joelniklaus commited on
Commit
0b7dcda
1 Parent(s): a20fb29

added language

Browse files
Files changed (1) hide show
  1. lextreme.py +15 -0
lextreme.py CHANGED
@@ -50,6 +50,7 @@ class LextremeConfig(datasets.BuilderConfig):
50
  url,
51
  citation,
52
  hf_hub_name,
 
53
  config_name="default",
54
  label_classes=None,
55
  task_type=None,
@@ -64,6 +65,7 @@ class LextremeConfig(datasets.BuilderConfig):
64
  url: `string`, url for the original project
65
  citation: `string`, citation for the data set
66
  hf_hub_name: `string`, huggingface dataset identifier
 
67
  config_name: `string`, huggingface dataset config name
68
  label_classes: `list[string]`, the list of classes if the label is
69
  categorical. If not provided, then the label will be of type
@@ -78,6 +80,7 @@ class LextremeConfig(datasets.BuilderConfig):
78
  self.description = textwrap.dedent(self.description)
79
  self.citation = textwrap.dedent(citation)
80
  self.hf_hub_name = hf_hub_name
 
81
  self.config_name = config_name
82
  self.label_classes = label_classes
83
  self.task_type = task_type
@@ -86,6 +89,7 @@ class LextremeConfig(datasets.BuilderConfig):
86
  _BRAZILIAN_COURT_DECISIONS_JUDGMENT = {
87
  "task_type": TaskType.SLTC,
88
  "hf_hub_name": "joelito/brazilian_court_decisions",
 
89
  "input_col": "decision_description",
90
  "label_col": "judgment_label",
91
  "url": "https://github.com/lagefreitas/predicting-brazilian-court-decisions",
@@ -150,6 +154,7 @@ _SWISS_JUDGMENT_PREDICTION = {
150
  _GERMAN_ARGUMENT_MINING = {
151
  "task_type": TaskType.SLTC,
152
  "hf_hub_name": "joelito/german_argument_mining",
 
153
  "input_col": "input_sentence",
154
  "label_col": "label",
155
  "url": "https://zenodo.org/record/3936490#.X1ed7ovgomK",
@@ -179,6 +184,7 @@ _GERMAN_ARGUMENT_MINING = {
179
  _GREEK_LEGAL_CODE_VOLUME = {
180
  "task_type": TaskType.SLTC,
181
  "hf_hub_name": "greek_legal_code",
 
182
  "config_name": "volume",
183
  "input_col": "text",
184
  "label_col": "label",
@@ -3766,6 +3772,7 @@ _MULTI_EURLEX_LEVEL_3 = {
3766
  _LENER_BR = {
3767
  "task_type": TaskType.NER,
3768
  "hf_hub_name": "lener_br",
 
3769
  "config_name": "lener_br",
3770
  "input_col": "tokens",
3771
  "label_col": "ner_tags",
@@ -3813,6 +3820,7 @@ _LENER_BR = {
3813
  _LEGAL_NERO = {
3814
  "task_type": TaskType.NER,
3815
  "hf_hub_name": "joelito/legalnero",
 
3816
  "input_col": "words",
3817
  "label_col": "ner",
3818
  "url": "https://zenodo.org/record/4922385",
@@ -3853,6 +3861,7 @@ _LEGAL_NERO = {
3853
  _GREEK_LEGAL_NER = {
3854
  "task_type": TaskType.NER,
3855
  "hf_hub_name": "joelito/greek_legal_ner",
 
3856
  "input_col": "words",
3857
  "label_col": "ner",
3858
  "url": "http://legislation.di.uoa.gr/publications?language=en",
@@ -4006,6 +4015,7 @@ class LEXTREME(datasets.GeneratorBasedBuilder):
4006
  features=datasets.Features({
4007
  "input": input_type,
4008
  "label": label_type,
 
4009
  }),
4010
  )
4011
 
@@ -4030,7 +4040,12 @@ class LEXTREME(datasets.GeneratorBasedBuilder):
4030
  label = dataset.features[label_col].int2str(label)
4031
  if self.config.task_type == TaskType.SLTC and label not in self.config.label_classes:
4032
  continue # filter out invalid classes to construct ClassLabel
 
 
 
 
4033
  yield id, {
4034
  "input": item[self.config.input_col],
4035
  "label": label,
 
4036
  }
 
50
  url,
51
  citation,
52
  hf_hub_name,
53
+ language="multilingual",
54
  config_name="default",
55
  label_classes=None,
56
  task_type=None,
 
65
  url: `string`, url for the original project
66
  citation: `string`, citation for the data set
67
  hf_hub_name: `string`, huggingface dataset identifier
68
+ hf_hub_name: `string`, the language of the dataset (either multilingual or a single language ISO code)
69
  config_name: `string`, huggingface dataset config name
70
  label_classes: `list[string]`, the list of classes if the label is
71
  categorical. If not provided, then the label will be of type
 
80
  self.description = textwrap.dedent(self.description)
81
  self.citation = textwrap.dedent(citation)
82
  self.hf_hub_name = hf_hub_name
83
+ self.language = language
84
  self.config_name = config_name
85
  self.label_classes = label_classes
86
  self.task_type = task_type
 
89
  _BRAZILIAN_COURT_DECISIONS_JUDGMENT = {
90
  "task_type": TaskType.SLTC,
91
  "hf_hub_name": "joelito/brazilian_court_decisions",
92
+ "language": "pt",
93
  "input_col": "decision_description",
94
  "label_col": "judgment_label",
95
  "url": "https://github.com/lagefreitas/predicting-brazilian-court-decisions",
 
154
  _GERMAN_ARGUMENT_MINING = {
155
  "task_type": TaskType.SLTC,
156
  "hf_hub_name": "joelito/german_argument_mining",
157
+ "language": "de",
158
  "input_col": "input_sentence",
159
  "label_col": "label",
160
  "url": "https://zenodo.org/record/3936490#.X1ed7ovgomK",
 
184
  _GREEK_LEGAL_CODE_VOLUME = {
185
  "task_type": TaskType.SLTC,
186
  "hf_hub_name": "greek_legal_code",
187
+ "language": "el",
188
  "config_name": "volume",
189
  "input_col": "text",
190
  "label_col": "label",
 
3772
  _LENER_BR = {
3773
  "task_type": TaskType.NER,
3774
  "hf_hub_name": "lener_br",
3775
+ "language": "pt",
3776
  "config_name": "lener_br",
3777
  "input_col": "tokens",
3778
  "label_col": "ner_tags",
 
3820
  _LEGAL_NERO = {
3821
  "task_type": TaskType.NER,
3822
  "hf_hub_name": "joelito/legalnero",
3823
+ "language": "ro",
3824
  "input_col": "words",
3825
  "label_col": "ner",
3826
  "url": "https://zenodo.org/record/4922385",
 
3861
  _GREEK_LEGAL_NER = {
3862
  "task_type": TaskType.NER,
3863
  "hf_hub_name": "joelito/greek_legal_ner",
3864
+ "language": "el",
3865
  "input_col": "words",
3866
  "label_col": "ner",
3867
  "url": "http://legislation.di.uoa.gr/publications?language=en",
 
4015
  features=datasets.Features({
4016
  "input": input_type,
4017
  "label": label_type,
4018
+ "language": datasets.Value("string"),
4019
  }),
4020
  )
4021
 
 
4040
  label = dataset.features[label_col].int2str(label)
4041
  if self.config.task_type == TaskType.SLTC and label not in self.config.label_classes:
4042
  continue # filter out invalid classes to construct ClassLabel
4043
+ if "language" in item.keys():
4044
+ language = item["language"]
4045
+ else:
4046
+ language = self.config.language
4047
  yield id, {
4048
  "input": item[self.config.input_col],
4049
  "label": label,
4050
+ "language": language,
4051
  }