change download url

fix split gen filepath
basic gen examples

citesum.py

@@ -16,8 +16,8 @@
 # Lint as: python3
 """CiteSum dataset"""
 
-import hashlib
 import os
+import json
 
 import datasets
 
@@ -49,7 +49,7 @@ _CITATION = """\
 
 """
 
-_DOWNLOAD_URL = "https://drive.google.com/file/d/1ndHCREXGSPnDUNllladh9qCtayqbXAfJ"
+_DOWNLOAD_URL = "https://drive.google.com/uc?export=download&id=1ndHCREXGSPnDUNllladh9qCtayqbXAfJ"
 
 
 class CiteSumConfig(datasets.BuilderConfig):
@@ -92,16 +92,19 @@ class CiteSum(datasets.GeneratorBasedBuilder):
         )
 
     def _split_generators(self, dl_manager):
-        dl_paths = dl_manager.download(_DOWNLOAD_URL)
+        dl_path = dl_manager.download_and_extract(_DOWNLOAD_URL)
+
+        file_mapping = {
+            datasets.Split.TRAIN: "train.json",
+            datasets.Split.VALIDATION: "val.json",
+            datasets.Split.TEST: "test.json",
+        }
+
         return [
             datasets.SplitGenerator(
                 name=split,
                 gen_kwargs={
-                    "urls_file": dl_paths[split],
-                    "files_per_archive": [
-                        dl_manager.iter_archive(dl_paths["cnn_stories"]),
-                        dl_manager.iter_archive(dl_paths["dm_stories"]),
-                    ],
+                    "filepath": os.path.join(dl_path,  file_mapping[split]),
                 },
             )
             for split in [
@@ -111,19 +114,8 @@ class CiteSum(datasets.GeneratorBasedBuilder):
             ]
         ]
 
-    def _generate_examples(self, urls_file, files_per_archive):
-        urls = _get_url_hashes(urls_file)
-        idx = 0
-        for files in files_per_archive:
-            for path, file in files:
-                hash_from_path = _get_hash_from_path(path)
-                if hash_from_path in urls:
-                    article, highlights = _get_art_abs(file, self.config.version)
-                    if not article or not highlights:
-                        continue
-                    yield idx, {
-                        _ARTICLE: article,
-                        _HIGHLIGHTS: highlights,
-                        "id": hash_from_path,
-                    }
-                    idx += 1
+    def _generate_examples(self, filepath):
+
+        with open(filepath, "r") as fp:
+            for idx, line in enumerate(fp.readlines()):
+                yield idx, json.loads(line)