albertvillanova HF staff commited on
Commit
63fb695
1 Parent(s): 5b1c424

Delete loading script

Browse files
Files changed (1) hide show
  1. kilt_tasks.py +0 -259
kilt_tasks.py DELETED
@@ -1,259 +0,0 @@
1
- # coding=utf-8
2
- # Copyright 2020 The TensorFlow Datasets Authors and the HuggingFace Datasets Authors.
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
-
16
- # Lint as: python3
17
- """KILT tasks training and evaluation data"""
18
-
19
-
20
- import json
21
-
22
- import datasets
23
-
24
-
25
- logger = datasets.logging.get_logger(__name__)
26
-
27
-
28
- _CITATION = """\
29
- @inproceedings{fb_kilt,
30
- author = {Fabio Petroni and
31
- Aleksandra Piktus and
32
- Angela Fan and
33
- Patrick Lewis and
34
- Majid Yazdani and
35
- Nicola De Cao and
36
- James Thorne and
37
- Yacine Jernite and
38
- Vassilis Plachouras and
39
- Tim Rockt\"aschel and
40
- Sebastian Riedel},
41
- title = {{KILT:} a {B}enchmark for {K}nowledge {I}ntensive {L}anguage {T}asks},
42
- journal = {CoRR},
43
- archivePrefix = {arXiv},
44
- year = {2020},
45
- """
46
-
47
- _DESCRIPTION = """\
48
- KILT tasks training and evaluation data.
49
- - [FEVER](https://fever.ai) | Fact Checking | fever
50
- - [AIDA CoNLL-YAGO](https://www.mpi-inf.mpg.de/departments/databases-and-information-systems/research/ambiverse-nlu/aida/downloads) | Entity Linking | aidayago2
51
- - [WNED-WIKI](https://github.com/U-Alberta/wned) | Entity Linking | wned
52
- - [WNED-CWEB](https://github.com/U-Alberta/wned) | Entity Linking | cweb
53
- - [T-REx](https://hadyelsahar.github.io/t-rex) | Slot Filling | trex
54
- - [Zero-Shot RE](http://nlp.cs.washington.edu/zeroshot) | Slot Filling | structured_zeroshot
55
- - [Natural Questions](https://ai.google.com/research/NaturalQuestions) | Open Domain QA | nq
56
- - [HotpotQA](https://hotpotqa.github.io) | Open Domain QA | hotpotqa
57
- - [TriviaQA](http://nlp.cs.washington.edu/triviaqa) | Open Domain QA | triviaqa
58
- - [ELI5](https://facebookresearch.github.io/ELI5/explore.html) | Open Domain QA | eli5
59
- - [Wizard of Wikipedia](https://parl.ai/projects/wizard_of_wikipedia) | Dialogue | wow
60
-
61
- To finish linking TriviaQA questions to the IDs provided, follow the instructions [here](http://github.com/huggingface/datasets/datasets/kilt_tasks/README.md).
62
- """
63
-
64
-
65
- _DATA_URLS = {
66
- "fever": {
67
- "train": "http://dl.fbaipublicfiles.com/KILT/fever-train-kilt.jsonl",
68
- "validation": "http://dl.fbaipublicfiles.com/KILT/fever-dev-kilt.jsonl",
69
- "test": "http://dl.fbaipublicfiles.com/KILT/fever-test_without_answers-kilt.jsonl",
70
- },
71
- "aidayago2": {
72
- "train": "http://dl.fbaipublicfiles.com/KILT/aidayago2-train-kilt.jsonl",
73
- "validation": "http://dl.fbaipublicfiles.com/KILT/aidayago2-dev-kilt.jsonl",
74
- "test": "http://dl.fbaipublicfiles.com/KILT/aidayago2-test_without_answers-kilt.jsonl",
75
- },
76
- "wned": {
77
- "validation": "http://dl.fbaipublicfiles.com/KILT/wned-dev-kilt.jsonl",
78
- "test": "http://dl.fbaipublicfiles.com/KILT/wned-test_without_answers-kilt.jsonl",
79
- },
80
- "cweb": {
81
- "validation": "http://dl.fbaipublicfiles.com/KILT/cweb-dev-kilt.jsonl",
82
- "test": "http://dl.fbaipublicfiles.com/KILT/cweb-test_without_answers-kilt.jsonl",
83
- },
84
- "trex": {
85
- "train": "http://dl.fbaipublicfiles.com/KILT/trex-train-kilt.jsonl",
86
- "validation": "http://dl.fbaipublicfiles.com/KILT/trex-dev-kilt.jsonl",
87
- "test": "http://dl.fbaipublicfiles.com/KILT/trex-test_without_answers-kilt.jsonl",
88
- },
89
- "structured_zeroshot": {
90
- "train": "http://dl.fbaipublicfiles.com/KILT/structured_zeroshot-train-kilt.jsonl",
91
- "validation": "http://dl.fbaipublicfiles.com/KILT/structured_zeroshot-dev-kilt.jsonl",
92
- "test": "http://dl.fbaipublicfiles.com/KILT/structured_zeroshot-test_without_answers-kilt.jsonl",
93
- },
94
- "nq": {
95
- "train": "http://dl.fbaipublicfiles.com/KILT/nq-train-kilt.jsonl",
96
- "validation": "http://dl.fbaipublicfiles.com/KILT/nq-dev-kilt.jsonl",
97
- "test": "http://dl.fbaipublicfiles.com/KILT/nq-test_without_answers-kilt.jsonl",
98
- },
99
- "hotpotqa": {
100
- "train": "http://dl.fbaipublicfiles.com/KILT/hotpotqa-train-kilt.jsonl",
101
- "validation": "http://dl.fbaipublicfiles.com/KILT/hotpotqa-dev-kilt.jsonl",
102
- "test": "http://dl.fbaipublicfiles.com/KILT/hotpotqa-test_without_answers-kilt.jsonl",
103
- },
104
- "triviaqa_support_only": {
105
- "train": "http://dl.fbaipublicfiles.com/KILT/triviaqa-train_id-kilt.jsonl",
106
- "validation": "http://dl.fbaipublicfiles.com/KILT/triviaqa-dev_id-kilt.jsonl",
107
- "test": "http://dl.fbaipublicfiles.com/KILT/triviaqa-test_id_without_answers-kilt.jsonl",
108
- },
109
- "eli5": {
110
- "train": "http://dl.fbaipublicfiles.com/KILT/eli5-train-kilt.jsonl",
111
- "validation": "http://dl.fbaipublicfiles.com/KILT/eli5-dev-kilt.jsonl",
112
- "test": "http://dl.fbaipublicfiles.com/KILT/eli5-test_without_answers-kilt.jsonl",
113
- },
114
- "wow": {
115
- "train": "http://dl.fbaipublicfiles.com/KILT/wow-train-kilt.jsonl",
116
- "validation": "http://dl.fbaipublicfiles.com/KILT/wow-dev-kilt.jsonl",
117
- "test": "http://dl.fbaipublicfiles.com/KILT/wow-test_without_answers-kilt.jsonl",
118
- },
119
- }
120
-
121
-
122
- class KiltTasks(datasets.GeneratorBasedBuilder):
123
-
124
- BUILDER_CONFIGS = [
125
- datasets.BuilderConfig(
126
- name="triviaqa_support_only",
127
- version=datasets.Version("1.0.0"),
128
- description="Supporting paragraphs information for the TriviaQA task",
129
- )
130
- ] + [
131
- datasets.BuilderConfig(
132
- name=k, version=datasets.Version("1.0.0"), description=f"Task data and supporting paragraphs for {k}"
133
- )
134
- for k in _DATA_URLS
135
- if k != "triviaqa_support_only"
136
- ]
137
-
138
- DEFAULT_CONFIG_NAME = "nq"
139
-
140
- def _info(self):
141
- return datasets.DatasetInfo(
142
- description=_DESCRIPTION,
143
- features=datasets.Features(
144
- {
145
- "id": datasets.Value("string"),
146
- "input": datasets.Value("string"),
147
- "meta": {
148
- "left_context": datasets.Value("string"),
149
- "mention": datasets.Value("string"),
150
- "right_context": datasets.Value("string"),
151
- "partial_evidence": [
152
- {
153
- "start_paragraph_id": datasets.Value("int32"),
154
- "end_paragraph_id": datasets.Value("int32"),
155
- "title": datasets.Value("string"),
156
- "section": datasets.Value("string"),
157
- "wikipedia_id": datasets.Value("string"),
158
- "meta": {"evidence_span": [datasets.Value("string")]},
159
- }
160
- ],
161
- "obj_surface": [datasets.Value("string")],
162
- "sub_surface": [datasets.Value("string")],
163
- "subj_aliases": [datasets.Value("string")],
164
- "template_questions": [datasets.Value("string")],
165
- },
166
- "output": [
167
- {
168
- "answer": datasets.Value("string"),
169
- "meta": {"score": datasets.Value("int32")},
170
- "provenance": [
171
- {
172
- "bleu_score": datasets.Value("float32"),
173
- "start_character": datasets.Value("int32"),
174
- "start_paragraph_id": datasets.Value("int32"),
175
- "end_character": datasets.Value("int32"),
176
- "end_paragraph_id": datasets.Value("int32"),
177
- "meta": {
178
- "fever_page_id": datasets.Value("string"),
179
- "fever_sentence_id": datasets.Value("int32"),
180
- "annotation_id": datasets.Value("string"), # int runs into overflow issues
181
- "yes_no_answer": datasets.Value("string"),
182
- "evidence_span": [datasets.Value("string")],
183
- },
184
- "section": datasets.Value("string"),
185
- "title": datasets.Value("string"),
186
- "wikipedia_id": datasets.Value("string"),
187
- }
188
- ],
189
- }
190
- ],
191
- }
192
- ),
193
- supervised_keys=None,
194
- homepage="https://github.com/facebookresearch/KILT",
195
- citation=_CITATION,
196
- )
197
-
198
- def _split_generators(self, dl_manager):
199
- file_paths = dl_manager.download_and_extract(_DATA_URLS[self.config.name])
200
- return [
201
- datasets.SplitGenerator(name=split, gen_kwargs={"filepath": downloaded_path})
202
- for split, downloaded_path in file_paths.items()
203
- ]
204
-
205
- def _generate_examples(self, filepath):
206
- logger.info("generating examples from = %s", filepath)
207
- with open(filepath, encoding="utf-8") as f:
208
- for idx, line in enumerate(f):
209
- article = json.loads(line.strip())
210
- article["input"] = article.get("input", "")
211
- # meta
212
- article["meta"] = article.get("meta", {})
213
- for k in ["left_context", "mention", "right_context"]:
214
- article["meta"][k] = article["meta"].get(k, "")
215
- for k in ["obj_surface", "sub_surface", "subj_aliases", "template_questions"]:
216
- article["meta"][k] = article["meta"].get(k, [])
217
- # partial evidence
218
- article["meta"]["partial_evidence"] = [
219
- {
220
- "start_paragraph_id": partial.get("start_paragraph_id", -1),
221
- "end_paragraph_id": partial.get("end_paragraph_id", -1),
222
- "title": partial.get("title", ""),
223
- "section": partial.get("section", ""),
224
- "wikipedia_id": partial.get("wikipedia_id", ""),
225
- "meta": {"evidence_span": partial.get("meta", {}).get("evidence_span", [])},
226
- }
227
- for partial in article["meta"].get("partial_evidence", [])
228
- ]
229
- # output
230
- article["output"] = [
231
- {
232
- "answer": output.get("answer", ""),
233
- "meta": output.get("meta", {"score": -1}),
234
- "provenance": [
235
- {
236
- "bleu_score": provenance.get("bleu_score", -1.0),
237
- "start_character": provenance.get("start_character", -1),
238
- "start_paragraph_id": provenance.get("start_paragraph_id", -1),
239
- "end_character": provenance.get("end_character", -1),
240
- "end_paragraph_id": provenance.get("end_paragraph_id", -1),
241
- "meta": {
242
- "fever_page_id": provenance.get("meta", {}).get("fever_page_id", ""),
243
- "fever_sentence_id": provenance.get("meta", {}).get("fever_sentence_id", -1),
244
- "annotation_id": str(
245
- provenance.get("meta", {}).get("annotation_id", -1)
246
- ), # int runs into overflow issues
247
- "yes_no_answer": provenance.get("meta", {}).get("yes_no_answer", ""),
248
- "evidence_span": provenance.get("meta", {}).get("evidence_span", []),
249
- },
250
- "section": provenance.get("section", ""),
251
- "title": provenance.get("title", ""),
252
- "wikipedia_id": provenance.get("wikipedia_id", ""),
253
- }
254
- for provenance in output.get("provenance", [])
255
- ],
256
- }
257
- for output in article.get("output", [])
258
- ]
259
- yield idx, article