albertvillanova HF staff commited on
Commit
eb28e6e
1 Parent(s): cdab174

Support streaming id_clickbait dataset (#4014)

Browse files

* Support streaming id_clickbait dataset

* Fix paths on Windows

* Simplify code

Commit from https://github.com/huggingface/datasets/commit/f56672b7adb8733cd9c914675f96736c196a9c0f

Files changed (1) hide show
  1. id_clickbait.py +13 -15
id_clickbait.py CHANGED
@@ -47,7 +47,7 @@ _HOMEPAGE = "https://github.com/feryandi/Dataset-Artikel"
47
 
48
  _LICENSE = "Creative Commons Attribution 4.0 International license"
49
 
50
- _URLs = ["https://md-datasets-cache-zipfiles-prod.s3.eu-west-1.amazonaws.com/k42j7x2kpn-1.zip"]
51
 
52
 
53
  class IdClickbaitConfig(datasets.BuilderConfig):
@@ -72,9 +72,9 @@ class IdClickbait(datasets.GeneratorBasedBuilder):
72
  version=VERSION,
73
  description="Annotated clickbait dataset",
74
  label_classes=["non-clickbait", "clickbait"],
75
- path="annotated/csv",
76
  ),
77
- IdClickbaitConfig(name="raw", version=VERSION, description="Raw dataset", path="raw/csv"),
78
  ]
79
 
80
  BUILDER_CONFIG_CLASS = IdClickbaitConfig
@@ -111,34 +111,32 @@ class IdClickbait(datasets.GeneratorBasedBuilder):
111
  )
112
 
113
  def _split_generators(self, dl_manager):
114
- my_urls = _URLs[0]
115
- data_dir = dl_manager.download_and_extract(my_urls)
116
  return [
117
  datasets.SplitGenerator(
118
  name=datasets.Split.TRAIN,
119
  gen_kwargs={
120
  "article_dir": os.path.join(data_dir, self.config.path),
121
- "split": "train",
122
  },
123
  )
124
  ]
125
 
126
- def _generate_examples(self, article_dir, split):
127
- logger.info("⏳ Generating %s examples from = %s", split, article_dir)
128
- id = 0
129
- for path in sorted(glob.glob(os.path.join(article_dir, "**/*.csv"), recursive=True)):
130
  with open(path, encoding="utf-8-sig", newline="") as f:
131
  reader = csv.DictReader(f)
132
  for row in reader:
133
  if self.config.name == "annotated":
134
- yield id, {
135
- "id": str(id),
136
  "title": row["title"],
137
  "label": row["label"],
138
  }
139
  else:
140
- yield id, {
141
- "id": str(id),
142
  "title": row["title"],
143
  "source": row["source"],
144
  "date": row["date"],
@@ -147,4 +145,4 @@ class IdClickbait(datasets.GeneratorBasedBuilder):
147
  "content": row["content"],
148
  "url": row["url"],
149
  }
150
- id += 1
 
47
 
48
  _LICENSE = "Creative Commons Attribution 4.0 International license"
49
 
50
+ _URL = "https://md-datasets-cache-zipfiles-prod.s3.eu-west-1.amazonaws.com/k42j7x2kpn-1.zip"
51
 
52
 
53
  class IdClickbaitConfig(datasets.BuilderConfig):
 
72
  version=VERSION,
73
  description="Annotated clickbait dataset",
74
  label_classes=["non-clickbait", "clickbait"],
75
+ path=os.path.join("annotated", "csv"),
76
  ),
77
+ IdClickbaitConfig(name="raw", version=VERSION, description="Raw dataset", path=os.path.join("raw", "csv")),
78
  ]
79
 
80
  BUILDER_CONFIG_CLASS = IdClickbaitConfig
 
111
  )
112
 
113
  def _split_generators(self, dl_manager):
114
+ data_dir = dl_manager.download_and_extract(_URL)
 
115
  return [
116
  datasets.SplitGenerator(
117
  name=datasets.Split.TRAIN,
118
  gen_kwargs={
119
  "article_dir": os.path.join(data_dir, self.config.path),
 
120
  },
121
  )
122
  ]
123
 
124
+ def _generate_examples(self, article_dir):
125
+ logger.info(f"⏳ Generating examples from = {article_dir}")
126
+ idx = 0
127
+ for path in sorted(glob.glob(os.path.join(article_dir, "*.csv"))):
128
  with open(path, encoding="utf-8-sig", newline="") as f:
129
  reader = csv.DictReader(f)
130
  for row in reader:
131
  if self.config.name == "annotated":
132
+ yield idx, {
133
+ "id": str(idx),
134
  "title": row["title"],
135
  "label": row["label"],
136
  }
137
  else:
138
+ yield idx, {
139
+ "id": str(idx),
140
  "title": row["title"],
141
  "source": row["source"],
142
  "date": row["date"],
 
145
  "content": row["content"],
146
  "url": row["url"],
147
  }
148
+ idx += 1