Update multiconer_v2.py

multiconer_v2.py

@@ -46,63 +46,8 @@ For more details see https://multiconer.github.io/
 * Shervin Malmasi, Anjie Fang, Besnik Fetahu, Sudipta Kar, Oleg Rokhlenko. 2022. SemEval-2022 Task 11: Multilingual Complex Named Entity Recognition (MultiCoNER).
 
 """
-
 _URL = "https://huggingface.co/datasets/MultiCoNER/multiconer_v2/raw/main"
 
-_BN_TRAINING_FILE = "BN-Bangla/bn_train.conll"
-_BN_DEV_FILE = "BN-Bangla/bn_dev.conll"
-_BN_TEST_FILE = "BN-Bangla/bn_test.conll"
-
-_DE_TRAINING_FILE = "DE-German/de_train.conll"
-_DE_DEV_FILE = "DE-German/de_dev.conll"
-_DE_TEST_FILE = "DE-German/de_test.conll"
-
-_EN_TRAINING_FILE = "EN-English/en_train.conll"
-_EN_DEV_FILE = "EN-English/en_dev.conll"
-_EN_TEST_FILE = "EN-English/en_test.conll"
-
-_ES_TRAINING_FILE = "ES-Spanish/es_train.conll"
-_ES_DEV_FILE = "ES-Spanish/es_dev.conll"
-_ES_TEST_FILE = "ES-Spanish/es_test.conll"
-
-_FA_TRAINING_FILE = "FA-Farsi/fa_train.conll"
-_FA_DEV_FILE = "FA-Farsi/fa_dev.conll"
-_FA_TEST_FILE = "FA-Farsi/fa_test.conll"
-
-_FR_TRAINING_FILE = "FR-French/fr_train.conll"
-_FR_DEV_FILE = "FR-French/fr_dev.conll"
-_FR_TEST_FILE = "FR-French/fr_test.conll"
-
-_HI_TRAINING_FILE = "HI-Hindi/hi_train.conll"
-_HI_DEV_FILE = "HI-Hindi/hi_dev.conll"
-_HI_TEST_FILE = "HI-Hindi/hi_test.conll"
-
-_IT_TRAINING_FILE = "IT-Italian/it_train.conll"
-_IT_DEV_FILE = "IT-Italian/it_dev.conll"
-_IT_TEST_FILE = "IT-Italian/it_test.conll"
-
-_MULTI_TRAINING_FILE = "MULTI-Multilingual/multi_train.conll"
-_MULTI_DEV_FILE = "MULTI-Multilingual/multi_dev.conll"
-_MULTI_TEST_FILE = "MULTI-Multilingual/multi_test.conll"
-
-_PT_TRAINING_FILE = "PT-Portuguese/pt_train.conll"
-_PT_DEV_FILE = "PT-Portuguese/pt_dev.conll"
-_PT_TEST_FILE = "PT-Portuguese/pt_test.conll"
-
-_SV_TRAINING_FILE = "SV-Swedish/sv_train.conll"
-_SV_DEV_FILE = "SV-Swedish/sv_dev.conll"
-_SV_TEST_FILE = "SV-Swedish/sv_test.conll"
-
-_UK_TRAINING_FILE = "UK-Ukrainian/uk_train.conll"
-_UK_DEV_FILE = "UK-Ukrainian/uk_dev.conll"
-_UK_TEST_FILE = "UK-Ukrainian/uk_test.conll"
-
-_ZH_TRAINING_FILE = "ZH-Chinese/zh_train.conll"
-_ZH_DEV_FILE = "ZH-Chinese/zh_dev.conll"
-_ZH_TEST_FILE = "ZH-Chinese/zh_test.conll"
-
-# label_vs_code_map[self.config.name].upper()-{code_vs_lang_map[label_vs_code_map[self.config.name]]}/{code_vs_lang_map[label_vs_code_map[self.config.name]]}_train.conll
-
 code_vs_lang_map = {"en": "English",
                     "es": "Spanish",
                     "pt": "Portuguese",
@@ -183,6 +128,7 @@ class MultiCoNER2(datasets.GeneratorBasedBuilder):
                     "id": datasets.Value("string"),
                     "sample_id": datasets.Value("string"),
                     "tokens": datasets.Sequence(datasets.Value("string")),
+                    "ner_tags_index": datasets.Sequence(datasets.Value("string")),
                     "ner_tags": datasets.Sequence(
                         datasets.features.ClassLabel(
                             names=[