add repo_data_only flag (#1)
Browse files- repo_data_only (1d885d21396697e083d6f832c81ed3e917e0cef5)
Co-authored-by: David Hall <dlwh@users.noreply.huggingface.co>
- Multi_Legal_Pile.py +40 -37
Multi_Legal_Pile.py
CHANGED
@@ -193,7 +193,7 @@ _JURISDICTONS = ["Austria", "Belgium", "Bulgaria", "Croatia", "Czechia", "Denmar
|
|
193 |
class MultiLegalPileConfig(datasets.BuilderConfig):
|
194 |
"""BuilderConfig for MultiLegalPile."""
|
195 |
|
196 |
-
def __init__(self, name: str, **kwargs):
|
197 |
"""BuilderConfig for MultiLegalPile.
|
198 |
Args:
|
199 |
name: combination of language and type with _
|
@@ -205,10 +205,12 @@ class MultiLegalPileConfig(datasets.BuilderConfig):
|
|
205 |
self.name = name
|
206 |
self.language = name.split("_")[0]
|
207 |
self.type = name.split("_")[1]
|
|
|
208 |
|
209 |
|
210 |
class MultiLegalPile(datasets.GeneratorBasedBuilder):
|
211 |
"""MultiLegalPile: A Corpus Covering Large-Scale Multilingual Legal Data"""
|
|
|
212 |
|
213 |
BUILDER_CONFIGS = [MultiLegalPileConfig(f"{language}_{type}")
|
214 |
for type in _TYPES + ["all"]
|
@@ -249,47 +251,48 @@ class MultiLegalPile(datasets.GeneratorBasedBuilder):
|
|
249 |
info["filepath"] = download_url("joelito/Multi_Legal_Pile", f"{language}/{type}/{data_name}")
|
250 |
data_infos.append(info.copy())
|
251 |
|
252 |
-
|
253 |
-
|
254 |
-
|
255 |
-
|
256 |
-
|
257 |
-
elif type == 'legislation':
|
258 |
-
for resource_type in ["decision", "directive", "regulation", "recommendation", "proposal"]:
|
259 |
-
info["filepath"] = download_url("joelito/eurlex_resources", f"{language}/{resource_type}")
|
260 |
data_infos.append(info.copy())
|
261 |
-
|
262 |
-
|
263 |
-
|
264 |
-
data_infos.append(info.copy())
|
265 |
-
|
266 |
-
# add mc4_legal data
|
267 |
-
if type == 'other':
|
268 |
-
info["jurisdiction"] = "N/A"
|
269 |
-
if language in ["de", "en", "es"]: # here we need to chunk because the files are too large
|
270 |
-
for idx in [0, 1]:
|
271 |
-
info["filepath"] = download_url("joelito/mc4_legal", f"{language}_{idx}")
|
272 |
data_infos.append(info.copy())
|
273 |
-
elif
|
274 |
-
|
275 |
-
|
276 |
-
info["filepath"] = download_url("joelito/mc4_legal", f"{language}")
|
277 |
data_infos.append(info.copy())
|
278 |
|
279 |
-
|
280 |
-
|
281 |
-
|
282 |
-
|
283 |
-
|
284 |
-
|
285 |
-
elif name in ['echr']:
|
286 |
-
info["jurisdiction"] = "EU"
|
287 |
-
elif name in ['constitutions']:
|
288 |
-
info["jurisdiction"] = "N/A"
|
289 |
-
for split in ["train", "validation"]:
|
290 |
-
for file_name in train_val_files[split]:
|
291 |
-
info["filepath"] = download_url("pile-of-law/pile-of-law", file_name)
|
292 |
data_infos.append(info.copy())
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
293 |
|
294 |
return [datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={"data_infos": data_infos})]
|
295 |
|
|
|
193 |
class MultiLegalPileConfig(datasets.BuilderConfig):
|
194 |
"""BuilderConfig for MultiLegalPile."""
|
195 |
|
196 |
+
def __init__(self, name: str, repo_data_only=False, **kwargs):
|
197 |
"""BuilderConfig for MultiLegalPile.
|
198 |
Args:
|
199 |
name: combination of language and type with _
|
|
|
205 |
self.name = name
|
206 |
self.language = name.split("_")[0]
|
207 |
self.type = name.split("_")[1]
|
208 |
+
self.repo_data_only = repo_data_only
|
209 |
|
210 |
|
211 |
class MultiLegalPile(datasets.GeneratorBasedBuilder):
|
212 |
"""MultiLegalPile: A Corpus Covering Large-Scale Multilingual Legal Data"""
|
213 |
+
BUILDER_CONFIG_CLASS = MultiLegalPileConfig
|
214 |
|
215 |
BUILDER_CONFIGS = [MultiLegalPileConfig(f"{language}_{type}")
|
216 |
for type in _TYPES + ["all"]
|
|
|
251 |
info["filepath"] = download_url("joelito/Multi_Legal_Pile", f"{language}/{type}/{data_name}")
|
252 |
data_infos.append(info.copy())
|
253 |
|
254 |
+
if not self.config.repo_data_only:
|
255 |
+
# add eurlex_resources data
|
256 |
+
info["jurisdiction"] = "EU"
|
257 |
+
if type == 'caselaw':
|
258 |
+
info["filepath"] = download_url("joelito/eurlex_resources", f"{language}/caselaw")
|
|
|
|
|
|
|
259 |
data_infos.append(info.copy())
|
260 |
+
elif type == 'legislation':
|
261 |
+
for resource_type in ["decision", "directive", "regulation", "recommendation", "proposal"]:
|
262 |
+
info["filepath"] = download_url("joelito/eurlex_resources", f"{language}/{resource_type}")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
263 |
data_infos.append(info.copy())
|
264 |
+
elif type == 'contracts':
|
265 |
+
# international agreements
|
266 |
+
info["filepath"] = download_url("joelito/eurlex_resources", f"{language}/intagr")
|
|
|
267 |
data_infos.append(info.copy())
|
268 |
|
269 |
+
# add mc4_legal data
|
270 |
+
if type == 'other':
|
271 |
+
info["jurisdiction"] = "N/A"
|
272 |
+
if language in ["de", "en", "es"]: # here we need to chunk because the files are too large
|
273 |
+
for idx in [0, 1]:
|
274 |
+
info["filepath"] = download_url("joelito/mc4_legal", f"{language}_{idx}")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
275 |
data_infos.append(info.copy())
|
276 |
+
elif language in ["hr"]:
|
277 |
+
pass # nothing there for Croatian
|
278 |
+
else:
|
279 |
+
info["filepath"] = download_url("joelito/mc4_legal", f"{language}")
|
280 |
+
data_infos.append(info.copy())
|
281 |
+
|
282 |
+
# add pile-of-law data
|
283 |
+
if language == 'en':
|
284 |
+
info["jurisdiction"] = "US"
|
285 |
+
for name, train_val_files in _PILE_OF_LAW_NAMES[type].items():
|
286 |
+
if name == 'canadian_decisions':
|
287 |
+
info["jurisdiction"] = "Canada"
|
288 |
+
elif name in ['echr']:
|
289 |
+
info["jurisdiction"] = "EU"
|
290 |
+
elif name in ['constitutions']:
|
291 |
+
info["jurisdiction"] = "N/A"
|
292 |
+
for split in ["train", "validation"]:
|
293 |
+
for file_name in train_val_files[split]:
|
294 |
+
info["filepath"] = download_url("pile-of-law/pile-of-law", file_name)
|
295 |
+
data_infos.append(info.copy())
|
296 |
|
297 |
return [datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={"data_infos": data_infos})]
|
298 |
|