holylovenia commited on
Commit
b3f5008
·
verified ·
1 Parent(s): 239ec94

Upload kde4.py with huggingface_hub

Browse files
Files changed (1) hide show
  1. kde4.py +574 -0
kde4.py ADDED
@@ -0,0 +1,574 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # coding=utf-8
2
+ # Copyright 2022 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
+
16
+ """
17
+ A parallel corpus of KDE4 localization files. The corpus is available in 92 languages in total, with 4099 bitexts.
18
+ """
19
+ import os
20
+ from pathlib import Path
21
+ from typing import Dict, List, Tuple
22
+
23
+ import datasets
24
+
25
+ from seacrowd.utils import schemas
26
+ from seacrowd.utils.configs import SEACrowdConfig
27
+ from seacrowd.utils.constants import Licenses, Tasks
28
+
29
+ _CITATION = """\
30
+ @inproceedings{tiedemann2012parallel,
31
+ title={Parallel Data, Tools and Interfaces in OPUS},
32
+ author={Tiedemann, J{\"o}rg},
33
+ booktitle={Proceedings of the Eighth International Conference on Language Resources and Evaluation (LREC'12)},
34
+ pages={2214--2218},
35
+ year={2012}
36
+ }
37
+ """
38
+
39
+ _DATASETNAME = "kde4"
40
+
41
+ _DESCRIPTION = """\
42
+ A parallel corpus of KDE4 localization files. The corpus is available in 92 languages in total, with 4099 bitexts.
43
+ """
44
+
45
+ _HOMEPAGE = "https://opus.nlpl.eu/KDE4/corpus/version/KDE4"
46
+
47
+ _LANGUAGES = ["ind", "khm", "zlm", "tha", "vie"]
48
+
49
+ _LICENSE = Licenses.UNKNOWN.value
50
+ _LOCAL = False
51
+
52
+ _URL_TEMPLATE = "https://object.pouta.csc.fi/OPUS-KDE4/v2/moses/{src}-{tgt}.txt.zip"
53
+
54
+ _SUPPORTED_TASKS = [Tasks.MACHINE_TRANSLATION]
55
+ _SOURCE_VERSION = "2.0.0"
56
+ _SEACROWD_VERSION = "2024.06.20"
57
+
58
+ kde4_language_codes = {
59
+ "aar": "aa",
60
+ "abk": "ab",
61
+ "ave": "ae",
62
+ "afr": "af",
63
+ "aka": "ak",
64
+ "amh": "am",
65
+ "arg": "an",
66
+ "ara": "ar",
67
+ "asm": "as",
68
+ "ava": "av",
69
+ "aym": "ay",
70
+ "aze": "az",
71
+ "bak": "ba",
72
+ "bel": "be",
73
+ "bul": "bg",
74
+ "bis": "bi",
75
+ "bam": "bm",
76
+ "ben": "bn",
77
+ "bod": "bo",
78
+ "bre": "br",
79
+ "bos": "bs",
80
+ "cat": "ca",
81
+ "che": "ce",
82
+ "cha": "ch",
83
+ "cos": "co",
84
+ "cre": "cr",
85
+ "ces": "cs",
86
+ "chu": "cu",
87
+ "chv": "cv",
88
+ "cym": "cy",
89
+ "dan": "da",
90
+ "deu": "de",
91
+ "div": "dv",
92
+ "dzo": "dz",
93
+ "ewe": "ee",
94
+ "ell": "el",
95
+ "eng": "en",
96
+ "epo": "eo",
97
+ "spa": "es",
98
+ "est": "et",
99
+ "eus": "eu",
100
+ "fas": "fa",
101
+ "ful": "ff",
102
+ "fin": "fi",
103
+ "fij": "fj",
104
+ "fao": "fo",
105
+ "fra": "fr",
106
+ "fry": "fy",
107
+ "gle": "ga",
108
+ "gla": "gd",
109
+ "glg": "gl",
110
+ "grn": "gn",
111
+ "guj": "gu",
112
+ "glv": "gv",
113
+ "hau": "ha",
114
+ "heb": "he",
115
+ "hin": "hi",
116
+ "hmo": "ho",
117
+ "hrv": "hr",
118
+ "hat": "ht",
119
+ "hun": "hu",
120
+ "hye": "hy",
121
+ "her": "hz",
122
+ "ina": "ia",
123
+ "ind": "id",
124
+ "ile": "ie",
125
+ "ibo": "ig",
126
+ "iii": "ii",
127
+ "ipk": "ik",
128
+ "ido": "io",
129
+ "isl": "is",
130
+ "ita": "it",
131
+ "iku": "iu",
132
+ "jpn": "ja",
133
+ "jav": "jv",
134
+ "kat": "ka",
135
+ "kon": "kg",
136
+ "kik": "ki",
137
+ "kua": "kj",
138
+ "kaz": "kk",
139
+ "kal": "kl",
140
+ "khm": "km",
141
+ "kan": "kn",
142
+ "kor": "ko",
143
+ "kau": "kr",
144
+ "kas": "ks",
145
+ "kur": "ku",
146
+ "kom": "kv",
147
+ "cor": "kw",
148
+ "kir": "ky",
149
+ "lat": "la",
150
+ "ltz": "lb",
151
+ "lug": "lg",
152
+ "lim": "li",
153
+ "lin": "ln",
154
+ "lao": "lo",
155
+ "lit": "lt",
156
+ "lub": "lu",
157
+ "lav": "lv",
158
+ "mlg": "mg",
159
+ "mah": "mh",
160
+ "mri": "mi",
161
+ "mkd": "mk",
162
+ "mal": "ml",
163
+ "mon": "mn",
164
+ "mar": "mr",
165
+ "msa": "ms",
166
+ "mlt": "mt",
167
+ "mya": "my",
168
+ "nau": "na",
169
+ "nob": "nb",
170
+ "nde": "nd",
171
+ "nep": "ne",
172
+ "ndo": "ng",
173
+ "nld": "nl",
174
+ "nno": "nn",
175
+ "nor": "no",
176
+ "nbl": "nr",
177
+ "nav": "nv",
178
+ "nya": "ny",
179
+ "oci": "oc",
180
+ "oji": "oj",
181
+ "orm": "om",
182
+ "ori": "or",
183
+ "oss": "os",
184
+ "pan": "pa",
185
+ "pli": "pi",
186
+ "pol": "pl",
187
+ "pus": "ps",
188
+ "por": "pt",
189
+ "que": "qu",
190
+ "roh": "rm",
191
+ "run": "rn",
192
+ "ron": "ro",
193
+ "rus": "ru",
194
+ "kin": "rw",
195
+ "san": "sa",
196
+ "srd": "sc",
197
+ "snd": "sd",
198
+ "sme": "se",
199
+ "sag": "sg",
200
+ "hbs": "sh",
201
+ "sin": "si",
202
+ "slk": "sk",
203
+ "slv": "sl",
204
+ "smo": "sm",
205
+ "sna": "sn",
206
+ "som": "so",
207
+ "sqi": "sq",
208
+ "srp": "sr",
209
+ "ssw": "ss",
210
+ "sot": "st",
211
+ "sun": "su",
212
+ "swe": "sv",
213
+ "swa": "sw",
214
+ "tam": "ta",
215
+ "tel": "te",
216
+ "tgk": "tg",
217
+ "tha": "th",
218
+ "tir": "ti",
219
+ "tuk": "tk",
220
+ "tgl": "tl",
221
+ "tsn": "tn",
222
+ "ton": "to",
223
+ "tur": "tr",
224
+ "tso": "ts",
225
+ "tat": "tt",
226
+ "twi": "tw",
227
+ "tah": "ty",
228
+ "uig": "ug",
229
+ "ukr": "uk",
230
+ "urd": "ur",
231
+ "uzb": "uz",
232
+ "ven": "ve",
233
+ "vie": "vi",
234
+ "vol": "vo",
235
+ "wln": "wa",
236
+ "wol": "wo",
237
+ "xho": "xh",
238
+ "yid": "yi",
239
+ "yor": "yo",
240
+ "zha": "za",
241
+ "zho": "zh",
242
+ "zul": "zu",
243
+ "nds": "nds",
244
+ "mai": "mai",
245
+ "nso": "nso",
246
+ "ast": "ast",
247
+ "crh": "crh",
248
+ "csb": "csb",
249
+ "hne": "hne",
250
+ "hsb": "hsb",
251
+ }
252
+
253
+ configs = {
254
+ "afr": ["msa", "tha", "khm", "ind", "vie"],
255
+ "ara": ["msa", "tha", "khm", "ind", "vie"],
256
+ "asm": ["msa", "tha", "khm", "ind", "vie"],
257
+ "ast": ["msa", "tha", "khm", "ind", "vie"],
258
+ "bel": ["msa", "tha", "khm", "ind", "vie"],
259
+ "bul": ["msa", "tha", "khm", "ind", "vie"],
260
+ "ben": ["msa", "tha", "khm", "ind", "vie"],
261
+ "bre": ["msa", "tha", "khm", "ind", "vie"],
262
+ "cat": ["msa", "tha", "khm", "ind", "vie"],
263
+ "crh": ["msa", "tha", "khm", "ind", "vie"],
264
+ "ces": ["msa", "tha", "khm", "ind", "vie"],
265
+ "csb": ["msa", "tha", "khm", "ind", "vie"],
266
+ "cym": ["msa", "tha", "khm", "ind", "vie"],
267
+ "dan": ["msa", "tha", "khm", "ind", "vie"],
268
+ "deu": ["msa", "tha", "khm", "ind", "vie"],
269
+ "ell": ["msa", "tha", "khm", "ind", "vie"],
270
+ "eng": ["msa", "tha", "khm", "ind", "vie"],
271
+ "epo": ["msa", "tha", "khm", "ind", "vie"],
272
+ "spa": ["msa", "tha", "khm", "ind", "vie"],
273
+ "est": ["msa", "tha", "khm", "ind", "vie"],
274
+ "eus": ["msa", "tha", "khm", "ind", "vie"],
275
+ "fas": ["msa", "tha", "khm", "ind", "vie"],
276
+ "fin": ["msa", "tha", "khm", "ind", "vie"],
277
+ "fra": ["msa", "tha", "khm", "ind", "vie"],
278
+ "fry": ["msa", "tha", "khm", "ind", "vie"],
279
+ "gle": ["msa", "tha", "khm", "ind", "vie"],
280
+ "glg": ["msa", "tha", "khm", "ind", "vie"],
281
+ "guj": ["msa", "tha", "khm", "ind", "vie"],
282
+ "hau": ["msa", "tha", "khm", "ind", "vie"],
283
+ "heb": ["msa", "tha", "khm", "ind", "vie"],
284
+ "hin": ["msa", "tha", "khm", "ind", "vie"],
285
+ "hne": ["msa", "tha", "khm", "ind", "vie"],
286
+ "hrv": ["msa", "tha", "khm", "ind", "vie"],
287
+ "hsb": ["msa", "tha", "khm", "ind", "vie"],
288
+ "hun": ["msa", "tha", "khm", "ind", "vie"],
289
+ "hye": ["msa", "tha", "khm", "ind", "vie"],
290
+ "ind": [
291
+ "kan",
292
+ "pus",
293
+ "msa",
294
+ "slv",
295
+ "tur",
296
+ "rus",
297
+ "nld",
298
+ "mkd",
299
+ "jpn",
300
+ "ori",
301
+ "nep",
302
+ "xho",
303
+ "nds",
304
+ "lav",
305
+ "ukr",
306
+ "vie",
307
+ "mai",
308
+ "tam",
309
+ "ltz",
310
+ "isl",
311
+ "uzb",
312
+ "sme",
313
+ "lit",
314
+ "tgk",
315
+ "kat",
316
+ "mal",
317
+ "srp",
318
+ "wln",
319
+ "por",
320
+ "oci",
321
+ "kur",
322
+ "mar",
323
+ "sin",
324
+ "slk",
325
+ "kor",
326
+ "kaz",
327
+ "ron",
328
+ "nno",
329
+ "tha",
330
+ "khm",
331
+ "tel",
332
+ "ita",
333
+ "pol",
334
+ "swe",
335
+ "pan",
336
+ "nob",
337
+ ],
338
+ "isl": ["khm", "msa", "tha", "vie"],
339
+ "ita": ["khm", "msa", "tha", "vie"],
340
+ "jpn": ["khm", "msa", "tha", "vie"],
341
+ "kat": ["khm", "msa", "tha", "vie"],
342
+ "kaz": ["khm", "msa", "tha", "vie"],
343
+ "khm": [
344
+ "kan",
345
+ "pus",
346
+ "msa",
347
+ "slv",
348
+ "tur",
349
+ "rus",
350
+ "kin",
351
+ "nld",
352
+ "mkd",
353
+ "ori",
354
+ "xho",
355
+ "nso",
356
+ "nep",
357
+ "nds",
358
+ "lav",
359
+ "ukr",
360
+ "vie",
361
+ "mai",
362
+ "tam",
363
+ "ltz",
364
+ "uzb",
365
+ "sme",
366
+ "lit",
367
+ "tgk",
368
+ "mlt",
369
+ "mal",
370
+ "srp",
371
+ "wln",
372
+ "por",
373
+ "oci",
374
+ "kur",
375
+ "mar",
376
+ "sin",
377
+ "slk",
378
+ "kor",
379
+ "ron",
380
+ "nno",
381
+ "tha",
382
+ "tel",
383
+ "pol",
384
+ "swe",
385
+ "pan",
386
+ "nob",
387
+ ],
388
+ "kan": ["msa", "tha", "vie"],
389
+ "kor": ["msa", "tha", "vie"],
390
+ "kur": ["msa", "tha", "vie"],
391
+ "ltz": ["msa", "tha", "vie"],
392
+ "lit": ["msa", "tha", "vie"],
393
+ "lav": ["msa", "tha", "vie"],
394
+ "mai": ["msa", "tha", "vie"],
395
+ "mkd": ["msa", "tha", "vie"],
396
+ "mal": ["msa", "tha", "vie"],
397
+ "mar": ["msa", "tha", "vie"],
398
+ "msa": [
399
+ "pus",
400
+ "slv",
401
+ "tur",
402
+ "rus",
403
+ "kin",
404
+ "nld",
405
+ "ori",
406
+ "xho",
407
+ "nso",
408
+ "nep",
409
+ "nds",
410
+ "ukr",
411
+ "vie",
412
+ "tam",
413
+ "uzb",
414
+ "sme",
415
+ "tgk",
416
+ "mlt",
417
+ "srp",
418
+ "wln",
419
+ "por",
420
+ "oci",
421
+ "slk",
422
+ "sin",
423
+ "ron",
424
+ "nno",
425
+ "tha",
426
+ "tel",
427
+ "pol",
428
+ "swe",
429
+ "pan",
430
+ "nob",
431
+ ],
432
+ "mlt": ["tha"],
433
+ "nob": ["tha", "vie"],
434
+ "nds": ["tha", "vie"],
435
+ "nep": ["tha", "vie"],
436
+ "nld": ["tha", "vie"],
437
+ "nno": ["tha", "vie"],
438
+ "oci": ["tha", "vie"],
439
+ "ori": ["tha", "vie"],
440
+ "pan": ["tha", "vie"],
441
+ "pol": ["tha", "vie"],
442
+ "pus": ["tha", "vie"],
443
+ "por": ["tha", "vie"],
444
+ "ron": ["tha", "vie"],
445
+ "rus": ["tha", "vie"],
446
+ "kin": ["tha"],
447
+ "sme": ["tha", "vie"],
448
+ "sin": ["tha", "vie"],
449
+ "slk": ["tha", "vie"],
450
+ "slv": ["tha", "vie"],
451
+ "srp": ["tha", "vie"],
452
+ "swe": ["tha", "vie"],
453
+ "tam": ["tha", "vie"],
454
+ "tel": ["tha", "vie"],
455
+ "tgk": ["tha", "vie"],
456
+ "tha": ["wln", "tur", "uzb", "xho", "vie", "ukr"],
457
+ "tur": ["vie"],
458
+ "ukr": ["vie"],
459
+ "uzb": ["vie"],
460
+ "vie": ["wln", "xho"],
461
+ }
462
+
463
+
464
+ class KDE4Dataset(datasets.GeneratorBasedBuilder):
465
+ """A parallel corpus of KDE4 localization files. The corpus is available in 92 languages in total, with 4099 bitexts."""
466
+
467
+ SOURCE_VERSION = datasets.Version(_SOURCE_VERSION)
468
+ SEACROWD_VERSION = datasets.Version(_SEACROWD_VERSION)
469
+
470
+ BUILDER_CONFIGS = (
471
+ [
472
+ SEACrowdConfig(
473
+ name="kde4_source",
474
+ version=datasets.Version(_SOURCE_VERSION),
475
+ description="kde4 source schema for afr to msa",
476
+ schema="source",
477
+ subset_id="afr_msa",
478
+ ),
479
+ SEACrowdConfig(
480
+ name="kde4_seacrowd_t2t",
481
+ version=datasets.Version(_SOURCE_VERSION),
482
+ description="kde4 seacrowd_t2t schema for afr to msa",
483
+ schema="seacrowd_t2t",
484
+ subset_id="afr_msa",
485
+ ),
486
+ ]
487
+ + [
488
+ SEACrowdConfig(
489
+ name=f"kde4_{src_lang}_{tgt_lang}_source",
490
+ version=datasets.Version(_SOURCE_VERSION),
491
+ description=f"kde4 source schema for {src_lang} to {tgt_lang}",
492
+ schema="source",
493
+ subset_id=f"{src_lang}_{tgt_lang}",
494
+ )
495
+ for src_lang in configs
496
+ for tgt_lang in configs[src_lang]
497
+ ]
498
+ + [
499
+ SEACrowdConfig(
500
+ name=f"kde4_{src_lang}_{tgt_lang}_seacrowd_t2t",
501
+ version=datasets.Version(_SEACROWD_VERSION),
502
+ description=f"kde4 seacrowd_t2t schema for {src_lang} to {tgt_lang}",
503
+ schema="seacrowd_t2t",
504
+ subset_id=f"{src_lang}_{tgt_lang}",
505
+ )
506
+ for src_lang in configs
507
+ for tgt_lang in configs[src_lang]
508
+ ]
509
+ )
510
+
511
+ DEFAULT_CONFIG_NAME = "kde4_source"
512
+
513
+ def _info(self) -> datasets.DatasetInfo:
514
+ if self.config.schema == "source":
515
+ features = datasets.Features(
516
+ {
517
+ "id": datasets.Value("string"),
518
+ "src_text": datasets.Value("string"),
519
+ "tgt_text": datasets.Value("string"),
520
+ }
521
+ )
522
+
523
+ elif self.config.schema == "seacrowd_t2t":
524
+ features = schemas.text2text_features
525
+
526
+ return datasets.DatasetInfo(
527
+ description=_DESCRIPTION,
528
+ features=features,
529
+ homepage=_HOMEPAGE,
530
+ license=_LICENSE,
531
+ citation=_CITATION,
532
+ )
533
+
534
+ def _split_generators(self, dl_manager: datasets.DownloadManager) -> List[datasets.SplitGenerator]:
535
+ """Returns SplitGenerators."""
536
+
537
+ src_lang, tgt_lang = self.config.subset_id.split("_")
538
+ kde4_src_lang, kde4_tgt_lang = kde4_language_codes[src_lang], kde4_language_codes[tgt_lang]
539
+
540
+ url = _URL_TEMPLATE.format(src=kde4_src_lang, tgt=kde4_tgt_lang)
541
+ data_dir = dl_manager.download_and_extract(url)
542
+
543
+ return [
544
+ datasets.SplitGenerator(
545
+ name=datasets.Split.TRAIN,
546
+ gen_kwargs={
547
+ "src_lang": src_lang,
548
+ "tgt_lang": tgt_lang,
549
+ "src_filepath": os.path.join(data_dir, f"KDE4.{kde4_src_lang}-{kde4_tgt_lang}.{kde4_src_lang}"),
550
+ "tgt_filepath": os.path.join(data_dir, f"KDE4.{kde4_src_lang}-{kde4_tgt_lang}.{kde4_tgt_lang}"),
551
+ },
552
+ )
553
+ ]
554
+
555
+ def _generate_examples(self, src_lang: str, tgt_lang: str, src_filepath: Path, tgt_filepath: Path) -> Tuple[int, Dict]:
556
+ """Yields examples as (key, example) tuples."""
557
+
558
+ if self.config.schema == "source":
559
+ for row_id, (src_text, tgt_text) in enumerate(zip(open(src_filepath), open(tgt_filepath))):
560
+ yield row_id, {
561
+ "id": row_id,
562
+ "src_text": src_text.strip(),
563
+ "tgt_text": tgt_text.strip(),
564
+ }
565
+
566
+ elif self.config.schema == "seacrowd_t2t":
567
+ for row_id, (src_text, tgt_text) in enumerate(zip(open(src_filepath), open(tgt_filepath))):
568
+ yield row_id, {
569
+ "id": row_id,
570
+ "text_1": src_text.strip(),
571
+ "text_2": tgt_text.strip(),
572
+ "text_1_name": src_lang,
573
+ "text_2_name": tgt_lang,
574
+ }