Text Classification for Spam Detection

src/apps/app.py

@@ -0,0 +1,16 @@
+import streamlit as st
+from utils.predict import predict_message
+
+st.title("Spam Detection System")
+st.write("Enter a message to classify it as Spam or Not Spam.")
+
+# Input field
+user_input = st.text_area("Message")
+
+if st.button("Classify"):
+    if user_input.strip():
+        prediction, confidence = predict_message(user_input)
+        st.subheader(f"Prediction: {prediction}")
+        st.write(f"Confidence: {confidence:.2%}")
+    else:
+        st.error("Please enter a valid message!")

src/apps/pages/models/model.py

@@ -0,0 +1,40 @@
+import pandas as pd
+from sklearn.model_selection import train_test_split
+from sklearn.naive_bayes import MultinomialNB
+from sklearn.metrics import classification_report
+import pickle
+from utils.preprocess import Preprocessor
+import os
+
+# Create dataset folder if it doesn't exist
+if not os.path.exists("dataset"):
+    os.makedirs("dataset")
+    print("Dataset folder created. Please place spam.csv in the dataset folder.")
+    exit()
+
+
+# Load dataset
+df = pd.read_csv("dataset/spam.csv", encoding="latin-1")[["v1", "v2"]]
+df.columns = ["label", "message"]
+df["label"] = df["label"].map({"ham": 0, "spam": 1})
+
+# Preprocess data
+preprocessor = Preprocessor()
+X = preprocessor.fit_transform(df["message"])
+y = df["label"]
+
+# Train-test split
+X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)
+
+# Train model
+model = MultinomialNB()
+model.fit(X_train, y_train)
+
+# Evaluate
+y_pred = model.predict(X_test)
+print(classification_report(y_test, y_pred))
+
+# Save model and vectorizer
+with open("utils/model.pkl", "wb") as model_file, open("utils/vectorizer.pkl", "wb") as vec_file:
+    pickle.dump(model, model_file)
+    pickle.dump(preprocessor.vectorizer, vec_file)

src/requirements.txt

@@ -0,0 +1,4 @@
+streamlit
+scikit-learn
+pandas
+numpy

src/utils/predict.py

@@ -0,0 +1,14 @@
+import pickle
+from utils.preprocess import Preprocessor
+
+# Load model and vectorizer
+with open("utils/model.pkl", "rb") as model_file, open("utils/vectorizer.pkl", "rb") as vec_file:
+    model = pickle.load(model_file)
+    vectorizer = pickle.load(vec_file)
+
+def predict_message(message):
+    preprocessor = Preprocessor()
+    transformed_message = vectorizer.transform([preprocessor.clean_text(message)])
+    prediction = model.predict(transformed_message)[0]
+    confidence = model.predict_proba(transformed_message).max()
+    return ("Spam" if prediction == 1 else "Not Spam", confidence)

src/utils/preprocess.py

@@ -0,0 +1,19 @@
+import re
+from sklearn.feature_extraction.text import TfidfVectorizer
+
+class Preprocessor:
+    def __init__(self):
+        self.vectorizer = TfidfVectorizer(stop_words="english", max_features=3000)
+
+    def clean_text(self, text):
+        text = re.sub(r"[^a-zA-Z\s]", "", text)  # Remove special characters
+        text = re.sub(r"\s+", " ", text)         # Remove extra spaces
+        return text.lower()
+
+    def fit_transform(self, messages):
+        cleaned_messages = [self.clean_text(msg) for msg in messages]
+        return self.vectorizer.fit_transform(cleaned_messages)
+
+    def transform(self, messages):
+        cleaned_messages = [self.clean_text(msg) for msg in messages]
+        return self.vectorizer.transform(cleaned_messages)