interscript 0.1.2 → 0.1.7

data/maps/odni-urd-Arab-Latn-2015.yaml

@@ -0,0 +1,221 @@
+---
+authority_id: odni
+id: 2015
+language: iso-639-2:urd
+source_script: Arab
+destination_script: Latn
+name: Office of the Director Of National Intelligence Hindi Urdu Personal Names 2015 System, ICS-630-01 Annex F
+#url:
+creation_date: 2015
+confirmation_date: 2015
+description: |
+  This system is the Intelligence Community (IC) standard for the transliteration of names
+  from Hindi and Urdu that will be applied to all final written reports and products for IC
+  consumers. It is not intended to eliminate variations of a name that can contribute
+  forensic information. Rather, it is to provide an IC standard Romanized (English)
+  transliteration from standard Hindi and Urdu that can then be linked to forensic
+  information in ways that will help identify the referent of the name.
+
+  There are typically a number of ways that names can be Romanized from either
+  Devanagari (Hindi) or modified Arabic (Urdu) scripts. Ambiguities can result from the
+  Romanization of Hindi and Urdu names for several reasons, including the fact that some
+  sounds in South Asian languages (e.g., retroflex consonants, voiced aspirates) have no
+  equivalent in English or other European languages. In the case of Urdu, as in the original
+  Arabic source of many Islamic names, short vowel markings, double consonant marks
+  and other diacritics that would clearly distinguish the name are almost always omitted
+  from standard written texts. And many Islamic names of Arabic or Persian origin reflect
+  spelling distinctions from those languages that are lost in modern Urdu pronunciation
+  (e.g., three distinct Arabic letters all represent the identical sound [s] for Urdu speakers);
+  transliterations might either maintain those spelling distinctions or ignore them entirely.
+  And, as in the Arabic source, names containing the Arabic definite article ‘al’ (‘ul’) show
+  anticipatory assimilation in pronunciation (e.g., Shams al Din > Shamsuddin);
+  transliterations may either reflect spelling or pronunciation in such cases.
+
+  Because Hindi and Urdu overlap so extensively, it is desirable to correlate Hindi and
+  Urdu transliterations as much as possible. In the area of names, this can become
+  problematic when the Urdu spelling accurately reflects original Arabic spelling while the
+  Hindi spelling was phonetically-based, in essence, already a transliteration. This would
+  argue for ignoring Arabic/Urdu spelling distinctions not reflected in pronunciation in
+  either Urdu or Hindi (e.g., letter sin versus letter sad) while maintaining Arabic/Urdu
+  spelling distinctions like Hindi also maintains (e.g., qaf as in Qutubbin versus kaf)).
+  However, this is not always possible (see Hindi va versus Urdu wau).
+
+  In cases where an individual’s name has already been transliterated, that is to be indicated
+  – as found – in parentheses immediately following its rendition in the transliteration
+  standard (e.g., Muhammad Khulud (Mohamed Khulood)). In addition, if the original
+  Devanagari or Arabic-script spelling is known, that spelling should also appear in
+  parentheses following the name, if possible, following best practices of the issuing
+  organization and taking into consideration information system capabilities. This
+  convention is designed to ensure that vital forensic information is not lost.
+
+  For names of persons who are known to not be part of the Hindi- or Urdu-speaking
+  community, use the relevant IC transliteration standard for names from that language
+  (e.g., Mikhail, Yitzhak). A translator’s note may be used to clarify the known origin of
+  the person. Spell names of individuals from languages that are written in Roman letters as
+  they are spelled in those languages (e.g., George Clooney, Jorge Garcia, Georges
+  Pompidou).
+
+  In the case of active senior government officials in the on-line CIA World Factbook and
+  the on-line directory of Chiefs of State and Cabinet Members of Foreign Governments,
+  the spellings given in these on-line reference works should be used in place of the IC
+  Standard. For any individual who has at one time been listed in the Factbook or Chiefs of
+  State directory but who no longer appears in those resources (i.e. is no longer a
+  government official), the IC Standard spelling should appear first, with the spelling, if
+  known, as it previously appeared in those resources listed within parentheses at the first
+  usage.
+
+  The primary goal of this system is to produce a consistent Romanized transcription of the
+  name that is readable to the non-specialist. The system uses the 26 letters of the standard
+  (English) Roman alphabet. Some ambiguities in the Romanized form will occur without
+  the use of diacritics. However, within the context of a report, where additional
+  information about the individual is provided, the referent will be clearly identified.
+  Additionally, this system will be used in conjunction with on-line tools, name dictionaries
+  and lists containing conventional spellings of names of well-known individuals
+
+notes:
+
+  - |
+    Long/Short Vowels: Long and short vowels are not distinguished in the system:
+    The borrowed Arabic name Samir could represent two distinct names, one with a
+    long /a/ (Saamir) and one with a long /i/ (Samiir). One solution would be to use
+    /ee/ to stand for the long /i/, as is often done (Sameer). The IC Standard will not
+    distinguish between these.
+
+  - |
+    No distinction is made between: retroflex and non-retroflex consonants; and
+    nasalized vowels and vowels followed by /n/.
+  - |
+    No distinction is made between the several Arabic letters with the same phonetic
+    value in Urdu: e.g., letters sin/svad, zal/ze/zoe.
+  - |
+    A distinction is drawn between Urdu letters qaf and kaf (and correspondingly,
+    Hindi qa and ka).
+  - |
+    A distinction is drawn between aspirated (e.g., /d/) and nonaspirated consonants
+    (e.g., /dh/), with the exception of ch/chh, both represented by /ch/.
+  - |
+    Double consonants: Double consonants represented by the tashdid (shaddah) are
+    shown in most cases (e.g., Hassan, Muhammad). Exceptions: consonants
+    represented by digraphs are not doubled (e.g., Mubashir [not Mubashshir]).
+  - |
+    Hamzah (glottal stop) and ayn: Unlike in the Arabic IC Standard, these are not
+    represented in the IC standard.
+  - |
+    Digraphs: No distinction is made between digraphs such as /sh/ and single
+    contiguous letters such as /s/ followed by /h/.
+  - |
+    Arabic definite article “al” (‘the’): Shows sun letter assimilation in the
+    Romanized form (e.g., Abdur Rahman rather than Abdal Rahman, rather than
+    Arabic IC standard 'Abd al-Rahman).
+  - |
+    Special Rules
+    - |
+      Hyphens: Hyphens (-) are NOT used to connect name elements within a name:
+      Abdur Rahman. The single exception to this is the izafat (i.e., linking vowel in
+      noun-link-modifier construction of Persian origin), which does show a hyphen
+      before the /e/ and a following space: Koh-e Nur (‘mountain of light’), “Jaish-e
+      xx” (‘Army of xx’ construction).
+    - |
+      Names incorporating “din” are written as one unit: Azermuddin, Badruddin,
+    Faizuddin, Salahuddin.
+    - |
+      Names that incorporate Allah as part of the name show the Arabic grammatical
+      marker /u/ rather than the /a/ of Allah: Abdullah (not Abdallah).
+    - |
+      Inherent short vowel /a/ in Devanagari is represented with an /a/ in Roman. Final
+      consonants are assumed not to have a short /a/ (e.g., masc. name Ram Lal, not
+      Rama Lala).
+    - |
+      As a general rule, Devanagari va is transcribed as a /v/: Vijay, Vishal, etc.
+      Exception: /sw/ combination: Saraswati, Krishnaswami. Urdu wau, however, is
+      transcribed as /w/: Wasim, Walid.
+
+tests:
+  - source: "وشال ٹھاکر"
+    expected: "ishal tsakr"
+  - source: "حسن محمود"
+    expected: "hasn mhamid"
+  - source: "یوسف خان"
+    expected: "iisf khan"
+  - source: "معین خان"
+    expected: "mein khan"
+  - source: "سعید اجمل"
+    expected: "seid ajml"
+  - source: "انضمام الحق"
+    expected: "anzmam alhaq"
+  - source: "فرہاد رضا"
+    expected: "frahad rza"
+  - source: "وسیم اکرام"
+    expected: "isim akram"
+  - source: "شکیب الحسن"
+    expected: "shkib alhasn"
+  - source: "حسن عبد اللہ"
+    expected: "hasn ebd allah"
+
+map:
+
+  characters:
+
+
+    '\u0622': 'a' # ARABIC ALEF W/MADDA ABOVE
+    '\u0627': 'a' # or u ARABIC ALEF
+    '\u06cc': 'i' # or ai or u ARABIC LETTER WAW
+    '\u0648': 'i' # or o,au ARABIC LETTER FARSI YEH
+    '\u06d2': 'e' # ARABIC LETTER YEH BARREE
+    '\u0650': 'e' # ARABIC ARABIC KASRA
+    '\u0639': 'e' # ARABIC LETTER AIN
+    '\u06c1': 'ah' # ARABIC LETTER HEH GOAL
+    '\u062d': 'ha' # ARABIC LETTER HAH
+    '\u0648\u0670': 'a' # ARABIC LETTER FARSI YEH + ARABIC LETTER SUPERSCRIPT ALEF
+    '\u0670': 'a' # ARABIC LETTER SUPERSCRIPT ALEF
+    '\u0627\u0650': 'i' # ARABIC ALEF + ARABIC KASRA
+    '\u0639\u0650': 'i' # ARABIC LETTER AIN + ARABIC KASRA
+    '\u064f': 'i' # ARABIC DAMMA
+    '\u06d3': 'i' #  ARABIC LETTER YEH BARREE WITH HAMZA ABOVE
+    '\u06c2': '-e' #  ARABIC LETTER HEH GOAL WITH HAMZA ABOVE
+    '\u06c1\u0650': 'o' #  ARABIC LETTER YEH  WITH HAMZA ABOVE
+    '\u0624': 'au' #  ARABIC LETTER WAW WITH HAMZA ABOVE
+
+
+
+    '\u06a9': 'k' #   ARABIC LETTER KEHEH
+    '\u0642': 'q' #   ARABIC LETTER QAF
+    '\u062e': 'kh' #   ARABIC LETTER KHAH
+    '\u06af': 'g' #   ARABIC LETTER GAF
+    '\u063a': 'gh' #   ARABIC LETTER GHAIN
+    '\u0686': 'ch' #   ARABIC LETTER TCHEH
+    '\u062c': 'j' #   Arabic Letter Jeem
+
+    '\u0632': 'z' #   Arabic Letter ZAIN
+    '\u0630': 'z' #   Arabic Letter THAL
+    '\u0636': 'z' #   Arabic Letter DAD
+    '\u0638': 'z' #   Arabic Letter ZAH
+    '\u0698': 'zh' #   Arabic Letter JEH
+
+
+    '\u0679': 't' #   ARABIC LETTER TTEH
+    '\u0688': 'd' #   Arabic Letter DDAL
+    '\u062f': 'd' #   Arabic Letter DAL
+    '\u0691': 'r' #   Arabic Letter RREH
+    '\u062a': 't' #   Arabic Letter TEH
+    '\u0637': 't' #   Arabic Letter TAH
+    '\u0646': 'n' #   Arabic Letter NOON
+    '\u06ba': 'n' #   Arabic Letter NOON GHUNNA
+
+
+    '\u067e': 'p' #   Arabic Letter PE
+    '\u0641': 'f' #   Arabic Letter FEH
+    '\u0628': 'b' #   Arabic Letter BEH
+    '\u0645': 'm' #   Arabic Letter MEEM
+    '\u0631': 'r' #   Arabic Letter REH
+    '\u0644': 'l' #   Arabic Letter LAM
+    '\u0634': 'sh' #   Arabic Letter SHEEN
+    '\u0633': 's' #   Arabic Letter SHEEN
+    '\u062b': 's' #   Arabic Letter THEH
+    '\u0635': 's' #   Arabic Letter SAD
+    '\u0621': '' #   Arabic Letter HAMZA
+    '\u06be': 's' #   Arabic Letter HEH DOACHASHMEE
+    '\u064b': 'n' #   Arabic Letter FATHATAN
+    '\u064e': 'n' #   Arabic sign FATHATAN
+    '\u0652': '' #   Arabic Letter SUKUN
+    '\u0651': '' #   Arabic Letter SHADDA

data/maps/odni-uzb-Cyrl-Latn-2015.yaml

@@ -0,0 +1,166 @@
+---
+authority_id: odni
+id: 2015
+language: iso-639-2:uzb
+source_script: Cyrl
+destination_script: Latn
+name: Office of the Director Of National Intelligence Uzbek Personal Names 2015, ICS-630-01 Annex V
+# url:
+source: ICS-630-01 Annex V
+creation_date: 2015
+confirmation_date: 2015
+description: |
+  This system is the Intelligence Community standard for the transliteration of Uzbek person names
+  that will be applied to all final written reports and products for IC consumers. It is not intended to
+  eliminate variations of a name that can contribute forensic information. Rather, it is to provide an
+  IC standard Romanized (English) transliteration from Uzbek that can then be linked to forensic
+  information in ways that will help identify the referent of the name.
+
+  In cases where an individual’s name has already been transliterated in a variant spelling, the IC
+  Standard spelling should appear first, followed by the variant spelling(s) in parentheses at the first
+  usage. In addition, if the original Cyrillic-script spelling is known, that spelling should also
+  appear in parentheses following the name, if possible, following best practices of the issuing
+  organization and taking into consideration information system capabilities. For example:
+  Farkhod Tojiev (also seen as Farhod Tadjiyev, Фарход Тожиев). This convention is designed to
+  ensure that vital forensic information is not lost.
+
+  For names of persons who are known to not be part of the Uzbek-speaking community, use the
+  relevant IC transliteration standard for names from that language (e.g., Yitzhak). A translator’s
+  note may be used to clarify the known origin of the person. Spell names of individuals from
+  languages that are written in Roman letters as they are spelled in those languages (e.g., George
+  Clooney, Jorge Garcia, Georges Pompidou).
+
+  In the case of active senior government officials in the on-line CIA World Factbook and the online directory of Chiefs of State and Cabinet Members of Foreign Governments, the spellings
+  given in these on-line reference works should be used in place of the IC Standard. For any
+  individual who has at one time been listed in the Factbook or Chiefs of State directory but who no
+  longer appears in those resources (i.e. is no longer a government official), the IC Standard
+  spelling should appear first, with the spelling, if known, as it previously appeared in those
+  resources listed within parentheses at the first usage.
+
+  The primary goal is to produce a consistent Romanized transcription of names that is specifically
+  readable to the English-speaking non-specialist. The system uses the 26 letters of the standard
+  (English) Roman alphabet. Some ambiguities in the Romanized form will occur without the use
+  of diacritics. However, within the context of a report, where additional information about the
+  individual is provided, the referent will be clearly identified. This system will be used in
+  conjunction with on-line tools, name dictionaries, and lists containing conventional spellings of
+  names of well-known individuals.
+
+notes:
+  - Transliterate double digraphs as a single digraph, i.e. шш -> sh, not shsh
+  - In the Roman, no distinction is made between digraphs such as 'sh' and single contiguous letters, (e.g. 's' followed by 'h').
+  - The Cyrillic ъ and ь are not transliterated, but instead are left out of the transliteration.
+
+tests:
+  - source: Фарход Тожиев
+    expected: Farkhod Tojiev
+  - source: Барча одамлар эркин, қадр-қиммат в ҳуқуқлард тенг бўлиб туғиладилар. Улар ақл в виждон соҳибидирлар в бир-бирлари ила биродарларча муомал қилишларь зарур.
+    expected: Barcha odamlar erkin, qadr-qimat v huquqlard teng bolib tughiladilar. Ular aql v vijdon sohibidirlar v bir-birlari ila birodarlarcha muomal qilishlar zarur.
+  - source: Тутук белгись
+    expected: Tutuk belgis
+  - source: Янги юл
+    expected: Yangi iul
+  - source: Ўзбек ёзуви
+    expected: Ozbek yozuvi
+  - source: Чиғатай гурунги
+    expected: Chighatay gurungi
+  - source: ъ
+    expected: ''
+  - source: шш
+    expected: sh
+  - source: ччччч
+    expected: ch
+
+map:
+  rules:
+    # note[1]
+    - pattern: "(.)\\1{1,}"
+      result: "\\1"
+
+  characters:
+    '\u0410': 'A'    # А
+    '\u0411': 'B'    # Б
+    '\u0412': 'V'    # В
+    '\u0413': 'G'    # Г
+    '\u0492': 'Gh'   # Ғ
+    '\u0414': 'D'    # Д
+    '\u0415': 'E'    # Е
+    '\u0401': 'Yo'   # Ё
+    '\u0416': 'J'    # Ж
+    '\u0417': 'Z'    # З
+    '\u0418': 'I'    # И
+    '\u0419': 'Y'    # Й
+    '\u041A': 'K'    # К
+    '\u049A': 'Q'    # Қ
+    '\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
+    '\u040E': 'O'    # Ў
+    '\u0424': 'F'    # Ф
+    '\u0425': 'Kh'   # Х
+    '\u04B2': 'H'    # Ҳ
+    '\u0426': 'Ts'   # Ц
+    '\u0427': 'Ch'   # Ч
+    '\u0428': 'Sh'   # Ш
+    '\u042D': 'E'    # Э
+    '\u042E': 'Yu'   # Ю
+    '\u042F': 'Ya'   # Я
+
+    '\u0430': 'a'    # а
+    '\u0431': 'b'    # б
+    '\u0432': 'v'    # в
+    '\u0433': 'g'    # г
+    '\u0493': 'gh'   # ғ
+    '\u0434': 'd'    # д
+    '\u0435': 'e'    # e
+    '\u0451': 'yo'   # ё
+    '\u0436': 'j'    # ж
+    '\u0437': 'z'    # з
+    '\u0438': 'i'    # и
+    '\u0439': 'y'    # й
+    '\u043A': 'k'    # к
+    '\u049B': 'q'    # қ
+    '\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': 'u'    # у
+    '\u045E': 'o'    # ў
+    '\u0444': 'f'    # ф
+    '\u044B': 'y'    # ы
+    '\u0447': 'ch'   # ч
+    '\u044F': 'ia'   # я
+    '\u044E': 'iu'   # ю
+    '\u0445': 'kh'   # х
+    '\u04B3': 'h'    # ҳ
+    '\u0448': 'sh'   # ш
+    '\u044D': 'e'    # э
+    '\u0449': 'shch' # щ
+    '\u0446': 'ts'   # ц
+    '\u0491': 'g'    # ґ
+    '\u046B': 'u'    # 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
+
+    # note[3]
+    '\u042a': ''     # Ъ
+    '\u042c': ''     # Ь
+    '\u044a': ''     # ъ
+    '\u044c': ''     # ь

data/maps/royin-tha-Thai-Latn-1939-generic.yaml

@@ -0,0 +1,90 @@
+---
+authority_id: royin
+id: 1939-generic
+language: iso-639-2:tha
+source_script: Thai
+destination_script: Latn
+name: Royal Thai General System of Transcription (1939) Generic
+url: http://www.siamese-heritage.org/jsspdf/1941/JSS_033_1d_RoyalInstituteTranscriptionOfThaiIntoRomanCharacters.pdf
+creation_date: 1939
+adoption_date:
+description: |
+  This map loads two external maps to convert Thai text first into phonemic Thai,
+  and then into IPA transcription.
+
+  The IPA transcription will then be handled by this map, and converted into
+  Royal Thai General System of Transcription (1939)
+
+  The first two parts are done via two external maps.
+
+
+notes: |
+  This is a draft for the map.
+  The conversion from Thai to Phonemic Thai is still work-in-progress.
+
+tests:
+  - source: "กษัตริย์"
+    expected: "kasat"
+  - source: "ประกาศ"
+    expected: "prakat"
+  # - source: "ราชบุรี่"
+  #   expected: "ratburi"
+  # - source: "ปากลัด"
+  #   expected: "pak-lat"
+
+
+chain: ["var-tha-Thai-Thai-phonemic" ,"var-tha-Thai-Zsym-ipa"]
+
+map:
+  title-case: false
+  word_separator: " "
+
+  rules:
+    - pattern: '[˩˨˧˦˥]'
+      result : ''
+
+  postrules:
+    - pattern: '\.'
+      result:  ''
+
+  characters:
+
+  dictionary:
+
+    '̯': ''
+    '̚': ''
+
+    'ʔ': ''
+    'ː': ''
+
+    't͡ɕʰ': 'ch'
+    't͡ɕ': 'čh'
+    'ŋ': 'ng'
+    'j': 'y'
+    'ɔ': 'o̦'
+    'ɤ': 'œ'
+    'ɛ': 'æ'
+    'ɯ': 'ư'
+    'ʰ': 'h'
+
+    'aːw': 'ao'
+    'aw': 'ao'
+    'a̯w': 'ao'
+    'eːw': 'eo'
+    'ew': 'eo'
+    'ɛːw': 'aeo'
+    'ɛw': 'æo'
+    'iːw': 'iu'
+    'iw': 'iu'
+
+    'aːj': 'ai'
+    'aj': 'ai'
+    'a̯j': 'ai'
+    'ɔːj': 'o̦i'
+    'ɔj': 'o̦i'
+    'oːj': 'oi'
+    'oj': 'oi'
+    'ɤːj': 'œi'
+    'ɤj': 'œi'
+    'uːj': 'ui'
+    'uj': 'ui'