nllb-vutuka-translation / flores200_codes.py
Svngoku's picture
Try to make it work
468fa7e
codes_as_string = '''
Afrikaans afr_Latn
Akan aka_Latn
Amharic amh_Ethi
Bambara bam_Latn
Bemba bem_Latn
English eng_Latn
Ewe ewe_Latn
Fijian fij_Latn
Finnish fin_Latn
Fon fon_Latn
French fra_Latn
Nigerian Fulfulde fuv_Latn
Haitian Creole hat_Latn
Hausa hau_Latn
Igbo ibo_Latn
Kikuyu kik_Latn
Kinyarwanda kin_Latn
Kimbundu kmb_Latn
Kikongo kon_Latn
Lingala lin_Latn
Luba-Kasai lua_Latn
Ganda lug_Latn
Luo luo_Latn
Mossi mos_Latn
Maori mri_Latn
Portuguese por_Latn
Russian rus_Cyrl
Sango sag_Latn
Sinhala sin_Sinh
Shona sna_Latn
Somali som_Latn
Swahili swh_Latn
Tigrinya tir_Ethi
Tswana tsn_Latn
Tsonga tso_Latn
Tumbuka tum_Latn
Twi twi_Latn
Central Atlas Tamazight tzm_Tfng
Umbundu umb_Latn
Wolof wol_Latn
Xhosa xho_Latn
Yoruba yor_Latn
Zulu zul_Latn'''
codes_as_string = codes_as_string.split('\n')
flores_codes = {}
for code in codes_as_string:
lang, lang_code = code.split('\t')
flores_codes[lang] = lang_code