interscript 0.1.1 → 0.1.2

data/maps/icao-bul-Cyrl-Latn-9303.yaml

@@ -0,0 +1,123 @@
+---
+authority_id: icao
+id: 9303
+language: bul
+source_script: Cyrl
+destination_script: Latn
+name: ICAO Doc 9303 Machine Readable Travel Documents Part 3 -- Bulgarian Cyrillic to Latin
+url: https://www.icao.int/publications/pages/publication.aspx?docnum=9303
+creation_date: 2015
+description: |
+  Part 3 defines specifications that are common to TD1, TD2 and TD3
+  size machine readable travel documents (MRTDs) including those
+  necessary for global interoperability using visual inspection and
+  machine readable (optical character recognition) means.
+
+  Since only Latin-alphabet characters are allowed in the VIZ, if
+  mandatory data elements are in a national language that does not use
+  the Latin alphabet, a transcription or transliteration shall also be
+  provided.
+
+  This document defines the transliteration mappings used to produce
+  this transcription or transliteration.
+tests:
+  - source:
+    expected:
+map:
+  # https://www.icao.int/publications/Documents/9303_p3_cons_en.pdf
+  characters:
+    "\u0027": ""    # '
+    "\u0410": "A"   # А
+    "\u0411": "B"   # Б
+    "\u0414": "D"   # Д
+    "\u0401": "E"   # Ё
+    "\u0415": "E"   # Е
+    "\u042D": "E"   # Э
+    "\u0424": "F"   # Ф
+    "\u0413": "G"   # Г
+    "\u0418": "I"   # И
+    "\u0419": "I"   # Й
+    "\u041A": "K"   # К
+    "\u041B": "L"   # Л
+    "\u041C": "M"   # М
+    "\u041D": "N"   # Н
+    "\u041E": "O"   # О
+    "\u041F": "P"   # П
+    "\u0420": "R"   # Р
+    "\u0421": "S"   # С
+    "\u0422": "T"   # Т
+    "\u0423": "U"   # У
+    "\u0412": "V"   # В
+    "\u042B": "Y"   # Ы
+    "\u0417": "Z"   # З
+    "\u0427": "CH"  # Ч
+    "\u042F": "IA"  # Я
+    "\u042E": "IU"  # Ю
+    "\u0425": "KH"  # Х
+    "\u0428": "SH"  # Ш
+    "\u0429": "SHT" # Щ
+    "\u0426": "TS"  # Ц
+    "\u0416": "ZH"  # Ж
+    "\u0490": "G"   # Ґ
+    "\u040E": "U"   # Ў
+    "\u046A": "U"   # Ѫ
+    "\u0402": "D"   # Ђ
+    "\u0405": "DZ"  # Ѕ
+    "\u0408": "J"   # Ј
+    "\u0409": "LJ"  # Љ
+    "\u040A": "NJ"  # Њ
+    "\u04BA": "C"   # Һ
+    "\u040F": "DZ"  # Џ
+    "\u0404": "IE"  # Є
+    "\u0407": "I"   # Ї
+    "\u0403": "G"   # Ѓ
+    "\u0027": ""    # '
+
+    "\u0430": "a"     # а
+    "\u0431": "b"     # б
+    "\u0434": "d"     # д
+    "\u0451": "e"     # ё
+    "\u0435": "e"     # e
+    "\u044D": "e"     # э
+    "\u0444": "f"     # ф
+    "\u0433": "g"     # г
+    "\u0438": "i"     # и
+    "\u0439": "i"     # й
+    "\u043A": "k"     # к
+    "\u043B": "l"     # л
+    "\u043C": "m"     # м
+    "\u043D": "n"     # н
+    "\u043E": "o"     # о
+    "\u043F": "p"     # п
+    "\u0440": "r"     # р
+    "\u0441": "s"     # с
+    "\u0442": "t"     # т
+    "\u0443": ""     # у
+    "\u0432": "v"     # в
+    "\u044B": "y"     # ы
+    "\u0437": "z"     # з
+    "\u0447": "ch"    # ч
+    "\u044F": "ia"    # я
+    "\u044E": "i"    # ю
+    "\u0445": "kh"    # х
+    "\u0448": "sh"    # ш
+    "\u0449": "sht"    # щ
+    "\u0446": "ts"    # ц
+    "\u0436": "zh"    # ж
+    "\u0491": "g"     # ґ
+    "\u045E": ""     # ў
+    "\u046B": ""     # ѫ
+    "\u0452": "d"     # ђ
+    "\u0455": "dz"    # ѕ
+    "\u0458": "j"     # ј
+    "\u0459": "lj"    # љ
+    "\u045A": "nj"    # њ
+    "\u04BB": "c"     # һ
+    "\u045F": "dz"    # џ
+    "\u0454": "ie"    # є
+    "\u0457": "i"     # ї
+    "\u0453": "g"     # ѓ
+
+
+
+

data/maps/icao-gre-Grek-Latn-9303.yaml

@@ -0,0 +1,101 @@
+---
+authority_id: icao
+id: 9303
+language: gre
+source_script: Grek
+destination_script: Latn
+name: ICAO Doc 9303 Machine Readable Travel Documents Part 3 -- Greek to Latin
+url: https://www.icao.int/publications/pages/publication.aspx?docnum=9303
+creation_date: 2015
+description: |
+  Part 3 defines specifications that are common to TD1, TD2 and TD3
+  size machine readable travel documents (MRTDs) including those
+  necessary for global interoperability using visual inspection and
+  machine readable (optical character recognition) means.
+
+  Since only Latin-alphabet characters are allowed in the VIZ, if
+  mandatory data elements are in a national language that does not use
+  the Latin alphabet, a transcription or transliteration shall also be
+  provided.
+
+  This document defines the transliteration mappings used to produce
+  this transcription or transliteration.
+tests:
+  - source:
+    expected:
+map:
+  # https://en.wikipedia.org/wiki/Romanization_of_Greek
+
+  characters:
+    "\u0027": "",
+    "\u0386": "A",  # Ά
+    "\u0391": "A",  # Α
+    "\u0392": "B",  # Β
+    "\u0393": "G",  # Γ
+    "\u0394": "D",  # Δ
+    "\u0395": "E",  # Ε
+    "\u0396": "Z",  # Ζ
+    "\u0397": "I",  # Η
+    "\u0398": "TH", # Θ
+    "\u0399": "I",  # Ι
+    "\u039A": "K",  # Κ
+    "\u039B": "L",  # Λ
+    "\u039C": "M",  # Μ
+    "\u039D": "N",  # Ν
+    "\u039E": "X",  # Ξ
+    "\u039F": "O",  # Ο
+    "\u03A0": "P",  # Π
+    "\u03A1": "P",  # Ρ
+    "\u03A3": "S",  # Σ
+    "\u03C2": "s",  # ς
+    "\u03A4": "T",  # Τ
+    "\u03A5": "Y",  # Υ
+    "\u03A7": "CH", # Χ
+    "\u03A8": "PS", # Ψ
+    "\u03A9": "O",  # Ω
+    "\u0388": "E",  # Έ
+    "\u0389": "I",  # Ή
+    "\u038A": "I",  # Ί
+    "\u038C": "O",  # Ό
+    "\u038E": "Y",  # Ύ
+    "\u038F": "O",  # Ώ
+    "\u03AA": "I",  # Ϊ
+    "\u03AB": "Y",  # Ϋ
+
+    "\u03AC": "a",  # ά
+    "\u03B1": "a",  # α
+    "\u03B2": "b",  # β
+    "\u03B3": "g",  # γ
+    "\u03B4": "d",  # δ
+    "\u03B5": "e",  # ε
+    "\u03B6": "z",  # ζ
+    "\u03B7": "i",  # η
+    "\u03B8": "th", # θ
+    "\u03B9": "i",  # ι
+    "\u03BA": "k",  # κ
+    "\u03BB": "l",  # λ
+    "\u03BC": "m",  # μ
+    "\u03BD": "n",  # ν
+    "\u03BE": "x",  # ξ
+    "\u03BF": "o",  # ο
+    "\u03C0": "p",  # π
+    "\u03C1": "p",  # ρ
+    "\u03C3": "s",  # σ
+
+    "\u03C4": "t",  # τ
+    "\u03C5": "y",  # υ
+    "\u03C7": "ch", # χ
+    "\u03C8": "ps", # ψ
+    "\u03C9": "o",  # ω
+    "\u03AD": "e",  # έ
+    "\u03AE": "i",  # ή
+    "\u03AF": "i",  # ί
+    "\u03CC": "o",  # ό
+    "\u03CD": "y",  # ύ
+    "\u03CE": "o",  # ώ
+    "\u03CA": "i",  # ϊ
+    "\u03CB": "y",  # ϋ
+
+
+
+

data/maps/icao-heb-Hebr-Latn-9303.yaml

@@ -0,0 +1,157 @@
+---
+authority_id: icao
+id: 9303
+language: heb
+source_script: Hebr
+destination_script: Latn
+name: ICAO Doc 9303 Machine Readable Travel Documents Part 3 -- Hebrew to Latin
+url: https://www.icao.int/publications/pages/publication.aspx?docnum=9303
+creation_date: 2015
+description: |
+  Part 3 defines specifications that are common to TD1, TD2 and TD3
+  size machine readable travel documents (MRTDs) including those
+  necessary for global interoperability using visual inspection and
+  machine readable (optical character recognition) means.
+
+  Since only Latin-alphabet characters are allowed in the VIZ, if
+  mandatory data elements are in a national language that does not use
+  the Latin alphabet, a transcription or transliteration shall also be
+  provided.
+
+  This document defines the transliteration mappings used to produce
+  this transcription or transliteration.
+tests:
+  - source:
+    expected:
+map:
+  # https://en.wikipedia.org/wiki/Romanization_of_Hebrew
+
+  characters:
+    "\u0027": "",                 # '
+
+    # Consonants
+    "\u05D0":             "",     # א
+    "\u05D1":             "V",    # ב
+    "\u05D1\u05BC": "B",    # בּ
+    "\uFB31":             "B",    # בּ
+    "\u05D2":             "G",    # ג
+    "\u05D2\u05BC": "G",    # גּ
+    "\uFB32":             "G",    # גּ‬‬
+    "\u05D2\u05F3": "J",    # ג׳
+    "\u05D3":             "D",    # ד
+    "\u05D3\u05BC": "D",    # דּ
+    "\uFB33":             "D",    # דּ
+    "\u05D3\u05F3": "DH",   # ד׳
+    "\u05D4":             "H",    # ה
+    "\u05D4\u05BC": "H",    # הּ
+    "\uFB34":             "H",    # הּ
+    "\u05D5":             "V",    # ו‬
+    "\u05D5\u202C": "V",    # ו‬
+    "\u05D5\u05BC": "V",    # וּ
+    # "\uFB35":             "V",  # וּ  # To vowels "U"
+    "\u05D6":             "Z",    # ז
+    "\u05D6\u05BC": "Z",    # זּ
+    "\uFB36":             "Z",    # זּ‬
+    "\u05D6\u05F3": "ZH",   # ז׳
+    "\u05D7":             "CH",   # ח
+    "\u05D8":             "T",    # ט
+    "\u05D8\u05BC": "T",    # טּ
+    "\uFB38":             "T",    # טּ
+    "\u05D9":             "Y",    # י
+    "\u05D9\u05BC": "Y",    # יּ
+    "\u05D9\u05BC" +
+    "\u202C":             "Y",    # יּ‬
+    "\uFB39":             "Y",    # יּ‬
+    "\u05DB":             "CH",   # כ
+    "\u05DB\u05BC": "CH",   # כּ
+    "\u05DB\u05BC" +
+    "\u202C":             "CH",   # כּ
+    "\uFB3B":             "C",    # כּ
+    "\u05DA":             "CH",   # ך
+    "\u05DA\u05BC": "CH",   # ךּ
+    "\u05DA\u05BC" +
+    "\u202C":             "CH",   # ךּ‬
+    "\uFB3A":             "CH",   # ךּ
+    "\u05DC":             "L",    # ל‬
+    "\u05DC\u05BC": "L",    # לּ
+    "\uFB3C":             "L",    # לּ
+    "\u05DD":             "M",    # ם
+    "\u05DE":             "M",    # מ‬
+    "\u05DE\u05BC": "M",    # מּ
+    "\uFB3E":             "M",    # מּ‬
+    "\u05DF":             "N",    # ן
+    "\u05E0":             "N",    # נ
+    "\u05E0\u05BC": "N",    # נּ
+    "\uFB40":             "N",    # נּ
+    "\u05E1":             "S",    # ס
+    "\u05E1\u05BC": "S",    # סּ
+    "\uFB41":             "S",    # סּ
+    "\u05E2":             "",     # ע
+    "\u05E3":             "F",    # ף
+    "\u05E3\u05BC": "P",    # Possible problem u05BC # ףּ
+    "\uFB43":             "P",    # ףּ
+    "\u05E4":             "F",    # פ‬
+    "\u05E4\u05BC": "P",    # פּ
+    "\uFB44":             "P",    # פּ
+    "\u05E5":             "TZ",   # ץ
+    "\u05E5\u05F3": "TSH",  # Possible problem u05F3  # ץ׳
+    "\u05E6":             "TZ",   # צ‬
+    "\u05E6\u05BC": "TZ",   # צּ
+    "\uFB46":             "TZ",   # צּ‬
+    "\u05E6\u05F3": "TSH",  # Possible problem u05F3  # צ׳
+    "\u05E7":             "Q",    # ק
+    "\u05E7\u05BC": "Q",    # קּ
+    "\uFB47":             "Q",    # קּ‬
+    "\u05E8":             "R",    # ר
+    "\u05E8\u05BC": "R",    # רּ
+    "\uFB48":             "R",    # רּ
+    "\u05E9":             "S",    # ש
+    "\u05E9\u05BC": "S",    # שּ
+    "\uFB49":             "S",    # שּ‬
+    "\u05E9\u05C2" +
+    "\u202C":             "S",    # שׂ
+    "\uFB2B":             "S",    # שׂ
+    "\u05E9\u05C1": "SH",   # שׁ
+    "\uFB2A":             "SH",   # שׁ
+    "\u05E9\u05BC" +
+    "\u05C2\u202C": "S",    # שּׂ‬
+    "\uFB2D":             "S",    # שּׂ
+    "\u05EA":             "T",    # ת
+    "\u05EA\u05BC": "T",    # תּ
+    "\uFB4A":             "T",    # תּ
+    "\u05EA\u05F3": "T",    # ת׳
+
+    # Niqqud vowels
+    "\u05B0":             "E",    # ( ְ‬ )
+    "\u05B1":             "E",    # ( ֱ )
+    "\u05B2":             "A",    # ( ֲ )
+    "\u05B3":             "O",    # ( ֲ )
+    "\u05B4":             "I",    # ( ִ )
+    "\u05B5":             "E",    # ( ֵ )
+    "\u05B6":             "E",    # ( ֶ )
+    "\u05B7":             "A",    # ( ַ )
+    "\u05B8":             "O",    # ( ָ ) # It could be "A" too
+    "\u05B9":             "O",    # ( ֹ )
+    "\u05BB":             "U",    # ( ֻ )
+    "\u05D5\u05BC": "U",    # ( וּ )
+    "\uFB35":             "U",    # ( וּ )
+
+    # Diphthongs
+    "\u05B5\u05D9": "EI",   # ( ֵי )
+    "\u05B6\u05D9": "EI",   # ( ֶי )
+    "\u05B7\u05D9": "AI",   # ( ַי )
+    "\u05B7\u05D9\u05B0":             "AI",   # ( ַיְ )
+    "\u05B7\u05D9\u05B0\u202C": "AI",   # ( ַיְ‬ )
+    "\u05B8\u05D9": "AI",   # ( ָי )
+    "\u05B8\u05D9\u202C":             "AI",   # ( ָי‬ )
+    "\u05B8\u05D9\u05B0":             "AI",   # ( ָיְ )
+    "\u05B8\u05D9\u05B0\u202C": "AI",   # ( ָיְ‬ )
+    "\u05B9\u05D9": "OI",   # ( ֹי )
+    "\u05B9\u05D9\u05B0":             "OI",   # ( ֹיְ )
+    "\u05B9\u05D9\u05B0\u202C": "OI",   # ( ֹיְ‬ )
+    "\u05BB\u05D9": "UI",   # ( ֻי )
+    "\u05BB\u05D9\u05B0":             "UI",   # ( ֻיְ )
+    "\u05BB\u05D9\u05B0\u202C": "UI",   # ( ֻיְ‬ )
+    "\u05D5\u05BC\u05D9":             "UI",   # ( וּי )
+    "\u05D5\u05BC\u05D9\u05B0": "UI",   # ( וּיְ )
+    "\u05D5\u05BC\u05D9\u05B0\u202C":             "UI",   # ( וּיְ‬ )

data/maps/icao-mkd-Cyrl-Latn-9303.yaml

@@ -0,0 +1,118 @@
+---
+authority_id: icao
+id: 9303
+language: mkd
+source_script: Cyrl
+destination_script: Latn
+name: ICAO Doc 9303 Machine Readable Travel Documents Part 3 -- Macedonian Cyrillic to Latin
+url: https://www.icao.int/publications/pages/publication.aspx?docnum=9303
+creation_date: 2015
+description: |
+  Part 3 defines specifications that are common to TD1, TD2 and TD3
+  size machine readable travel documents (MRTDs) including those
+  necessary for global interoperability using visual inspection and
+  machine readable (optical character recognition) means.
+
+  Since only Latin-alphabet characters are allowed in the VIZ, if
+  mandatory data elements are in a national language that does not use
+  the Latin alphabet, a transcription or transliteration shall also be
+  provided.
+
+  This document defines the transliteration mappings used to produce
+  this transcription or transliteration.
+tests:
+  - source:
+    expected:
+map:
+  # https://www.icao.int/publications/Documents/9303_p3_cons_en.pdf
+  characters:
+    "\u0027": ""     # '
+    "\u0410": "A"    # А
+    "\u0411": "B"    # Б
+    "\u0414": "D"    # Д
+    "\u0401": "E"    # Ё
+    "\u0415": "E"    # Е
+    "\u042D": "E"    # Э
+    "\u0424": "F"    # Ф
+    "\u0413": "G"    # Г
+    "\u0418": "I"    # И
+    "\u0419": "I"    # Й
+    "\u041A": "KJ"   # К
+    "\u041B": "L"    # Л
+    "\u041C": "M"    # М
+    "\u041D": "N"    # Н
+    "\u041E": "O"    # О
+    "\u041F": "P"    # П
+    "\u0420": "R"    # Р
+    "\u0421": "S"    # С
+    "\u0422": "T"    # Т
+    "\u0423": "U"    # У
+    "\u0412": "V"    # В
+    "\u042B": "Y"    # Ы
+    "\u0417": "Z"    # З
+    "\u0427": "CH"   # Ч
+    "\u042F": "IA"   # Я
+    "\u042E": "IU"   # Ю
+    "\u0425": "H"    # Х
+    "\u0428": "SH"   # Ш
+    "\u0429": "SHCH" # Щ
+    "\u0426": "C"    # Ц
+    "\u0416": "ZH"   # Ж
+    "\u0490": "G"    # Ґ
+    "\u040E": "U"    # Ў
+    "\u046A": "U"    # Ѫ
+    "\u0402": "D"    # Ђ
+    "\u0405": "DZ"   # Ѕ
+    "\u0408": "J"    # Ј
+    "\u0409": "LJ"   # Љ
+    "\u040A": "NJ"   # Њ
+    "\u04BA": "C"    # Һ
+    "\u040F": "DJ"   # Џ
+    "\u0404": "IE"   # Є
+    "\u0407": "I"    # Ї
+    "\u0403": "GJ"   # Ѓ
+
+    "\u0430": "a"     # а
+    "\u0431": "b"     # б
+    "\u0434": "d"     # д
+    "\u0451": "e"     # ё
+    "\u0435": "e"     # e
+    "\u044D": "e"     # э
+    "\u0444": "f"     # ф
+    "\u0433": "g"     # г
+    "\u0438": "i"     # и
+    "\u0439": "i"     # й
+    "\u043A": "kj"    # к
+    "\u043B": "l"     # л
+    "\u043C": "m"     # м
+    "\u043D": "n"     # н
+    "\u043E": "o"     # о
+    "\u043F": "p"     # п
+    "\u0440": "r"     # р
+    "\u0441": "s"     # с
+    "\u0442": "t"     # т
+    "\u0443": ""      # у
+    "\u0432": "v"     # в
+    "\u044B": "y"     # ы
+    "\u0437": "z"     # з
+    "\u0447": "ch"    # ч
+    "\u044F": "ia"    # я
+    "\u044E": "i"     # ю
+    "\u0445": "h"     # х
+    "\u0448": "sh"    # ш
+    "\u0449": "shch"  # щ
+    "\u0446": "c"     # ц
+    "\u0436": "zh"    # ж
+    "\u0491": "g"     # ґ
+    "\u045E": ""      # ў
+    "\u046B": ""      # ѫ
+    "\u0452": "d"     # ђ
+    "\u0455": "dz"    # ѕ
+    "\u0458": "j"     # ј
+    "\u0459": "lj"    # љ
+    "\u045A": "nj"    # њ
+    "\u04BB": "c"     # һ
+    "\u045F": "dj"    # џ
+    "\u0454": "ie"    # є
+    "\u0457": "i"     # ї
+    "\u0453": "gj"    # ѓ