Create movielens.py

movielens.py

@@ -0,0 +1,90 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+import os
+from functools import partial
+from pathlib import Path
+from typing import Dict, Iterable
+
+import datasets
+from datasets import DatasetDict, DownloadManager, load_dataset
+import pandas as pd
+
+
+AVAILABLE_DATASETS = {
+    'small': 'https://files.grouplens.org/datasets/movielens/ml-latest-small.zip',
+    'full': 'https://files.grouplens.org/datasets/movielens/ml-latest.zip',
+}
+
+
+VERSION = datasets.Version("0.0.1")
+
+
+class MovielensDataset(datasets.GeneratorBasedBuilder):
+    """MovielensDataset dataset."""
+
+    BUILDER_CONFIGS = [
+        datasets.BuilderConfig(
+            name=data_name, version=VERSION, description=f"{data_name} movielens dataset"
+        )
+        for data_name in AVAILABLE_DATASETS
+    ]
+
+    def _info(self) -> datasets.DatasetInfo:
+        return datasets.DatasetInfo(
+            description="",
+            features=datasets.Features(
+                {
+                    "movieId": datasets.Value("string"),
+                    "title": datasets.Value("string"),
+                    "genres": datasets.Sequence(datasets.Value("string")),
+                    "tag": datasets.Sequence(datasets.Value("string")),
+                }
+            ),
+            supervised_keys=None,
+            homepage="https://grouplens.org/datasets/movielens/latest/",
+            citation="",
+        )
+
+    def _split_generators(
+        self, dl_manager: DownloadManager
+    ) -> Iterable[datasets.SplitGenerator]:
+        downloader = partial(
+            lambda split: dl_manager.download_and_extract(
+                AVAILABLE_DATASETS[self.config.name]
+            )
+        )
+
+        folder = os.path.splitext(os.path.basename(AVAILABLE_DATASETS[self.config.name]))[
+            0
+        ]
+
+        # There is no predefined train/val/test split for this dataset.
+        return [
+            datasets.SplitGenerator(
+                name=datasets.Split.TRAIN,
+                gen_kwargs={
+                    "root_path": downloader("train"),
+                    "split": "train",
+                    "folder": folder,
+                },
+            ),
+        ]
+
+    def _generate_examples(
+        self, root_path: str, split: str, folder: str
+    ) -> Iterable[Dict]:
+        split_path = Path(root_path) / folder
+        movies_file = split_path / "movies.csv"
+        movies = pd.read_csv(movies_file, sep=',', encoding='utf-8')
+        movies['genres'] = movies['genres'].str.split('|')
+        tags_file = split_path / "tags.csv"
+        tags = pd.read_csv(tags_file, sep=',', encoding='utf-8')
+        tags = tags.groupby('movieId').agg({'tag': list}).reset_index()
+        movies = movies.merge(tags, on='movieId')
+        for idx, row in movies.iterrows():
+            yield idx, {
+                'movieId': row['movieId'],
+                'title': row['title'],
+                'genres': row['genres'],
+                'tag': row['tag'],
+            }