Merge pull request #124 from KonstantinUshenin/add_miccai_dataset

Toy data from miccai2015 dataset

nndt/datasets/__init__.py

@@ -1,8 +1,10 @@
 from .acdc import ACDC
+from .miccai2015 import MICCAI2015
 from .utils import *
 
 list_of_datasets = [
     "ACDC",
+    "MICCAI2015",
 ]
 
 source_url = "https://github.com/KonstantinUshenin/nndt"

nndt/datasets/acdc.py

@@ -1,11 +1,11 @@
-from nndt.datasets.dataset import dataset
+from nndt.datasets.dataset import Dataset
 
 
-class ACDC(dataset):
+class ACDC(Dataset):
     def __init__(self, name="ACDC_5", to_path=None):
-        super().__init__()
+        super().__init__(name=name, to_path=to_path)
 
-        __dict = {
+        self._dict = {
             "ACDC_5": [
                 [
                     "https://drive.google.com/file/d/1UzC2WPkjMQSxzI5sj1rMT47URuZbQhYb/view?usp=sharing",
@@ -36,17 +36,14 @@ def __init__(self, name="ACDC_5", to_path=None):
             ],
         }
 
-        if name in __dict.keys():
+        if name in self._dict.keys():
             if "test" in name:
                 self.name = "ACDC_5"
             else:
                 self.name = name
             self.to_path = to_path
-            self.urls, self.hash = __dict[name]
+            self.urls, self.hash = self._dict[name]
         else:
             raise ValueError(
-                f'name must be in {[key for key in __dict if "test" not in key]}'
+                f'name must be in {[key for key in self._dict if "_test" not in key]}'
             )
-
-    def load(self):
-        super(ACDC, self).load()

nndt/datasets/dataset.py

@@ -1,4 +1,5 @@
 import os
+import warnings
 from pathlib import Path
 
 from nndt import datasets
@@ -10,12 +11,16 @@
 )
 
 
-class dataset:
+class Dataset:
     def __init__(self, name=None, to_path=None):
         self.name = name
         self.to_path = to_path
         self.hash = None
         self.urls = None
+        self._dict = None
+
+    def dataset_list(self):
+        return [key for key in self._dict if "_test" not in key]
 
     def load(self) -> None:
 
@@ -32,7 +37,7 @@ def load(self) -> None:
                     assert _check_md5(z, self.hash)
                     _extract_7z_file(z, self.to_path)
                 except Exception as e:
-                    print(url, str(e))
+                    warnings.warn(str(e))
                     continue
             else:
                 try:
@@ -41,7 +46,7 @@ def load(self) -> None:
                     assert _check_md5(z, self.hash)
                     _extract_7z_file(z, self.to_path)
                 except Exception as e:
-                    print(url, str(e))
+                    warnings.warn(str(e))
                     continue
             complete = True
             os.remove(z)

nndt/datasets/miccai2015.py

@@ -0,0 +1,56 @@
+from nndt.datasets.dataset import Dataset
+
+
+class MICCAI2015(Dataset):
+    def __init__(self, name="left_adrenal_gland_10", to_path=None):
+        super().__init__(name=name, to_path=to_path)
+
+        self._dict = {
+            "left_adrenal_gland_10": [
+                [
+                    "https://drive.google.com/file/d/1LFJs5tW-acZOXTVVHh-WpWRNkAgtKaDd/view?usp=sharing",
+                    "https://www.dropbox.com/s/xyarraopmj069de/left_adrenal_gland.7z?raw=1",
+                ],
+                "03c5f5c7e33a57ed71f497725d6925db",
+            ],
+            "stomach_10": [
+                [
+                    "https://drive.google.com/file/d/189tTsVX89qUijClkPkZfBiX5hRMupW03/view?usp=sharing",
+                    "https://www.dropbox.com/s/uzhurjucqqnansh/stomach.7z?raw=1",
+                ],
+                "3a7a817748c43194fa75b1e5ab798d56",
+            ],
+            "wrong_url_test": [
+                [
+                    "https://drive.google.com/file",
+                    "a",
+                    "b",
+                    "https://a.com",
+                    "https://www.dropbox.com/s/m",
+                ],
+                "03c5f5c7e33a57ed71f497725d6925db",
+            ],
+            "wrong_hash_test": [
+                [
+                    "https://drive.google.com/file/d/1UzC2WPkjMQSxzI5sj1rMT47URuZbQhYb/view?usp=sharing",
+                    "https://www.dropbox.com/s/6fomqxbjs0iu79m/ACDC_5.7z?raw=1",
+                ],
+                "34d007546353673899c73337d38c9c12  ",
+            ],
+            "dropbox_test": [
+                [
+                    "https://www.dropbox.com/s/xyarraopmj069de/left_adrenal_gland.7z?raw=1"
+                ],
+                "03c5f5c7e33a57ed71f497725d6925db",
+            ],
+        }
+
+        if name in self._dict.keys():
+            self.name = "MICCAI2015"
+            self.to_path = to_path
+            self.urls, self.hash = self._dict[name]
+
+        else:
+            raise ValueError(
+                f'name must be in {[key for key in self._dict if "_test" not in key]}'
+            )

tests/test_load_datasets.py

@@ -1,7 +1,7 @@
 import os
 import shutil
 
-from nndt.datasets import ACDC
+from nndt.datasets import ACDC, MICCAI2015
 from tests.base import BaseTestCase
 
 DATASETS_FOLDER = ".datasets"
@@ -38,3 +38,9 @@ def test_random_name(self):
     def test_wrong_hash(self):
         with self.assertRaises(ConnectionError):
             ACDC("wrong_hash_test").load()
+
+    def test_dataset_list(self):
+        _list = MICCAI2015().dataset_list()
+        self.assertTrue(_list is not None)
+        self.assertTrue(len(_list) > 0)
+        self.assertTrue(len([item for item in _list if "_test" in item]) == 0)