Upload WISDM.py

WISDM.py

@@ -36,6 +36,7 @@ year={2020}
 # TODO: Add description of the dataset here
 # You can copy an official description
 _DATASET = "WISDM"
+_SHAPE = (3, 100)
 _DESCRIPTION = """\
 This new dataset is designed to solve this great NLP task and is crafted with a lot of care.
 """
@@ -50,6 +51,15 @@ _LICENSE = ""
 # The HuggingFace Datasets library doesn't host the datasets but only points to the original files.
 # This can be an arbitrary nested dict/list of URLs (see below in `_split_generators` method)
 _URL = [f"https://huggingface.co/datasets/lynn-miller/{_DATASET}"]
+_URLS = {
+    'data': f"{_DATASET}_X.npy",
+    'labels': f"{_DATASET}_y.npy",
+    'fold_0': "test_indices_fold_0.txt",
+    'fold_1': "test_indices_fold_1.txt",
+    'fold_2': "test_indices_fold_2.txt",
+    'fold_3': "test_indices_fold_3.txt",
+    'fold_4': "test_indices_fold_4.txt",
+    }
 
 
 # TODO: Name of the dataset usually matches the script name with CamelCase instead of snake_case
@@ -82,19 +92,19 @@ class Monster(datasets.GeneratorBasedBuilder):
 
     def _info(self):
         # TODO: This method specifies the datasets.DatasetInfo object which contains informations and typings for the dataset
-        data_dir = f"G:\My Drive\Postdoc\Other papers\Time Series Archive\{_DATASET}"
-        data_dir = _URL[0]
-        metadata_file = os.path.join(data_dir, "metadata", f"{_DATASET}_metadata.txt")
-        try:
-            metadata = np.loadtxt(metadata_file, delimiter=',', dtype=str)
-            shape = (int(metadata[1,1]), int(metadata[2,1]))
-            print(shape)
-        except:
-            print("Error reading metadata")
-            shape = (3, 100)
+        # data_dir = f"G:\My Drive\Postdoc\Other papers\Time Series Archive\{_DATASET}"
+        # data_dir = _URL[0]
+        # metadata_file = os.path.join(data_dir, "metadata", f"{_DATASET}_metadata.txt")
+        # try:
+        #     metadata = np.loadtxt(metadata_file, delimiter=',', dtype=str)
+        #     shape = (int(metadata[1,1]), int(metadata[2,1]))
+        #     print(shape)
+        # except:
+        #     print("Error reading metadata")
+        #     shape = (3, 100)
         features = datasets.Features(
             {
-                "X": datasets.Array2D(shape, "float32"),
+                "X": datasets.Array2D(_SHAPE, "float32"),
                 "y": datasets.Value("int64")
                 # These are the features of your dataset like images, labels ...
             }
@@ -124,27 +134,31 @@ class Monster(datasets.GeneratorBasedBuilder):
         # By default the archives will be extracted and a path to a cached folder where they are extracted is returned instead of the archive
 #        urls = _URLS[self.config.name]
 #        data_dir = f"G:\My Drive\Postdoc\Other papers\Time Series Archive\{_DATASET}"
-        data_dir = dl_manager.download_and_extract(_URL[0])
+        data = dl_manager.download_and_extract(_URLS['data'])
+        labels = dl_manager.download_and_extract(_URLS['labels'])
         if self.config.name == "full":
             return [
                 datasets.SplitGenerator(
                     name=datasets.Split.TRAIN,
                     # These kwargs will be passed to _generate_examples
                     gen_kwargs={
-                        "filepath": data_dir,
-                        "fold": self.config.name,
+                        "data": data,
+                        "labels": labels,
+                        "fold": None,
                         "split": "train",
                     },
                 ),
             ]
-        else:            
+        else: 
+            fold = dl_manager.download_and_extract(_URLS[self.config.name])
             return [
                 datasets.SplitGenerator(
                     name=datasets.Split.TRAIN,
                     # These kwargs will be passed to _generate_examples
                     gen_kwargs={
-                        "filepath": data_dir,
-                        "fold": self.config.name,
+                        "data": data,
+                        "labels": labels,
+                        "fold": fold,
                         "split": "train",
                     },
                 ),
@@ -152,7 +166,8 @@ class Monster(datasets.GeneratorBasedBuilder):
                     name=datasets.Split.TEST,
                     # These kwargs will be passed to _generate_examples
                     gen_kwargs={
-                        "filepath": data_dir,
+                        "data": data,
+                        "labels": labels,
                         "fold": self.config.name,
                         "split": "test"
                     },
@@ -160,19 +175,19 @@ class Monster(datasets.GeneratorBasedBuilder):
             ]
 
     # method parameters are unpacked from `gen_kwargs` as given in `_split_generators`
-    def _generate_examples(self, filepath, fold, split):
+    def _generate_examples(self, data, labels, fold, split):
         # TODO: This method handles input defined in _split_generators to yield (key, example) tuples from the dataset.
         # The `key` is for legacy reasons (tfds) and is not important in itself, but must be unique for each example.
-        X_file = f'{_DATASET}_X.npy' #os.path.join(filepath, f'{_DATASET}_X.npy')
-        y_file = f'{_DATASET}_y.npy' #os.path.join(filepath, f'{_DATASET}_y.npy')
-        X = np.load(X_file) #, 'r')
-        y = np.load(y_file) #, 'r')
+#        X_file = f'{_DATASET}_X.npy' #os.path.join(filepath, f'{_DATASET}_X.npy')
+#        y_file = f'{_DATASET}_y.npy' #os.path.join(filepath, f'{_DATASET}_y.npy')
+        X = np.load(data) #, 'r')
+        y = np.load(labels) #, 'r')
         if self.config.name == "full":
             for row in range(y.shape[0]):
                 yield(row, {"X": X[row], "y": y[row]})
         else:
-            indices_file = os.path.join(filepath, f"test_indices_{fold}.txt")
-            test_indices = np.loadtxt(indices_file, dtype='int')
+#            indices_file = os.path.join(filepath, f"test_indices_{fold}.txt")
+            test_indices = np.loadtxt(fold, dtype='int')
             if split == "test":
                 for row in test_indices:
                     yield(int(row), {"X": X[row], "y": y[row]})