Datasets:

Languages:
English
Multilinguality:
monolingual
Size Categories:
1K<n<10K
10K<n<100K
Language Creators:
found
Annotations Creators:
expert-generated
Source Datasets:
original
ArXiv:
Tags:
License:
shannons commited on
Commit
8e0586b
1 Parent(s): 5130967

Add v20230518 release file

Browse files
README.md CHANGED
@@ -75,7 +75,7 @@ Please check the exemplar usage below for loading the data:
75
  ```python
76
  from datasets import load_dataset
77
 
78
- multi_lexsum = load_dataset("allenai/multi_lexsum", name="v20220616")
79
  # Download multi_lexsum locally and load it as a Dataset object
80
 
81
  example = multi_lexsum["validation"][0] # The first instance of the dev set
@@ -83,6 +83,8 @@ example["sources"] # A list of source document text for the case
83
 
84
  for sum_len in ["long", "short", "tiny"]:
85
  print(example["summary/" + sum_len]) # Summaries of three lengths
 
 
86
  ```
87
 
88
  ### Data Splits
@@ -125,7 +127,7 @@ The corresponding code for downloading and loading the dataset is licensed under
125
  title = {Multi-LexSum: Real-World Summaries of Civil Rights Lawsuits at Multiple Granularities},
126
  journal = {CoRR},
127
  volume = {abs/2206.10883},
128
- year = {2022},
129
  url = {https://doi.org/10.48550/arXiv.2206.10883},
130
  doi = {10.48550/arXiv.2206.10883}
131
  }
@@ -134,6 +136,7 @@ The corresponding code for downloading and loading the dataset is licensed under
134
 
135
  ## Release History
136
 
137
- | Version | Description |
138
- | ----------: | -----------------------: |
139
- | `v20220616` | The initial v1.0 release |
 
75
  ```python
76
  from datasets import load_dataset
77
 
78
+ multi_lexsum = load_dataset("allenai/multi_lexsum", name="v20230518")
79
  # Download multi_lexsum locally and load it as a Dataset object
80
 
81
  example = multi_lexsum["validation"][0] # The first instance of the dev set
83
 
84
  for sum_len in ["long", "short", "tiny"]:
85
  print(example["summary/" + sum_len]) # Summaries of three lengths
86
+
87
+ print(example['case_metadata']) # The corresponding metadata for a case in a dict
88
  ```
89
 
90
  ### Data Splits
127
  title = {Multi-LexSum: Real-World Summaries of Civil Rights Lawsuits at Multiple Granularities},
128
  journal = {CoRR},
129
  volume = {abs/2206.10883},
130
+ year = {2022},****
131
  url = {https://doi.org/10.48550/arXiv.2206.10883},
132
  doi = {10.48550/arXiv.2206.10883}
133
  }
136
 
137
  ## Release History
138
 
139
+ | Version | Description |
140
+ | ----------: | -----------------------------------------------------------: |
141
+ | `v20230518` | The v1.1 release including case and source document metadata |
142
+ | `v20220616` | The initial v1.0 release |
multi_lexsum.py CHANGED
@@ -1,6 +1,6 @@
1
- from typing import List, Union, Dict, Any, Tuple
2
  import json
3
  import os
 
4
 
5
  import datasets
6
  from datasets.tasks import Summarization
@@ -17,7 +17,6 @@ def _load_jsonl(filename):
17
 
18
 
19
  def _load_json(filepath):
20
-
21
  with open(filepath, "r") as fp:
22
  res = json.load(fp)
23
  return res
@@ -42,7 +41,7 @@ _CITATION = """
42
 
43
  _DESCRIPTION = """
44
  Multi-LexSum is a multi-doc summarization dataset for civil rights litigation lawsuits with summaries of three granularities.
45
- """ # TODO: Update with full abstract
46
 
47
  _HOMEPAGE = "https://multilexsum.github.io"
48
 
@@ -68,8 +67,8 @@ class MultiLexsumConfig(datasets.BuilderConfig):
68
 
69
 
70
  class MultiLexsum(datasets.GeneratorBasedBuilder):
71
- """MultiLexSum Dataset: a multi-doc summarization dataset for
72
- civil rights litigation lawsuits with summaries of three granularities.
73
  """
74
 
75
  BUILDER_CONFIGS = [
@@ -78,30 +77,90 @@ class MultiLexsum(datasets.GeneratorBasedBuilder):
78
  version=datasets.Version("1.0.0", "Public v1.0 release."),
79
  description="The v1.0 Multi-LexSum dataset",
80
  ),
 
 
 
 
 
81
  ]
82
 
83
  def _info(self):
84
- return datasets.DatasetInfo(
85
- description=_DESCRIPTION,
86
- features=datasets.Features(
87
- {
88
- "id": datasets.Value("string"),
89
- "sources": datasets.Sequence(datasets.Value("string")),
90
- "summary/long": datasets.Value("string"),
91
- "summary/short": datasets.Value("string"),
92
- "summary/tiny": datasets.Value("string"),
93
- }
94
- ),
95
- supervised_keys=None,
96
- homepage=_HOMEPAGE,
97
- citation=_CITATION,
98
- task_templates=[
99
- Summarization(text_column="source", summary_column="summary/long")
100
- ],
101
- )
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
102
 
103
  def _split_generators(self, dl_manager):
104
-
105
  base_url = _BASE_URL if self.config.data_dir is None else self.config.data_dir
106
  downloaded_files = dl_manager.download_and_extract(
107
  {
@@ -140,16 +199,67 @@ class MultiLexsum(datasets.GeneratorBasedBuilder):
140
  """This function returns the examples in the raw (text) form."""
141
  logger.info(f"generating examples from = {subset_file}")
142
 
143
- subset_cases = _load_jsonl(subset_file)
144
- for case_data in subset_cases:
145
- case_sources = [
146
- sources[source_id]["doc_text"]
147
- for source_id in case_data["case_documents"]
148
- ]
149
- yield case_data["case_id"], {
150
- "id": case_data["case_id"],
151
- "sources": case_sources,
152
- "summary/long": case_data["summary/long"],
153
- "summary/short": case_data["summary/short"],
154
- "summary/tiny": case_data["summary/tiny"],
155
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  import json
2
  import os
3
+ from typing import Any, Dict, List, Tuple, Union
4
 
5
  import datasets
6
  from datasets.tasks import Summarization
17
 
18
 
19
  def _load_json(filepath):
 
20
  with open(filepath, "r") as fp:
21
  res = json.load(fp)
22
  return res
41
 
42
  _DESCRIPTION = """
43
  Multi-LexSum is a multi-doc summarization dataset for civil rights litigation lawsuits with summaries of three granularities.
44
+ """ # TODO: Update with full abstract
45
 
46
  _HOMEPAGE = "https://multilexsum.github.io"
47
 
67
 
68
 
69
  class MultiLexsum(datasets.GeneratorBasedBuilder):
70
+ """MultiLexSum Dataset: a multi-doc summarization dataset for
71
+ civil rights litigation lawsuits with summaries of three granularities.
72
  """
73
 
74
  BUILDER_CONFIGS = [
77
  version=datasets.Version("1.0.0", "Public v1.0 release."),
78
  description="The v1.0 Multi-LexSum dataset",
79
  ),
80
+ MultiLexsumConfig(
81
+ name="v20230518",
82
+ version=datasets.Version("1.1.0", "Public v1.1 release."),
83
+ description="It adds additional metadata for documents and cases",
84
+ ),
85
  ]
86
 
87
  def _info(self):
88
+ if self.config.name == "v20220616":
89
+ return datasets.DatasetInfo(
90
+ description=_DESCRIPTION,
91
+ features=datasets.Features(
92
+ {
93
+ "id": datasets.Value("string"),
94
+ "sources": datasets.Sequence(datasets.Value("string")),
95
+ "summary/long": datasets.Value("string"),
96
+ "summary/short": datasets.Value("string"),
97
+ "summary/tiny": datasets.Value("string"),
98
+ }
99
+ ),
100
+ supervised_keys=None,
101
+ homepage=_HOMEPAGE,
102
+ citation=_CITATION,
103
+ task_templates=[
104
+ Summarization(text_column="source", summary_column="summary/long")
105
+ ],
106
+ )
107
+ elif self.config.name == "v20230518":
108
+ return datasets.DatasetInfo(
109
+ description=_DESCRIPTION,
110
+ features=datasets.Features(
111
+ {
112
+ "id": datasets.Value("string"),
113
+ "sources": datasets.Sequence(datasets.Value("string")),
114
+ "sources_metadata": datasets.Sequence(
115
+ {
116
+ "doc_id": datasets.Value("string"),
117
+ "doc_type": datasets.Value("string"),
118
+ "doc_title": datasets.Value("string"),
119
+ "parser": datasets.Value("string"),
120
+ "is_ocr": datasets.Value("bool"),
121
+ "url": datasets.Value("string"),
122
+ }
123
+ ),
124
+ "summary/long": datasets.Value("string"),
125
+ "summary/short": datasets.Value("string"),
126
+ "summary/tiny": datasets.Value("string"),
127
+ "case_metadata": datasets.Features(
128
+ {
129
+ # fmt: off
130
+ "case_name": datasets.Value("string"),
131
+ "case_type": datasets.Value("string"),
132
+ "filing_date": datasets.Value("string"),
133
+ "filing_year": datasets.Value("string"),
134
+ "case_ongoing": datasets.Value("string"),
135
+ "case_ongoing_record_time": datasets.Value("string"),
136
+ "closing_year": datasets.Value("string"),
137
+ "order_start_year": datasets.Value("string"),
138
+ "order_end_year": datasets.Value("string"),
139
+ "defendant_payment": datasets.Value("string"),
140
+ "class_action_sought": datasets.Value("string"),
141
+ "class_action_granted": datasets.Value("string"),
142
+ "attorney_orgs": [datasets.Value("string")],
143
+ "prevailing_party": datasets.Value("string"),
144
+ "plaintiff_types": [datasets.Value("string")],
145
+ "plaintiff_description": datasets.Value("string"),
146
+ "constitutional_clauses": [datasets.Value("string")],
147
+ "causes_of_action": [datasets.Value("string")],
148
+ "summary_authors": [datasets.Value("string")],
149
+ "case_url": datasets.Value("string"),
150
+ # fmt: on
151
+ }
152
+ ),
153
+ }
154
+ ),
155
+ supervised_keys=None,
156
+ homepage=_HOMEPAGE,
157
+ citation=_CITATION,
158
+ task_templates=[
159
+ Summarization(text_column="source", summary_column="summary/long")
160
+ ],
161
+ )
162
 
163
  def _split_generators(self, dl_manager):
 
164
  base_url = _BASE_URL if self.config.data_dir is None else self.config.data_dir
165
  downloaded_files = dl_manager.download_and_extract(
166
  {
199
  """This function returns the examples in the raw (text) form."""
200
  logger.info(f"generating examples from = {subset_file}")
201
 
202
+ if self.config.name == "v20220616":
203
+ subset_cases = _load_jsonl(subset_file)
204
+ for case_data in subset_cases:
205
+ case_sources = [
206
+ sources[source_id]["doc_text"]
207
+ for source_id in case_data["case_documents"]
208
+ ]
209
+ yield case_data["case_id"], {
210
+ "id": case_data["case_id"],
211
+ "sources": case_sources,
212
+ "summary/long": case_data["summary/long"],
213
+ "summary/short": case_data["summary/short"],
214
+ "summary/tiny": case_data["summary/tiny"],
215
+ }
216
+ elif self.config.name == "v20230518":
217
+ subset_cases = _load_jsonl(subset_file)
218
+
219
+ for idx, case_data in enumerate(subset_cases):
220
+ case_sources = [
221
+ sources[source_id]["doc_text"]
222
+ for source_id in case_data["case_documents"]
223
+ ]
224
+
225
+ case_source_metadata = [
226
+ {
227
+ key: val
228
+ for key, val in sources[source_id].items()
229
+ if key != "doc_text"
230
+ }
231
+ for source_id in case_data["case_documents"]
232
+ ]
233
+
234
+ case_metadata = {
235
+ "case_name": case_data["case_name"],
236
+ "case_type": case_data["case_type"],
237
+ "filing_date": case_data["filing_date"],
238
+ "filing_year": case_data["filing_year"],
239
+ "case_ongoing": case_data["case_ongoing"],
240
+ "case_ongoing_record_time": case_data["case_ongoing_record_time"],
241
+ "closing_year": case_data["closing_year"],
242
+ "order_start_year": case_data["order_start_year"],
243
+ "order_end_year": case_data["order_end_year"],
244
+ "defendant_payment": case_data["defendant_payment"],
245
+ "class_action_sought": case_data["class_action_sought"],
246
+ "class_action_granted": case_data["class_action_granted"],
247
+ "attorney_orgs": case_data["attorney_org"],
248
+ "prevailing_party": case_data["prevailing_party"],
249
+ "plaintiff_types": case_data["plaintiff_types"],
250
+ "plaintiff_description": case_data["plaintiff_description"],
251
+ "constitutional_clauses": case_data["constitutional_clauses"],
252
+ "causes_of_action": case_data["causes_of_action"],
253
+ "summary_authors": case_data["summary_authors"],
254
+ "case_url": case_data["case_url"],
255
+ }
256
+
257
+ yield case_data["case_id"], {
258
+ "id": case_data["case_id"],
259
+ "sources": case_sources,
260
+ "sources_metadata": case_source_metadata,
261
+ "summary/long": case_data["summary/long"],
262
+ "summary/short": case_data["summary/short"],
263
+ "summary/tiny": case_data["summary/tiny"],
264
+ "case_metadata": case_metadata,
265
+ }
releases/v20230518/dev.json ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0ae6fe46c5557975b29ed5bbfe780e9d188623320ac6c94d75ea8412ffe64b31
3
+ size 2555645
releases/v20230518/sources.json ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:805fd58ac5f72c40c9b575fa1e1b0be1943ea3c07a1a3089827444a3e0abc858
3
+ size 2221892187
releases/v20230518/test.json ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:56ffcfc1add9fbd895b024d1220317acb25109508ea76f275952ca8419594e61
3
+ size 4821719
releases/v20230518/train.json ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:168c1c22fb4739e4cb0405bfc29499744357218a386d9bc234c0dfc39f0c1302
3
+ size 17628135