holylovenia
commited on
Commit
•
e1ea39e
1
Parent(s):
ffe75db
Upload tico_19.py with huggingface_hub
Browse files- tico_19.py +71 -83
tico_19.py
CHANGED
@@ -14,18 +14,17 @@
|
|
14 |
# limitations under the License.
|
15 |
|
16 |
import csv
|
17 |
-
from fnmatch import translate
|
18 |
import os
|
19 |
import re
|
20 |
from pathlib import Path
|
21 |
from typing import Dict, List, Tuple
|
22 |
-
from translate.storage.tmx import tmxfile
|
23 |
|
24 |
import datasets
|
|
|
25 |
|
26 |
-
from
|
27 |
-
from
|
28 |
-
from
|
29 |
|
30 |
_CITATION = """\
|
31 |
@inproceedings{anastasopoulos-etal-2020-tico,
|
@@ -59,34 +58,50 @@ _CITATION = """\
|
|
59 |
"""
|
60 |
|
61 |
# We follow ISO639-3 language code (https://iso639-3.sil.org/code_tables/639/data)
|
62 |
-
_LANGUAGES = ["ind", "ara", "spa", "fra", "hin", "por", "rus", "zho", "eng"]
|
63 |
_LOCAL = False
|
64 |
_SUPPORTED_LANG_PAIRS = [
|
65 |
-
("ind", "ara"),
|
66 |
-
("
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
67 |
]
|
68 |
|
69 |
-
_LANG_CODE_MAP = {
|
70 |
-
|
71 |
-
|
72 |
-
"spa": "es-LA",
|
73 |
-
"fra": "fr",
|
74 |
-
"hin": "hi",
|
75 |
-
"por": "pt-BR",
|
76 |
-
"rus": "ru",
|
77 |
-
"zho": "zh",
|
78 |
-
"eng": "en"
|
79 |
-
}
|
80 |
|
81 |
_DATASETNAME = "tico_19"
|
82 |
|
83 |
_DESCRIPTION = """\
|
84 |
-
TICO-19 (Translation Initiative for COVID-19) is sampled from a variety of public sources containing
|
85 |
-
COVID-19 related content, representing different domains (e.g., news, wiki articles, and others). TICO-19
|
86 |
-
includes 30 documents (3071 sentences, 69.7k words) translated from English into 36 languages: Amharic,
|
87 |
-
Arabic (Modern Standard), Bengali, Chinese (Simplified), Dari, Dinka, Farsi, French (European), Hausa,
|
88 |
-
Hindi, Indonesian, Kanuri, Khmer (Central), Kinyarwanda, Kurdish Kurmanji, Kurdish Sorani, Lingala,
|
89 |
-
Luganda, Malay, Marathi, Myanmar, Nepali, Nigerian Fulfulde, Nuer, Oromo, Pashto, Portuguese (Brazilian),
|
90 |
Russian, Somali, Spanish (Latin American), Swahili, Congolese Swahili, Tagalog, Tamil, Tigrinya, Urdu, Zulu.
|
91 |
"""
|
92 |
|
@@ -94,25 +109,22 @@ _HOMEPAGE = "https://tico-19.github.io"
|
|
94 |
|
95 |
_LICENSE = "CC0"
|
96 |
|
97 |
-
_URLS = {
|
98 |
-
"evaluation": "https://tico-19.github.io/data/tico19-testset.zip",
|
99 |
-
"all": "https://tico-19.github.io/data/TM/all.{lang_pairs}.tmx.zip"
|
100 |
-
}
|
101 |
|
102 |
_SUPPORTED_TASKS = [Tasks.MACHINE_TRANSLATION]
|
103 |
|
104 |
_SOURCE_VERSION = "1.0.0"
|
105 |
|
106 |
-
|
107 |
|
108 |
|
109 |
-
def
|
110 |
-
"""Construct
|
111 |
-
if schema != "source" and schema != "
|
112 |
raise ValueError(f"Invalid schema: {schema}")
|
113 |
|
114 |
if lang_source == "" and lang_target == "":
|
115 |
-
return
|
116 |
name="tico_19_{schema}".format(schema=schema),
|
117 |
version=datasets.Version(version),
|
118 |
description="tico_19 {schema} schema for default language pair (eng-ind)".format(schema=schema),
|
@@ -120,7 +132,7 @@ def nusantara_config_constructor(lang_source, lang_target, schema, version):
|
|
120 |
subset_id="tico_19",
|
121 |
)
|
122 |
else:
|
123 |
-
return
|
124 |
name="tico_19_{src}_{tgt}_{schema}".format(src=lang_source, tgt=lang_target, schema=schema),
|
125 |
version=datasets.Version(version),
|
126 |
description="tico_19 {schema} schema for {src}-{tgt} language pair".format(src=lang_source, tgt=lang_target, schema=schema),
|
@@ -128,16 +140,14 @@ def nusantara_config_constructor(lang_source, lang_target, schema, version):
|
|
128 |
subset_id="tico_19",
|
129 |
)
|
130 |
|
|
|
131 |
class Tico19(datasets.GeneratorBasedBuilder):
|
132 |
"""TICO-19 is MT dataset sampled from a variety of public sources containing COVID-19 related content"""
|
133 |
|
134 |
SOURCE_VERSION = datasets.Version(_SOURCE_VERSION)
|
135 |
-
|
136 |
|
137 |
-
BUILDER_CONFIGS = [
|
138 |
-
nusantara_config_constructor(src, tgt, schema, version)
|
139 |
-
for src, tgt in [("", "")] + _SUPPORTED_LANG_PAIRS for schema, version in zip(["source", "nusantara_t2t"], [_SOURCE_VERSION, _NUSANTARA_VERSION])
|
140 |
-
]
|
141 |
|
142 |
DEFAULT_CONFIG_NAME = "tico_19_source"
|
143 |
|
@@ -155,7 +165,7 @@ class Tico19(datasets.GeneratorBasedBuilder):
|
|
155 |
"translatorId": datasets.Value("string"),
|
156 |
}
|
157 |
)
|
158 |
-
elif self.config.schema == "
|
159 |
features = schemas.text2text_features
|
160 |
|
161 |
return datasets.DatasetInfo(
|
@@ -168,34 +178,28 @@ class Tico19(datasets.GeneratorBasedBuilder):
|
|
168 |
|
169 |
def _split_generators(self, dl_manager: datasets.DownloadManager) -> List[datasets.SplitGenerator]:
|
170 |
"""Returns SplitGenerators."""
|
171 |
-
|
172 |
try:
|
173 |
-
lang_pairs_config = re.search("tico_19_(.+?)_(source|
|
174 |
lang_src, lang_tgt = lang_pairs_config.split("_")
|
175 |
except AttributeError:
|
176 |
lang_src, lang_tgt = "eng", "ind"
|
177 |
|
178 |
lang_pairs = _LANG_CODE_MAP[lang_src] + "-" + _LANG_CODE_MAP[lang_tgt]
|
179 |
|
180 |
-
# dev & test split only applicable to eng-
|
181 |
-
if lang_pairs in
|
|
|
|
|
182 |
data_dir = dl_manager.download_and_extract(_URLS["evaluation"])
|
183 |
return [
|
184 |
datasets.SplitGenerator(
|
185 |
name=datasets.Split.TEST,
|
186 |
-
gen_kwargs={
|
187 |
-
"filepath": os.path.join(data_dir, "tico19-testset", "test", f"test.en-id.tsv"),
|
188 |
-
"lang_source": lang_src,
|
189 |
-
"lang_target": lang_tgt
|
190 |
-
},
|
191 |
),
|
192 |
datasets.SplitGenerator(
|
193 |
name=datasets.Split.VALIDATION,
|
194 |
-
gen_kwargs={
|
195 |
-
"filepath": os.path.join(data_dir, "tico19-testset", "dev", f"dev.en-id.tsv"),
|
196 |
-
"lang_source": lang_src,
|
197 |
-
"lang_target": lang_tgt
|
198 |
-
},
|
199 |
),
|
200 |
]
|
201 |
else:
|
@@ -203,20 +207,16 @@ class Tico19(datasets.GeneratorBasedBuilder):
|
|
203 |
return [
|
204 |
datasets.SplitGenerator(
|
205 |
name=datasets.Split.TRAIN,
|
206 |
-
gen_kwargs={
|
207 |
-
"filepath": os.path.join(data_dir, f"all.{lang_pairs}.tmx"),
|
208 |
-
"lang_source": lang_src,
|
209 |
-
"lang_target": lang_tgt
|
210 |
-
},
|
211 |
)
|
212 |
]
|
213 |
|
214 |
def _generate_examples(self, filepath: Path, lang_source: str, lang_target: str) -> Tuple[int, Dict]:
|
215 |
"""Yields examples as (key, example) tuples."""
|
216 |
-
|
217 |
if self.config.schema == "source":
|
218 |
-
# eng-
|
219 |
-
if
|
220 |
with open(filepath, encoding="utf-8") as f:
|
221 |
reader = csv.reader(f, delimiter="\t", quotechar='"')
|
222 |
for id_, row in enumerate(reader):
|
@@ -242,7 +242,7 @@ class Tico19(datasets.GeneratorBasedBuilder):
|
|
242 |
"license": row[6],
|
243 |
"translatorId": row[7],
|
244 |
}
|
245 |
-
|
246 |
# all language pairs except eng-ind dataset provided in .tmx format
|
247 |
else:
|
248 |
with open(filepath, "rb") as f:
|
@@ -250,8 +250,8 @@ class Tico19(datasets.GeneratorBasedBuilder):
|
|
250 |
|
251 |
for id_, node in enumerate(tmx_file.unit_iter()):
|
252 |
try:
|
253 |
-
url = [text for text in node.xmlelement.itertext(
|
254 |
-
except:
|
255 |
url = ""
|
256 |
yield id_, {
|
257 |
"sourceLang": _LANG_CODE_MAP[lang_source],
|
@@ -264,8 +264,8 @@ class Tico19(datasets.GeneratorBasedBuilder):
|
|
264 |
"translatorId": "",
|
265 |
}
|
266 |
|
267 |
-
elif self.config.schema == "
|
268 |
-
if
|
269 |
with open(filepath, encoding="utf-8") as f:
|
270 |
reader = csv.reader(f, delimiter="\t", quotechar='"')
|
271 |
for id_, row in enumerate(reader):
|
@@ -277,22 +277,10 @@ class Tico19(datasets.GeneratorBasedBuilder):
|
|
277 |
else:
|
278 |
source_string = row[3]
|
279 |
target_string = row[2]
|
280 |
-
yield id_, {
|
281 |
-
"id": row[4],
|
282 |
-
"text_1": source_string,
|
283 |
-
"text_2": target_string,
|
284 |
-
"text_1_name": lang_source,
|
285 |
-
"text_2_name": lang_target
|
286 |
-
}
|
287 |
else:
|
288 |
with open(filepath, "rb") as f:
|
289 |
tmx_file = tmxfile(f)
|
290 |
-
|
291 |
for id_, node in enumerate(tmx_file.unit_iter()):
|
292 |
-
yield id_, {
|
293 |
-
"id": node.getid(),
|
294 |
-
"text_1": node.source,
|
295 |
-
"text_2": node.target,
|
296 |
-
"text_1_name": lang_source,
|
297 |
-
"text_2_name": lang_target
|
298 |
-
}
|
|
|
14 |
# limitations under the License.
|
15 |
|
16 |
import csv
|
|
|
17 |
import os
|
18 |
import re
|
19 |
from pathlib import Path
|
20 |
from typing import Dict, List, Tuple
|
|
|
21 |
|
22 |
import datasets
|
23 |
+
from translate.storage.tmx import tmxfile
|
24 |
|
25 |
+
from seacrowd.utils import schemas
|
26 |
+
from seacrowd.utils.configs import SEACrowdConfig
|
27 |
+
from seacrowd.utils.constants import Tasks
|
28 |
|
29 |
_CITATION = """\
|
30 |
@inproceedings{anastasopoulos-etal-2020-tico,
|
|
|
58 |
"""
|
59 |
|
60 |
# We follow ISO639-3 language code (https://iso639-3.sil.org/code_tables/639/data)
|
61 |
+
_LANGUAGES = ["ind", "ara", "spa", "fra", "hin", "por", "rus", "zho", "eng", "khm", "zlm", "mya", "tgl", "tam"]
|
62 |
_LOCAL = False
|
63 |
_SUPPORTED_LANG_PAIRS = [
|
64 |
+
("ind", "ara"),
|
65 |
+
("ind", "spa"),
|
66 |
+
("ind", "fra"),
|
67 |
+
("ind", "hin"),
|
68 |
+
("ind", "por"),
|
69 |
+
("ind", "rus"),
|
70 |
+
("ind", "zho"),
|
71 |
+
("ind", "eng"),
|
72 |
+
("ara", "ind"),
|
73 |
+
("spa", "ind"),
|
74 |
+
("fra", "ind"),
|
75 |
+
("hin", "ind"),
|
76 |
+
("por", "ind"),
|
77 |
+
("rus", "ind"),
|
78 |
+
("zho", "ind"),
|
79 |
+
("eng", "ind"),
|
80 |
+
("khm", "eng"),
|
81 |
+
("eng", "khm"),
|
82 |
+
("mya", "eng"),
|
83 |
+
("eng", "mya"),
|
84 |
+
("zlm", "eng"),
|
85 |
+
("eng", "zlm"),
|
86 |
+
("tgl", "eng"),
|
87 |
+
("eng", "tgl"),
|
88 |
+
("tam", "eng"),
|
89 |
+
("eng", "tam"),
|
90 |
]
|
91 |
|
92 |
+
_LANG_CODE_MAP = {"ind": "id", "ara": "ar", "spa": "es-LA", "fra": "fr", "hin": "hi", "por": "pt-BR", "rus": "ru", "zho": "zh", "eng": "en", "khm": "km", "zlm": "ms", "mya": "my", "tgl": "tl", "tam": "ta"}
|
93 |
+
|
94 |
+
_DEVTEST_LANG_PAIRS = [_LANG_CODE_MAP[source_lang] + "-" + _LANG_CODE_MAP[target_lang] for (source_lang, target_lang) in _SUPPORTED_LANG_PAIRS if (source_lang == "eng" or target_lang == "eng")]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
95 |
|
96 |
_DATASETNAME = "tico_19"
|
97 |
|
98 |
_DESCRIPTION = """\
|
99 |
+
TICO-19 (Translation Initiative for COVID-19) is sampled from a variety of public sources containing
|
100 |
+
COVID-19 related content, representing different domains (e.g., news, wiki articles, and others). TICO-19
|
101 |
+
includes 30 documents (3071 sentences, 69.7k words) translated from English into 36 languages: Amharic,
|
102 |
+
Arabic (Modern Standard), Bengali, Chinese (Simplified), Dari, Dinka, Farsi, French (European), Hausa,
|
103 |
+
Hindi, Indonesian, Kanuri, Khmer (Central), Kinyarwanda, Kurdish Kurmanji, Kurdish Sorani, Lingala,
|
104 |
+
Luganda, Malay, Marathi, Myanmar, Nepali, Nigerian Fulfulde, Nuer, Oromo, Pashto, Portuguese (Brazilian),
|
105 |
Russian, Somali, Spanish (Latin American), Swahili, Congolese Swahili, Tagalog, Tamil, Tigrinya, Urdu, Zulu.
|
106 |
"""
|
107 |
|
|
|
109 |
|
110 |
_LICENSE = "CC0"
|
111 |
|
112 |
+
_URLS = {"evaluation": "https://tico-19.github.io/data/tico19-testset.zip", "all": "https://tico-19.github.io/data/TM/all.{lang_pairs}.tmx.zip"}
|
|
|
|
|
|
|
113 |
|
114 |
_SUPPORTED_TASKS = [Tasks.MACHINE_TRANSLATION]
|
115 |
|
116 |
_SOURCE_VERSION = "1.0.0"
|
117 |
|
118 |
+
_SEACROWD_VERSION = "2024.06.20"
|
119 |
|
120 |
|
121 |
+
def seacrowd_config_constructor(lang_source, lang_target, schema, version):
|
122 |
+
"""Construct SEACrowdConfig with tico_19_{lang_source}_{lang_target}_{schema} as the name format"""
|
123 |
+
if schema != "source" and schema != "seacrowd_t2t":
|
124 |
raise ValueError(f"Invalid schema: {schema}")
|
125 |
|
126 |
if lang_source == "" and lang_target == "":
|
127 |
+
return SEACrowdConfig(
|
128 |
name="tico_19_{schema}".format(schema=schema),
|
129 |
version=datasets.Version(version),
|
130 |
description="tico_19 {schema} schema for default language pair (eng-ind)".format(schema=schema),
|
|
|
132 |
subset_id="tico_19",
|
133 |
)
|
134 |
else:
|
135 |
+
return SEACrowdConfig(
|
136 |
name="tico_19_{src}_{tgt}_{schema}".format(src=lang_source, tgt=lang_target, schema=schema),
|
137 |
version=datasets.Version(version),
|
138 |
description="tico_19 {schema} schema for {src}-{tgt} language pair".format(src=lang_source, tgt=lang_target, schema=schema),
|
|
|
140 |
subset_id="tico_19",
|
141 |
)
|
142 |
|
143 |
+
|
144 |
class Tico19(datasets.GeneratorBasedBuilder):
|
145 |
"""TICO-19 is MT dataset sampled from a variety of public sources containing COVID-19 related content"""
|
146 |
|
147 |
SOURCE_VERSION = datasets.Version(_SOURCE_VERSION)
|
148 |
+
SEACROWD_VERSION = datasets.Version(_SEACROWD_VERSION)
|
149 |
|
150 |
+
BUILDER_CONFIGS = [seacrowd_config_constructor(src, tgt, schema, version) for src, tgt in [("", "")] + _SUPPORTED_LANG_PAIRS for schema, version in zip(["source", "seacrowd_t2t"], [_SOURCE_VERSION, _SEACROWD_VERSION])]
|
|
|
|
|
|
|
151 |
|
152 |
DEFAULT_CONFIG_NAME = "tico_19_source"
|
153 |
|
|
|
165 |
"translatorId": datasets.Value("string"),
|
166 |
}
|
167 |
)
|
168 |
+
elif self.config.schema == "seacrowd_t2t":
|
169 |
features = schemas.text2text_features
|
170 |
|
171 |
return datasets.DatasetInfo(
|
|
|
178 |
|
179 |
def _split_generators(self, dl_manager: datasets.DownloadManager) -> List[datasets.SplitGenerator]:
|
180 |
"""Returns SplitGenerators."""
|
181 |
+
|
182 |
try:
|
183 |
+
lang_pairs_config = re.search("tico_19_(.+?)_(source|seacrowd_t2t)", self.config.name).group(1)
|
184 |
lang_src, lang_tgt = lang_pairs_config.split("_")
|
185 |
except AttributeError:
|
186 |
lang_src, lang_tgt = "eng", "ind"
|
187 |
|
188 |
lang_pairs = _LANG_CODE_MAP[lang_src] + "-" + _LANG_CODE_MAP[lang_tgt]
|
189 |
|
190 |
+
# dev & test split only applicable to eng-[sea language] language pair
|
191 |
+
if lang_pairs in set(_DEVTEST_LANG_PAIRS):
|
192 |
+
lang_sea = _LANG_CODE_MAP[lang_tgt] if lang_src == "eng" else _LANG_CODE_MAP[lang_src]
|
193 |
+
|
194 |
data_dir = dl_manager.download_and_extract(_URLS["evaluation"])
|
195 |
return [
|
196 |
datasets.SplitGenerator(
|
197 |
name=datasets.Split.TEST,
|
198 |
+
gen_kwargs={"filepath": os.path.join(data_dir, "tico19-testset", "test", f"test.en-{lang_sea}.tsv"), "lang_source": lang_src, "lang_target": lang_tgt},
|
|
|
|
|
|
|
|
|
199 |
),
|
200 |
datasets.SplitGenerator(
|
201 |
name=datasets.Split.VALIDATION,
|
202 |
+
gen_kwargs={"filepath": os.path.join(data_dir, "tico19-testset", "dev", f"dev.en-{lang_sea}.tsv"), "lang_source": lang_src, "lang_target": lang_tgt},
|
|
|
|
|
|
|
|
|
203 |
),
|
204 |
]
|
205 |
else:
|
|
|
207 |
return [
|
208 |
datasets.SplitGenerator(
|
209 |
name=datasets.Split.TRAIN,
|
210 |
+
gen_kwargs={"filepath": os.path.join(data_dir, f"all.{lang_pairs}.tmx"), "lang_source": lang_src, "lang_target": lang_tgt},
|
|
|
|
|
|
|
|
|
211 |
)
|
212 |
]
|
213 |
|
214 |
def _generate_examples(self, filepath: Path, lang_source: str, lang_target: str) -> Tuple[int, Dict]:
|
215 |
"""Yields examples as (key, example) tuples."""
|
216 |
+
|
217 |
if self.config.schema == "source":
|
218 |
+
# eng-[sea language] language pair dataset provided in .tsv format
|
219 |
+
if f"{_LANG_CODE_MAP[lang_source]}-{_LANG_CODE_MAP[lang_target]}" in set(_DEVTEST_LANG_PAIRS):
|
220 |
with open(filepath, encoding="utf-8") as f:
|
221 |
reader = csv.reader(f, delimiter="\t", quotechar='"')
|
222 |
for id_, row in enumerate(reader):
|
|
|
242 |
"license": row[6],
|
243 |
"translatorId": row[7],
|
244 |
}
|
245 |
+
|
246 |
# all language pairs except eng-ind dataset provided in .tmx format
|
247 |
else:
|
248 |
with open(filepath, "rb") as f:
|
|
|
250 |
|
251 |
for id_, node in enumerate(tmx_file.unit_iter()):
|
252 |
try:
|
253 |
+
url = [text for text in node.xmlelement.itertext("prop")][0]
|
254 |
+
except Exception:
|
255 |
url = ""
|
256 |
yield id_, {
|
257 |
"sourceLang": _LANG_CODE_MAP[lang_source],
|
|
|
264 |
"translatorId": "",
|
265 |
}
|
266 |
|
267 |
+
elif self.config.schema == "seacrowd_t2t":
|
268 |
+
if f"{_LANG_CODE_MAP[lang_source]}-{_LANG_CODE_MAP[lang_target]}" in set(_DEVTEST_LANG_PAIRS):
|
269 |
with open(filepath, encoding="utf-8") as f:
|
270 |
reader = csv.reader(f, delimiter="\t", quotechar='"')
|
271 |
for id_, row in enumerate(reader):
|
|
|
277 |
else:
|
278 |
source_string = row[3]
|
279 |
target_string = row[2]
|
280 |
+
yield id_, {"id": row[4], "text_1": source_string, "text_2": target_string, "text_1_name": lang_source, "text_2_name": lang_target}
|
|
|
|
|
|
|
|
|
|
|
|
|
281 |
else:
|
282 |
with open(filepath, "rb") as f:
|
283 |
tmx_file = tmxfile(f)
|
284 |
+
|
285 |
for id_, node in enumerate(tmx_file.unit_iter()):
|
286 |
+
yield id_, {"id": node.getid(), "text_1": node.source, "text_2": node.target, "text_1_name": lang_source, "text_2_name": lang_target}
|
|
|
|
|
|
|
|
|
|
|
|