Create dataset.py

dataset.py

@@ -0,0 +1,38 @@
+from datasets import Dataset, DatasetInfo, load_dataset_builder, GeneratorBasedBuilder
+
+class FreshWikiDataset(GeneratorBasedBuilder):
+    def _info(self):
+        return DatasetInfo(
+            description="The FreshWiki Dataset is a collection of high-quality Wikipedia articles focusing on the most-edited pages from February 2022 to September 2023.",
+            features=datasets.Features({
+                "json_file": datasets.Value("string"),
+                "txt_file": datasets.Value("string"),
+                "topics": datasets.Sequence(datasets.Value("string")),
+            }),
+        )
+
+    def _split_generators(self, dl_manager):
+        # Access the folder where all files are stored
+        data_dir = dl_manager.download_and_extract("https://huggingface.co/datasets/EchoShao8899/FreshWiki")
+        return [
+            datasets.SplitGenerator(
+                name=datasets.Split.TRAIN,
+                gen_kwargs={
+                    "json_files": os.path.join(data_dir, "json"),
+                    "txt_files": os.path.join(data_dir, "txt"),
+                    "topic_list_csv": os.path.join(data_dir, "topic_list.csv"),
+                },
+            ),
+        ]
+
+    def _generate_examples(self, json_files, txt_files, topic_list_csv):
+        # Read the CSV file, and pair it with JSON and TXT files
+        with open(topic_list_csv, "r") as csv_file:
+            reader = csv.DictReader(csv_file)
+            for idx, row in enumerate(reader):
+                topic_name = row['topic'].replace(' ', '_').replace('/', '_')
+                yield idx, {
+                    "json_file": open(os.path.join(json_files, f'{topic_name}.json'), "r").read(),
+                    "txt_file": open(os.path.join(txt_files, f'{topic_name}.txt'), "r").read(),
+                    "topic": row['topic'],
+                }