVHMbOvgdOe9IHSsAIws+LAvcBFJM8gBkQ9zOxfoIlzbo2/YzlTZrYL6OOc+8LfsaSXmb1P2t4XfbMgnExhZo/jXSIfD7zunCtqZlWAWc65S/0a3GmY2Va841EW2HHS6nOAac65AVkeWDqZ2Q6ggXNuSwrrKgErnXNlsjywM2BmzYGngUsAwzs+/wNGOucW+zM2SZ365EpO82aix1WBm4F3gc3AuUA/4C0/xJURU4E8wEzgmJ9jOVN7gU3+DiKDooFl/g7iDG30dwBnwY3Axc65f83sNd+yf4Aq/gspzXrjNaQWAn0SLXdAhHNuvV+iSr8ipH4r1m1Asq4Y2ZGZ1cM7Fm8C9+P98CgHdATmm1lT59yPfgxRUqFMruRYZvYd0D/xXVTMrDbenVUa+C+y9DGzQ0BJ51yUv2M5U2Z2PdALL+u2O/G6lLJA2ZGZ9QPqAaOdc//6O56czpdFrOKcizzRh9XMCgFrnHMV/B1fWphZfudcoP5wxcxWAfekdHXDzK4Dxjnnzs/6yNLHzGYA65xzj6Sw7hGgjnOua9ZHJqejRq7kWGZ2EAh3zkUmWpYP2OmcK+q3wNLJzL4B+jnn/vJ3LGfKzNoBk/BGXyfmnHOhfggp3czsCmA6UD7xYgKoDif47hVfE+942InlzrklfgsqnczsTbzs+l14A57OAV4AcgdKP1DfcbiRlLsiZfuuI2Z2IzAGuB2Y7ZyLN7MQvAzoy8ADzrm3/RhimpjZFrxuFyd3HcHMygI/OOfKJy8p/qZGruRYZjYP7/L+w3iXzioAjwKFnHNt/BhaupjZY3iDa94Gkkwr5JwLiK4XZrYdeASvkZhkMIdzLs4vQaWTmW0EpuENqDm5DgHzA8TMGgGz8LrAFAYOAYWArc65QLjUDyQM0HoXbyR/LiAS+Bzo65w77M/Y0srMpgMXAPM5qSuSc260X4JKJzO7B+9zNQ/e9GElgCjgMefcs34MLc1Sm+Ui0fpkAwQle1AjV3Is34CtCXhZhVAgFvgYGBpIl5vNbGkqq5xzrkWWBnOGzCwCKBsoDdqUmNl+oLgL8A9VM/sRmOqce8HM9jvnipnZKOCYc26cv+NLLzMrhTfCf2ugzC17gu+cquycO+DvWDLC103kSv6bJ3elc+6Qf6NKuzQ0ck+5XvxHjVzJ8XyXz0oCe5xz8f6OJycys/uA3MBTgdpINLPngV+dc+/5O5aM8HXjKea7tHyikZsb+Mc5V87f8aWVmV0LbHLObUi0rCZwbqDMgGFmvwHXOuci/B1LTmZmcaQ+gM6AMs45DeTPhtTIlRzPl2U4ue/h3/6LKP3MrBjQBm/E73ZgvnNuv3+jSjvflEml8fpQ7k28LlDmmPX1ja6PN4I/SaMkUKZyg4T+hxc45w6Y2RqgM94x2eCcK+Lf6NLOzP7Em5ZuZ6JlZYFlzrka/oss7XyX+rsAL5L8nAqY/tGBzsyanm4b59xXWRGLpI8auZJj+WZSmIJ3ly3Hf3MfEkgDhXwDnhYA6/hvKrTzgFbOuZX+jC2tTvUlEihfHr7ZFVLknHs3K2PJCDMbjzeQZqqZ3QvcB8QAi5xz/f0aXDqY2cGTG+W+G0QcDJRLy2b2TyqrXCD1jw5mviuBLZ1zC/0diySnRq7kWGa2DG8y78fwsm+V8EYCr3DOfeC/yNLHzL4HXnDOTU+0rBtwr3PuMv9FJsHANxCtEPBZIHUlMbNf8KavWpJoWXNgvHPuQv9FJsHAzC7Am1e9FxDqnCvp55AkBWrkSo7lG9RRyjkXY2YHfHdEKgD84Zyr7O/40spXj3MS9yc2s1DgX+dcMf9Fdmpm9qBz7knf48dS2845NyrrokofM+vjnHvf9/jm1LYLlFkuAMzsJefcHSksH++cG+aHkM6Ib1q6d4HJwF94N3+5CbjJOTfXn7Glh5nlAhrgDcyc4fuMwiW6Ta5kDd8gxl5AX7xZLxxwB97c6pGnKiv+EeLvAET8KBJvaiGAf83sXLz3xDn+C+mM/Al0P2lZF7wv9uws8bySFU7xLzvrkehxn1T+9fZDXBlxYyrL+6SyPFvyNWSvBQoArXz/vy7AGrjnAxvw5pCe7FvclMC7K2NAM7MuZjYfb7zDTXjTBFYB9gAfqoGbfSmTKzmWmc0EFjrn3jGzsXgDt6KALc659n4NLh3M7ErgE7wvw8143S6qA62dcyv8GJoEkESZ6FfwJu9PrArQxTlXM2ujytl8gxknOufeTzTTRQG8QYABM9NFoDOzeLzBlwOdc7MTLd8JXOic251qYfErNXJFSBg80BOv7+F7gXYp0De7QiugLN591Rc65/b5N6r0C4aZLiBhgFPiOmT7qekSzbfcGPg60SqHN7L/Refcd1ke2BkK9LuFQdK5l0/cmti3POGxnH2+O7f1A5oAP+INWJ4B/ApcpEZu9qVGruR4vgZueOKphiRrBcNMF2ZWDi8L2gQomnhdoNQBwMyecM495O84MsrMpuGdT4F8t7BfgFuccz+daNiaWX3gFedcfX/Hl9P4urT19f07MW7jFuD9QL6RTTDT5MWSY/myn6/izQMaAxQws7ZA/ez+JW9mnznnWvoef42vQXiyAJqfdQKwFGjOSTNd+DGm9HodrzF1FfAVXmP3USDQphZqBSQ7/83sJ+dcPT/Ec6ZaEvh3C3sYWGBmrwO5zWwkcCtew0qyiJnlx3tP1AV+wRt0diledvcF4Em8Ocolm1EmV3Is333h9+NNIbbG19+tJN4UYtX9G92pmVlP59xU3+OAn581GGa6MLO9eHfTOpqoDsXxzqda/o4vrVK6Ramv+8XeQLpEHix3CzOzi/EatRWBrcAk59zP/o0qZzGzt4F6wKfADcBS59xQ37o8QHvn3Aw/hiipUCNXciwz24M3LU/MSf3dkk0in52Z2eXOue9TWF7fOfeDP2JKL98AjqrOuWNmthFogfcDZHsATdy/G6jgnIsys03AZcAhvKncCvk1uDQwsxO3I+6G198wsUp43xeNszSoDNDdwiSz+D6fLnHO7TSzCsDyQPnxndOpu4LkZAfxBjklvu3nuYmfB4gvgJQagp8BgZJ5+xroCrwDfIiXMYkCAqkx8j1elmc2sAivoXgc+MmfQaXDX6k8dsC3wKysDSfDTswQ8dRJyx3ebBHZnpmF4U1TdzHJB88N9EtQOVOBE2M2nHNbzSxgkiA5nRq5kpO9CXxkZg8CIb7b4z6F17cy2/MNmDPvYdLR/HgT38f6JbAz4JzrmujpA8BqvC/191IukS314b+5x4cB9+LVYbyf4kmXE4OxzOw759wif8eTUUGSafsAOB/vR19Ad7sIcGG+u+VZKs91dSCbUncFybF8DcM7gEF4/d224E24XjS7DzyDhLkbU3sDxwNPOucezbqIJBiYWYvU1gXaF3mg3y3MzA7gdYE57O9YcjJf96NTNZaccy4grg7kNGrkiiRiZnmBo4Ew5ZOZVcTLJJwYyX+CA/Y45477JbA0MrP3OfUXB5C95zQ91e2IE8vOtyY+mZn9c9KikkBuYFsgfZH77hY2D6/bS3nnXEEzuwHo55zr5t/o0sbMvgV6Ouc2+zsWkUCk7goiSZ2YozXbS/TFV9GvgZy5jf4OIBNk99sOp9vJl/nNLBRv+qRAyya+Bow6cbcw37Kv8G6RGyj6AG+a2eckHzwXSF15RPxCmVyRRHzTwRzL7plcM3vjxMCTRKPik8nOWVAAM2sItHXO3Z/CuqeB2YF0l61g5RsAtc05V9rfsaRVMNwtzHelYDjwB94gxhNcAM2BLeI3yuRKjnOqPod4l2UDQeJLyn+lulX29wDejSBSshR4EGiTdeGcOTPrC/zqnFuVaNmFwAXOuff9F1mmuAavn3cg2YQ3YX/C7Ba+u4UF0hWEO4GLnXNr/R2ISCBSJldynBT6HCYTJCOzsz0z2453A4Vkt8T0ZQ+3OOfKZn1k6Wdmm/HuY78/0bLiwC/OuYDpUmJmW0naVzo/kBcYEkiXyM2sNd5A0teBe/DuSnUr3m1yP/dnbGllZhvwGrkBMVBOJLtRI1ckwPmmstnknPvHzEoDT+Nl3UY653b5N7pTM7PDeHc6SzZIzszyAbsD4UYKkHB5vETiBruvP+u+ALu5SNOTFh0FNjjnDvkjnowI9LuFmdlg4Fq89/TuxOucc3/7JSiRAKJGrkiAM7O1wHXOuS1mNtW3+DhQ0jnX1o+hnZaZ/Qg84Zybm8K6dsBDzrnLsj6y9PONhH/ROTcz0bLOwL3OuQb+i+zM+KbYK4F3xzZ9UfiBb5rAlLjsPm5AJDtQn1yRwFfO18ANA67Dy1pFAzv8G1aavABM9GU85zjn4n03uWgPvArc7c/g0ul+YKGZdcPrJ10NuArvLmgBw8yKAi/j3RI3FxBjZrOAO51z+/wZ2+kE25RuzrmQ028lIqlRI1ck8B0ys3CgLrDGOXfEzHLjNVCyNefcVF8Xi3eBPGb2L172MAp4xDk3za8Bps8KvGPQE29qsR/wGoZb/RpV+r0NxOHdSnYz3o+m0cBbeD8+srPEU7rlBToBP+LV41ygPvCRH+LKEDOrgPdjVjONiKSDuiuIBDgzux+4DW9miGHOuem+frpjnXOX+ze6tDGzwsAVwDnAXmBlIPUB9WWij+DdLS/K3/FkhJkdBEon7idtZvmBHc65on4LLJ3MbDowyzn3UaJlHYEuzrke/oss7czsXGAacBFeF4WCvi4wLZ1zA/wanEgAUCNXJAiYWQ0gzjn3V6LneZxzv/s3spzDzH4DrnfOBUI3kVSZ2ffAjYmnrTKzWsC7gfKjCRIa68VPGggYBuwNlIGAZvYp8DUwFi/uYmZWBFgVSDN2iPiLuiuIBIe/gSt984BuB1Y452L9HFNOMwX4xMxeBLaRaBou59wSv0WVBmZ2c6Kni4HPfbdd3orXBaA3EGhz/W7Eu8LxUqJlgwmseaXrA618fdUdgHPuoK+hKyKnoUyuSIDzZdnmA/n4r1ESCbTRJPJZ5xTzLzvnXJUsDSadzGxpGjZzzrlT3UglW/FNHzYbL5mzHSgHxAIdnXP/82dsaWVma4D2zrkNJ+7UZma1genOuQv8HZ9IdqdGrkiAM7MlwKfAuBNTPZnZvXgZoOZ+DU7Ej8wsF15f7zLATry+3jH+jSrtfBn2EcAY4EVgEN5dAsc656b4MzaRQKBGrkiAM7N9eHPintz3cI9zrpj/Ist5fI2qBkBZ59wMMysAEGh3rDKzc/CmPivtnHvWzMoCIc65bX4OLcfxzRc9CG+Wiy3AROfcHL8GJRIg1MgVCXBm9gdwR+J+n77ZFV5xztXxX2Q5i5mdD8zDm/6svG8k/A1AP+dcN/9Gl3a+O559BPwENHTOFfItu9c518a/0Z2ama11zp3ne3zy7YkTOOfOzdLAzpCZXe6c+z6F5fWdcz/4IyaRQKJGrkiAM7M2eNMMfcJ/85q2AnqndCcxOTvM7Bu8LNv7ZrbfNxK+AN4tccv5O760MrNf8Bq0ixPVIy+w2TkX7u/4TsXMGjnnvvE9Pvn2xAmcc19lXVRnzswOOecKp7B8n3OuuD9iEgkkauSKBCjf3KUP4d2A4DCwFiiLd6ezmc65DX4ML8cxs/14U1a5xI2QQGuQnGjY+h6fGOwUgtf95Rw/h5dmZlbCOfevv+M4E76/twEHgMK+xydUBb51zpXyQ2giAUVTiIkErleBeniDzm4A9jvnhvg3pBxtE3Ap3mV+wLusjDeVVSBZY2bXOecWJVp2NRBocy5vMbNlwFRgdoD1i47lv64WJ08FGA88mbXhiAQmZXJFApSZ7QQucc7t9N32c7lzrrK/48qpzKw1MBl4HbgHryFyK3CLc+5zf8aWHmbWAK/rywKgK/Ae0AZo55z70Z+xpYeZlcCLvydwIV6dpgKfZvc5pM2sIl729iugSaJVDi+jfjzFgiKShBq5IgHq5P56gXZZPBj55ma9Ba9f9FZgknPuZ/9GlX6+2RR689+I/pXAUOdcF78GdoZ8jcYeeA3eMs65kn4OSUSygLoriASuMN8sCpbK82x/p61gYmZdnHOzgCEnLe/snPvQT2Glma+P90jgIuBP4FGgJDAOr+/3e/6KLROUAsKBEnj9XAOGmbUFmuLFnvi93ddvQYkECGVyRQKUmW0ilSmSfLL9nbaCSaCPhDezt4GLgUXA9UAEUAt4B3gx0AZx+e4M1sP3Lx8wE5gWSFNvmdkjeF1epuPNlTsRLxs9wzl3hz9jEwkEauSKiGSAmZ34IbEKOJ+kI+GrAO8558pmeWDpZGY7gIucc7vNrDxeN4Vmzrnlfg7tjPhmu/gIb3q9pc65eD+HlG5mthnvzoV/mNkB51xR32DGh5xzbf0dn0h2p0auiEgGmFk8XkbdUli9C3jUOfdG1kaVfin08U4xM53dmdlLzrk7zCy3cy7azPo75yYnWv+Rc66TP2NMKzM76Jwr4nu8GyjnnItJvFxEUqdGrohIJjCzr5xzqd6AILszs2N4NxE50VifA7QjwPp4n25AZiA13s3sf0Af59xqM1uCd0z2A4875yr5MzaRQKBGroiIBE0fbzM77JwrlOh5ws0tfM8DqZF7A3DEObfc101hKlAQuM0595F/oxPJ/tTIFRHJBGZWGW9u3IvwGiIJnHPn+iOmnCiYMrkikjGaQkxEJHNMBf7CuxHEMT/HkpOdbmq9UP+ElX6pzcxhZrt1W1+R01MmV0QkE5jZIaBoII7iDyZp6HZBoNwZ8OSuF75luYBdzrlz/BSWSMBQJldEJHMsx5tnNuDucBZMgmFAlpl9jddQz2tmJ0/hVh5YkfVRiQQeNXJFRDLHJuAzM5uNN3VYAufcKL9EJIHqTbzuFZcBkxMtd3g36cj2s1yIZAfqriAikgl8dwxLkXPupqyMRYKDmdVyzq3zdxwigUqNXBERkWzEzC4Fopxzf/ielwTGA3WBlcC9zrkj/otQJDCokSsikgkS3d43Gefc31kZiwQ2X5/c0c65L33P5wJlgXeAHsAq59wQ/0UoEhjUyBURyQSp3N7XATjnAmbaKvE/M/sX7xa+UWZWFNgN1HXObTCzCsAK51wFvwYpEgA08ExEJBM450ISPzez0sAjwNf+iUgCWBgQ7XvcAG/KsA0AzrmtvoaviJxGyOk3ERGR9HLO7QKGAWP8HIoEntVAF9/j7sCXJ1aYWTngoD+CEgk0yuSKiJw9NYH8/g5CAs79wHwzex2IAxolWtcN+NYvUYkEGPXJFRHJBIkm8D8hP1AHeMw5p2yupIuZFQJqABucc4cTLa8JHHbO7fA9L++c2+anMEWyNTVyRUQygZn1O2nRUeA359yf/ohHcgYzO+ScK+zvOESyIzVyRUREApSZHXbOFfJ3HCLZkQaeiYhkAjPLZWajzexvM4v0/X+0meX2d2wS1JSpEkmFBp6JiGSOZ4D6wK3AZqAi8DBQGLjLj3GJiORI6q4gIpIJzGwbcKFzbm+iZSXw+uWW819kEszUJ1ckdequICKSOSydy0Uyg84vkVSokSsikjlm4c1tep2ZnWdmLYE5vuUiZ0ttfwcgkl2pu4KISCbwDTB7COgJlAW2A9OAJ5xz0acqK5ISM9tKygPLooBtwMfAa8652CwNTCRAKJMrIpIBZtbQzJ52zkU750Y556o55/I756oDeYBL/B2jBKyXgP3AaGAA8BiwF3gbmAHcATzlt+hEsjllckVEMsDMFgATnHMLUljXErjNOdcm6yOTQGdmq4FrTtzdzLesHPC5c66O7+5nXzrnKvgtSJFsTI1cEZEMMLPtwLnOubgU1oUBW5xzZbM+Mgl0ZrYPqOScO5RoWVHgH+dcMTMz4JBuBiGSMnVXEBHJmMJAajd8yAWoASJnaj4w18yuNrNaZnY18JFvOcAVwCZ/BSeS3amRKyKSMeuAa1NZd61vvciZGAR8D0wEfgHeAH7Eu+EIwN9AK/+EJpL9qbuCiEgGmFlP4HlgCDDHORdvZiFAe+BV4G7n3DQ/higikiPptr4iIhngnJtqZqWBd4E8ZvYvUAJvmqdH1MCVjPANLrsQKJh4uXPuLf9EJBI4lMkVEckEZlYYr4/kOXjTPK1MPGBIJL3M7AFgFPAbcCzRKueca+GfqEQChxq5IiIi2ZCZ7Qauds6t8ncsIoFIA89ERESyp+No4KLIGVMjV0REJHt6GHjZzMqYWUjif/4OTCQQqLuCiIhINmRm8b6Hib+oDa9PbqgfQhIJKJpdQUREJHuq7O8ARAKZMrkiIiIiEnSUyRUREckmzOwN59xA3+P3SdpVIYFzrm+WBiYSgNTIFRERyT7+SfR4o9+iEAkC6q4gIiIiIkFHmVwREZFswszSdCcz59ySsx2LSKBTJldERCSbMLN/TlpUDq9f7l68W0YbsM05VyWrYxMJNMrkioiIZBPOuYRpw8zsAbyG7cPOuWNmlh94DK/BKyKnoUyuiIhINmRme4CyzrmYRMtyATuccyX9F5lIYNCtAUVERLKno0D9k5ZdBhzzQywiAUfdFURERLKnh4HPzGw+sBWoALQGbvNrVCIBQt0VREREsikzqw10AsoCO4EPnXNr/BuVSGBQI1dEREREgo66K4iIiGRDuq2vSMaokSsiIpI9nXxb39JAZ2CKH2IRCTjqriAiIhIgzKwe8Ihzro2/YxHJ7tTIFRERCRBmFgbsc84V9ncsItmduiuIiIhkQ2bW4qRF+YHugGZXEEkDNXJFRESyp8knPT8K/Ar0yPpQRAKPuiuIiIiISNBRJldERCSbMLOQtGznnIs/27GIBDo1ckVERLKPWFKZG9fHfOtDsyYckcClRq6IiEj2UdnfAYgEC/XJFRERyYbMLA8Q75yLSbQsFxDy//buPeSyqozj+PfXqFOWE1PjbcyxMSMsLUmoZCKKGLUrGvmPFBpMKGSmBglqXiAipTQqu/xhaJexlGrK0FRQkcor5NhFo9Qcx6J0zJxhdEbHpz/2eetwPOe9qPPufc58P7A5511r77Wf9/xzHtZ59lpVtaW9yKTxMKvaH0mSNO+uBw4daDsUuLaFWKSx40yuJEkdlOTfwKuq74u692Dahqpa3F5k0nhwJleSpG76D7DnQNueNOvlSpqBSa4kSd30E2B1n8fDBgAABl1JREFUkoOS7JrkYOB7wBUtxyWNBZNcSZK66UzgHuB2YCNwK/Bn4Iw2g5LGhTW5kiR1WJIAS4BHyy9tadZMciVJ6qgkrwTeALyiv72qbmgnIml8uBmEJEkdlOR44GJgE7C5r6uA/duISRonzuRKktRBSR4GVlXVNW3HIo0jk1xJkjooyT+BpVW1re1YpHHk6gqSJHXT+cBZvQ0gJM2RM7mSJHVQkoeAvYCtwIb+vqpa1kpQ0hjxwTNJkrrpY20HII0zZ3IlSZI0cazzkSSpg5LsnOS8JPcnear3el6SXdqOTRoHlitIktRNFwBvA04EHgT2Az4PLAJObTEuaSxYriBJUgclWQ+8pao29LUtAdZW1T7tRSaNB8sVJEnqpsyxXVIfk1xJkrrpSuCqJEckOTDJkcCaXrukGViuIElSB/UeMDsLOBZYCjwMXA58oaq2thmbNA6cyZUkqUOSrEhyflVtraqzq+qAqtq1ql4PLATe2naM0jgwyZUkqVvOAG4e0XcjcOY8xiKNLcsVJEnqkCQPA8uqatuQvp2AdVW1dP4jk8aLM7mSJHXLImDUhg87A7vNYyzS2DLJlSSpW+4FDh/Rd3ivX9IM3PFMkqRuuQj4TpIFwJqqejbJS4CjgIuB09oMThoXJrmSJHVIVa1OshdwGbAwyaPAEmALcE5VXd5qgNKY8MEzSZI6KMki4DDg1cAG4JaqeqLdqKTxYZIrSZKkieODZ5IkSZo4JrmSJEmaOCa5kjSPkhyfpPqOjUnWJjmpt9A/SW4acs5vknx4yHiXDpzbf6zpO+/cIWP+JcnqJEcMGfe1vfNWDek7LMkVSf6eZGuSDUmuT3JckgVJ/jZNTP87Rnwe/cfjffd890Dfk0nWJ7k6yaoko9aVlbSDcnUFSWrHMcB6moX/jwG+DuwBnN3rvxs4ofd+Gc1Wrz9NsqKqbhsY6xHgOQkw8NiQtncC24BdgeXAR4FfJfkBcFxVPTtd0ElOAS4EbgBOBx4EFtOs3/ot4HHgaGBh32XfBBb0/T/DTH0e/Z4Zct7JwB00myIsBVbSLKt1UpKVVfXIdPFL2nGY5EpSO+6qqr/23l+X5ADgM/w/yd1YVbf23t+a5LfAOuATwGCSu7Xv3JncVlX9yeMlSU6lSVzvAr4y6sIk7+qd942qOnmg++dJLgReXlV3D1z3BLDTDDH2fx7TuWdgnB8nuYQm6f4u8KFZjCFpB2C5giR1wx3AoiR7DOusqvU0M7bLXuwbV9VFwO9okuzpnE4zO/y5EePcN5jgzoequgX4NvDBJK+b7/tL6iaTXEnqhuU0ZQSbhnUm2Y1mvdT7RvTvNOTIHO5/DbBvkqFJdG/3rfcA11XVU3MYd7YWDIl/Lt9RV/deV2yH2CSNIZNcSWrHVFK3OMkJwEeAX1bV5qkT+pK95TQ/xT9Gs+XroH2Ap4ccn51DPOt6r3uP6F8CvIymBnd7uJfnxv+LOVw/U/ySdjDW5EpSO+7te/8s8EPglL62FTSJ3pQtwMqqun/IWP8CPjCk/aE5xDM169vWDkFH89wHzx6fw/Vtxy+pY0xyJakdU0ndRuDBISUAa4FVNKsSvAk4H7gyycFDVhB4uqrufIHx7Nt7/ceI/g3Ak8B+L/A+o/xhlg+ejTJT/JJ2MCa5ktSOmZK6TX2J621JHqBZQeBc4FPbIZ73A+uqaujsb1U9k+QmYGWShVW1ZTvE8EJMzWT/utUoJHWGNbmSNAaq6kbgZ8CqJK95McfuLSF2CM3yYNP5Es3DbxeMGGd5kje/mLHNRpLDgBOBNVX1wHzfX1I3OZMrSePjHOAomqW8Pt3XvkuSdww5f/OQJb3enmQb8FJgf5rNIN4HXAZ8bbqbV9XNSU4DLkzyRuBSmge+FgPvpSmvOJZmI4u5OiTJkiHtdw6s63tgkk00319702xC8XHgT8Ann8d9JU0ok1xJGhNV9fskP6KZzf1iVU3Vn+4O3DLkkj8CBw20Tf2cv5mmfvV24MiqunaWMXw1ye3AqcCXaVZd2AjcSbOj2VVz+Jf6XTmifXfg0b6/pxLxLTR1wmtpyje+X1Vbn+e9JU2gVPkgqiRJkiaLNbmSJEmaOCa5kiRJmjgmuZIkSZo4JrmSJEmaOCa5kiRJmjgmuZIkSZo4JrmSJEmaOCa5kiRJmjgmuZIkSZo4/wX0cl9AifKiNgAAAABJRU5ErkJggg==\n",
+      "text/plain": [
+       "<Figure size 720x720 with 2 Axes>"
+      ]
+     },
+     "metadata": {
+      "needs_background": "light"
+     },
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "show_confusion_matrix(y_true,y_pred)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "f03e95e4",
+   "metadata": {
+    "hidden": true
+   },
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "accelerator": "GPU",
+  "colab": {
+   "machine_shape": "hm",
+   "name": "training_v2.ipynb",
+   "provenance": [],
+   "toc_visible": true
+  },
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.8.12"
+  },
+  "toc": {
+   "base_numbering": 1,
+   "nav_menu": {},
+   "number_sections": true,
+   "sideBar": true,
+   "skip_h1_title": false,
+   "title_cell": "Table of Contents",
+   "title_sidebar": "Contents",
+   "toc_cell": true,
+   "toc_position": {
+    "height": "calc(100% - 180px)",
+    "left": "10px",
+    "top": "150px",
+    "width": "288px"
+   },
+   "toc_section_display": true,
+   "toc_window_display": true
+  },
+  "varInspector": {
+   "cols": {
+    "lenName": 16,
+    "lenType": 16,
+    "lenVar": 40
+   },
+   "kernels_config": {
+    "python": {
+     "delete_cmd_postfix": "",
+     "delete_cmd_prefix": "del ",
+     "library": "var_list.py",
+     "varRefreshCmd": "print(var_dic_list())"
+    },
+    "r": {
+     "delete_cmd_postfix": ") ",
+     "delete_cmd_prefix": "rm(",
+     "library": "var_list.r",
+     "varRefreshCmd": "cat(var_dic_list()) "
+    }
+   },
+   "types_to_exclude": [
+    "module",
+    "function",
+    "builtin_function_or_method",
+    "instance",
+    "_Feature"
+   ],
+   "window_display": false
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}
diff --git a/notebooks/arthur/utils.ipynb b/notebooks/arthur/utils.ipynb
new file mode 100644
index 0000000..69551cd
--- /dev/null
+++ b/notebooks/arthur/utils.ipynb
@@ -0,0 +1,373 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "9b1e0525",
+   "metadata": {},
+   "source": [
+    "# Imports and data loading"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "id": "d09b1154",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-14T17:47:42.080464Z",
+     "start_time": "2022-02-14T17:47:39.447048Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "\n",
+    "import pandas as pd\n",
+    "import numpy as np \n",
+    "import matplotlib.pyplot as plt\n",
+    "import seaborn as sns\n",
+    "\n",
+    "from IPython.display import HTML"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 12,
+   "id": "36a2c14a",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-14T17:48:09.482598Z",
+     "start_time": "2022-02-14T17:48:09.468757Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "# loading only a sample for testing purposes\n",
+    "df = pd.read_csv('../raw_data/train.csv',nrows=300)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "ee8eeff1",
+   "metadata": {},
+   "source": [
+    "# Showing ground truth and prediction in text"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 112,
+   "id": "4a3f06ab",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-15T14:28:01.286117Z",
+     "start_time": "2022-02-15T14:28:01.183939Z"
+    }
+   },
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<style>\n",
+       ".column {\n",
+       "  float: left;\n",
+       "  width: 50%;\n",
+       "  /*margin-top:10px;*/\n",
+       "  border-right: 2px dotted black;\n",
+       "  text-align:justify;\n",
+       "  }\n",
+       ".column:first-child {\n",
+       "  padding-right:16px;\n",
+       "  padding-left:0px;\n",
+       "  }\n",
+       "\n",
+       ".column:last-child {\n",
+       "  padding-right:0px;\n",
+       "  padding-left:15px;\n",
+       "  border: none;\n",
+       "  }\n",
+       "\n",
+       ".row:after {\n",
+       "  content: \"\";\n",
+       "  display: table;\n",
+       "  clear: both;\n",
+       "  text-align:left\n",
+       "  }\n",
+       "\n",
+       ".title {\n",
+       "  text-align:center;\n",
+       "  padding-bottom: 15px;\n",
+       "  padding-top: 0;\n",
+       "  margin-top: 0;\n",
+       "  border-bottom: 2px dotted black\n",
+       " }\n",
+       "\n",
+       "\n",
+       "div.content > * {\n",
+       "  text-align: center;\n",
+       "  font-weight: bold;\n",
+       "  padding-right: 10px;\n",
+       "  padding-left: 10px;\n",
+       "  padding-top: 5px ;\n",
+       "  padding-bottom:5px;\n",
+       " }\n",
+       "\n",
+       ".content {\n",
+       "  display:flex;\n",
+       "  justify-content:space-evenly;\n",
+       "  margin: 0px auto;\n",
+       "  margin-bottom: 5px;\n",
+       "  background: #FFFFFF;\n",
+       "  padding-block: 10px;\n",
+       "  max-width:auto;\n",
+       "  border: 2px solid black;\n",
+       "  }\n",
+       "\n",
+       "/*styles for each discourse type*/\n",
+       "\n",
+       "Lead {\n",
+       "    background-color:#ff8585;\n",
+       "  }\n",
+       "Position {\n",
+       "  background-color: #7cf0ff;\n",
+       "  }\n",
+       "Evidence {\n",
+       "  background-color: #badcfc;\n",
+       "  }\n",
+       "Claim {\n",
+       "  background-color: #3a3dff;\n",
+       "  }\n",
+       "Concluding_Statement {\n",
+       "  background-color: #ff7df9;\n",
+       "  }\n",
+       "Counterclaim {\n",
+       "  background-color: #4d92e0;\n",
+       "  }\n",
+       "Rebuttal {\n",
+       "  background-color: #ffd57c;\n",
+       "  }\n",
+       "</style>"
+      ],
+      "text/plain": [
+       "<IPython.core.display.HTML object>"
+      ]
+     },
+     "execution_count": 112,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "#import custom.css into this notebook\n",
+    "\n",
+    "def css():\n",
+    "    styles = open(\"./styles/custom.css\", \"r\").read()\n",
+    "    return HTML('<style>'+styles+'</style>')\n",
+    "css()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 104,
+   "id": "703baf88",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-14T20:22:58.368090Z",
+     "start_time": "2022-02-14T20:22:58.289671Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "def render_html(df):\n",
+    "    return \"<{0} style='padding: 2px'>{1} <strong> [{0}] </strong></{0}>\".format(df['discourse_type'],df['discourse_text'])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 106,
+   "id": "eafb22ad",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-14T20:24:18.459300Z",
+     "start_time": "2022-02-14T20:24:18.452825Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "def comparison_text(prediction=pred, ground_truth=true):\n",
+    "    html = f\"\"\"\n",
+    "    <div class=\"content\">\n",
+    "     <span style=\"font-size:16px\">Legend --></span>\n",
+    "      <lead>Lead</lead>\n",
+    "      <Position>Position</Position>\n",
+    "      <Claim>Claim</Claim>\n",
+    "      <Counterclaim>Counterclaim</Counterclaim>\n",
+    "      <Rebuttal>Rebuttal</Rebuttal>\n",
+    "      <Evidence>Evidence</Evidence>\n",
+    "      <Concluding_Statement>Concluding_Statement</Concluding_Statement>\n",
+    "    </div>\n",
+    "\n",
+    "    <div class=\"row\">\n",
+    "      <div class=\"column\">\n",
+    "        <h2 class=\"title\">Prediction</h2>\n",
+    "        <p style=\"text-align:justify\">{pred}</p>\n",
+    "      </div>\n",
+    "      <div class=\"column\">\n",
+    "        <h2 class=\"title\">Ground Truth</h2>\n",
+    "        <p style=\"text-align:justify\">{true}</p>\n",
+    "      </div>\n",
+    "    </div>\n",
+    "    \n",
+    "    \"\"\"\n",
+    "    \n",
+    "    return HTML(html)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 102,
+   "id": "b513750b",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-14T18:15:43.568970Z",
+     "start_time": "2022-02-14T18:15:43.554821Z"
+    }
+   },
+   "outputs": [],
+   "source": [
+    "# as if the output were post processed ; just to see \n",
+    "\n",
+    "df['html'] = df.apply(render_html, axis=1) #transform discourse_text in html txt with formating \n",
+    "df_essays = df.groupby('id').agg({'html':' '.join,}).reset_index() #groupby essay id\n",
+    "\n",
+    "true = df_essays.loc[7,'html'].replace('Concluding Statement','Concluding_Statement')\n",
+    "pred = df_essays.loc[15,'html'].replace('Concluding Statement','Concluding_Statement')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 115,
+   "id": "756fcffb",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2022-02-15T18:32:31.663131Z",
+     "start_time": "2022-02-15T18:32:31.657046Z"
+    }
+   },
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "\n",
+       "    <div class=\"content\">\n",
+       "     <span style=\"font-size:16px\">Legend --></span>\n",
+       "      <lead>Lead</lead>\n",
+       "      <Position>Position</Position>\n",
+       "      <Claim>Claim</Claim>\n",
+       "      <Counterclaim>Counterclaim</Counterclaim>\n",
+       "      <Rebuttal>Rebuttal</Rebuttal>\n",
+       "      <Evidence>Evidence</Evidence>\n",
+       "      <Concluding_Statement>Concluding_Statement</Concluding_Statement>\n",
+       "    </div>\n",
+       "\n",
+       "    <div class=\"row\">\n",
+       "      <div class=\"column\">\n",
+       "        <h2 class=\"title\">Prediction</h2>\n",
+       "        <p style=\"text-align:justify\"><Lead style='padding: 2px'>The ability to stay connected to people we know despite distance was originally brought to fruition by the use of letters. This system was found to be rather slow and new pathways were searched for until the invention of the telegram; the people thought it was an invention of the millennia, but after time it too was thought of as slow until the invention of the telephone. Today, a telephone is in the hand or pocket of a majority of the seven billion people on planet earth <strong> [Lead] </strong></Lead> <Position style='padding: 2px'> However, this device is taken to areas that it is irresponsible and dangerous.  <strong> [Position] </strong></Position> <Claim style='padding: 2px'>Within a vehicle capable of traveling upwards of one hundred miles per hour any possible distraction can become fatal spontaneously  <strong> [Claim] </strong></Claim> <Evidence style='padding: 2px'>The most common of these distractions is a cell phone, with its capabilities to connect us to anyone also in ownership of one, it is easy to pick it up whenever it sounds. In that split second of reaching over for a phone, eyes no longer on the road, it is impossible to no an exact location of anything, making an extremely dangerous action. For the myriad of possibilities that lead to serious injury cell phones should stay in the current state they are in regards of the law, but taken as a more serious offense.  <strong> [Evidence] </strong></Evidence> <Counterclaim style='padding: 2px'>Conversely people may believe that laws in present need to change, becoming less restrictive. People have the right to communicate with whom they wish, when or wherever they may choose to do so. <strong> [Counterclaim] </strong></Counterclaim> <Rebuttal style='padding: 2px'>The problem becomes apparent that this is a selfish process of thought; people aren't thinking of those they share the road with. Laws currently in place are not to punish people making poor choices, they are an attempt to keep people safe.  <strong> [Rebuttal] </strong></Rebuttal> <Concluding_Statement style='padding: 2px'>The creation of telecommunication devices was to keep connected to others without regard to the obsession that would encompass the human mind that was bound to follow. The safety of people is top priority without exemption.  <strong> [Concluding_Statement] </strong></Concluding_Statement></p>\n",
+       "      </div>\n",
+       "      <div class=\"column\">\n",
+       "        <h2 class=\"title\">Ground Truth</h2>\n",
+       "        <p style=\"text-align:justify\"><Position style='padding: 2px'>Drivers should absolutely never be able to use a cell phone while driving. <strong> [Position] </strong></Position> <Evidence style='padding: 2px'>This is evident as in recent years there has been a big problem in our society with drivers using their cell phones while driving. It has gotten so bad that it has become illegal because driving while using a cell phone is more likely to be the cause of a fatal accident, than when a person drives under the influence of drugs or alcohol.\n",
+       " <strong> [Evidence] </strong></Evidence> <Evidence style='padding: 2px'>Use a hands free device at all times while driving. I say this because a cell phone will take your attention off of the task at hand. A driver's focus should be maintained on only driving to ensure safety <strong> [Evidence] </strong></Evidence> <Claim style='padding: 2px'> It has been proven that driving while using a cell phone is the cause of more vehicular accidents resulting in a fatality than those caused by drivers who are under the influence of drugs or alcohol.  <strong> [Claim] </strong></Claim> <Evidence style='padding: 2px'>In two thousand and thirteen there were three thousand one hundred fifty-four people killed in distracted driver related crashes. In just one year's time, there are three hundred and ninety thousand people injured in accidents caused by texting and driving. I personally do not use my cell phone when I drive, because it distracts me very easily. <strong> [Evidence] </strong></Evidence> <Claim style='padding: 2px'> I suggest that drivers should be able to use their cell phone for an emergency purpose only. Even in a crisis situation, the driver should pull over to the side of the road before making a call.  <strong> [Claim] </strong></Claim> <Claim style='padding: 2px'>I also encourage that the cell phone companies should be made to have the cell phone come from the factory to where if your cellphone is moving, it will not be allowed to ring or give you any notifications.  <strong> [Claim] </strong></Claim> <Evidence style='padding: 2px'>In order to get your notifications, you will have to manually tell your cellphone to read that you are not driving for it to allow you to open the phone. It will then give you the notifications and or missed calls you have missed. In addition, it will not allow you to access absolutely any of your social media platforms or let you get social media notifications while you are driving.  <strong> [Evidence] </strong></Evidence> <Concluding_Statement style='padding: 2px'>In conclusion, my opinion on whether drivers should be using their cell phones while driving is that they should refrain from using their cell phone while driving. This is because driving is a task that requires a lot of focus and a cell phone is very distracting.\n",
+       "\n",
+       "Driving while using a cell phone is now illegal because of how dangerous it is.     <strong> [Concluding_Statement] </strong></Concluding_Statement></p>\n",
+       "      </div>\n",
+       "    </div>\n",
+       "    \n",
+       "    "
+      ],
+      "text/plain": [
+       "<IPython.core.display.HTML object>"
+      ]
+     },
+     "execution_count": 115,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "comparison_text(pred,true)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "7eeb7b82",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.8.12"
+  },
+  "toc": {
+   "base_numbering": 1,
+   "nav_menu": {},
+   "number_sections": true,
+   "sideBar": true,
+   "skip_h1_title": false,
+   "title_cell": "Table of Contents",
+   "title_sidebar": "Contents",
+   "toc_cell": false,
+   "toc_position": {},
+   "toc_section_display": true,
+   "toc_window_display": false
+  },
+  "varInspector": {
+   "cols": {
+    "lenName": 16,
+    "lenType": 16,
+    "lenVar": 40
+   },
+   "kernels_config": {
+    "python": {
+     "delete_cmd_postfix": "",
+     "delete_cmd_prefix": "del ",
+     "library": "var_list.py",
+     "varRefreshCmd": "print(var_dic_list())"
+    },
+    "r": {
+     "delete_cmd_postfix": ") ",
+     "delete_cmd_prefix": "rm(",
+     "library": "var_list.r",
+     "varRefreshCmd": "cat(var_dic_list()) "
+    }
+   },
+   "types_to_exclude": [
+    "module",
+    "function",
+    "builtin_function_or_method",
+    "instance",
+    "_Feature"
+   ],
+   "window_display": false
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}