Update URL for datasets

#2
Files changed (1) hide show
  1. indonlu.py +49 -49
indonlu.py CHANGED
@@ -38,7 +38,7 @@ and analyzing natural language understanding systems for Bahasa Indonesia.
38
 
39
  _INDONLU_HOMEPAGE = "https://www.indobenchmark.com/"
40
 
41
- _INDONLU_LICENSE = "https://raw.githubusercontent.com/indobenchmark/indonlu/master/LICENSE"
42
 
43
 
44
  class IndonluConfig(datasets.BuilderConfig):
@@ -93,12 +93,12 @@ class Indonlu(datasets.GeneratorBasedBuilder):
93
  different emotion labels: sadness, anger, love, fear, and happy."""
94
  ),
95
  text_features={"tweet": "tweet"},
96
- # label classes sorted refer to https://github.com/indobenchmark/indonlu/blob/master/utils/data_utils.py
97
  label_classes=["sadness", "anger", "love", "fear", "happy"],
98
  label_column="label",
99
- train_url="https://raw.githubusercontent.com/indobenchmark/indonlu/master/dataset/emot_emotion-twitter/train_preprocess.csv",
100
- valid_url="https://raw.githubusercontent.com/indobenchmark/indonlu/master/dataset/emot_emotion-twitter/valid_preprocess.csv",
101
- test_url="https://raw.githubusercontent.com/indobenchmark/indonlu/master/dataset/emot_emotion-twitter/test_preprocess_masked_label.csv",
102
  citation=textwrap.dedent(
103
  """\
104
  @inproceedings{saputri2018emotion,
@@ -122,12 +122,12 @@ class Indonlu(datasets.GeneratorBasedBuilder):
122
  dataset: positive, negative, and neutral."""
123
  ),
124
  text_features={"text": "text"},
125
- # label classes sorted refer to https://github.com/indobenchmark/indonlu/blob/master/utils/data_utils.py
126
  label_classes=["positive", "neutral", "negative"],
127
  label_column="label",
128
- train_url="https://raw.githubusercontent.com/indobenchmark/indonlu/master/dataset/smsa_doc-sentiment-prosa/train_preprocess.tsv",
129
- valid_url="https://raw.githubusercontent.com/indobenchmark/indonlu/master/dataset/smsa_doc-sentiment-prosa/valid_preprocess.tsv",
130
- test_url="https://raw.githubusercontent.com/indobenchmark/indonlu/master/dataset/smsa_doc-sentiment-prosa/test_preprocess_masked_label.tsv",
131
  citation=textwrap.dedent(
132
  """\
133
  @inproceedings{purwarianti2019improving,
@@ -151,12 +151,12 @@ class Indonlu(datasets.GeneratorBasedBuilder):
151
  negative, and neutral."""
152
  ),
153
  text_features={"sentence": "sentence"},
154
- # label classes sorted refer to https://github.com/indobenchmark/indonlu/blob/master/utils/data_utils.py
155
  label_classes=["negative", "neutral", "positive"],
156
  label_column=["fuel", "machine", "others", "part", "price", "service"],
157
- train_url="https://raw.githubusercontent.com/indobenchmark/indonlu/master/dataset/casa_absa-prosa/train_preprocess.csv",
158
- valid_url="https://raw.githubusercontent.com/indobenchmark/indonlu/master/dataset/casa_absa-prosa/valid_preprocess.csv",
159
- test_url="https://raw.githubusercontent.com/indobenchmark/indonlu/master/dataset/casa_absa-prosa/test_preprocess_masked_label.csv",
160
  citation=textwrap.dedent(
161
  """\
162
  @inproceedings{ilmania2018aspect,
@@ -181,7 +181,7 @@ class Indonlu(datasets.GeneratorBasedBuilder):
181
  of the same aspect but for different objects (e.g., cleanliness of bed and toilet)."""
182
  ),
183
  text_features={"sentence": "sentence"},
184
- # label classes sorted refer to https://github.com/indobenchmark/indonlu/blob/master/utils/data_utils.py
185
  label_classes=["neg", "neut", "pos", "neg_pos"],
186
  label_column=[
187
  "ac",
@@ -195,9 +195,9 @@ class Indonlu(datasets.GeneratorBasedBuilder):
195
  "tv",
196
  "wifi",
197
  ],
198
- train_url="https://raw.githubusercontent.com/indobenchmark/indonlu/master/dataset/hoasa_absa-airy/train_preprocess.csv",
199
- valid_url="https://raw.githubusercontent.com/indobenchmark/indonlu/master/dataset/hoasa_absa-airy/valid_preprocess.csv",
200
- test_url="https://raw.githubusercontent.com/indobenchmark/indonlu/master/dataset/hoasa_absa-airy/test_preprocess_masked_label.csv",
201
  citation=textwrap.dedent(
202
  """\
203
  @inproceedings{azhar2019multi,
@@ -223,12 +223,12 @@ class Indonlu(datasets.GeneratorBasedBuilder):
223
  "hypothesis": "hypothesis",
224
  "category": "category",
225
  },
226
- # label classes sorted refer to https://github.com/indobenchmark/indonlu/blob/master/utils/data_utils.py
227
  label_classes=["NotEntail", "Entail_or_Paraphrase"],
228
  label_column="label",
229
- train_url="https://raw.githubusercontent.com/indobenchmark/indonlu/master/dataset/wrete_entailment-ui/train_preprocess.csv",
230
- valid_url="https://raw.githubusercontent.com/indobenchmark/indonlu/master/dataset/wrete_entailment-ui/valid_preprocess.csv",
231
- test_url="https://raw.githubusercontent.com/indobenchmark/indonlu/master/dataset/wrete_entailment-ui/test_preprocess_masked_label.csv",
232
  citation=textwrap.dedent(
233
  """\
234
  @inproceedings{setya2018semi,
@@ -248,7 +248,7 @@ class Indonlu(datasets.GeneratorBasedBuilder):
248
  Indonesian Association of Computational Linguistics (INACL) POS Tagging Convention."""
249
  ),
250
  text_features={"tokens": "tokens"},
251
- # label classes sorted refer to https://github.com/indobenchmark/indonlu/blob/master/utils/data_utils.py
252
  label_classes=[
253
  "B-PPO",
254
  "B-KUA",
@@ -278,9 +278,9 @@ class Indonlu(datasets.GeneratorBasedBuilder):
278
  "B-VBE",
279
  ],
280
  label_column="pos_tags",
281
- train_url="https://raw.githubusercontent.com/indobenchmark/indonlu/master/dataset/posp_pos-prosa/train_preprocess.txt",
282
- valid_url="https://raw.githubusercontent.com/indobenchmark/indonlu/master/dataset/posp_pos-prosa/valid_preprocess.txt",
283
- test_url="https://raw.githubusercontent.com/indobenchmark/indonlu/master/dataset/posp_pos-prosa/test_preprocess_masked_label.txt",
284
  citation=textwrap.dedent(
285
  """\
286
  @inproceedings{hoesen2018investigating,
@@ -302,7 +302,7 @@ class Indonlu(datasets.GeneratorBasedBuilder):
302
  the experimental setting used by Kurniawan and Aji (2018)"""
303
  ),
304
  text_features={"tokens": "tokens"},
305
- # label classes sorted refer to https://github.com/indobenchmark/indonlu/blob/master/utils/data_utils.py
306
  label_classes=[
307
  "B-PR",
308
  "B-CD",
@@ -347,9 +347,9 @@ class Indonlu(datasets.GeneratorBasedBuilder):
347
  "B-X",
348
  ],
349
  label_column="pos_tags",
350
- train_url="https://raw.githubusercontent.com/indobenchmark/indonlu/master/dataset/bapos_pos-idn/train_preprocess.txt",
351
- valid_url="https://raw.githubusercontent.com/indobenchmark/indonlu/master/dataset/bapos_pos-idn/valid_preprocess.txt",
352
- test_url="https://raw.githubusercontent.com/indobenchmark/indonlu/master/dataset/bapos_pos-idn/test_preprocess_masked_label.txt",
353
  citation=textwrap.dedent(
354
  """\
355
  @inproceedings{dinakaramani2014designing,
@@ -380,12 +380,12 @@ class Indonlu(datasets.GeneratorBasedBuilder):
380
  Inside-Outside-Beginning (IOB) tagging representation with two kinds of tags, aspect and sentiment."""
381
  ),
382
  text_features={"tokens": "tokens"},
383
- # label classes sorted refer to https://github.com/indobenchmark/indonlu/blob/master/utils/data_utils.py
384
  label_classes=["I-SENTIMENT", "O", "I-ASPECT", "B-SENTIMENT", "B-ASPECT"],
385
  label_column="seq_label",
386
- train_url="https://raw.githubusercontent.com/indobenchmark/indonlu/master/dataset/terma_term-extraction-airy/train_preprocess.txt",
387
- valid_url="https://raw.githubusercontent.com/indobenchmark/indonlu/master/dataset/terma_term-extraction-airy/valid_preprocess.txt",
388
- test_url="https://raw.githubusercontent.com/indobenchmark/indonlu/master/dataset/terma_term-extraction-airy/test_preprocess_masked_label.txt",
389
  citation=textwrap.dedent(
390
  """\
391
  @article{winatmoko2019aspect,
@@ -413,12 +413,12 @@ class Indonlu(datasets.GeneratorBasedBuilder):
413
  which represents the position of the keyphrase."""
414
  ),
415
  text_features={"tokens": "tokens"},
416
- # label classes sorted refer to https://github.com/indobenchmark/indonlu/blob/master/utils/data_utils.py
417
  label_classes=["O", "B", "I"],
418
  label_column="seq_label",
419
- train_url="https://raw.githubusercontent.com/indobenchmark/indonlu/master/dataset/keps_keyword-extraction-prosa/train_preprocess.txt",
420
- valid_url="https://raw.githubusercontent.com/indobenchmark/indonlu/master/dataset/keps_keyword-extraction-prosa/valid_preprocess.txt",
421
- test_url="https://raw.githubusercontent.com/indobenchmark/indonlu/master/dataset/keps_keyword-extraction-prosa/test_preprocess_masked_label.txt",
422
  citation=textwrap.dedent(
423
  """\
424
  @inproceedings{mahfuzh2019improving,
@@ -440,12 +440,12 @@ class Indonlu(datasets.GeneratorBasedBuilder):
440
  ORGANIZATION (name of organization)."""
441
  ),
442
  text_features={"tokens": "tokens"},
443
- # label classes sorted refer to https://github.com/indobenchmark/indonlu/blob/master/utils/data_utils.py
444
  label_classes=["I-PERSON", "B-ORGANISATION", "I-ORGANISATION", "B-PLACE", "I-PLACE", "O", "B-PERSON"],
445
  label_column="ner_tags",
446
- train_url="https://raw.githubusercontent.com/indobenchmark/indonlu/master/dataset/nergrit_ner-grit/train_preprocess.txt",
447
- valid_url="https://raw.githubusercontent.com/indobenchmark/indonlu/master/dataset/nergrit_ner-grit/valid_preprocess.txt",
448
- test_url="https://raw.githubusercontent.com/indobenchmark/indonlu/master/dataset/nergrit_ner-grit/test_preprocess_masked_label.txt",
449
  citation=textwrap.dedent(
450
  """\
451
  @online{nergrit2019,
@@ -465,7 +465,7 @@ class Indonlu(datasets.GeneratorBasedBuilder):
465
  EVT (name of the event), and FNB (name of food and beverage). The NERP dataset uses the IOB chunking format."""
466
  ),
467
  text_features={"tokens": "tokens"},
468
- # label classes sorted refer to https://github.com/indobenchmark/indonlu/blob/master/utils/data_utils.py
469
  label_classes=[
470
  "I-PPL",
471
  "B-EVT",
@@ -480,9 +480,9 @@ class Indonlu(datasets.GeneratorBasedBuilder):
480
  "I-FNB",
481
  ],
482
  label_column="ner_tags",
483
- train_url="https://raw.githubusercontent.com/indobenchmark/indonlu/master/dataset/nerp_ner-prosa/train_preprocess.txt",
484
- valid_url="https://raw.githubusercontent.com/indobenchmark/indonlu/master/dataset/nerp_ner-prosa/valid_preprocess.txt",
485
- test_url="https://raw.githubusercontent.com/indobenchmark/indonlu/master/dataset/nerp_ner-prosa/test_preprocess_masked_label.txt",
486
  citation=textwrap.dedent(
487
  """\
488
  @inproceedings{hoesen2018investigating,
@@ -505,12 +505,12 @@ class Indonlu(datasets.GeneratorBasedBuilder):
505
  There are six categories of questions: date, location, name, organization, person, and quantitative."""
506
  ),
507
  text_features={"question": "question", "passage": "passage"},
508
- # label classes sorted refer to https://github.com/indobenchmark/indonlu/blob/master/utils/data_utils.py
509
  label_classes=["O", "B", "I"],
510
  label_column="seq_label",
511
- train_url="https://raw.githubusercontent.com/indobenchmark/indonlu/master/dataset/facqa_qa-factoid-itb/train_preprocess.csv",
512
- valid_url="https://raw.githubusercontent.com/indobenchmark/indonlu/master/dataset/facqa_qa-factoid-itb/valid_preprocess.csv",
513
- test_url="https://raw.githubusercontent.com/indobenchmark/indonlu/master/dataset/facqa_qa-factoid-itb/test_preprocess_masked_label.csv",
514
  citation=textwrap.dedent(
515
  """\
516
  @inproceedings{purwarianti2007machine,
 
38
 
39
  _INDONLU_HOMEPAGE = "https://www.indobenchmark.com/"
40
 
41
+ _INDONLU_LICENSE = "https://raw.githubusercontent.com/IndoNLP/indonlu/master/LICENSE"
42
 
43
 
44
  class IndonluConfig(datasets.BuilderConfig):
 
93
  different emotion labels: sadness, anger, love, fear, and happy."""
94
  ),
95
  text_features={"tweet": "tweet"},
96
+ # label classes sorted refer to https://github.com/IndoNLP/indonlu/blob/master/utils/data_utils.py
97
  label_classes=["sadness", "anger", "love", "fear", "happy"],
98
  label_column="label",
99
+ train_url="https://raw.githubusercontent.com/IndoNLP/indonlu/master/dataset/emot_emotion-twitter/train_preprocess.csv",
100
+ valid_url="https://raw.githubusercontent.com/IndoNLP/indonlu/master/dataset/emot_emotion-twitter/valid_preprocess.csv",
101
+ test_url="https://raw.githubusercontent.com/IndoNLP/indonlu/master/dataset/emot_emotion-twitter/test_preprocess.csv",
102
  citation=textwrap.dedent(
103
  """\
104
  @inproceedings{saputri2018emotion,
 
122
  dataset: positive, negative, and neutral."""
123
  ),
124
  text_features={"text": "text"},
125
+ # label classes sorted refer to https://github.com/IndoNLP/indonlu/blob/master/utils/data_utils.py
126
  label_classes=["positive", "neutral", "negative"],
127
  label_column="label",
128
+ train_url="https://raw.githubusercontent.com/IndoNLP/indonlu/master/dataset/smsa_doc-sentiment-prosa/train_preprocess.tsv",
129
+ valid_url="https://raw.githubusercontent.com/IndoNLP/indonlu/master/dataset/smsa_doc-sentiment-prosa/valid_preprocess.tsv",
130
+ test_url="https://raw.githubusercontent.com/IndoNLP/indonlu/master/dataset/smsa_doc-sentiment-prosa/test_preprocess.tsv",
131
  citation=textwrap.dedent(
132
  """\
133
  @inproceedings{purwarianti2019improving,
 
151
  negative, and neutral."""
152
  ),
153
  text_features={"sentence": "sentence"},
154
+ # label classes sorted refer to https://github.com/IndoNLP/indonlu/blob/master/utils/data_utils.py
155
  label_classes=["negative", "neutral", "positive"],
156
  label_column=["fuel", "machine", "others", "part", "price", "service"],
157
+ train_url="https://raw.githubusercontent.com/IndoNLP/indonlu/master/dataset/casa_absa-prosa/train_preprocess.csv",
158
+ valid_url="https://raw.githubusercontent.com/IndoNLP/indonlu/master/dataset/casa_absa-prosa/valid_preprocess.csv",
159
+ test_url="https://raw.githubusercontent.com/IndoNLP/indonlu/master/dataset/casa_absa-prosa/test_preprocess.csv",
160
  citation=textwrap.dedent(
161
  """\
162
  @inproceedings{ilmania2018aspect,
 
181
  of the same aspect but for different objects (e.g., cleanliness of bed and toilet)."""
182
  ),
183
  text_features={"sentence": "sentence"},
184
+ # label classes sorted refer to https://github.com/IndoNLP/indonlu/blob/master/utils/data_utils.py
185
  label_classes=["neg", "neut", "pos", "neg_pos"],
186
  label_column=[
187
  "ac",
 
195
  "tv",
196
  "wifi",
197
  ],
198
+ train_url="https://raw.githubusercontent.com/IndoNLP/indonlu/master/dataset/hoasa_absa-airy/train_preprocess.csv",
199
+ valid_url="https://raw.githubusercontent.com/IndoNLP/indonlu/master/dataset/hoasa_absa-airy/valid_preprocess.csv",
200
+ test_url="https://raw.githubusercontent.com/IndoNLP/indonlu/master/dataset/hoasa_absa-airy/test_preprocess.csv",
201
  citation=textwrap.dedent(
202
  """\
203
  @inproceedings{azhar2019multi,
 
223
  "hypothesis": "hypothesis",
224
  "category": "category",
225
  },
226
+ # label classes sorted refer to https://github.com/IndoNLP/indonlu/blob/master/utils/data_utils.py
227
  label_classes=["NotEntail", "Entail_or_Paraphrase"],
228
  label_column="label",
229
+ train_url="https://raw.githubusercontent.com/IndoNLP/indonlu/master/dataset/wrete_entailment-ui/train_preprocess.csv",
230
+ valid_url="https://raw.githubusercontent.com/IndoNLP/indonlu/master/dataset/wrete_entailment-ui/valid_preprocess.csv",
231
+ test_url="https://raw.githubusercontent.com/IndoNLP/indonlu/master/dataset/wrete_entailment-ui/test_preprocess.csv",
232
  citation=textwrap.dedent(
233
  """\
234
  @inproceedings{setya2018semi,
 
248
  Indonesian Association of Computational Linguistics (INACL) POS Tagging Convention."""
249
  ),
250
  text_features={"tokens": "tokens"},
251
+ # label classes sorted refer to https://github.com/IndoNLP/indonlu/blob/master/utils/data_utils.py
252
  label_classes=[
253
  "B-PPO",
254
  "B-KUA",
 
278
  "B-VBE",
279
  ],
280
  label_column="pos_tags",
281
+ train_url="https://raw.githubusercontent.com/IndoNLP/indonlu/master/dataset/posp_pos-prosa/train_preprocess.txt",
282
+ valid_url="https://raw.githubusercontent.com/IndoNLP/indonlu/master/dataset/posp_pos-prosa/valid_preprocess.txt",
283
+ test_url="https://raw.githubusercontent.com/IndoNLP/indonlu/master/dataset/posp_pos-prosa/test_preprocess.txt",
284
  citation=textwrap.dedent(
285
  """\
286
  @inproceedings{hoesen2018investigating,
 
302
  the experimental setting used by Kurniawan and Aji (2018)"""
303
  ),
304
  text_features={"tokens": "tokens"},
305
+ # label classes sorted refer to https://github.com/IndoNLP/indonlu/blob/master/utils/data_utils.py
306
  label_classes=[
307
  "B-PR",
308
  "B-CD",
 
347
  "B-X",
348
  ],
349
  label_column="pos_tags",
350
+ train_url="https://raw.githubusercontent.com/IndoNLP/indonlu/master/dataset/bapos_pos-idn/train_preprocess.txt",
351
+ valid_url="https://raw.githubusercontent.com/IndoNLP/indonlu/master/dataset/bapos_pos-idn/valid_preprocess.txt",
352
+ test_url="https://raw.githubusercontent.com/IndoNLP/indonlu/master/dataset/bapos_pos-idn/test_preprocess.txt",
353
  citation=textwrap.dedent(
354
  """\
355
  @inproceedings{dinakaramani2014designing,
 
380
  Inside-Outside-Beginning (IOB) tagging representation with two kinds of tags, aspect and sentiment."""
381
  ),
382
  text_features={"tokens": "tokens"},
383
+ # label classes sorted refer to https://github.com/IndoNLP/indonlu/blob/master/utils/data_utils.py
384
  label_classes=["I-SENTIMENT", "O", "I-ASPECT", "B-SENTIMENT", "B-ASPECT"],
385
  label_column="seq_label",
386
+ train_url="https://raw.githubusercontent.com/IndoNLP/indonlu/master/dataset/terma_term-extraction-airy/train_preprocess.txt",
387
+ valid_url="https://raw.githubusercontent.com/IndoNLP/indonlu/master/dataset/terma_term-extraction-airy/valid_preprocess.txt",
388
+ test_url="https://raw.githubusercontent.com/IndoNLP/indonlu/master/dataset/terma_term-extraction-airy/test_preprocess.txt",
389
  citation=textwrap.dedent(
390
  """\
391
  @article{winatmoko2019aspect,
 
413
  which represents the position of the keyphrase."""
414
  ),
415
  text_features={"tokens": "tokens"},
416
+ # label classes sorted refer to https://github.com/IndoNLP/indonlu/blob/master/utils/data_utils.py
417
  label_classes=["O", "B", "I"],
418
  label_column="seq_label",
419
+ train_url="https://raw.githubusercontent.com/IndoNLP/indonlu/master/dataset/keps_keyword-extraction-prosa/train_preprocess.txt",
420
+ valid_url="https://raw.githubusercontent.com/IndoNLP/indonlu/master/dataset/keps_keyword-extraction-prosa/valid_preprocess.txt",
421
+ test_url="https://raw.githubusercontent.com/IndoNLP/indonlu/master/dataset/keps_keyword-extraction-prosa/test_preprocess.txt",
422
  citation=textwrap.dedent(
423
  """\
424
  @inproceedings{mahfuzh2019improving,
 
440
  ORGANIZATION (name of organization)."""
441
  ),
442
  text_features={"tokens": "tokens"},
443
+ # label classes sorted refer to https://github.com/IndoNLP/indonlu/blob/master/utils/data_utils.py
444
  label_classes=["I-PERSON", "B-ORGANISATION", "I-ORGANISATION", "B-PLACE", "I-PLACE", "O", "B-PERSON"],
445
  label_column="ner_tags",
446
+ train_url="https://raw.githubusercontent.com/IndoNLP/indonlu/master/dataset/nergrit_ner-grit/train_preprocess.txt",
447
+ valid_url="https://raw.githubusercontent.com/IndoNLP/indonlu/master/dataset/nergrit_ner-grit/valid_preprocess.txt",
448
+ test_url="https://raw.githubusercontent.com/IndoNLP/indonlu/master/dataset/nergrit_ner-grit/test_preprocess.txt",
449
  citation=textwrap.dedent(
450
  """\
451
  @online{nergrit2019,
 
465
  EVT (name of the event), and FNB (name of food and beverage). The NERP dataset uses the IOB chunking format."""
466
  ),
467
  text_features={"tokens": "tokens"},
468
+ # label classes sorted refer to https://github.com/IndoNLP/indonlu/blob/master/utils/data_utils.py
469
  label_classes=[
470
  "I-PPL",
471
  "B-EVT",
 
480
  "I-FNB",
481
  ],
482
  label_column="ner_tags",
483
+ train_url="https://raw.githubusercontent.com/IndoNLP/indonlu/master/dataset/nerp_ner-prosa/train_preprocess.txt",
484
+ valid_url="https://raw.githubusercontent.com/IndoNLP/indonlu/master/dataset/nerp_ner-prosa/valid_preprocess.txt",
485
+ test_url="https://raw.githubusercontent.com/IndoNLP/indonlu/master/dataset/nerp_ner-prosa/test_preprocess.txt",
486
  citation=textwrap.dedent(
487
  """\
488
  @inproceedings{hoesen2018investigating,
 
505
  There are six categories of questions: date, location, name, organization, person, and quantitative."""
506
  ),
507
  text_features={"question": "question", "passage": "passage"},
508
+ # label classes sorted refer to https://github.com/IndoNLP/indonlu/blob/master/utils/data_utils.py
509
  label_classes=["O", "B", "I"],
510
  label_column="seq_label",
511
+ train_url="https://raw.githubusercontent.com/IndoNLP/indonlu/master/dataset/facqa_qa-factoid-itb/train_preprocess.csv",
512
+ valid_url="https://raw.githubusercontent.com/IndoNLP/indonlu/master/dataset/facqa_qa-factoid-itb/valid_preprocess.csv",
513
+ test_url="https://raw.githubusercontent.com/IndoNLP/indonlu/master/dataset/facqa_qa-factoid-itb/test_preprocess.csv",
514
  citation=textwrap.dedent(
515
  """\
516
  @inproceedings{purwarianti2007machine,