feat: add loading script

.gitattributes

@@ -56,3 +56,5 @@ reutersidf10k_test.csv filter=lfs diff=lfs merge=lfs -text
 test.csv filter=lfs diff=lfs merge=lfs -text
 reutersidf10k_train.csv filter=lfs diff=lfs merge=lfs -text
 train.csv filter=lfs diff=lfs merge=lfs -text
+test.npy filter=lfs diff=lfs merge=lfs -text
+train.npy filter=lfs diff=lfs merge=lfs -text

.gitignore

@@ -0,0 +1 @@
+__pycache__

reuters10k.py

@@ -0,0 +1,65 @@
+import os
+
+import datasets
+from sklearn.preprocessing import MinMaxScaler, LabelEncoder, StandardScaler
+import numpy as np
+
+
+# TODO: Add BibTeX citation
+# Find for instance the citation on arxiv or on the dataset repo/website
+_CITATION = """\
+@InProceedings{huggingface:dataset,
+title = {A great new dataset},
+author={huggingface, Inc.
+},
+year={2020}
+}
+"""
+
+_URL = "https://huggingface.co/datasets/wwydmanski/reuters10k/raw/main/"
+
+class Reuters10K(datasets.GeneratorBasedBuilder):
+    """TODO: Short description of my dataset."""
+
+    VERSION = datasets.Version("0.0.1")
+
+    def _info(self):
+        return datasets.DatasetInfo(
+            description="Reuters10K dataset",
+            version=Reuters10K.VERSION,
+        )
+
+    def _split_generators(self, dl_manager):
+        train_url = _URL + "train.npy"
+        test_url = _URL + "test.npy"
+
+        data_dir = dl_manager.download_and_extract([train_url, test_url])
+        return [
+            datasets.SplitGenerator(
+                name=datasets.Split.TRAIN,
+                gen_kwargs={
+                    "filepath": os.path.join(data_dir, "train.npy")
+                },
+            ),
+            datasets.SplitGenerator(
+                name=datasets.Split.TEST,
+                gen_kwargs={
+                    "filepath": os.path.join(data_dir, "dev.npy")
+                },
+            )
+        ]
+
+    def _generate_examples(self, filepath):
+        """Yields examples."""
+        train_dataset = np.load(filepath, allow_pickle=True)
+
+        X_train = train_dataset.item()['data']
+        Y_train = train_dataset.item()['label']
+
+        scaler = MinMaxScaler()
+        X_train = scaler.fit_transform(X_train)
+
+        # yield "key", {"text": text, "label": label}
+        for i, (x, y) in enumerate(zip(X_train, Y_train)):
+            yield i, {"features": x, "label": y}
+        

test.csv → test.npy

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:389a149bd99b460c8b37ac6ade993d7b99134d425a5912a0defa9fa4d319c7fb
-size 100050000
+oid sha256:5f85c13eac62427cfbbd8d7f5de51e91d70daa16868be2caa128cb036eb82e81
+size 32392405

train.csv → train.npy

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fca58a1599dd252c228704f1cf1b966a30088f0795c23fded9e400e91cdfb2c8
-size 500250000
+oid sha256:d30184c950cc71f8fa943c0c61564db94ee090206508ea6279cf7032f4c53a07
+size 161971527