Datasets:

License:
theyorubayesian commited on
Commit
cbaa143
1 Parent(s): 21e7529

add translated dataset

Browse files
.gitattributes CHANGED
@@ -58,3 +58,7 @@ passages-v1.0/yoruba_passages.jsonl filter=lfs diff=lfs merge=lfs -text
58
  passages-v1.0/hausa_passages.jsonl filter=lfs diff=lfs merge=lfs -text
59
  passages-v1.0/somali_passages.jsonl filter=lfs diff=lfs merge=lfs -text
60
  passages-v1.0/swahili_passages.jsonl filter=lfs diff=lfs merge=lfs -text
 
 
 
 
 
58
  passages-v1.0/hausa_passages.jsonl filter=lfs diff=lfs merge=lfs -text
59
  passages-v1.0/somali_passages.jsonl filter=lfs diff=lfs merge=lfs -text
60
  passages-v1.0/swahili_passages.jsonl filter=lfs diff=lfs merge=lfs -text
61
+ translated-passages-v1.0/hausa_passages.jsonl filter=lfs diff=lfs merge=lfs -text
62
+ translated-passages-v1.0/somali_passages.jsonl filter=lfs diff=lfs merge=lfs -text
63
+ translated-passages-v1.0/swahili_passages.jsonl filter=lfs diff=lfs merge=lfs -text
64
+ translated-passages-v1.0/yoruba_passages.jsonl filter=lfs diff=lfs merge=lfs -text
README.md CHANGED
@@ -45,6 +45,16 @@ language = "hausa"
45
  dataset = load_dataset("ciral/ciral-corpus", language)
46
  ```
47
 
 
 
 
 
 
 
 
 
 
 
48
  ## Citation
49
 
50
  ...
 
45
  dataset = load_dataset("ciral/ciral-corpus", language)
46
  ```
47
 
48
+ ## Translated Dataset
49
+
50
+ We also include a translated version of the dataset for all the languages. Translation was done using Google Machine Translation (GMT).
51
+
52
+ ```python
53
+ language = "hausa"
54
+
55
+ dataset = load_dataset("ciral/ciral-corpus", language, translated=True)
56
+ ```
57
+
58
  ## Citation
59
 
60
  ...
ciral-corpus.py CHANGED
@@ -1,4 +1,5 @@
1
  import json
 
2
  from string import Template
3
 
4
  import datasets
@@ -18,12 +19,27 @@ _LANGUAGES = [
18
  "yoruba",
19
  "combined"
20
  ]
21
- _DATASET_URL = Template("./passages-v1.0/${language}_passages.jsonl")
22
 
23
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
24
  class CiralPassages(datasets.GeneratorBasedBuilder):
25
  BUILDER_CONFIGS = [
26
- datasets.BuilderConfig(
27
  version=datasets.Version(_VERSION),
28
  name=language,
29
  description=f"CIRAL passages for language: {language}"
@@ -49,7 +65,7 @@ class CiralPassages(datasets.GeneratorBasedBuilder):
49
  language = self.config.name
50
  if language == "combined":
51
  language_file = dl_manager.download_and_extract({
52
- _language: _DATASET_URL.substitute(language=_language)
53
  for _language in _LANGUAGES[:-1]
54
  })
55
 
@@ -58,10 +74,9 @@ class CiralPassages(datasets.GeneratorBasedBuilder):
58
  name=_language, gen_kwargs={"filepath": language_file[_language]}
59
  ) for _language in _LANGUAGES[:-1]
60
  ]
61
- return splits
62
  else:
63
  language_file = dl_manager.download_and_extract(
64
- _DATASET_URL.substitute(language=language))
65
 
66
  splits = [
67
  datasets.SplitGenerator(
@@ -69,7 +84,8 @@ class CiralPassages(datasets.GeneratorBasedBuilder):
69
  gen_kwargs={"filepath": language_file}
70
  )
71
  ]
72
- return splits
 
73
 
74
  def _generate_examples(self, filepath: str):
75
  with open(filepath, encoding="utf-8") as f:
 
1
  import json
2
+ from dataclasses import dataclass
3
  from string import Template
4
 
5
  import datasets
 
19
  "yoruba",
20
  "combined"
21
  ]
22
+ _DATASET_URL = Template("./${mode}passages-v1.0/${language}_passages.jsonl")
23
 
24
 
25
+ @dataclass
26
+ class CiralConfig(datasets.BuilderConfig):
27
+ translated: bool = False
28
+ file_stub_dict = {
29
+ None: "",
30
+ True: "translated-",
31
+ False: ""
32
+ }
33
+
34
+ def get_file_url(self, language: str) -> str:
35
+ return _DATASET_URL.substitute(
36
+ mode=self.file_stub_dict.get(self.translated),
37
+ language=language
38
+ )
39
+
40
  class CiralPassages(datasets.GeneratorBasedBuilder):
41
  BUILDER_CONFIGS = [
42
+ CiralConfig(
43
  version=datasets.Version(_VERSION),
44
  name=language,
45
  description=f"CIRAL passages for language: {language}"
 
65
  language = self.config.name
66
  if language == "combined":
67
  language_file = dl_manager.download_and_extract({
68
+ _language: self.config.get_file_url(language=_language)
69
  for _language in _LANGUAGES[:-1]
70
  })
71
 
 
74
  name=_language, gen_kwargs={"filepath": language_file[_language]}
75
  ) for _language in _LANGUAGES[:-1]
76
  ]
 
77
  else:
78
  language_file = dl_manager.download_and_extract(
79
+ self.config.get_file_url(language=language))
80
 
81
  splits = [
82
  datasets.SplitGenerator(
 
84
  gen_kwargs={"filepath": language_file}
85
  )
86
  ]
87
+
88
+ return splits
89
 
90
  def _generate_examples(self, filepath: str):
91
  with open(filepath, encoding="utf-8") as f:
translated-passages-v1.0/hausa_passages.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e9c32f6f5d45557fb141dd34fdd0a6ffdf1b2a46d3d38e7cb2a63c7d35ab9737
3
+ size 600720600
translated-passages-v1.0/somali_passages.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:95ef93c66e93d71e724ee633339f0829cf21d059705ec2353d4ec9b065701ed8
3
+ size 731277640
translated-passages-v1.0/swahili_passages.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7d3f13d4f0c4e5b231325f5ccb96c9c87d9df137803fabce98a65136c9ac9ade
3
+ size 860356926
translated-passages-v1.0/yoruba_passages.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:723033805f1bdde16d217efa23d489fd6622c251cb3f34a2f5ff6fb460abaaca
3
+ size 73743549