knkarthick commited on
Commit
2152969
1 Parent(s): f15cd6e

Update dialogsum_reformat.py

Browse files
Files changed (1) hide show
  1. dialogsum_reformat.py +186 -65
dialogsum_reformat.py CHANGED
@@ -1,9 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  import json
2
- import pandas as pd
 
 
3
  import datasets
4
- import os
5
 
6
- logger = datasets.logging.get_logger(__name__)
7
 
8
  _CITATION = """
9
  @inproceedings{chen-etal-2021-dialogsum,
@@ -32,40 +49,30 @@ There are two features:
32
  """
33
  _HOMEPAGE = "hhttps://aclanthology.org/2021.findings-acl.449"
34
  _LICENSE = "CC BY-NC-ND 4.0"
35
- # _URL = "https://huggingface.co/datasets/knkarthick/dialogsum_reformat/tree/main/"
36
- _URL = "https://huggingface.co/datasets/knkarthick/dialogsum_reformat/resolve/main/"
37
- # _URL = "https://huggingface.co/datasets/knkarthick/dialogsum_reformat/blob/main/"
38
-
39
- _URLS = {
40
- "train": _URL + "train.json",
41
- "test": _URL + "test.json",
42
- "val": _URL + "val.json",
43
- }
44
-
45
 
46
 
47
  class Dialogsum(datasets.GeneratorBasedBuilder):
48
- """Dialogsum Corpus dataset."""
49
 
50
  VERSION = datasets.Version("1.1.0")
51
 
52
  BUILDER_CONFIGS = [
53
- datasets.BuilderConfig(name="dialogsum_reformat"),
54
  ]
55
 
56
  def _info(self):
 
 
 
 
 
 
 
 
57
  return datasets.DatasetInfo(
58
  description=_DESCRIPTION,
59
- features=datasets.Features(
60
- {
61
- "id": datasets.Value("string"),
62
- "dialogue": datasets.Value("string"),
63
- "summary": datasets.Value("string"),
64
- "topic": datasets.Value("string"),
65
- }
66
- ),
67
- # No default supervised_keys (as we have to pass both question
68
- # and context as input).
69
  supervised_keys=None,
70
  homepage=_HOMEPAGE,
71
  license=_LICENSE,
@@ -73,47 +80,161 @@ class Dialogsum(datasets.GeneratorBasedBuilder):
73
  )
74
 
75
  def _split_generators(self, dl_manager):
76
- downloaded_files = dl_manager.download_and_extract(_URLS)
 
77
  return [
78
- datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={"filepath": downloaded_files["train"]}),
79
- datasets.SplitGenerator(name=datasets.Split.TEST, gen_kwargs={"filepath": downloaded_files["test"]}),
80
- datasets.SplitGenerator(name=datasets.Split.VALIDATION, gen_kwargs={"filepath": downloaded_files["val"]}),
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
81
  ]
82
 
83
- def _generate_examples(self, filepath):
84
- """This function returns the examples in the raw (text) form."""
85
- logger.info("generating examples from = %s", filepath)
86
- with open(filepath) as f :
87
- data = json.load(f)
88
-
89
- for info in data :
90
- dialogue_id = info['id']
91
- dialogue_name = info['dialogue']
92
- dialogue_summary = info['summary']
93
- dialogue_topic = info['topic']
94
-
95
- yield {
96
- "id" : dialogue_id,
97
- "dialogue" : dialogue_name,
98
- "summary" : dialogue_summary,
99
- "topic" : dialogue_topic,
100
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
101
 
102
- # def _generate_examples(self, filepath, split):
103
- # """This function returns the examples in the raw (text) form."""
104
- # logger.info("generating examples from = %s", filepath)
105
- # with open(os.path.join(filepath, split)) as f :
106
- # data = json.load(f)
107
-
108
- # for info in data :
109
- # dialogue_id = info['id']
110
- # dialogue_name = info['dialogue']
111
- # dialogue_summary = info['summary']
112
- # dialogue_topic = info['topic']
113
-
114
- # yield key, {
115
- # "id" : dialogue_id,
116
- # "dialogue" : dialogue_name,
117
- # "summary" : dialogue_summary,
118
- # "topic" : dialogue_topic,
119
- # }
 
1
+ # coding=utf-8
2
+ # Copyright 2020 The HuggingFace Datasets Authors and the current dataset script contributor.
3
+ #
4
+ # Licensed under the Apache License, Version 2.0 (the "License");
5
+ # you may not use this file except in compliance with the License.
6
+ # You may obtain a copy of the License at
7
+ #
8
+ # http://www.apache.org/licenses/LICENSE-2.0
9
+ #
10
+ # Unless required by applicable law or agreed to in writing, software
11
+ # distributed under the License is distributed on an "AS IS" BASIS,
12
+ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13
+ # See the License for the specific language governing permissions and
14
+ # limitations under the License.
15
+ """DIALOGSum dataset."""
16
+
17
+
18
  import json
19
+
20
+ import py7zr
21
+
22
  import datasets
 
23
 
 
24
 
25
  _CITATION = """
26
  @inproceedings{chen-etal-2021-dialogsum,
 
49
  """
50
  _HOMEPAGE = "hhttps://aclanthology.org/2021.findings-acl.449"
51
  _LICENSE = "CC BY-NC-ND 4.0"
52
+ _URL = "https://huggingface.co/datasets/knkarthick/dialogsum_reformat/resolve/main/corpus.7z"
 
 
 
 
 
 
 
 
 
53
 
54
 
55
  class Dialogsum(datasets.GeneratorBasedBuilder):
56
+ """DIALOGSum Corpus dataset."""
57
 
58
  VERSION = datasets.Version("1.1.0")
59
 
60
  BUILDER_CONFIGS = [
61
+ datasets.BuilderConfig(name="dialogsum"),
62
  ]
63
 
64
  def _info(self):
65
+ features = datasets.Features(
66
+ {
67
+ "id": datasets.Value("string"),
68
+ "dialogue": datasets.Value("string"),
69
+ "summary": datasets.Value("string"),
70
+ "topic": datasets.Value("string"),
71
+ }
72
+ )
73
  return datasets.DatasetInfo(
74
  description=_DESCRIPTION,
75
+ features=features,
 
 
 
 
 
 
 
 
 
76
  supervised_keys=None,
77
  homepage=_HOMEPAGE,
78
  license=_LICENSE,
 
80
  )
81
 
82
  def _split_generators(self, dl_manager):
83
+ """Returns SplitGenerators."""
84
+ path = dl_manager.download(_URL)
85
  return [
86
+ datasets.SplitGenerator(
87
+ name=datasets.Split.TRAIN,
88
+ gen_kwargs={
89
+ "filepath": (path, "train.json"),
90
+ "split": "train",
91
+ },
92
+ ),
93
+ datasets.SplitGenerator(
94
+ name=datasets.Split.TEST,
95
+ gen_kwargs={
96
+ "filepath": (path, "test.json"),
97
+ "split": "test",
98
+ },
99
+ ),
100
+ datasets.SplitGenerator(
101
+ name=datasets.Split.VALIDATION,
102
+ gen_kwargs={
103
+ "filepath": (path, "val.json"),
104
+ "split": "val",
105
+ },
106
+ ),
107
  ]
108
 
109
+ def _generate_examples(self, filepath, split):
110
+ """Yields examples."""
111
+ path, fname = filepath
112
+ with open(path, "rb") as f:
113
+ with py7zr.SevenZipFile(f, "r") as z:
114
+ for name, bio in z.readall().items():
115
+ if name == fname:
116
+ data = json.load(bio)
117
+ for example in data:
118
+ yield example["id"], example
119
+
120
+ ###################### OLD #####################
121
+
122
+ # import json
123
+ # import pandas as pd
124
+ # import datasets
125
+ # import os
126
+
127
+ # logger = datasets.logging.get_logger(__name__)
128
+
129
+ # _CITATION = """
130
+ # @inproceedings{chen-etal-2021-dialogsum,
131
+ # title={{D}ialog{S}um: {A} Real-Life Scenario Dialogue Summarization Dataset},
132
+ # author={Chen, Yulong and Liu, Yang and Chen, Liang and Zhang, Yue},
133
+ # journal={arXiv preprint arXiv:1911.12237},
134
+ # year={2021},
135
+ # booktitle ={Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021"},
136
+ # month = {aug},
137
+ # address = {Online},
138
+ # publisher = {Association for Computational Linguistics},
139
+ # url = {https://aclanthology.org/2021.findings-acl.449},
140
+ # doi = {10.18653/v1/2021.findings-acl.449},
141
+ # pages = {5062--5074}
142
+ # }
143
+ # """
144
+
145
+ # _DESCRIPTION = """
146
+ # DialogSUM Corpus contains 13460 chat dialogues with manually annotated
147
+ # summaries.
148
+ # There are two features:
149
+ # - dialogue: text of dialogue.
150
+ # - summary: human written summary of the dialogue.
151
+ # - topic: one liner summary of the dialogue.
152
+ # - id: id of a example.
153
+ # """
154
+ # _HOMEPAGE = "hhttps://aclanthology.org/2021.findings-acl.449"
155
+ # _LICENSE = "CC BY-NC-ND 4.0"
156
+ # # _URL = "https://huggingface.co/datasets/knkarthick/dialogsum_reformat/tree/main/"
157
+ # _URL = "https://huggingface.co/datasets/knkarthick/dialogsum_reformat/resolve/main/"
158
+ # # _URL = "https://huggingface.co/datasets/knkarthick/dialogsum_reformat/blob/main/"
159
+
160
+ # _URLS = {
161
+ # "train": _URL + "train.json",
162
+ # "test": _URL + "test.json",
163
+ # "val": _URL + "val.json",
164
+ # }
165
+
166
+
167
+
168
+ # class Dialogsum(datasets.GeneratorBasedBuilder):
169
+ # """Dialogsum Corpus dataset."""
170
+
171
+ # VERSION = datasets.Version("1.1.0")
172
+
173
+ # BUILDER_CONFIGS = [
174
+ # datasets.BuilderConfig(name="dialogsum_reformat"),
175
+ # ]
176
+
177
+ # def _info(self):
178
+ # return datasets.DatasetInfo(
179
+ # description=_DESCRIPTION,
180
+ # features=datasets.Features(
181
+ # {
182
+ # "id": datasets.Value("string"),
183
+ # "dialogue": datasets.Value("string"),
184
+ # "summary": datasets.Value("string"),
185
+ # "topic": datasets.Value("string"),
186
+ # }
187
+ # ),
188
+ # # No default supervised_keys (as we have to pass both question
189
+ # # and context as input).
190
+ # supervised_keys=None,
191
+ # homepage=_HOMEPAGE,
192
+ # license=_LICENSE,
193
+ # citation=_CITATION,
194
+ # )
195
+
196
+ # def _split_generators(self, dl_manager):
197
+ # downloaded_files = dl_manager.download_and_extract(_URLS)
198
+ # return [
199
+ # datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={"filepath": downloaded_files["train"]}),
200
+ # datasets.SplitGenerator(name=datasets.Split.TEST, gen_kwargs={"filepath": downloaded_files["test"]}),
201
+ # datasets.SplitGenerator(name=datasets.Split.VALIDATION, gen_kwargs={"filepath": downloaded_files["val"]}),
202
+ # ]
203
+
204
+ # def _generate_examples(self, filepath):
205
+ # """This function returns the examples in the raw (text) form."""
206
+ # logger.info("generating examples from = %s", filepath)
207
+ # with open(filepath) as f :
208
+ # data = json.load(f)
209
+
210
+ # for info in data :
211
+ # dialogue_id = info['id']
212
+ # dialogue_name = info['dialogue']
213
+ # dialogue_summary = info['summary']
214
+ # dialogue_topic = info['topic']
215
+
216
+ # yield {
217
+ # "id" : dialogue_id,
218
+ # "dialogue" : dialogue_name,
219
+ # "summary" : dialogue_summary,
220
+ # "topic" : dialogue_topic,
221
+ # }
222
 
223
+ # # def _generate_examples(self, filepath, split):
224
+ # # """This function returns the examples in the raw (text) form."""
225
+ # # logger.info("generating examples from = %s", filepath)
226
+ # # with open(os.path.join(filepath, split)) as f :
227
+ # # data = json.load(f)
228
+
229
+ # # for info in data :
230
+ # # dialogue_id = info['id']
231
+ # # dialogue_name = info['dialogue']
232
+ # # dialogue_summary = info['summary']
233
+ # # dialogue_topic = info['topic']
234
+
235
+ # # yield key, {
236
+ # # "id" : dialogue_id,
237
+ # # "dialogue" : dialogue_name,
238
+ # # "summary" : dialogue_summary,
239
+ # # "topic" : dialogue_topic,
240
+ # # }