Datasets:

ArXiv:
akshitab commited on
Commit
c36967a
1 Parent(s): a6efe24

- add ccmatrix bitext (f883c6a1196bdc8284a17c71ceab78ec7ccd6af4)
- cleanup language list (69397c259a1bd82049b6466c08826b3c9b1f292b)
- fix ccmatrix loading (7248137579c471c4cb98840ab3536274f5148c8b)
- remove \n from line before parsing it (b5a508f5285b20f0d082c25da4227f80ccd31662)
- add dataset_infos for ccmatrix (5de2d46792bea01e07a3ba789666221a2866bf28)
- use CC-Matrix citation for CCMatrix data (24a8ba0735d8375ed9ae099fb25b2d311553ffba)
- add CCMatrix to Readme (cbd3655e2fdaea23391d4a768ad1d69bdb7079b7)
- move ccmatrix_lang_pairs.py into nllb_lang_pairs.py (0c26393ca23e1828ec6cfcbdeec4da30487ffc5a)

Files changed (4) hide show
  1. README.md +2 -1
  2. dataset_infos.json +0 -0
  3. nllb.py +109 -32
  4. nllb_lang_pairs.py +1144 -1
README.md CHANGED
@@ -188,9 +188,10 @@ The dataset is released under the terms of [ODC-BY](https://opendatacommons.org/
188
 
189
  ### Citation Information
190
 
 
191
  Hefferman et al, Bitext Mining Using Distilled Sentence Representations for Low-Resource Languages. Arxiv https://arxiv.org/abs/2205.12654, 2022.<br>
192
  NLLB Team et al, No Language Left Behind: Scaling Human-Centered Machine Translation, Arxiv https://arxiv.org/abs/2207.04672, 2022.
193
 
194
  ### Contributions
195
 
196
- We thank the NLLB Meta AI team for open sourcing the meta data and instructions on how to use it with special thanks to Bapi Akula, Pierre Andrews, Onur Çelebi, Sergey Edunov, Kenneth Heafield, Philipp Koehn, Alex Mourachko, Safiyyah Saleem, Holger Schwenk, and Guillaume Wenzek. We also thank the AllenNLP team at AI2 for hosting and releasing this data, including Akshita Bhagia (for engineering efforts to host the data, and create the huggingface dataset), and Jesse Dodge (for organizing the connection).
 
188
 
189
  ### Citation Information
190
 
191
+ Schwenk et al, CCMatrix: Mining Billions of High-Quality Parallel Sentences on the Web. ACL https://aclanthology.org/2021.acl-long.507/
192
  Hefferman et al, Bitext Mining Using Distilled Sentence Representations for Low-Resource Languages. Arxiv https://arxiv.org/abs/2205.12654, 2022.<br>
193
  NLLB Team et al, No Language Left Behind: Scaling Human-Centered Machine Translation, Arxiv https://arxiv.org/abs/2207.04672, 2022.
194
 
195
  ### Contributions
196
 
197
+ We thank the NLLB Meta AI team for open sourcing the meta data and instructions on how to use it with special thanks to Bapi Akula, Pierre Andrews, Onur Çelebi, Sergey Edunov, Kenneth Heafield, Philipp Koehn, Alex Mourachko, Safiyyah Saleem, Holger Schwenk, and Guillaume Wenzek. We also thank the AllenNLP team at AI2 for hosting and releasing this data, including Akshita Bhagia (for engineering efforts to host the data, and create the huggingface dataset), and Jesse Dodge (for organizing the connection).
dataset_infos.json CHANGED
The diff for this file is too large to render. See raw diff
 
nllb.py CHANGED
@@ -12,22 +12,36 @@
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
- """No Language Left Behind (NLLB)"""
 
 
 
 
 
 
16
 
17
  import datasets
18
- import csv
19
- import json
20
 
21
- _CITATION = (
22
  "@article{team2022NoLL,"
23
  "title={No Language Left Behind: Scaling Human-Centered Machine Translation},"
24
- "author={Nllb team and Marta Ruiz Costa-juss{\`a} and James Cross and Onur cCelebi and Maha Elbayad and Kenneth Heafield and Kevin Heffernan and Elahe Kalbassi and Janice Lam and Daniel Licht and Jean Maillard and Anna Sun and Skyler Wang and Guillaume Wenzek and Alison Youngblood and Bapi Akula and Lo{\"i}c Barrault and Gabriel Mejia Gonzalez and Prangthip Hansanti and John Hoffman and Semarley Jarrett and Kaushik Ram Sadagopan and Dirk Rowe and Shannon L. Spruit and C. Tran and Pierre Andrews and Necip Fazil Ayan and Shruti Bhosale and Sergey Edunov and Angela Fan and Cynthia Gao and Vedanuj Goswami and Francisco Guzm'an and Philipp Koehn and Alexandre Mourachko and Christophe Ropers and Safiyyah Saleem and Holger Schwenk and Jeff Wang},"
25
  "journal={ArXiv},"
26
  "year={2022},"
27
  "volume={abs/2207.04672}"
28
  "}"
29
  )
30
 
 
 
 
 
 
 
 
 
 
31
 
32
  _DESCRIPTION = "" # TODO
33
 
@@ -35,42 +49,72 @@ _HOMEPAGE = "" # TODO
35
 
36
  _LICENSE = "https://opendatacommons.org/licenses/by/1-0/"
37
 
38
- from .nllb_lang_pairs import LANG_PAIRS as _LANGUAGE_PAIRS
39
-
40
- _URL_BASE = "https://storage.googleapis.com/allennlp-data-bucket/nllb/"
41
 
42
- _URLs = {
43
- f"{src_lg}-{trg_lg}": f"{_URL_BASE}{src_lg}-{trg_lg}.gz"
44
- for src_lg, trg_lg in _LANGUAGE_PAIRS
45
- }
46
 
47
 
48
  class NLLBTaskConfig(datasets.BuilderConfig):
49
  """BuilderConfig for No Language Left Behind Dataset."""
50
 
51
- def __init__(self, src_lg, tgt_lg, **kwargs):
52
  super(NLLBTaskConfig, self).__init__(**kwargs)
53
  self.src_lg = src_lg
54
  self.tgt_lg = tgt_lg
 
 
55
 
 
 
 
56
 
57
- class NLLB(datasets.GeneratorBasedBuilder):
58
- """No Language Left Behind Dataset."""
 
 
59
 
60
- BUILDER_CONFIGS = [
61
- NLLBTaskConfig(
 
62
  name=f"{src_lg}-{tgt_lg}",
63
  version=datasets.Version("1.0.0"),
64
  description=f"No Language Left Behind (NLLB): {src_lg} - {tgt_lg}",
65
  src_lg=src_lg,
66
  tgt_lg=tgt_lg,
 
67
  )
68
- for (src_lg, tgt_lg) in _LANGUAGE_PAIRS
69
- ]
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
70
  BUILDER_CONFIG_CLASS = NLLBTaskConfig
71
 
72
  def _info(self):
73
  # define feature types
 
74
  features = datasets.Features(
75
  {
76
  "translation": datasets.Translation(
@@ -85,6 +129,17 @@ class NLLB(datasets.GeneratorBasedBuilder):
85
  "target_sentence_url": datasets.Value("string"),
86
  }
87
  )
 
 
 
 
 
 
 
 
 
 
 
88
 
89
  return datasets.DatasetInfo(
90
  description=_DESCRIPTION,
@@ -92,22 +147,20 @@ class NLLB(datasets.GeneratorBasedBuilder):
92
  supervised_keys=None,
93
  homepage=_HOMEPAGE,
94
  license=_LICENSE,
95
- citation=_CITATION,
96
  )
97
 
98
  def _split_generators(self, dl_manager):
99
- """Returns SplitGenerators."""
100
- pair = f"{self.config.src_lg}-{self.config.tgt_lg}" # string identifier for language pair
101
- url = _URLs[pair] # url for download of pair-specific file
102
- data_file = dl_manager.download_and_extract(
103
- url
104
- ) # extract downloaded data and store path in data_file
105
 
 
 
 
106
  return [
107
  datasets.SplitGenerator(
108
  name=datasets.Split.TRAIN,
109
  gen_kwargs={
110
- "filepath": data_file,
111
  "source_lg": self.config.src_lg,
112
  "target_lg": self.config.tgt_lg,
113
  },
@@ -115,16 +168,24 @@ class NLLB(datasets.GeneratorBasedBuilder):
115
  ]
116
 
117
  def _generate_examples(self, filepath, source_lg, target_lg):
 
 
 
 
 
 
 
118
  with open(filepath, encoding="utf-8") as f:
119
  # reader = csv.reader(f, delimiter="\t")
120
  for id_, example in enumerate(f):
121
  try:
122
- datarow = example.split("\t")
123
  row = {}
 
124
  row["translation"] = {
125
  source_lg: datarow[0],
126
  target_lg: datarow[1],
127
- } # create translation json
128
  row["laser_score"] = float(datarow[2])
129
  row["source_sentence_lid"] = float(datarow[3])
130
  row["target_sentence_lid"] = float(datarow[4])
@@ -132,14 +193,30 @@ class NLLB(datasets.GeneratorBasedBuilder):
132
  row["source_sentence_url"] = datarow[6]
133
  row["target_sentence_source"] = datarow[7]
134
  row["target_sentence_url"] = datarow[8]
135
- row = {
136
- k: None if not v else v for k, v in row.items()
137
- } # replace empty values
138
  except:
139
  print(datarow)
140
  raise
141
  yield id_, row
142
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
143
 
144
  # to test the script, go to the root folder of the repo (nllb) and run:
145
  # datasets-cli test nllb --save_infos --all_configs
 
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
+ """No Language Left Behind (NLLB)
16
+
17
+ The "No Language Left Behind" paper is a dataset with translation examples across 200 languages.
18
+ This paper reuses prior work, and for some language pairs is just reusing CC-Matrix published on statmt.org.
19
+ Depending on the language pair chosen, this script will fetch either the original
20
+ version from statmt.org, or the new one from AllenAI.
21
+ """
22
 
23
  import datasets
24
+ import typing as tp
 
25
 
26
+ NLLB_CITATION = (
27
  "@article{team2022NoLL,"
28
  "title={No Language Left Behind: Scaling Human-Centered Machine Translation},"
29
+ r"author={Nllb team and Marta Ruiz Costa-juss{\`a} and James Cross and Onur Celebi and Maha Elbayad and Kenneth Heafield and Kevin Heffernan and Elahe Kalbassi and Janice Lam and Daniel Licht and Jean Maillard and Anna Sun and Skyler Wang and Guillaume Wenzek and Alison Youngblood and Bapi Akula and Lo{\"i}c Barrault and Gabriel Mejia Gonzalez and Prangthip Hansanti and John Hoffman and Semarley Jarrett and Kaushik Ram Sadagopan and Dirk Rowe and Shannon L. Spruit and C. Tran and Pierre Andrews and Necip Fazil Ayan and Shruti Bhosale and Sergey Edunov and Angela Fan and Cynthia Gao and Vedanuj Goswami and Francisco Guzm'an and Philipp Koehn and Alexandre Mourachko and Christophe Ropers and Safiyyah Saleem and Holger Schwenk and Jeff Wang},"
30
  "journal={ArXiv},"
31
  "year={2022},"
32
  "volume={abs/2207.04672}"
33
  "}"
34
  )
35
 
36
+ CCMATRIX_CITATION = (
37
+ "@inproceedings{schwenk2021ccmatrix,"
38
+ "title={CCMatrix: Mining Billions of High-Quality Parallel Sentences on the Web},"
39
+ "author={Schwenk, Holger and Wenzek, Guillaume and Edunov, Sergey and Grave, {'E}douard and Joulin, Armand and Fan, Angela},"
40
+ "booktitle={Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers)},"
41
+ "pages={6490--6500},"
42
+ "year={2021}"
43
+ )
44
+
45
 
46
  _DESCRIPTION = "" # TODO
47
 
 
49
 
50
  _LICENSE = "https://opendatacommons.org/licenses/by/1-0/"
51
 
52
+ from .nllb_lang_pairs import NLLB_PAIRS, CCMATRIX_PAIRS, CCMATRIX_MAPPING
 
 
53
 
54
+ _ALLENAI_URL = "https://storage.googleapis.com/allennlp-data-bucket/nllb/"
55
+ _STATMT_URL = "http://data.statmt.org/cc-matrix/"
 
 
56
 
57
 
58
  class NLLBTaskConfig(datasets.BuilderConfig):
59
  """BuilderConfig for No Language Left Behind Dataset."""
60
 
61
+ def __init__(self, src_lg, tgt_lg, url, **kwargs):
62
  super(NLLBTaskConfig, self).__init__(**kwargs)
63
  self.src_lg = src_lg
64
  self.tgt_lg = tgt_lg
65
+ self.url = url
66
+ self.source = "statmt" if url.startswith(_STATMT_URL) else "allenai"
67
 
68
+ def _builder_configs() -> tp.List[NLLBTaskConfig]:
69
+ """
70
+ Creates the dataset used for training NLLB model.
71
 
72
+ Note we always return data from AllenAI if possible because CC-Matrix data
73
+ is older, and most language pairs have been improved between the two versions.
74
+ """
75
+ configs = {}
76
 
77
+ for (src_lg, tgt_lg) in NLLB_PAIRS:
78
+ assert (src_lg, tgt_lg) not in configs
79
+ configs[(src_lg, tgt_lg)] = NLLBTaskConfig(
80
  name=f"{src_lg}-{tgt_lg}",
81
  version=datasets.Version("1.0.0"),
82
  description=f"No Language Left Behind (NLLB): {src_lg} - {tgt_lg}",
83
  src_lg=src_lg,
84
  tgt_lg=tgt_lg,
85
+ url=f"{_ALLENAI_URL}{src_lg}-{tgt_lg}.gz",
86
  )
87
+
88
+ for (src_lg, tgt_lg) in CCMATRIX_PAIRS:
89
+ # Prevent accidental override
90
+ # Note: the lang pairs are not consistently alphabetically sorted,
91
+ # because CCMatrix was using other language code which may swap the order.
92
+ # Let's keep the original order because they correspond to the column order in the file.
93
+ assert (src_lg, tgt_lg) not in configs
94
+ assert (tgt_lg, src_lg) not in configs
95
+ src_cc, tgt_cc = CCMATRIX_MAPPING[src_lg], CCMATRIX_MAPPING[tgt_lg]
96
+ configs[(src_lg, tgt_lg)] = NLLBTaskConfig(
97
+ name=f"{src_lg}-{tgt_lg}",
98
+ version=datasets.Version("1.0.0"),
99
+ description=f"No Language Left Behind (NLLB): {src_lg} - {tgt_lg}",
100
+ src_lg=src_lg,
101
+ tgt_lg=tgt_lg,
102
+ # Use CCMatrix language code to fetch from statmt
103
+ url=f"{_STATMT_URL}{src_cc}-{tgt_cc}.bitextf.tsv.gz",
104
+ )
105
+
106
+ return list(configs.values())
107
+
108
+
109
+ class NLLB(datasets.GeneratorBasedBuilder):
110
+ """No Language Left Behind Dataset."""
111
+
112
+ BUILDER_CONFIGS = _builder_configs()
113
  BUILDER_CONFIG_CLASS = NLLBTaskConfig
114
 
115
  def _info(self):
116
  # define feature types
117
+ citation = NLLB_CITATION
118
  features = datasets.Features(
119
  {
120
  "translation": datasets.Translation(
 
129
  "target_sentence_url": datasets.Value("string"),
130
  }
131
  )
132
+ if self.config.source == "statmt":
133
+ citation = CCMATRIX_CITATION
134
+ # MT stats didn't published all the metadata
135
+ features = datasets.Features(
136
+ {
137
+ "translation": datasets.Translation(
138
+ languages=(self.config.src_lg, self.config.tgt_lg)
139
+ ),
140
+ "laser_score": datasets.Value("float32"),
141
+ }
142
+ )
143
 
144
  return datasets.DatasetInfo(
145
  description=_DESCRIPTION,
 
147
  supervised_keys=None,
148
  homepage=_HOMEPAGE,
149
  license=_LICENSE,
150
+ citation=citation,
151
  )
152
 
153
  def _split_generators(self, dl_manager):
154
+ """Returns one training generator. NLLB200 is meant for training.
 
 
 
 
 
155
 
156
+ If you're interested in evaluation look at https://huggingface.co/datasets/facebook/flores
157
+ """
158
+ local_file = dl_manager.download_and_extract(self.config.url)
159
  return [
160
  datasets.SplitGenerator(
161
  name=datasets.Split.TRAIN,
162
  gen_kwargs={
163
+ "filepath": local_file,
164
  "source_lg": self.config.src_lg,
165
  "target_lg": self.config.tgt_lg,
166
  },
 
168
  ]
169
 
170
  def _generate_examples(self, filepath, source_lg, target_lg):
171
+ if self.config.source == "statmt":
172
+ # MT stats didn't published all the metadata
173
+ return self._generate_minimal_examples(filepath, source_lg, target_lg)
174
+
175
+ return self._generate_full_examples(filepath, source_lg, target_lg)
176
+
177
+ def _generate_full_examples(self, filepath, source_lg, target_lg):
178
  with open(filepath, encoding="utf-8") as f:
179
  # reader = csv.reader(f, delimiter="\t")
180
  for id_, example in enumerate(f):
181
  try:
182
+ datarow = example.rstrip("\n").split("\t")
183
  row = {}
184
+ # create translation json
185
  row["translation"] = {
186
  source_lg: datarow[0],
187
  target_lg: datarow[1],
188
+ }
189
  row["laser_score"] = float(datarow[2])
190
  row["source_sentence_lid"] = float(datarow[3])
191
  row["target_sentence_lid"] = float(datarow[4])
 
193
  row["source_sentence_url"] = datarow[6]
194
  row["target_sentence_source"] = datarow[7]
195
  row["target_sentence_url"] = datarow[8]
196
+ # replace empty values
197
+ row = {k: None if not v else v for k, v in row.items()}
 
198
  except:
199
  print(datarow)
200
  raise
201
  yield id_, row
202
 
203
+ def _generate_minimal_examples(self, filepath, source_lg, target_lg):
204
+ with open(filepath, encoding="utf-8") as f:
205
+ for i, example in enumerate(f):
206
+ try:
207
+ (score, src, tgt) = example.rstrip("\n").split("\t")
208
+ row = {
209
+ "translation": {
210
+ source_lg: src,
211
+ target_lg: tgt,
212
+ },
213
+ "laser_score": score,
214
+ }
215
+ except:
216
+ print(example)
217
+ raise
218
+ yield i, row
219
+
220
 
221
  # to test the script, go to the root folder of the repo (nllb) and run:
222
  # datasets-cli test nllb --save_infos --all_configs
nllb_lang_pairs.py CHANGED
@@ -1,4 +1,11 @@
1
- LANG_PAIRS = [
 
 
 
 
 
 
 
2
  ("ace_Latn", "ban_Latn"),
3
  ("ace_Latn", "bjn_Latn"),
4
  ("ace_Latn", "bug_Latn"),
@@ -1611,3 +1618,1139 @@ LANG_PAIRS = [
1611
  ("xho_Latn", "zul_Latn"),
1612
  ("yor_Latn", "zul_Latn"),
1613
  ]
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ """
2
+ List of languages pairs with bitext used in NLLB paper.
3
+
4
+ The list is split in two:
5
+ - NLLB_PAIRS for pairs released with NLLB paper,
6
+ - CCMATRIX_PAIRS pairs for which NLLB paper reused previous CCMatrix dataset.
7
+ """
8
+ NLLB_PAIRS = [
9
  ("ace_Latn", "ban_Latn"),
10
  ("ace_Latn", "bjn_Latn"),
11
  ("ace_Latn", "bug_Latn"),
 
1618
  ("xho_Latn", "zul_Latn"),
1619
  ("yor_Latn", "zul_Latn"),
1620
  ]
1621
+
1622
+ CCMATRIX_PAIRS = [
1623
+ ("afr_Latn", "dan_Latn"),
1624
+ ("afr_Latn", "deu_Latn"),
1625
+ ("afr_Latn", "isl_Latn"),
1626
+ ("afr_Latn", "ltz_Latn"),
1627
+ ("afr_Latn", "nld_Latn"),
1628
+ ("afr_Latn", "nob_Latn"),
1629
+ ("afr_Latn", "swe_Latn"),
1630
+ ("afr_Latn", "tur_Latn"),
1631
+ ("afr_Latn", "ydd_Hebr"),
1632
+ ("als_Latn", "srp_Cyrl"),
1633
+ ("als_Latn", "swe_Latn"),
1634
+ ("als_Latn", "tur_Latn"),
1635
+ ("als_Latn", "ukr_Cyrl"),
1636
+ ("arb_Arab", "als_Latn"),
1637
+ ("arb_Arab", "azj_Latn"),
1638
+ ("arb_Arab", "ben_Beng"),
1639
+ ("arb_Arab", "bul_Cyrl"),
1640
+ ("arb_Arab", "cat_Latn"),
1641
+ ("arb_Arab", "ces_Latn"),
1642
+ ("arb_Arab", "dan_Latn"),
1643
+ ("arb_Arab", "deu_Latn"),
1644
+ ("arb_Arab", "ell_Grek"),
1645
+ ("arb_Arab", "eng_Latn"),
1646
+ ("arb_Arab", "epo_Latn"),
1647
+ ("arb_Arab", "est_Latn"),
1648
+ ("arb_Arab", "fin_Latn"),
1649
+ ("arb_Arab", "fra_Latn"),
1650
+ ("arb_Arab", "gaz_Latn"),
1651
+ ("arb_Arab", "heb_Hebr"),
1652
+ ("arb_Arab", "hin_Deva"),
1653
+ ("arb_Arab", "hrv_Latn"),
1654
+ ("arb_Arab", "hun_Latn"),
1655
+ ("arb_Arab", "hye_Armn"),
1656
+ ("arb_Arab", "ind_Latn"),
1657
+ ("arb_Arab", "ita_Latn"),
1658
+ ("arb_Arab", "jpn_Jpan"),
1659
+ ("arb_Arab", "kat_Geor"),
1660
+ ("arb_Arab", "kaz_Cyrl"),
1661
+ ("arb_Arab", "khm_Khmr"),
1662
+ ("arb_Arab", "kor_Hang"),
1663
+ ("arb_Arab", "lit_Latn"),
1664
+ ("arb_Arab", "lug_Latn"),
1665
+ ("arb_Arab", "lvs_Latn"),
1666
+ ("arb_Arab", "mal_Mlym"),
1667
+ ("arb_Arab", "mar_Deva"),
1668
+ ("arb_Arab", "mya_Mymr"),
1669
+ ("arb_Arab", "nld_Latn"),
1670
+ ("arb_Arab", "nob_Latn"),
1671
+ ("arb_Arab", "npi_Deva"),
1672
+ ("arb_Arab", "ory_Orya"),
1673
+ ("arb_Arab", "pes_Arab"),
1674
+ ("arb_Arab", "pol_Latn"),
1675
+ ("arb_Arab", "por_Latn"),
1676
+ ("arb_Arab", "ron_Latn"),
1677
+ ("arb_Arab", "rus_Cyrl"),
1678
+ ("arb_Arab", "sin_Sinh"),
1679
+ ("arb_Arab", "slk_Latn"),
1680
+ ("arb_Arab", "slv_Latn"),
1681
+ ("arb_Arab", "snd_Arab"),
1682
+ ("arb_Arab", "spa_Latn"),
1683
+ ("arb_Arab", "srp_Cyrl"),
1684
+ ("arb_Arab", "swe_Latn"),
1685
+ ("arb_Arab", "swh_Latn"),
1686
+ ("arb_Arab", "tam_Taml"),
1687
+ ("arb_Arab", "tgl_Latn"),
1688
+ ("arb_Arab", "tur_Latn"),
1689
+ ("arb_Arab", "ukr_Cyrl"),
1690
+ ("arb_Arab", "uzn_Latn"),
1691
+ ("arb_Arab", "vie_Latn"),
1692
+ ("arb_Arab", "xho_Latn"),
1693
+ ("arb_Arab", "zho_Hans"),
1694
+ ("arb_Arab", "zsm_Latn"),
1695
+ ("arb_Arab", "zul_Latn"),
1696
+ ("ast_Latn", "cat_Latn"),
1697
+ ("ast_Latn", "eng_Latn"),
1698
+ ("ast_Latn", "fra_Latn"),
1699
+ ("ast_Latn", "glg_Latn"),
1700
+ ("ast_Latn", "ita_Latn"),
1701
+ ("ast_Latn", "oci_Latn"),
1702
+ ("ast_Latn", "por_Latn"),
1703
+ ("ast_Latn", "ron_Latn"),
1704
+ ("ast_Latn", "spa_Latn"),
1705
+ ("azj_Latn", "hye_Armn"),
1706
+ ("azj_Latn", "ind_Latn"),
1707
+ ("azj_Latn", "kat_Geor"),
1708
+ ("azj_Latn", "kaz_Cyrl"),
1709
+ ("azj_Latn", "khm_Khmr"),
1710
+ ("azj_Latn", "mya_Mymr"),
1711
+ ("azj_Latn", "tat_Cyrl"),
1712
+ ("azj_Latn", "tur_Latn"),
1713
+ ("azj_Latn", "vie_Latn"),
1714
+ ("bel_Cyrl", "als_Latn"),
1715
+ ("bel_Cyrl", "ben_Beng"),
1716
+ ("bel_Cyrl", "bul_Cyrl"),
1717
+ ("bel_Cyrl", "ces_Latn"),
1718
+ ("bel_Cyrl", "dan_Latn"),
1719
+ ("bel_Cyrl", "deu_Latn"),
1720
+ ("bel_Cyrl", "ell_Grek"),
1721
+ ("bel_Cyrl", "epo_Latn"),
1722
+ ("bel_Cyrl", "est_Latn"),
1723
+ ("bel_Cyrl", "fin_Latn"),
1724
+ ("bel_Cyrl", "hin_Deva"),
1725
+ ("bel_Cyrl", "hrv_Latn"),
1726
+ ("bel_Cyrl", "hun_Latn"),
1727
+ ("bel_Cyrl", "ita_Latn"),
1728
+ ("bel_Cyrl", "lit_Latn"),
1729
+ ("bel_Cyrl", "lvs_Latn"),
1730
+ ("bel_Cyrl", "nld_Latn"),
1731
+ ("bel_Cyrl", "nob_Latn"),
1732
+ ("bel_Cyrl", "pol_Latn"),
1733
+ ("bel_Cyrl", "ron_Latn"),
1734
+ ("bel_Cyrl", "slk_Latn"),
1735
+ ("bel_Cyrl", "slv_Latn"),
1736
+ ("bel_Cyrl", "srp_Cyrl"),
1737
+ ("bel_Cyrl", "swe_Latn"),
1738
+ ("bel_Cyrl", "swh_Latn"),
1739
+ ("bel_Cyrl", "tur_Latn"),
1740
+ ("bel_Cyrl", "ukr_Cyrl"),
1741
+ ("bel_Cyrl", "vie_Latn"),
1742
+ ("ben_Beng", "cat_Latn"),
1743
+ ("ben_Beng", "ces_Latn"),
1744
+ ("ben_Beng", "deu_Latn"),
1745
+ ("ben_Beng", "ell_Grek"),
1746
+ ("ben_Beng", "epo_Latn"),
1747
+ ("ben_Beng", "est_Latn"),
1748
+ ("ben_Beng", "fin_Latn"),
1749
+ ("ben_Beng", "fra_Latn"),
1750
+ ("ben_Beng", "heb_Hebr"),
1751
+ ("ben_Beng", "hun_Latn"),
1752
+ ("ben_Beng", "ilo_Latn"),
1753
+ ("ben_Beng", "ind_Latn"),
1754
+ ("ben_Beng", "ita_Latn"),
1755
+ ("ben_Beng", "jpn_Jpan"),
1756
+ ("ben_Beng", "kor_Hang"),
1757
+ ("ben_Beng", "lit_Latn"),
1758
+ ("ben_Beng", "lvs_Latn"),
1759
+ ("ben_Beng", "mkd_Cyrl"),
1760
+ ("ben_Beng", "nld_Latn"),
1761
+ ("ben_Beng", "pes_Arab"),
1762
+ ("ben_Beng", "plt_Latn"),
1763
+ ("ben_Beng", "pol_Latn"),
1764
+ ("ben_Beng", "por_Latn"),
1765
+ ("ben_Beng", "ron_Latn"),
1766
+ ("ben_Beng", "rus_Cyrl"),
1767
+ ("ben_Beng", "slk_Latn"),
1768
+ ("ben_Beng", "slv_Latn"),
1769
+ ("ben_Beng", "spa_Latn"),
1770
+ ("ben_Beng", "srp_Cyrl"),
1771
+ ("ben_Beng", "swe_Latn"),
1772
+ ("ben_Beng", "swh_Latn"),
1773
+ ("ben_Beng", "tgl_Latn"),
1774
+ ("ben_Beng", "tur_Latn"),
1775
+ ("ben_Beng", "ukr_Cyrl"),
1776
+ ("ben_Beng", "vie_Latn"),
1777
+ ("ben_Beng", "zho_Hans"),
1778
+ ("ben_Beng", "zsm_Latn"),
1779
+ ("bul_Cyrl", "als_Latn"),
1780
+ ("bul_Cyrl", "ben_Beng"),
1781
+ ("bul_Cyrl", "cat_Latn"),
1782
+ ("bul_Cyrl", "ces_Latn"),
1783
+ ("bul_Cyrl", "dan_Latn"),
1784
+ ("bul_Cyrl", "deu_Latn"),
1785
+ ("bul_Cyrl", "ell_Grek"),
1786
+ ("bul_Cyrl", "eng_Latn"),
1787
+ ("bul_Cyrl", "epo_Latn"),
1788
+ ("bul_Cyrl", "est_Latn"),
1789
+ ("bul_Cyrl", "fin_Latn"),
1790
+ ("bul_Cyrl", "fra_Latn"),
1791
+ ("bul_Cyrl", "heb_Hebr"),
1792
+ ("bul_Cyrl", "hin_Deva"),
1793
+ ("bul_Cyrl", "hrv_Latn"),
1794
+ ("bul_Cyrl", "hun_Latn"),
1795
+ ("bul_Cyrl", "ind_Latn"),
1796
+ ("bul_Cyrl", "ita_Latn"),
1797
+ ("bul_Cyrl", "jpn_Jpan"),
1798
+ ("bul_Cyrl", "kor_Hang"),
1799
+ ("bul_Cyrl", "lit_Latn"),
1800
+ ("bul_Cyrl", "lvs_Latn"),
1801
+ ("bul_Cyrl", "mkd_Cyrl"),
1802
+ ("bul_Cyrl", "nld_Latn"),
1803
+ ("bul_Cyrl", "nob_Latn"),
1804
+ ("bul_Cyrl", "pol_Latn"),
1805
+ ("bul_Cyrl", "por_Latn"),
1806
+ ("bul_Cyrl", "ron_Latn"),
1807
+ ("bul_Cyrl", "rus_Cyrl"),
1808
+ ("bul_Cyrl", "slk_Latn"),
1809
+ ("bul_Cyrl", "slv_Latn"),
1810
+ ("bul_Cyrl", "spa_Latn"),
1811
+ ("bul_Cyrl", "srp_Cyrl"),
1812
+ ("bul_Cyrl", "swe_Latn"),
1813
+ ("bul_Cyrl", "swh_Latn"),
1814
+ ("bul_Cyrl", "tam_Taml"),
1815
+ ("bul_Cyrl", "tur_Latn"),
1816
+ ("bul_Cyrl", "ukr_Cyrl"),
1817
+ ("bul_Cyrl", "vie_Latn"),
1818
+ ("bul_Cyrl", "zho_Hans"),
1819
+ ("cat_Latn", "als_Latn"),
1820
+ ("cat_Latn", "dan_Latn"),
1821
+ ("cat_Latn", "ell_Grek"),
1822
+ ("cat_Latn", "eng_Latn"),
1823
+ ("cat_Latn", "epo_Latn"),
1824
+ ("cat_Latn", "est_Latn"),
1825
+ ("cat_Latn", "fin_Latn"),
1826
+ ("cat_Latn", "fra_Latn"),
1827
+ ("cat_Latn", "glg_Latn"),
1828
+ ("cat_Latn", "heb_Hebr"),
1829
+ ("cat_Latn", "hin_Deva"),
1830
+ ("cat_Latn", "hun_Latn"),
1831
+ ("cat_Latn", "ind_Latn"),
1832
+ ("cat_Latn", "ita_Latn"),
1833
+ ("cat_Latn", "lit_Latn"),
1834
+ ("cat_Latn", "lvs_Latn"),
1835
+ ("cat_Latn", "nob_Latn"),
1836
+ ("cat_Latn", "oci_Latn"),
1837
+ ("cat_Latn", "pes_Arab"),
1838
+ ("cat_Latn", "por_Latn"),
1839
+ ("cat_Latn", "ron_Latn"),
1840
+ ("cat_Latn", "slk_Latn"),
1841
+ ("cat_Latn", "slv_Latn"),
1842
+ ("cat_Latn", "spa_Latn"),
1843
+ ("cat_Latn", "srp_Cyrl"),
1844
+ ("cat_Latn", "swh_Latn"),
1845
+ ("cat_Latn", "tam_Taml"),
1846
+ ("cat_Latn", "ukr_Cyrl"),
1847
+ ("ceb_Latn", "deu_Latn"),
1848
+ ("ceb_Latn", "hin_Deva"),
1849
+ ("ceb_Latn", "ind_Latn"),
1850
+ ("ceb_Latn", "jpn_Jpan"),
1851
+ ("ceb_Latn", "mal_Mlym"),
1852
+ ("ceb_Latn", "tgl_Latn"),
1853
+ ("ceb_Latn", "zsm_Latn"),
1854
+ ("ces_Latn", "als_Latn"),
1855
+ ("ces_Latn", "dan_Latn"),
1856
+ ("ces_Latn", "deu_Latn"),
1857
+ ("ces_Latn", "ell_Grek"),
1858
+ ("ces_Latn", "eng_Latn"),
1859
+ ("ces_Latn", "epo_Latn"),
1860
+ ("ces_Latn", "est_Latn"),
1861
+ ("ces_Latn", "fin_Latn"),
1862
+ ("ces_Latn", "fra_Latn"),
1863
+ ("ces_Latn", "heb_Hebr"),
1864
+ ("ces_Latn", "hin_Deva"),
1865
+ ("ces_Latn", "hrv_Latn"),
1866
+ ("ces_Latn", "hun_Latn"),
1867
+ ("ces_Latn", "ind_Latn"),
1868
+ ("ces_Latn", "ita_Latn"),
1869
+ ("ces_Latn", "jpn_Jpan"),
1870
+ ("ces_Latn", "kor_Hang"),
1871
+ ("ces_Latn", "lit_Latn"),
1872
+ ("ces_Latn", "lvs_Latn"),
1873
+ ("ces_Latn", "mkd_Cyrl"),
1874
+ ("ces_Latn", "nld_Latn"),
1875
+ ("ces_Latn", "nob_Latn"),
1876
+ ("ces_Latn", "pes_Arab"),
1877
+ ("ces_Latn", "pol_Latn"),
1878
+ ("ces_Latn", "por_Latn"),
1879
+ ("ces_Latn", "ron_Latn"),
1880
+ ("ces_Latn", "rus_Cyrl"),
1881
+ ("ces_Latn", "slk_Latn"),
1882
+ ("ces_Latn", "slv_Latn"),
1883
+ ("ces_Latn", "spa_Latn"),
1884
+ ("ces_Latn", "srp_Cyrl"),
1885
+ ("ces_Latn", "swe_Latn"),
1886
+ ("ces_Latn", "swh_Latn"),
1887
+ ("ces_Latn", "tam_Taml"),
1888
+ ("ces_Latn", "tur_Latn"),
1889
+ ("ces_Latn", "ukr_Cyrl"),
1890
+ ("ces_Latn", "vie_Latn"),
1891
+ ("ces_Latn", "zho_Hans"),
1892
+ ("cym_Latn", "nld_Latn"),
1893
+ ("dan_Latn", "als_Latn"),
1894
+ ("dan_Latn", "deu_Latn"),
1895
+ ("dan_Latn", "ell_Grek"),
1896
+ ("dan_Latn", "eng_Latn"),
1897
+ ("dan_Latn", "epo_Latn"),
1898
+ ("dan_Latn", "est_Latn"),
1899
+ ("dan_Latn", "fin_Latn"),
1900
+ ("dan_Latn", "fra_Latn"),
1901
+ ("dan_Latn", "gla_Latn"),
1902
+ ("dan_Latn", "heb_Hebr"),
1903
+ ("dan_Latn", "hrv_Latn"),
1904
+ ("dan_Latn", "hun_Latn"),
1905
+ ("dan_Latn", "ind_Latn"),
1906
+ ("dan_Latn", "isl_Latn"),
1907
+ ("dan_Latn", "ita_Latn"),
1908
+ ("dan_Latn", "jpn_Jpan"),
1909
+ ("dan_Latn", "kor_Hang"),
1910
+ ("dan_Latn", "lit_Latn"),
1911
+ ("dan_Latn", "ltz_Latn"),
1912
+ ("dan_Latn", "lvs_Latn"),
1913
+ ("dan_Latn", "mkd_Cyrl"),
1914
+ ("dan_Latn", "nld_Latn"),
1915
+ ("dan_Latn", "nob_Latn"),
1916
+ ("dan_Latn", "pes_Arab"),
1917
+ ("dan_Latn", "pol_Latn"),
1918
+ ("dan_Latn", "por_Latn"),
1919
+ ("dan_Latn", "ron_Latn"),
1920
+ ("dan_Latn", "rus_Cyrl"),
1921
+ ("dan_Latn", "slk_Latn"),
1922
+ ("dan_Latn", "slv_Latn"),
1923
+ ("dan_Latn", "spa_Latn"),
1924
+ ("dan_Latn", "srp_Cyrl"),
1925
+ ("dan_Latn", "swe_Latn"),
1926
+ ("dan_Latn", "tam_Taml"),
1927
+ ("dan_Latn", "tur_Latn"),
1928
+ ("dan_Latn", "ukr_Cyrl"),
1929
+ ("dan_Latn", "vie_Latn"),
1930
+ ("dan_Latn", "ydd_Hebr"),
1931
+ ("dan_Latn", "zho_Hans"),
1932
+ ("deu_Latn", "als_Latn"),
1933
+ ("deu_Latn", "ell_Grek"),
1934
+ ("deu_Latn", "eng_Latn"),
1935
+ ("deu_Latn", "epo_Latn"),
1936
+ ("deu_Latn", "est_Latn"),
1937
+ ("deu_Latn", "fin_Latn"),
1938
+ ("deu_Latn", "fra_Latn"),
1939
+ ("deu_Latn", "hau_Latn"),
1940
+ ("deu_Latn", "heb_Hebr"),
1941
+ ("deu_Latn", "hin_Deva"),
1942
+ ("deu_Latn", "hrv_Latn"),
1943
+ ("deu_Latn", "hun_Latn"),
1944
+ ("deu_Latn", "ilo_Latn"),
1945
+ ("deu_Latn", "ind_Latn"),
1946
+ ("deu_Latn", "isl_Latn"),
1947
+ ("deu_Latn", "ita_Latn"),
1948
+ ("deu_Latn", "jav_Latn"),
1949
+ ("deu_Latn", "jpn_Jpan"),
1950
+ ("deu_Latn", "kor_Hang"),
1951
+ ("deu_Latn", "lit_Latn"),
1952
+ ("deu_Latn", "ltz_Latn"),
1953
+ ("deu_Latn", "lvs_Latn"),
1954
+ ("deu_Latn", "mal_Mlym"),
1955
+ ("deu_Latn", "mkd_Cyrl"),
1956
+ ("deu_Latn", "nld_Latn"),
1957
+ ("deu_Latn", "nob_Latn"),
1958
+ ("deu_Latn", "pes_Arab"),
1959
+ ("deu_Latn", "plt_Latn"),
1960
+ ("deu_Latn", "pol_Latn"),
1961
+ ("deu_Latn", "por_Latn"),
1962
+ ("deu_Latn", "ron_Latn"),
1963
+ ("deu_Latn", "rus_Cyrl"),
1964
+ ("deu_Latn", "slk_Latn"),
1965
+ ("deu_Latn", "slv_Latn"),
1966
+ ("deu_Latn", "spa_Latn"),
1967
+ ("deu_Latn", "srp_Cyrl"),
1968
+ ("deu_Latn", "sun_Latn"),
1969
+ ("deu_Latn", "swe_Latn"),
1970
+ ("deu_Latn", "swh_Latn"),
1971
+ ("deu_Latn", "tam_Taml"),
1972
+ ("deu_Latn", "tgl_Latn"),
1973
+ ("deu_Latn", "tur_Latn"),
1974
+ ("deu_Latn", "ukr_Cyrl"),
1975
+ ("deu_Latn", "vie_Latn"),
1976
+ ("deu_Latn", "ydd_Hebr"),
1977
+ ("deu_Latn", "yor_Latn"),
1978
+ ("deu_Latn", "zho_Hans"),
1979
+ ("deu_Latn", "zsm_Latn"),
1980
+ ("ell_Grek", "als_Latn"),
1981
+ ("ell_Grek", "eng_Latn"),
1982
+ ("ell_Grek", "epo_Latn"),
1983
+ ("ell_Grek", "est_Latn"),
1984
+ ("ell_Grek", "fin_Latn"),
1985
+ ("ell_Grek", "fra_Latn"),
1986
+ ("ell_Grek", "heb_Hebr"),
1987
+ ("ell_Grek", "hin_Deva"),
1988
+ ("ell_Grek", "hun_Latn"),
1989
+ ("ell_Grek", "ind_Latn"),
1990
+ ("ell_Grek", "ita_Latn"),
1991
+ ("ell_Grek", "jpn_Jpan"),
1992
+ ("ell_Grek", "kor_Hang"),
1993
+ ("ell_Grek", "lit_Latn"),
1994
+ ("ell_Grek", "lvs_Latn"),
1995
+ ("ell_Grek", "nld_Latn"),
1996
+ ("ell_Grek", "nob_Latn"),
1997
+ ("ell_Grek", "pes_Arab"),
1998
+ ("ell_Grek", "pol_Latn"),
1999
+ ("ell_Grek", "por_Latn"),
2000
+ ("ell_Grek", "ron_Latn"),
2001
+ ("ell_Grek", "rus_Cyrl"),
2002
+ ("ell_Grek", "slk_Latn"),
2003
+ ("ell_Grek", "slv_Latn"),
2004
+ ("ell_Grek", "spa_Latn"),
2005
+ ("ell_Grek", "srp_Cyrl"),
2006
+ ("ell_Grek", "swe_Latn"),
2007
+ ("ell_Grek", "swh_Latn"),
2008
+ ("ell_Grek", "tam_Taml"),
2009
+ ("ell_Grek", "tur_Latn"),
2010
+ ("ell_Grek", "ukr_Cyrl"),
2011
+ ("ell_Grek", "vie_Latn"),
2012
+ ("ell_Grek", "zho_Hans"),
2013
+ ("eng_Latn", "est_Latn"),
2014
+ ("eng_Latn", "fin_Latn"),
2015
+ ("eng_Latn", "fra_Latn"),
2016
+ ("eng_Latn", "glg_Latn"),
2017
+ ("eng_Latn", "heb_Hebr"),
2018
+ ("eng_Latn", "hrv_Latn"),
2019
+ ("eng_Latn", "hun_Latn"),
2020
+ ("eng_Latn", "ind_Latn"),
2021
+ ("eng_Latn", "isl_Latn"),
2022
+ ("eng_Latn", "ita_Latn"),
2023
+ ("eng_Latn", "jpn_Jpan"),
2024
+ ("eng_Latn", "kor_Hang"),
2025
+ ("eng_Latn", "lit_Latn"),
2026
+ ("eng_Latn", "lvs_Latn"),
2027
+ ("eng_Latn", "mkd_Cyrl"),
2028
+ ("eng_Latn", "nld_Latn"),
2029
+ ("eng_Latn", "nob_Latn"),
2030
+ ("eng_Latn", "oci_Latn"),
2031
+ ("eng_Latn", "pes_Arab"),
2032
+ ("eng_Latn", "pol_Latn"),
2033
+ ("eng_Latn", "por_Latn"),
2034
+ ("eng_Latn", "ron_Latn"),
2035
+ ("eng_Latn", "rus_Cyrl"),
2036
+ ("eng_Latn", "slk_Latn"),
2037
+ ("eng_Latn", "slv_Latn"),
2038
+ ("eng_Latn", "spa_Latn"),
2039
+ ("eng_Latn", "srp_Cyrl"),
2040
+ ("eng_Latn", "swe_Latn"),
2041
+ ("eng_Latn", "tur_Latn"),
2042
+ ("eng_Latn", "ukr_Cyrl"),
2043
+ ("eng_Latn", "vie_Latn"),
2044
+ ("eng_Latn", "zho_Hans"),
2045
+ ("epo_Latn", "als_Latn"),
2046
+ ("epo_Latn", "est_Latn"),
2047
+ ("epo_Latn", "fin_Latn"),
2048
+ ("epo_Latn", "heb_Hebr"),
2049
+ ("epo_Latn", "hrv_Latn"),
2050
+ ("epo_Latn", "hun_Latn"),
2051
+ ("epo_Latn", "ita_Latn"),
2052
+ ("epo_Latn", "lit_Latn"),
2053
+ ("epo_Latn", "lvs_Latn"),
2054
+ ("epo_Latn", "mkd_Cyrl"),
2055
+ ("epo_Latn", "nld_Latn"),
2056
+ ("epo_Latn", "nob_Latn"),
2057
+ ("epo_Latn", "pes_Arab"),
2058
+ ("epo_Latn", "ron_Latn"),
2059
+ ("epo_Latn", "slk_Latn"),
2060
+ ("epo_Latn", "slv_Latn"),
2061
+ ("epo_Latn", "spa_Latn"),
2062
+ ("epo_Latn", "srp_Cyrl"),
2063
+ ("epo_Latn", "swe_Latn"),
2064
+ ("epo_Latn", "swh_Latn"),
2065
+ ("epo_Latn", "ukr_Cyrl"),
2066
+ ("est_Latn", "als_Latn"),
2067
+ ("est_Latn", "fin_Latn"),
2068
+ ("est_Latn", "fra_Latn"),
2069
+ ("est_Latn", "heb_Hebr"),
2070
+ ("est_Latn", "hin_Deva"),
2071
+ ("est_Latn", "hrv_Latn"),
2072
+ ("est_Latn", "hun_Latn"),
2073
+ ("est_Latn", "ind_Latn"),
2074
+ ("est_Latn", "ita_Latn"),
2075
+ ("est_Latn", "kor_Hang"),
2076
+ ("est_Latn", "lit_Latn"),
2077
+ ("est_Latn", "lvs_Latn"),
2078
+ ("est_Latn", "mkd_Cyrl"),
2079
+ ("est_Latn", "nld_Latn"),
2080
+ ("est_Latn", "nob_Latn"),
2081
+ ("est_Latn", "pes_Arab"),
2082
+ ("est_Latn", "pol_Latn"),
2083
+ ("est_Latn", "por_Latn"),
2084
+ ("est_Latn", "ron_Latn"),
2085
+ ("est_Latn", "rus_Cyrl"),
2086
+ ("est_Latn", "slk_Latn"),
2087
+ ("est_Latn", "slv_Latn"),
2088
+ ("est_Latn", "srp_Cyrl"),
2089
+ ("est_Latn", "swe_Latn"),
2090
+ ("est_Latn", "swh_Latn"),
2091
+ ("est_Latn", "tam_Taml"),
2092
+ ("est_Latn", "tur_Latn"),
2093
+ ("est_Latn", "ukr_Cyrl"),
2094
+ ("est_Latn", "vie_Latn"),
2095
+ ("est_Latn", "zho_Hans"),
2096
+ ("fin_Latn", "als_Latn"),
2097
+ ("fin_Latn", "fra_Latn"),
2098
+ ("fin_Latn", "heb_Hebr"),
2099
+ ("fin_Latn", "hin_Deva"),
2100
+ ("fin_Latn", "hrv_Latn"),
2101
+ ("fin_Latn", "hun_Latn"),
2102
+ ("fin_Latn", "ind_Latn"),
2103
+ ("fin_Latn", "ita_Latn"),
2104
+ ("fin_Latn", "jpn_Jpan"),
2105
+ ("fin_Latn", "kor_Hang"),
2106
+ ("fin_Latn", "lit_Latn"),
2107
+ ("fin_Latn", "lvs_Latn"),
2108
+ ("fin_Latn", "mkd_Cyrl"),
2109
+ ("fin_Latn", "nld_Latn"),
2110
+ ("fin_Latn", "nob_Latn"),
2111
+ ("fin_Latn", "pol_Latn"),
2112
+ ("fin_Latn", "por_Latn"),
2113
+ ("fin_Latn", "ron_Latn"),
2114
+ ("fin_Latn", "rus_Cyrl"),
2115
+ ("fin_Latn", "slk_Latn"),
2116
+ ("fin_Latn", "slv_Latn"),
2117
+ ("fin_Latn", "srp_Cyrl"),
2118
+ ("fin_Latn", "swe_Latn"),
2119
+ ("fin_Latn", "swh_Latn"),
2120
+ ("fin_Latn", "tam_Taml"),
2121
+ ("fin_Latn", "tur_Latn"),
2122
+ ("fin_Latn", "ukr_Cyrl"),
2123
+ ("fin_Latn", "vie_Latn"),
2124
+ ("fin_Latn", "zho_Hans"),
2125
+ ("fra_Latn", "heb_Hebr"),
2126
+ ("fra_Latn", "hin_Deva"),
2127
+ ("fra_Latn", "hrv_Latn"),
2128
+ ("fra_Latn", "hun_Latn"),
2129
+ ("fra_Latn", "ind_Latn"),
2130
+ ("fra_Latn", "ita_Latn"),
2131
+ ("fra_Latn", "jpn_Jpan"),
2132
+ ("fra_Latn", "kor_Hang"),
2133
+ ("fra_Latn", "lit_Latn"),
2134
+ ("fra_Latn", "lvs_Latn"),
2135
+ ("fra_Latn", "mal_Mlym"),
2136
+ ("fra_Latn", "nld_Latn"),
2137
+ ("fra_Latn", "nob_Latn"),
2138
+ ("fra_Latn", "pol_Latn"),
2139
+ ("fra_Latn", "por_Latn"),
2140
+ ("fra_Latn", "ron_Latn"),
2141
+ ("fra_Latn", "rus_Cyrl"),
2142
+ ("fra_Latn", "slk_Latn"),
2143
+ ("fra_Latn", "slv_Latn"),
2144
+ ("fra_Latn", "srp_Cyrl"),
2145
+ ("fra_Latn", "swe_Latn"),
2146
+ ("fra_Latn", "tam_Taml"),
2147
+ ("fra_Latn", "tgl_Latn"),
2148
+ ("fra_Latn", "tur_Latn"),
2149
+ ("fra_Latn", "ukr_Cyrl"),
2150
+ ("fra_Latn", "vie_Latn"),
2151
+ ("fra_Latn", "zho_Hans"),
2152
+ ("fra_Latn", "zsm_Latn"),
2153
+ ("gla_Latn", "nld_Latn"),
2154
+ ("gla_Latn", "nob_Latn"),
2155
+ ("gle_Latn", "nld_Latn"),
2156
+ ("glg_Latn", "ita_Latn"),
2157
+ ("glg_Latn", "oci_Latn"),
2158
+ ("glg_Latn", "ron_Latn"),
2159
+ ("glg_Latn", "tur_Latn"),
2160
+ ("hau_Latn", "nld_Latn"),
2161
+ ("heb_Hebr", "als_Latn"),
2162
+ ("heb_Hebr", "hin_Deva"),
2163
+ ("heb_Hebr", "hun_Latn"),
2164
+ ("heb_Hebr", "ind_Latn"),
2165
+ ("heb_Hebr", "jpn_Jpan"),
2166
+ ("heb_Hebr", "kor_Hang"),
2167
+ ("heb_Hebr", "lit_Latn"),
2168
+ ("heb_Hebr", "lvs_Latn"),
2169
+ ("heb_Hebr", "mal_Mlym"),
2170
+ ("heb_Hebr", "mar_Deva"),
2171
+ ("heb_Hebr", "nld_Latn"),
2172
+ ("heb_Hebr", "nob_Latn"),
2173
+ ("heb_Hebr", "npi_Deva"),
2174
+ ("heb_Hebr", "plt_Latn"),
2175
+ ("heb_Hebr", "pol_Latn"),
2176
+ ("heb_Hebr", "por_Latn"),
2177
+ ("heb_Hebr", "ron_Latn"),
2178
+ ("heb_Hebr", "rus_Cyrl"),
2179
+ ("heb_Hebr", "sin_Sinh"),
2180
+ ("heb_Hebr", "slk_Latn"),
2181
+ ("heb_Hebr", "slv_Latn"),
2182
+ ("heb_Hebr", "srp_Cyrl"),
2183
+ ("heb_Hebr", "swe_Latn"),
2184
+ ("heb_Hebr", "swh_Latn"),
2185
+ ("heb_Hebr", "tam_Taml"),
2186
+ ("heb_Hebr", "tgl_Latn"),
2187
+ ("heb_Hebr", "tur_Latn"),
2188
+ ("heb_Hebr", "ukr_Cyrl"),
2189
+ ("heb_Hebr", "urd_Arab"),
2190
+ ("heb_Hebr", "vie_Latn"),
2191
+ ("heb_Hebr", "zho_Hans"),
2192
+ ("heb_Hebr", "zsm_Latn"),
2193
+ ("hin_Deva", "hun_Latn"),
2194
+ ("hin_Deva", "ilo_Latn"),
2195
+ ("hin_Deva", "ind_Latn"),
2196
+ ("hin_Deva", "ita_Latn"),
2197
+ ("hin_Deva", "jav_Latn"),
2198
+ ("hin_Deva", "jpn_Jpan"),
2199
+ ("hin_Deva", "kor_Hang"),
2200
+ ("hin_Deva", "lit_Latn"),
2201
+ ("hin_Deva", "lvs_Latn"),
2202
+ ("hin_Deva", "nld_Latn"),
2203
+ ("hin_Deva", "plt_Latn"),
2204
+ ("hin_Deva", "pol_Latn"),
2205
+ ("hin_Deva", "por_Latn"),
2206
+ ("hin_Deva", "ron_Latn"),
2207
+ ("hin_Deva", "rus_Cyrl"),
2208
+ ("hin_Deva", "slk_Latn"),
2209
+ ("hin_Deva", "slv_Latn"),
2210
+ ("hin_Deva", "srp_Cyrl"),
2211
+ ("hin_Deva", "swe_Latn"),
2212
+ ("hin_Deva", "swh_Latn"),
2213
+ ("hin_Deva", "tgl_Latn"),
2214
+ ("hin_Deva", "tur_Latn"),
2215
+ ("hin_Deva", "ukr_Cyrl"),
2216
+ ("hin_Deva", "vie_Latn"),
2217
+ ("hin_Deva", "zho_Hans"),
2218
+ ("hin_Deva", "zsm_Latn"),
2219
+ ("hrv_Latn", "als_Latn"),
2220
+ ("hrv_Latn", "hun_Latn"),
2221
+ ("hrv_Latn", "lit_Latn"),
2222
+ ("hrv_Latn", "lvs_Latn"),
2223
+ ("hrv_Latn", "mkd_Cyrl"),
2224
+ ("hrv_Latn", "nob_Latn"),
2225
+ ("hrv_Latn", "pol_Latn"),
2226
+ ("hrv_Latn", "ron_Latn"),
2227
+ ("hrv_Latn", "rus_Cyrl"),
2228
+ ("hrv_Latn", "slk_Latn"),
2229
+ ("hrv_Latn", "slv_Latn"),
2230
+ ("hrv_Latn", "srp_Cyrl"),
2231
+ ("hrv_Latn", "tam_Taml"),
2232
+ ("hrv_Latn", "ukr_Cyrl"),
2233
+ ("hrv_Latn", "zho_Hans"),
2234
+ ("hun_Latn", "als_Latn"),
2235
+ ("hun_Latn", "gaz_Latn"),
2236
+ ("hun_Latn", "ind_Latn"),
2237
+ ("hun_Latn", "ita_Latn"),
2238
+ ("hun_Latn", "jpn_Jpan"),
2239
+ ("hun_Latn", "kor_Hang"),
2240
+ ("hun_Latn", "lit_Latn"),
2241
+ ("hun_Latn", "lvs_Latn"),
2242
+ ("hun_Latn", "mkd_Cyrl"),
2243
+ ("hun_Latn", "nld_Latn"),
2244
+ ("hun_Latn", "nob_Latn"),
2245
+ ("hun_Latn", "pol_Latn"),
2246
+ ("hun_Latn", "por_Latn"),
2247
+ ("hun_Latn", "ron_Latn"),
2248
+ ("hun_Latn", "rus_Cyrl"),
2249
+ ("hun_Latn", "slk_Latn"),
2250
+ ("hun_Latn", "slv_Latn"),
2251
+ ("hun_Latn", "srp_Cyrl"),
2252
+ ("hun_Latn", "swe_Latn"),
2253
+ ("hun_Latn", "swh_Latn"),
2254
+ ("hun_Latn", "tam_Taml"),
2255
+ ("hun_Latn", "tur_Latn"),
2256
+ ("hun_Latn", "ukr_Cyrl"),
2257
+ ("hun_Latn", "vie_Latn"),
2258
+ ("hun_Latn", "xho_Latn"),
2259
+ ("hun_Latn", "zho_Hans"),
2260
+ ("hun_Latn", "zul_Latn"),
2261
+ ("hye_Armn", "kat_Geor"),
2262
+ ("hye_Armn", "kaz_Cyrl"),
2263
+ ("hye_Armn", "khm_Khmr"),
2264
+ ("hye_Armn", "mya_Mymr"),
2265
+ ("hye_Armn", "tat_Cyrl"),
2266
+ ("hye_Armn", "tur_Latn"),
2267
+ ("hye_Armn", "vie_Latn"),
2268
+ ("ibo_Latn", "nld_Latn"),
2269
+ ("ilo_Latn", "mal_Mlym"),
2270
+ ("ilo_Latn", "tgl_Latn"),
2271
+ ("ilo_Latn", "zsm_Latn"),
2272
+ ("ind_Latn", "ilo_Latn"),
2273
+ ("ind_Latn", "ita_Latn"),
2274
+ ("ind_Latn", "jpn_Jpan"),
2275
+ ("ind_Latn", "kor_Hang"),
2276
+ ("ind_Latn", "lit_Latn"),
2277
+ ("ind_Latn", "lvs_Latn"),
2278
+ ("ind_Latn", "mal_Mlym"),
2279
+ ("ind_Latn", "mar_Deva"),
2280
+ ("ind_Latn", "nld_Latn"),
2281
+ ("ind_Latn", "nob_Latn"),
2282
+ ("ind_Latn", "npi_Deva"),
2283
+ ("ind_Latn", "plt_Latn"),
2284
+ ("ind_Latn", "pol_Latn"),
2285
+ ("ind_Latn", "por_Latn"),
2286
+ ("ind_Latn", "ron_Latn"),
2287
+ ("ind_Latn", "rus_Cyrl"),
2288
+ ("ind_Latn", "sin_Sinh"),
2289
+ ("ind_Latn", "slk_Latn"),
2290
+ ("ind_Latn", "slv_Latn"),
2291
+ ("ind_Latn", "srp_Cyrl"),
2292
+ ("ind_Latn", "swe_Latn"),
2293
+ ("ind_Latn", "swh_Latn"),
2294
+ ("ind_Latn", "tam_Taml"),
2295
+ ("ind_Latn", "tgl_Latn"),
2296
+ ("ind_Latn", "tur_Latn"),
2297
+ ("ind_Latn", "ukr_Cyrl"),
2298
+ ("ind_Latn", "urd_Arab"),
2299
+ ("ind_Latn", "vie_Latn"),
2300
+ ("ind_Latn", "zho_Hans"),
2301
+ ("ind_Latn", "zsm_Latn"),
2302
+ ("isl_Latn", "ltz_Latn"),
2303
+ ("isl_Latn", "nld_Latn"),
2304
+ ("isl_Latn", "nob_Latn"),
2305
+ ("isl_Latn", "swe_Latn"),
2306
+ ("isl_Latn", "ydd_Hebr"),
2307
+ ("ita_Latn", "als_Latn"),
2308
+ ("ita_Latn", "jpn_Jpan"),
2309
+ ("ita_Latn", "kor_Hang"),
2310
+ ("ita_Latn", "lit_Latn"),
2311
+ ("ita_Latn", "lvs_Latn"),
2312
+ ("ita_Latn", "nld_Latn"),
2313
+ ("ita_Latn", "nob_Latn"),
2314
+ ("ita_Latn", "oci_Latn"),
2315
+ ("ita_Latn", "pol_Latn"),
2316
+ ("ita_Latn", "por_Latn"),
2317
+ ("ita_Latn", "ron_Latn"),
2318
+ ("ita_Latn", "rus_Cyrl"),
2319
+ ("ita_Latn", "slk_Latn"),
2320
+ ("ita_Latn", "slv_Latn"),
2321
+ ("ita_Latn", "srp_Cyrl"),
2322
+ ("ita_Latn", "swe_Latn"),
2323
+ ("ita_Latn", "swh_Latn"),
2324
+ ("ita_Latn", "tam_Taml"),
2325
+ ("ita_Latn", "tur_Latn"),
2326
+ ("ita_Latn", "ukr_Cyrl"),
2327
+ ("ita_Latn", "vie_Latn"),
2328
+ ("ita_Latn", "zho_Hans"),
2329
+ ("jav_Latn", "mal_Mlym"),
2330
+ ("jav_Latn", "tgl_Latn"),
2331
+ ("jav_Latn", "zsm_Latn"),
2332
+ ("jpn_Jpan", "kor_Hang"),
2333
+ ("jpn_Jpan", "lit_Latn"),
2334
+ ("jpn_Jpan", "mal_Mlym"),
2335
+ ("jpn_Jpan", "nld_Latn"),
2336
+ ("jpn_Jpan", "nob_Latn"),
2337
+ ("jpn_Jpan", "pol_Latn"),
2338
+ ("jpn_Jpan", "por_Latn"),
2339
+ ("jpn_Jpan", "ron_Latn"),
2340
+ ("jpn_Jpan", "rus_Cyrl"),
2341
+ ("jpn_Jpan", "slk_Latn"),
2342
+ ("jpn_Jpan", "slv_Latn"),
2343
+ ("jpn_Jpan", "srp_Cyrl"),
2344
+ ("jpn_Jpan", "swe_Latn"),
2345
+ ("jpn_Jpan", "swh_Latn"),
2346
+ ("jpn_Jpan", "tam_Taml"),
2347
+ ("jpn_Jpan", "tgl_Latn"),
2348
+ ("jpn_Jpan", "tur_Latn"),
2349
+ ("jpn_Jpan", "ukr_Cyrl"),
2350
+ ("jpn_Jpan", "vie_Latn"),
2351
+ ("jpn_Jpan", "zho_Hans"),
2352
+ ("jpn_Jpan", "zsm_Latn"),
2353
+ ("kat_Geor", "kaz_Cyrl"),
2354
+ ("kat_Geor", "tat_Cyrl"),
2355
+ ("kat_Geor", "tur_Latn"),
2356
+ ("kat_Geor", "vie_Latn"),
2357
+ ("kaz_Cyrl", "tat_Cyrl"),
2358
+ ("kaz_Cyrl", "tur_Latn"),
2359
+ ("kaz_Cyrl", "vie_Latn"),
2360
+ ("khm_Khmr", "tur_Latn"),
2361
+ ("kor_Hang", "lit_Latn"),
2362
+ ("kor_Hang", "lvs_Latn"),
2363
+ ("kor_Hang", "mkd_Cyrl"),
2364
+ ("kor_Hang", "nld_Latn"),
2365
+ ("kor_Hang", "nob_Latn"),
2366
+ ("kor_Hang", "pol_Latn"),
2367
+ ("kor_Hang", "por_Latn"),
2368
+ ("kor_Hang", "rus_Cyrl"),
2369
+ ("kor_Hang", "slk_Latn"),
2370
+ ("kor_Hang", "slv_Latn"),
2371
+ ("kor_Hang", "srp_Cyrl"),
2372
+ ("kor_Hang", "swe_Latn"),
2373
+ ("kor_Hang", "swh_Latn"),
2374
+ ("kor_Hang", "tam_Taml"),
2375
+ ("kor_Hang", "tur_Latn"),
2376
+ ("kor_Hang", "ukr_Cyrl"),
2377
+ ("kor_Hang", "vie_Latn"),
2378
+ ("kor_Hang", "zho_Hans"),
2379
+ ("lit_Latn", "als_Latn"),
2380
+ ("lit_Latn", "lvs_Latn"),
2381
+ ("lit_Latn", "mkd_Cyrl"),
2382
+ ("lit_Latn", "nld_Latn"),
2383
+ ("lit_Latn", "nob_Latn"),
2384
+ ("lit_Latn", "plt_Latn"),
2385
+ ("lit_Latn", "pol_Latn"),
2386
+ ("lit_Latn", "por_Latn"),
2387
+ ("lit_Latn", "ron_Latn"),
2388
+ ("lit_Latn", "rus_Cyrl"),
2389
+ ("lit_Latn", "slk_Latn"),
2390
+ ("lit_Latn", "slv_Latn"),
2391
+ ("lit_Latn", "srp_Cyrl"),
2392
+ ("lit_Latn", "swe_Latn"),
2393
+ ("lit_Latn", "swh_Latn"),
2394
+ ("lit_Latn", "tam_Taml"),
2395
+ ("lit_Latn", "tur_Latn"),
2396
+ ("lit_Latn", "ukr_Cyrl"),
2397
+ ("lit_Latn", "vie_Latn"),
2398
+ ("lit_Latn", "zho_Hans"),
2399
+ ("ltz_Latn", "nld_Latn"),
2400
+ ("ltz_Latn", "nob_Latn"),
2401
+ ("ltz_Latn", "zho_Hans"),
2402
+ ("lvs_Latn", "als_Latn"),
2403
+ ("lvs_Latn", "mkd_Cyrl"),
2404
+ ("lvs_Latn", "nld_Latn"),
2405
+ ("lvs_Latn", "nob_Latn"),
2406
+ ("lvs_Latn", "pol_Latn"),
2407
+ ("lvs_Latn", "por_Latn"),
2408
+ ("lvs_Latn", "ron_Latn"),
2409
+ ("lvs_Latn", "rus_Cyrl"),
2410
+ ("lvs_Latn", "slk_Latn"),
2411
+ ("lvs_Latn", "slv_Latn"),
2412
+ ("lvs_Latn", "srp_Cyrl"),
2413
+ ("lvs_Latn", "swe_Latn"),
2414
+ ("lvs_Latn", "swh_Latn"),
2415
+ ("lvs_Latn", "tam_Taml"),
2416
+ ("lvs_Latn", "tur_Latn"),
2417
+ ("lvs_Latn", "ukr_Cyrl"),
2418
+ ("lvs_Latn", "vie_Latn"),
2419
+ ("lvs_Latn", "zho_Hans"),
2420
+ ("mal_Mlym", "rus_Cyrl"),
2421
+ ("mal_Mlym", "sun_Latn"),
2422
+ ("mal_Mlym", "tgl_Latn"),
2423
+ ("mal_Mlym", "tur_Latn"),
2424
+ ("mal_Mlym", "zho_Hans"),
2425
+ ("mal_Mlym", "zsm_Latn"),
2426
+ ("mar_Deva", "rus_Cyrl"),
2427
+ ("mar_Deva", "tur_Latn"),
2428
+ ("mkd_Cyrl", "als_Latn"),
2429
+ ("mkd_Cyrl", "nld_Latn"),
2430
+ ("mkd_Cyrl", "nob_Latn"),
2431
+ ("mkd_Cyrl", "pol_Latn"),
2432
+ ("mkd_Cyrl", "ron_Latn"),
2433
+ ("mkd_Cyrl", "rus_Cyrl"),
2434
+ ("mkd_Cyrl", "slk_Latn"),
2435
+ ("mkd_Cyrl", "slv_Latn"),
2436
+ ("mkd_Cyrl", "swe_Latn"),
2437
+ ("mkd_Cyrl", "ukr_Cyrl"),
2438
+ ("mkd_Cyrl", "vie_Latn"),
2439
+ ("mkd_Cyrl", "zho_Hans"),
2440
+ ("mya_Mymr", "tur_Latn"),
2441
+ ("nld_Latn", "nob_Latn"),
2442
+ ("nld_Latn", "pol_Latn"),
2443
+ ("nld_Latn", "por_Latn"),
2444
+ ("nld_Latn", "ron_Latn"),
2445
+ ("nld_Latn", "rus_Cyrl"),
2446
+ ("nld_Latn", "slk_Latn"),
2447
+ ("nld_Latn", "slv_Latn"),
2448
+ ("nld_Latn", "som_Latn"),
2449
+ ("nld_Latn", "srp_Cyrl"),
2450
+ ("nld_Latn", "swe_Latn"),
2451
+ ("nld_Latn", "swh_Latn"),
2452
+ ("nld_Latn", "tam_Taml"),
2453
+ ("nld_Latn", "tur_Latn"),
2454
+ ("nld_Latn", "ukr_Cyrl"),
2455
+ ("nld_Latn", "vie_Latn"),
2456
+ ("nld_Latn", "ydd_Hebr"),
2457
+ ("nld_Latn", "yor_Latn"),
2458
+ ("nld_Latn", "zho_Hans"),
2459
+ ("nob_Latn", "als_Latn"),
2460
+ ("nob_Latn", "pol_Latn"),
2461
+ ("nob_Latn", "por_Latn"),
2462
+ ("nob_Latn", "ron_Latn"),
2463
+ ("nob_Latn", "rus_Cyrl"),
2464
+ ("nob_Latn", "slk_Latn"),
2465
+ ("nob_Latn", "slv_Latn"),
2466
+ ("nob_Latn", "srp_Cyrl"),
2467
+ ("nob_Latn", "swe_Latn"),
2468
+ ("nob_Latn", "tam_Taml"),
2469
+ ("nob_Latn", "tur_Latn"),
2470
+ ("nob_Latn", "ukr_Cyrl"),
2471
+ ("nob_Latn", "vie_Latn"),
2472
+ ("nob_Latn", "ydd_Hebr"),
2473
+ ("nob_Latn", "zho_Hans"),
2474
+ ("npi_Deva", "rus_Cyrl"),
2475
+ ("npi_Deva", "tur_Latn"),
2476
+ ("oci_Latn", "ron_Latn"),
2477
+ ("pes_Arab", "als_Latn"),
2478
+ ("pes_Arab", "fin_Latn"),
2479
+ ("pes_Arab", "fra_Latn"),
2480
+ ("pes_Arab", "heb_Hebr"),
2481
+ ("pes_Arab", "hin_Deva"),
2482
+ ("pes_Arab", "hrv_Latn"),
2483
+ ("pes_Arab", "hun_Latn"),
2484
+ ("pes_Arab", "ind_Latn"),
2485
+ ("pes_Arab", "ita_Latn"),
2486
+ ("pes_Arab", "jpn_Jpan"),
2487
+ ("pes_Arab", "kor_Hang"),
2488
+ ("pes_Arab", "lit_Latn"),
2489
+ ("pes_Arab", "lvs_Latn"),
2490
+ ("pes_Arab", "mal_Mlym"),
2491
+ ("pes_Arab", "mar_Deva"),
2492
+ ("pes_Arab", "nld_Latn"),
2493
+ ("pes_Arab", "nob_Latn"),
2494
+ ("pes_Arab", "npi_Deva"),
2495
+ ("pes_Arab", "ory_Orya"),
2496
+ ("pes_Arab", "pol_Latn"),
2497
+ ("pes_Arab", "por_Latn"),
2498
+ ("pes_Arab", "ron_Latn"),
2499
+ ("pes_Arab", "rus_Cyrl"),
2500
+ ("pes_Arab", "sin_Sinh"),
2501
+ ("pes_Arab", "slk_Latn"),
2502
+ ("pes_Arab", "slv_Latn"),
2503
+ ("pes_Arab", "snd_Arab"),
2504
+ ("pes_Arab", "srp_Cyrl"),
2505
+ ("pes_Arab", "swe_Latn"),
2506
+ ("pes_Arab", "swh_Latn"),
2507
+ ("pes_Arab", "tam_Taml"),
2508
+ ("pes_Arab", "tgl_Latn"),
2509
+ ("pes_Arab", "tur_Latn"),
2510
+ ("pes_Arab", "ukr_Cyrl"),
2511
+ ("pes_Arab", "urd_Arab"),
2512
+ ("pes_Arab", "uzn_Latn"),
2513
+ ("pes_Arab", "vie_Latn"),
2514
+ ("pes_Arab", "zho_Hans"),
2515
+ ("pes_Arab", "zsm_Latn"),
2516
+ ("plt_Latn", "mal_Mlym"),
2517
+ ("plt_Latn", "rus_Cyrl"),
2518
+ ("plt_Latn", "tam_Taml"),
2519
+ ("plt_Latn", "tgl_Latn"),
2520
+ ("plt_Latn", "zho_Hans"),
2521
+ ("plt_Latn", "zsm_Latn"),
2522
+ ("pol_Latn", "als_Latn"),
2523
+ ("pol_Latn", "por_Latn"),
2524
+ ("pol_Latn", "ron_Latn"),
2525
+ ("pol_Latn", "rus_Cyrl"),
2526
+ ("pol_Latn", "slk_Latn"),
2527
+ ("pol_Latn", "slv_Latn"),
2528
+ ("pol_Latn", "srp_Cyrl"),
2529
+ ("pol_Latn", "swe_Latn"),
2530
+ ("pol_Latn", "swh_Latn"),
2531
+ ("pol_Latn", "tam_Taml"),
2532
+ ("pol_Latn", "tur_Latn"),
2533
+ ("pol_Latn", "ukr_Cyrl"),
2534
+ ("pol_Latn", "vie_Latn"),
2535
+ ("pol_Latn", "zho_Hans"),
2536
+ ("por_Latn", "ron_Latn"),
2537
+ ("por_Latn", "rus_Cyrl"),
2538
+ ("por_Latn", "slk_Latn"),
2539
+ ("por_Latn", "slv_Latn"),
2540
+ ("por_Latn", "srp_Cyrl"),
2541
+ ("por_Latn", "swe_Latn"),
2542
+ ("por_Latn", "swh_Latn"),
2543
+ ("por_Latn", "tam_Taml"),
2544
+ ("por_Latn", "tur_Latn"),
2545
+ ("por_Latn", "ukr_Cyrl"),
2546
+ ("por_Latn", "vie_Latn"),
2547
+ ("por_Latn", "zho_Hans"),
2548
+ ("ron_Latn", "als_Latn"),
2549
+ ("ron_Latn", "slk_Latn"),
2550
+ ("ron_Latn", "slv_Latn"),
2551
+ ("ron_Latn", "srp_Cyrl"),
2552
+ ("ron_Latn", "swe_Latn"),
2553
+ ("ron_Latn", "tam_Taml"),
2554
+ ("ron_Latn", "tur_Latn"),
2555
+ ("ron_Latn", "ukr_Cyrl"),
2556
+ ("ron_Latn", "vie_Latn"),
2557
+ ("ron_Latn", "zho_Hans"),
2558
+ ("rus_Cyrl", "als_Latn"),
2559
+ ("rus_Cyrl", "sin_Sinh"),
2560
+ ("rus_Cyrl", "slk_Latn"),
2561
+ ("rus_Cyrl", "slv_Latn"),
2562
+ ("rus_Cyrl", "srp_Cyrl"),
2563
+ ("rus_Cyrl", "swe_Latn"),
2564
+ ("rus_Cyrl", "swh_Latn"),
2565
+ ("rus_Cyrl", "tam_Taml"),
2566
+ ("rus_Cyrl", "tgl_Latn"),
2567
+ ("rus_Cyrl", "tur_Latn"),
2568
+ ("rus_Cyrl", "ukr_Cyrl"),
2569
+ ("rus_Cyrl", "urd_Arab"),
2570
+ ("rus_Cyrl", "vie_Latn"),
2571
+ ("rus_Cyrl", "zho_Hans"),
2572
+ ("sin_Sinh", "tur_Latn"),
2573
+ ("slk_Latn", "als_Latn"),
2574
+ ("slk_Latn", "slv_Latn"),
2575
+ ("slk_Latn", "srp_Cyrl"),
2576
+ ("slk_Latn", "swe_Latn"),
2577
+ ("slk_Latn", "swh_Latn"),
2578
+ ("slk_Latn", "tam_Taml"),
2579
+ ("slk_Latn", "tur_Latn"),
2580
+ ("slk_Latn", "ukr_Cyrl"),
2581
+ ("slk_Latn", "vie_Latn"),
2582
+ ("slk_Latn", "zho_Hans"),
2583
+ ("slv_Latn", "als_Latn"),
2584
+ ("slv_Latn", "srp_Cyrl"),
2585
+ ("slv_Latn", "swe_Latn"),
2586
+ ("slv_Latn", "swh_Latn"),
2587
+ ("slv_Latn", "tam_Taml"),
2588
+ ("slv_Latn", "tur_Latn"),
2589
+ ("slv_Latn", "ukr_Cyrl"),
2590
+ ("slv_Latn", "vie_Latn"),
2591
+ ("slv_Latn", "zho_Hans"),
2592
+ ("spa_Latn", "als_Latn"),
2593
+ ("spa_Latn", "est_Latn"),
2594
+ ("spa_Latn", "fin_Latn"),
2595
+ ("spa_Latn", "fra_Latn"),
2596
+ ("spa_Latn", "glg_Latn"),
2597
+ ("spa_Latn", "heb_Hebr"),
2598
+ ("spa_Latn", "hin_Deva"),
2599
+ ("spa_Latn", "hun_Latn"),
2600
+ ("spa_Latn", "ind_Latn"),
2601
+ ("spa_Latn", "ita_Latn"),
2602
+ ("spa_Latn", "jpn_Jpan"),
2603
+ ("spa_Latn", "kor_Hang"),
2604
+ ("spa_Latn", "lit_Latn"),
2605
+ ("spa_Latn", "lvs_Latn"),
2606
+ ("spa_Latn", "mkd_Cyrl"),
2607
+ ("spa_Latn", "nld_Latn"),
2608
+ ("spa_Latn", "nob_Latn"),
2609
+ ("spa_Latn", "oci_Latn"),
2610
+ ("spa_Latn", "pes_Arab"),
2611
+ ("spa_Latn", "pol_Latn"),
2612
+ ("spa_Latn", "por_Latn"),
2613
+ ("spa_Latn", "ron_Latn"),
2614
+ ("spa_Latn", "rus_Cyrl"),
2615
+ ("spa_Latn", "slk_Latn"),
2616
+ ("spa_Latn", "slv_Latn"),
2617
+ ("spa_Latn", "srp_Cyrl"),
2618
+ ("spa_Latn", "swe_Latn"),
2619
+ ("spa_Latn", "swh_Latn"),
2620
+ ("spa_Latn", "tam_Taml"),
2621
+ ("spa_Latn", "tur_Latn"),
2622
+ ("spa_Latn", "ukr_Cyrl"),
2623
+ ("spa_Latn", "vie_Latn"),
2624
+ ("spa_Latn", "zho_Hans"),
2625
+ ("srp_Cyrl", "swe_Latn"),
2626
+ ("srp_Cyrl", "swh_Latn"),
2627
+ ("srp_Cyrl", "tam_Taml"),
2628
+ ("srp_Cyrl", "tur_Latn"),
2629
+ ("srp_Cyrl", "ukr_Cyrl"),
2630
+ ("srp_Cyrl", "vie_Latn"),
2631
+ ("srp_Cyrl", "zho_Hans"),
2632
+ ("sun_Latn", "tgl_Latn"),
2633
+ ("swe_Latn", "swh_Latn"),
2634
+ ("swe_Latn", "tam_Taml"),
2635
+ ("swe_Latn", "tur_Latn"),
2636
+ ("swe_Latn", "ukr_Cyrl"),
2637
+ ("swe_Latn", "vie_Latn"),
2638
+ ("swe_Latn", "zho_Hans"),
2639
+ ("swh_Latn", "tam_Taml"),
2640
+ ("swh_Latn", "tur_Latn"),
2641
+ ("swh_Latn", "ukr_Cyrl"),
2642
+ ("swh_Latn", "vie_Latn"),
2643
+ ("swh_Latn", "zho_Hans"),
2644
+ ("tam_Taml", "tgl_Latn"),
2645
+ ("tam_Taml", "tur_Latn"),
2646
+ ("tam_Taml", "ukr_Cyrl"),
2647
+ ("tam_Taml", "vie_Latn"),
2648
+ ("tam_Taml", "zho_Hans"),
2649
+ ("tgl_Latn", "tur_Latn"),
2650
+ ("tgl_Latn", "vie_Latn"),
2651
+ ("tgl_Latn", "zho_Hans"),
2652
+ ("tur_Latn", "ukr_Cyrl"),
2653
+ ("tur_Latn", "urd_Arab"),
2654
+ ("tur_Latn", "uzn_Latn"),
2655
+ ("tur_Latn", "vie_Latn"),
2656
+ ("tur_Latn", "zho_Hans"),
2657
+ ("ukr_Cyrl", "vie_Latn"),
2658
+ ("ukr_Cyrl", "zho_Hans"),
2659
+ ("vie_Latn", "zho_Hans"),
2660
+ ("ydd_Hebr", "zho_Hans"),
2661
+ ("zsm_Latn", "rus_Cyrl"),
2662
+ ("zsm_Latn", "sun_Latn"),
2663
+ ("zsm_Latn", "tam_Taml"),
2664
+ ("zsm_Latn", "tgl_Latn"),
2665
+ ("zsm_Latn", "tur_Latn"),
2666
+ ("zsm_Latn", "vie_Latn"),
2667
+ ("zsm_Latn", "zho_Hans"),
2668
+ ]
2669
+ CCMATRIX_MAPPING = {
2670
+ "afr_Latn": "af",
2671
+ "als_Latn": "sq",
2672
+ "amh_Ethi": "am",
2673
+ "arb_Arab": "ar",
2674
+ "ast_Latn": "ast",
2675
+ "azj_Latn": "az",
2676
+ "bel_Cyrl": "be",
2677
+ "ben_Beng": "bn",
2678
+ "bul_Cyrl": "bg",
2679
+ "cat_Latn": "ca",
2680
+ "ceb_Latn": "ceb",
2681
+ "ces_Latn": "cs",
2682
+ "cym_Latn": "cy",
2683
+ "dan_Latn": "da",
2684
+ "deu_Latn": "de",
2685
+ "ell_Grek": "el",
2686
+ "eng_Latn": "en",
2687
+ "epo_Latn": "eo",
2688
+ "est_Latn": "et",
2689
+ "fin_Latn": "fi",
2690
+ "fra_Latn": "fr",
2691
+ "gaz_Latn": "om",
2692
+ "gla_Latn": "gd",
2693
+ "gle_Latn": "ga",
2694
+ "glg_Latn": "gl",
2695
+ "hau_Latn": "ha",
2696
+ "heb_Hebr": "he",
2697
+ "hin_Deva": "hi",
2698
+ "hrv_Latn": "hr",
2699
+ "hun_Latn": "hu",
2700
+ "hye_Armn": "hy",
2701
+ "ibo_Latn": "ig",
2702
+ "ilo_Latn": "ilo",
2703
+ "ind_Latn": "id",
2704
+ "isl_Latn": "is",
2705
+ "ita_Latn": "it",
2706
+ "jav_Latn": "jv",
2707
+ "jpn_Jpan": "ja",
2708
+ "kat_Geor": "ka",
2709
+ "kaz_Cyrl": "kk",
2710
+ "khm_Khmr": "km",
2711
+ "kor_Hang": "ko",
2712
+ "lit_Latn": "lt",
2713
+ "ltz_Latn": "lb",
2714
+ "lug_Latn": "lg",
2715
+ "lvs_Latn": "lv",
2716
+ "mal_Mlym": "ml",
2717
+ "mar_Deva": "mr",
2718
+ "mkd_Cyrl": "mk",
2719
+ "mya_Mymr": "my",
2720
+ "nld_Latn": "nl",
2721
+ "nob_Latn": "no",
2722
+ "npi_Deva": "ne",
2723
+ "oci_Latn": "oc",
2724
+ "ory_Orya": "or",
2725
+ "pes_Arab": "fa",
2726
+ "plt_Latn": "mg",
2727
+ "pol_Latn": "pl",
2728
+ "por_Latn": "pt",
2729
+ "ron_Latn": "ro",
2730
+ "rus_Cyrl": "ru",
2731
+ "sin_Sinh": "si",
2732
+ "slk_Latn": "sk",
2733
+ "slv_Latn": "sl",
2734
+ "snd_Arab": "sd",
2735
+ "som_Latn": "so",
2736
+ "spa_Latn": "es",
2737
+ "srp_Cyrl": "sr",
2738
+ "sun_Latn": "su",
2739
+ "swe_Latn": "sv",
2740
+ "swh_Latn": "sw",
2741
+ "tam_Taml": "ta",
2742
+ "tat_Cyrl": "tt",
2743
+ "tgl_Latn": "tl",
2744
+ "tur_Latn": "tr",
2745
+ "ukr_Cyrl": "uk",
2746
+ "urd_Arab": "ur",
2747
+ "uzn_Latn": "uz",
2748
+ "vie_Latn": "vi",
2749
+ "wol_Latn": "wo",
2750
+ "xho_Latn": "xh",
2751
+ "ydd_Hebr": "yi",
2752
+ "yor_Latn": "yo",
2753
+ "zho_Hans": "zh",
2754
+ "zsm_Latn": "ms",
2755
+ "zul_Latn": "zu",
2756
+ }