Alphebatize the OCR language dropdown

This commit is contained in:
Micah Lee 2020-01-09 09:47:33 -08:00
parent 2e7a8d2ecd
commit 48fd8ece40
No known key found for this signature in database
GPG key ID: 403C2657CD994F73

View file

@ -43,165 +43,165 @@ class Common(object):
# Languages supported by tesseract # Languages supported by tesseract
self.ocr_languages = { self.ocr_languages = {
"Afrikaans": "ar", "Afrikaans": "ar",
"Albanian": "sqi",
"Amharic": "amh", "Amharic": "amh",
"Arabic": "ara", "Arabic": "ara",
"Arabic script": "Arabic",
"Armenian": "hye",
"Armenian script": "Armenian",
"Assamese": "asm", "Assamese": "asm",
"Azerbaijani": "aze", "Azerbaijani": "aze",
"Azerbaijani (Cyrillic)": "aze_cyrl", "Azerbaijani (Cyrillic)": "aze_cyrl",
"Basque": "eus",
"Belarusian": "bel", "Belarusian": "bel",
"Bengali": "ben", "Bengali": "ben",
"Tibetan Standard": "bod", "Bengali script": "Bengali",
"Bosnian": "bos", "Bosnian": "bos",
"Breton": "bre", "Breton": "bre",
"Bulgarian": "bul", "Bulgarian": "bul",
"Burmese": "mya",
"Canadian Aboriginal script": "Canadian_Aboriginal",
"Catalan": "cat", "Catalan": "cat",
"Cebuano": "ceb", "Cebuano": "ceb",
"Czech": "ces", "Cherokee": "chr",
"Cherokee script": "Cherokee",
"Chinese - Simplified": "chi_sim", "Chinese - Simplified": "chi_sim",
"Chinese - Simplified (vertical)": "chi_sim_vert", "Chinese - Simplified (vertical)": "chi_sim_vert",
"Chinese - Traditional": "chi_tra", "Chinese - Traditional": "chi_tra",
"Chinese - Traditional (vertical)": "chi_tra_vert", "Chinese - Traditional (vertical)": "chi_tra_vert",
"Cherokee": "chr",
"Corsican": "cos", "Corsican": "cos",
"Welsh": "cym", "Croatian": "hrv",
"Cyrillic script": "Cyrillic",
"Czech": "ces",
"Danish": "dan", "Danish": "dan",
"German": "deu", "Devanagari script": "Devanagari",
"Divehi": "div", "Divehi": "div",
"Dutch": "nld",
"Dzongkha": "dzo", "Dzongkha": "dzo",
"Greek": "ell",
"English": "eng", "English": "eng",
"English, Middle (1100-1500)": "enm", "English, Middle (1100-1500)": "enm",
"Esperanto": "epo", "Esperanto": "epo",
"Estonian": "est", "Estonian": "est",
"Basque": "eus", "Ethiopic script": "Ethiopic",
"Faroese": "fao", "Faroese": "fao",
"Persian": "fas",
"Filipino": "fil", "Filipino": "fil",
"Finnish": "fin", "Finnish": "fin",
"French": "fra", "Fraktur script": "Fraktur",
"Frankish": "frk", "Frankish": "frk",
"French": "fra",
"French, Middle (ca.1400-1600)": "frm", "French, Middle (ca.1400-1600)": "frm",
"Frisian (Western)": "fry", "Frisian (Western)": "fry",
"Gaelic (Scots)": "gla", "Gaelic (Scots)": "gla",
"Irish": "gle",
"Galician": "glg", "Galician": "glg",
"Gujarati": "guj",
"Hatian": "hat",
"Hebrew": "heb",
"Hindi": "hin",
"Croatian": "hrv",
"Hungarian": "hun",
"Armenian": "hye",
"Inuktitut": "iku",
"Indonesian": "ind",
"Icelandic": "isl",
"Italian": "ita",
"Italian - Old": "ita_old",
"Javanese": "jav",
"Japanese": "jpn",
"Japanese (vertical)": "jpn_vert",
"Kannada": "kan",
"Georgian": "kat", "Georgian": "kat",
"Old Georgian": "kat_old",
"Kazakh": "kaz",
"Khmer": "khm",
"Kyrgyz": "kir",
"Korean": "kor",
"Korean (vertical)": "kor_vert",
"Kurdish (Arabic)": "kur_ara",
"Lao": "lao",
"Latin": "lat",
"Latvian": "lav",
"Lithuanian": "lit",
"Luxembourgish": "ltz",
"Malayalam": "mal",
"Marathi": "mar",
"Macedonian": "mkd",
"Maltese": "mlt",
"Mongolian": "mon",
"Maori": "mri",
"Malay": "msa",
"Burmese": "mya",
"Nepali": "nep",
"Dutch": "nld",
"Norwegian": "nor",
"Occitan (post 1500)": "oci",
"Oriya": "ori",
"script and orientation": "osd",
"Punjabi": "pan",
"Polish": "pol",
"Portuguese": "por",
"Pashto": "pus",
"Quechua": "que",
"Romanian": "ron",
"Russian": "rus",
"Sanskrit": "san",
"Sinhala": "sin",
"Slovakian": "slk",
"Slovenian": "slv",
"Sindhi": "snd",
"Spanish": "spa",
"Spanish, Castilian - Old": "spa_old",
"Albanian": "sqi",
"Serbian": "srp",
"Serbian (Latin)": "srp_latn",
"Sundanese": "sun",
"Swahili": "swa",
"Swedish": "swe",
"Syriac": "syr",
"Tamil": "tam",
"Tatar": "tat",
"Telugu": "tel",
"Tajik": "tgk",
"Thai": "tha",
"Tigrinya": "tir",
"Tonga": "ton",
"Turkish": "tur",
"Uyghur": "uig",
"Ukrainian": "ukr",
"Urdu": "urd",
"Uzbek": "uzb",
"Uzbek (Cyrillic)": "uzb_cyrl",
"Vietnamese": "vie",
"Yiddish": "yid",
"Yoruba": "yor",
"Arabic script": "Arabic",
"Armenian script": "Armenian",
"Bengali script": "Bengali",
"Canadian Aboriginal script": "Canadian_Aboriginal",
"Cherokee script": "Cherokee",
"Cyrillic script": "Cyrillic",
"Devanagari script": "Devanagari",
"Ethiopic script": "Ethiopic",
"Fraktur script": "Fraktur",
"Georgian script": "Georgian", "Georgian script": "Georgian",
"German": "deu",
"Greek": "ell",
"Greek script": "Greek", "Greek script": "Greek",
"Gujarati": "guj",
"Gujarati script": "Gujarati", "Gujarati script": "Gujarati",
"Gurmukhi script": "Gurmukhi", "Gurmukhi script": "Gurmukhi",
"Hangul script": "Hangul",
"Hangul (vertical) script": "Hangul_vert",
"Han - Simplified script": "HanS", "Han - Simplified script": "HanS",
"Han - Simplified (vertical) script": "HanS_vert", "Han - Simplified (vertical) script": "HanS_vert",
"Han - Traditional script": "HanT", "Han - Traditional script": "HanT",
"Han - Traditional (vertical) script": "HanT_vert", "Han - Traditional (vertical) script": "HanT_vert",
"Hangul script": "Hangul", "Hatian": "hat",
"Hangul (vertical) script": "Hangul_vert", "Hebrew": "heb",
"Hebrew script": "Hebrew", "Hebrew script": "Hebrew",
"Hindi": "hin",
"Hungarian": "hun",
"Icelandic": "isl",
"Indonesian": "ind",
"Inuktitut": "iku",
"Irish": "gle",
"Italian": "ita",
"Italian - Old": "ita_old",
"Japanese": "jpn",
"Japanese script": "Japanese", "Japanese script": "Japanese",
"Japanese (vertical)": "jpn_vert",
"Japanese (vertical) script": "Japanese_vert", "Japanese (vertical) script": "Japanese_vert",
"Javanese": "jav",
"Kannada": "kan",
"Kannada script": "Kannada", "Kannada script": "Kannada",
"Kazakh": "kaz",
"Khmer": "khm",
"Khmer script": "Khmer", "Khmer script": "Khmer",
"Korean": "kor",
"Korean (vertical)": "kor_vert",
"Kurdish (Arabic)": "kur_ara",
"Kyrgyz": "kir",
"Lao": "lao",
"Lao script": "Lao", "Lao script": "Lao",
"Latin": "lat",
"Latin script": "Latin", "Latin script": "Latin",
"Latvian": "lav",
"Lithuanian": "lit",
"Luxembourgish": "ltz",
"Macedonian": "mkd",
"Malayalam": "mal",
"Malayalam script": "Malayalam", "Malayalam script": "Malayalam",
"Malay": "msa",
"Maltese": "mlt",
"Maori": "mri",
"Marathi": "mar",
"Mongolian": "mon",
"Myanmar script": "Myanmar", "Myanmar script": "Myanmar",
"Nepali": "nep",
"Norwegian": "nor",
"Occitan (post 1500)": "oci",
"Old Georgian": "kat_old",
"Oriya (Odia) script": "Oriya", "Oriya (Odia) script": "Oriya",
"Oriya": "ori",
"Pashto": "pus",
"Persian": "fas",
"Polish": "pol",
"Portuguese": "por",
"Punjabi": "pan",
"Quechua": "que",
"Romanian": "ron",
"Russian": "rus",
"Sanskrit": "san",
"script and orientation": "osd",
"Serbian (Latin)": "srp_latn",
"Serbian": "srp",
"Sindhi": "snd",
"Sinhala script": "Sinhala", "Sinhala script": "Sinhala",
"Sinhala": "sin",
"Slovakian": "slk",
"Slovenian": "slv",
"Spanish, Castilian - Old": "spa_old",
"Spanish": "spa",
"Sundanese": "sun",
"Swahili": "swa",
"Swedish": "swe",
"Syriac script": "Syriac", "Syriac script": "Syriac",
"Syriac": "syr",
"Tajik": "tgk",
"Tamil script": "Tamil", "Tamil script": "Tamil",
"Tamil": "tam",
"Tatar": "tat",
"Telugu script": "Telugu", "Telugu script": "Telugu",
"Telugu": "tel",
"Thaana script": "Thaana", "Thaana script": "Thaana",
"Thai script": "Thai", "Thai script": "Thai",
"Thai": "tha",
"Tibetan script": "Tibetan", "Tibetan script": "Tibetan",
"Tibetan Standard": "bod",
"Tigrinya": "tir",
"Tonga": "ton",
"Turkish": "tur",
"Ukrainian": "ukr",
"Urdu": "urd",
"Uyghur": "uig",
"Uzbek (Cyrillic)": "uzb_cyrl",
"Uzbek": "uzb",
"Vietnamese script": "Vietnamese", "Vietnamese script": "Vietnamese",
"Vietnamese": "vie",
"Welsh": "cym",
"Yiddish": "yid",
"Yoruba": "yor",
} }
# Load settings # Load settings