interscript-maps 2.4.3 → 2.5.0

data/maps/bgnpcgn-kur-Arab-Latn-2007.isc

@@ -0,0 +1,186 @@
+system "BGN-PCGN:kur-Arab:Latn:2007" {
+
+metadata {
+  authority_id bgnpcgn
+  id 2007
+  language iso-639-2:kur
+  source_script Arab
+  destination_script Latn
+  name Romanization of Kurdish (2007)
+  url https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/693727/ROMANIZATION_OF_KURDISH.pdf
+  creation_date 2007
+  confirmation_date 2017-12
+  description {    The tabulation below is applicable to the Kurdish language as a
+    whole. It is based for the most part on the Hawar Roman alphabet used
+    in the Library of Congress Standard Kurdish Orthography Table, but it
+    also incorporates certain non-Hawar elements found in A Kurdish-English
+    Dictionary (Taufiq Wahby & C J Edmonds, OUP, 1966). The tabulation
+    covers both major varieties of the Kurdish language: Kurmanji and
+    Sorani. Kurmanji is spoken principally in Turkey and in Iraq north of
+    the Great Zab River (Dahūk/Dihok Governorate). It is generally written
+    in Roman script, and usually employs the Roman orthography. Sorani is
+    spoken principally in Iraq south of the Great Zab river (Arbīl/Hewlêr
+    and As Sulaymānīyah/Slêmanî governorates). It is generally written in
+    Perso-Arabic script, and usually employs the Perso-Arabic script
+    orthography.
+
+    Kurdish forms of geographical names in Turkey will usually be found
+    in Roman script, and so no romanization process will be required. The
+    digraph options for consonant letters '\\u0686', '\\u0634', and '\\u063A'
+    will not be encountered for such names. In Iraq, Syria, and Iran,
+    Kurdish will usually be encountered in Perso-Arabic script, in which
+    case it should be romanized into the corresponding Roman script form.
+    Kurdish geographical names for places and features outside Turkey,
+    found in Roman script form, should, where necessary and if possible, be
+    tailored to fit the orthography of the Romanization shown below and
+    should employ the digraph options for consonant letters '\\u0686',
+    '\\u0634', and '\\u063A'.
+  }
+
+  notes {
+
+    note "In pure Kurdish words hamza is borne by yā’ ( ئ ) and occurs only\n before initial vowels; it is not romanized. Medial and final hamza in\n Arabic borrowings are romanized by ’ (apostrophe – Unicode encoding\n 2019)."
+
+    note "The letters ث ذ ص ض ط ظ do not occur in pure Kurdish words. In Arabic\n borrowings some writers retain these letters, others substitute س ز س ز\n ت ز respectively. Only the letters ط ض and ص are catered for in the\n Library of Congress tabulation, as reflected in lines 16-18 of the\n above Consonant table. Words of obvious Arabic origin occurring in a\n Kurdish toponymic environment will be treated as Kurdish rather than\n Arabic, as will words of other non-Kurdish origins."
+
+    note "The digraph options appearing in rows 6, 15 and 20 of the consonants\n table should be used for Kurdish geographical names in Iraq, Iran, and\n Syria. The single character options should be used for Kurdish\n geographical names in Turkey."
+
+    note "ڨ is used to represent v in foreign words. Some southern Kurdish\n writers use it to represent the v in borrowings from northern Kurdish\n dialects. و is pronounced as a v in the north and as a w elsewhere."
+
+    note "Hā’ can be used as a vowel or a consonant. The initial (ه) and medial\n (forms are used for the consonant h, Consonant table, row 31, while the\n final (ه) and independent (forms are used to represent the vowel e,\n Vowel table, row 1. Therefore, when used as a consonant, the final and\n independent forms of hā’ will be seen as ‘ه’ instead of ‘and ‘ه’,\n respectively. For example, مهه meh, (“month”). When used as ‘e’, the\n hā’ behaves like the letters alif (ا) , wāw, dāl (د) , and rā (ر) , in\n that it never joins to the following letter (i.e., it has no medial\n form). Consequently, the following letter will display the initial\n form, e.g. هەولێر Hewlêr (unless there is only one following letter, in\n which case it will be written in the independent form, e.g. ماوەت\n Mawet). As with other vowels (see special rules 2 and 3), initial e is\n preceded by the kursî hamza, yielding initial ئه , e.g. ئهني enî\n “forehead”."
+
+    note "In pure Kurdish words, the vowel ى is always long î, e.g. كانى ماسێ\n Kanî Masê. When it represents îzafe, it is also romanized î and joined\n by means of a hyphen to its preceding word e.g. پارێزگاى دهۆك Parêzga-î\n Dihok."
+
+    note "      An inventory of letter-diacritic combinations, used in addition to\n      the unmodified letters of the basic Roman script in the Romanization of\n      Kurdish, with their Unicode encoding, is:\n\n        '‘': '\\u2018' , '’': '2019'\n        'Ç': '00C7' , 'ç': '00E7'\n        'Ḍ': '1E0C' , 'ḍ': '1E0D'\n        'Ê': '00CA' , 'ê': '00EA'\n\n        # There is no single Unicode encoding for these letter-diacritic combinations.\n        'Ḧ': '0048+0308' , 'ḧ': '0068+0308'\n        'Î': '00CE' , 'î': '00EE'\n        'Ł': '0141' , 'ł': '0142'\n        'Ö': '00D6' , 'ö': '00F6'\n        'Ṟ': '1E5E' , 'ṟ': '1E5F'\n        'Ş': '015E' , 'ş': '015F'\n        'Ṣ': '1E62' , 'ṣ': '1E63'\n        'Ṭ': '1E6C' , 'ṭ': '1E6D'\n        'Û': '00DB' , 'û': '00FB'\n        'Ü': '00DC' , 'ü': '00FC'\n        'Ẍ': '1E8C' , 'ẍ': '1E8D'"
+
+    note "The Romanization column shows only lowercase forms but, when\n romanizing, uppercase and lowercase Roman letters as appropriate should\n be used."
+    # Special Rules
+    note "The conjunction و (and) should be rendered u if the\n preceding word ends in a consonant, and w if the preceding\n word ends in a vowel. It should be separated by spaces from\n the preceding and following words."
+
+    note "In the Perso-Arabic orthography for Kurdish, all vowels are\n written, with the exception of the short i, which is\n expressed with a kasrah under the preceding consonant (ِ).\n In Perso-Arabic script, the kasrah will rarely be written (\n e.g., كرن kirin “to do”). Like all Kurdish vowels, the\n short i will be preceded by a kursî hamza )ئ )if it appears\n at the beginning of a word (see 3 below; see row 4 of vowel\n table)."
+
+    note "In the Perso-Arabic orthography for Kurdish, when a vowel\n comes at the beginning of a word, or when a vowel directly\n follows another vowel, a kursî hamza )ئ )precedes it (e.g.,\n ئاگر agir “fire”)."
+
+    note "A Kurdish word will never start with alif )ا .)A Kurdish\n word may begin with a yā’ (ي) or wāw )و ,)but only when\n they are used as a consonant, when they will be romanized\n as y and w, respectively."
+
+    note "When preceded by a consonant, yā’ (ي )and wāw )و )should be\n romanized î and u, respectively. When preceded by a vowel (\n including short i, which is not written), yā’ (ي )and wāw (\n و )should be romanized y and w, respectively."
+
+    note "The Arabic sign shaddah ( ّ ) denoting a doubled consonant\n is not used in Kurdish; doubled consonants, which are rare,\n are written twice e.g. موحەممەد Muḧemmed; ننا موسه Musanna.\n Shaddah might be used in Arabic borrowings but, as in\n unpointed Arabic, would generally be omitted."
+
+    note "Particles such as له le (= at, in, on) and به be (= to,\n for, by, with) should be written separately from their\n following word, e.g. كوردستانێ له Le Kurdistanê “in\n Kurdistan”"
+
+    note "Occasionally the character sequences چه ,سه and گه occur.\n They may be romanized c·h, s·h, and g·h in order to\n differentiate those romanizations from the digraphs ch, sh,\n and gh."
+  }
+}
+
+tests {
+   "كاني ماسێ" ->  "Kanî Masê"
+   "كِرِن" ->  "Kirin"
+   "ئاگِر" ->  "Agir"
+   "موحەممەد" ->  "Muḧemmed"
+  # - source: موسەننا # issue 604
+  #   expected: Musanna
+   "لەكوردِستانێ" ->  "Le Kurdistanê"
+}
+
+stage main {
+
+  # CHARACTERS
+  parallel {
+
+    sub "\u0650" "i"
+    sub "\u0644\u06d5" "le "
+    sub "\u0628\u06d5" "be "
+    # Note 1
+    sub "\u0621" "’"
+    sub "\u0624" "’"
+    sub "\u0626" "’"
+
+    sub {
+    from "و"
+    to "u"
+    before any("ء|ب|پ|ت|ج|چ|ح|خ|د|ر|ڕ|ز|ژ|س|ش|ص|ض|ط|ع|غ|́|ڨ|ق|ک|ك|گ|ل|ڵ|م|ن|و|ه|ي")
+  }
+    sub {
+    from "ي"
+    to "î"
+    before any("ء|ب|پ|ت|ج|چ|ح|خ|د|ر|ڕ|ز|ژ|س|ش|ص|ض|ط|ع|غ|́|ڨ|ق|ک|ك|گ|ل|ڵ|م|ن|و|ه|ي")
+  }
+    sub "\u0621" "’"
+    sub "\u0628" "b"
+    sub "\u067E" "p"
+    sub "\u062A" "t"
+    sub "\u062C" "c"
+    sub {
+    from "\u0686"
+    to any(["ch", "ç"])
+  }
+    sub "\u062D" "ḧ"
+    sub "\u062E" "x"
+    sub "\u062F" "d"
+    sub "\u0631" "r"
+    sub "\u0695" "ṟ"
+    sub "\u0632" "z"
+    sub "\u0698" "j"
+    sub "\u0633" "s"
+    sub {
+    from "\u0634"
+    to any(["sh", "ş"])
+  }
+    sub "\u0635" "ṣ"
+    sub "\u0636" "ḍ"
+    sub "\u0637" "ṭ"
+    sub "\u0639" "‘"
+    sub {
+    from "\u063A"
+    to any(["gh", "ẍ"])
+  }
+    sub "\u0341" "f"
+    sub "\u06A8" "v"
+    sub "\u0642" "q"
+    sub "\u06A9" "k"
+    sub "\u0643" "k"
+    sub "\u06AF" "g"
+    sub "\u0644" "l"
+    sub "\u06B5" "ł"
+    sub "\u0645" "m"
+    sub "\u0646" "n"
+    sub "\u0648" "w"
+    sub "\u0647" "h"
+    sub "\u064A" "y"
+
+    # VOWELS
+    sub {
+    from "\u0647" + boundary
+    to "e"
+  }
+    sub "\u06D5" "e"
+    sub "\u0626\u06D5" "e"
+    sub "\u0627" "a"
+    sub "\u0626\u0627" "a"
+    sub "\u064A" "î"
+    sub "\u0626\u064A" "î"
+    sub "\u0626" "i"
+    sub "\u06CE" "ê"
+    sub "\u0626\u06CE" "ê"
+    sub "\u0648" "u"
+    sub "\u0626\u0648" "u"
+    sub "\u0648\u0648" "û"
+    sub "\u0626\u0648\u0648" "û"
+    sub "\u06C6" "o"
+    sub "\u0626\u06C6" "o"
+    sub "\u0648" "ö"
+    sub "\u06CA" "ü"
+  }
+
+  # POSTRULES
+  sub {
+    from any("\u0061".."\uFFFF")
+    to upcase
+    before boundary
+    not_before boundary + any("‘’'")
+  }
+
+
+}
+}

data/maps/bgnpcgn-mkd-Cyrl-Latn-1981.isc

@@ -0,0 +1,158 @@
+system "BGN-PCGN:mkd-Cyrl:Latn:1981" {
+
+metadata {
+  authority_id bgnpcgn
+  id 1981
+  language iso-639-2:mkd
+  source_script Cyrl
+  destination_script Latn
+  name Romanization for Macedonian (1981)
+  url https://github.com/riboseinc/interscript/files/4247920/USBGN_romanization_Macedonian_1981.pdf
+  creation_date 1981
+  description BGN/PCGN Romanization table for Macedonian.
+
+  notes {
+    note "The character ѓ should be romanized g when it occurs before е and и. In other\n instances, it should be romanized ǵ (Ǵ)."
+    note "The character ќ should be romanized k when it occurs before е and и. In other\n instances, it should be romanized ć."
+  }
+}
+
+tests {
+   "Ѓол" ->  "Đol"
+   "Јусек Тепеси" ->  "Jusek Tepesi"
+   "Љуги Ќарит" ->  "Ljugi Ćarit"
+   "Ќафа Сан" ->  "Ćafa San"
+   "Агроплод Ресен" ->  "Agroplod Resen"
+   "Алта Чука" ->  "Alta Čuka"
+   "Баш Тепе" ->  "Baš Tepe"
+   "Браќам" ->  "Braćam"
+   "Винарска Визба Агропин" ->  "Vinarska Vizba Agropin"
+   "Галичица" ->  "Galičica"
+   "Дрењево" ->  "Drenjevo"
+   "Енешево" ->  "Eneševo"
+   "Иберлија" ->  "Iberlija"
+   "Крмзи Су" ->  "Krmzi Su"
+   "Лесноски Рид" ->  "Lesnoski Rid"
+   "Мала Корабска Врата" ->  "Mala Korabska Vrata"
+   "Низок Врв" ->  "Nizok Vrv"
+   "Охридско Езеро" ->  "Ohridsko Ezero"
+   "Прлиќ" ->  "Prlić"
+   "Равна Гора" ->  "Ravna Gora"
+   "Сеѓавечкиот Рид" ->  "Seđavečkiot Rid"
+   "Трновите Њиве" ->  "Trnovite Njive"
+   "Фасов Рид" ->  "Fasov Rid"
+   "Црни Камен" ->  "Crni Kamen"
+   "Чатал Чешми" ->  "Čatal Češmi"
+   "Шехово" ->  "Šehovo"
+}
+
+dependency "var-Cyrl"
+
+stage main {
+
+  # RULES
+  sub {
+    from "Ѓ"
+    to "G"
+    after "еЕиИ"
+  }
+  sub {
+    from "ѓ"
+    to "g"
+    after "еЕиИ"
+  }
+  sub {
+    from "Ќ"
+    to "K"
+    after "еЕиИ"
+  }
+  sub {
+    from "ќ"
+    to "k"
+    after "еЕиИ"
+  }
+  # Those aren't tested.
+
+
+  # CHARACTERS
+  parallel {
+    sub "А" "A"
+    sub "Б" "B"
+    sub "В" "V"
+    sub "Г" "G"
+    sub "Д" "D"
+    sub "Ѓ" "Đ"
+    sub "Е" "E"
+    sub "Ж" "Ž"
+    sub "З" "Z"
+    sub "Ѕ" "Dz"
+    sub "И" "I"
+    sub "Ј" "J"
+    sub "К" "K"
+    sub "Л" "L"
+    sub "Љ" "Lj"
+    sub "М" "M"
+    sub "Н" "N"
+    sub "Њ" "Nj"
+    sub "О" "O"
+    sub "П" "P"
+    sub "Р" "R"
+    sub "С" "S"
+    sub "Т" "T"
+    sub "Ќ" "Ć"
+    sub "У" "U"
+    sub "Ф" "F"
+    sub "Х" "H"
+    sub "Ц" "C"
+    sub "Ч" "Č"
+    sub "Џ" "Dž"
+    sub "Ш" "Š"
+    sub "а" "a"
+    sub "б" "b"
+    sub "в" "v"
+    sub "г" "g"
+    sub "д" "d"
+    sub "ѓ" "đ"
+    sub "е" "e"
+    sub "ж" "ž"
+    sub "з" "z"
+    sub "ѕ" "dz"
+    sub "и" "i"
+    sub "ј" "j"
+    sub "к" "k"
+    sub "л" "l"
+    sub "љ" "lj"
+    sub "м" "m"
+    sub "н" "n"
+    sub "њ" "nj"
+    sub "о" "o"
+    sub "п" "p"
+    sub "р" "r"
+    sub "с" "s"
+    sub "т" "t"
+    sub "ќ" "ć"
+    sub "у" "u"
+    sub "ф" "f"
+    sub "х" "h"
+    sub "ц" "c"
+    sub "ч" "č"
+    sub "џ" "dž"
+    sub "ш" "š"
+  }
+
+  # POSTRULES
+  sub {
+    from any(["Dz", "Lj", "Nj", "Dž"])
+    to upcase
+    before any("A".."Z")
+  }
+  sub {
+    from any(["Dz", "Lj", "Nj", "Dž"])
+    to upcase
+    after any("A".."Z")
+  }
+  # Those aren't tested
+
+  compose
+}
+}

data/maps/bgnpcgn-mkd-Cyrl-Latn-2013.isc

@@ -0,0 +1,181 @@
+system "BGN-PCGN:mkd-Cyrl:Latn:2013" {
+
+metadata {
+  authority_id bgnpcgn
+  id 2013
+  language iso-639-2:mkd
+  source_script Cyrl
+  destination_script Latn
+  name Romanization for Macedonian (2013)
+  url https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/811514/ROMANIZATION_OF_MACEDONIAN.pdf
+  creation_date 2013
+  description {    Macedonian was officially established as a written language in Yugoslavia during
+    World War II and is now the official language of North Macedonia.
+    This romanization system replaces the BGN/PCGN 1981 agreement and adheres
+    to the most widely-accepted standardization of Macedonian orthography.
+  }
+
+  notes {
+    note "The Macedonian Cyrillic lowercase italic Д may sometimes be seen as g.\n There is no specific Unicode encoding for this variant form so a comparable character\n has been used here for illustrative purposes."
+    note "The Macedonian Cyrillic lowercase italic Ѓ may sometimes be seen as ī.\n There is no specific Unicode encoding for this variant form so a comparable character\n has been used here for illustrative purposes."
+    note "The Macedonian Cyrillic lowercase italic П may sometimes be seen as ū.\n There is no specific Unicode encoding for this variant form so a comparable character\n has been used here for illustrative purposes."
+    note "The Macedonian Cyrillic lowercase italic Т may sometimes be seen as w̄.\n There is no specific Unicode encoding for this variant form so a comparable character\n has been used here for illustrative purposes."
+    note "      An inventory of letter-diacritic combinations, with their Unicode encoding,\n      in addition to the unmodified letters of the basic Roman script is:\n      | Ǵ (U+01F4)   | ǵ (U+01F5)   |\n      | Ž (U+017D)   | ž (U+017E)   |\n      | Dz (U+01F2)* | dz (U+01F3)* |\n      | Lj (U+01C8)* | lj (U+01C9)* |\n      | Nj (U+01CB)* | nj(U+01CC)*  |\n      | Ḱ (U+IE30)   | ḱ (U+IE31)   |\n      | Č (U+010C)   | č (U+010D)   |\n      | Dž (U+01C5)* | dž (U+01C6)* |\n      | Š (U+0160)   | š (U+0161)   |\n      * Note that these characters can also be reproduced with individual letters (e.g. l+j)."
+    note "The Romanization column shows only lowercase forms but, when romanizing,\n uppercase and lowercase Roman letters as appropriate should be used."
+  }
+}
+
+tests {
+   "Ѓенови Ливаѓе" ->  "Ǵenovi Livaǵe"
+   "ЛУЃЕ луѓе" ->  "LUǴE luǵe"
+   "ЅВЕЗДА ѕвезда Ѕвезда" ->  "DZVEZDA dzvezda Dzvezda"
+   "Јабежица" ->  "Jabežica"
+   "Љиќен и Бард" ->  "Ljiḱen i Bard"
+   "Ќамилов Чукар" ->  "Ḱamilov Čukar"
+   "Џавидин Кајнак" ->  "Džavidin Kajnak"
+   "Џамалџи" ->  "Džamaldži"
+   "Џибра Гури и Зи" ->  "Džibra Guri i Zi"
+   "Абазова Куќарица" ->  "Abazova Kuḱarica"
+   "Баба Анѓина Маала" ->  "Baba Anǵina Maala"
+   "Ваљановец" ->  "Valjanovec"
+   "Галал Једи Дереш" ->  "Galal Jedi Dereš"
+   "Дванаесет Клајнци" ->  "Dvanaeset Klajnci"
+   "Електродистрибуција Струга" ->  "Elektrodistribucija Struga"
+   "Железничка Станица Рајко Жинзифов" ->  "Železnička Stanica Rajko Žinzifov"
+   "Заедничко Речиште" ->  "Zaedničko Rečište"
+   "Испраена Плоча" ->  "Ispraena Ploča"
+   "Казнено-Поправна Установа Идризово" ->  "Kazneno-Popravna Ustanova Idrizovo"
+   "Лази и Зејнелит" ->  "Lazi i Zejnelit"
+   "Мавровско Езеро" ->  "Mavrovsko Ezero"
+   "Национален Парк Галичица" ->  "Nacionalen Park Galičica"
+   "Одморалиште Свети Стефан" ->  "Odmoralište Sveti Stefan"
+   "Планинарски Дом Караџица" ->  "Planinarski Dom Karadžica"
+   "Раса е Лисењит" ->  "Rasa e Lisenjit"
+   "Скочивирска Клисура" ->  "Skočivirska Klisura"
+   "Термо-електроцентрала Неготино" ->  "Termo-elektrocentrala Negotino"
+   "Узуновско Бресје" ->  "Uzunovsko Bresje"
+   "Фабрика Југохром" ->  "Fabrika Jugohrom"
+   "Хидроелектрана Сапунџица" ->  "Hidroelektrana Sapundžica"
+   "Цветковско Рамниште" ->  "Cvetkovsko Ramnište"
+   "Чалтанова Пештера" ->  "Čaltanova Peštera"
+   "Шкемби Вишнејц" ->  "Škembi Višnejc"
+}
+
+dependency "posix"
+
+stage main {
+
+
+
+  # CHARACTERS
+  parallel {
+    sub "А" "A"
+    sub "Б" "B"
+    sub "В" "V"
+    sub "Г" "G"
+    sub "Д" "D"
+    sub "Ѓ" "Ǵ"
+    sub "Е" "E"
+    sub "Ж" "Ž"
+    sub "З" "Z"
+    sub "Ѕ" "Dz"
+    sub "И" "I"
+    sub "Ј" "J"
+    sub "К" "K"
+    sub "Л" "L"
+    sub "Љ" "Lj"
+    sub "М" "M"
+    sub "Н" "N"
+    sub "Њ" "Nj"
+    sub "О" "O"
+    sub "П" "P"
+    sub "Р" "R"
+    sub "С" "S"
+    sub "Т" "T"
+    sub "Ќ" "Ḱ"
+    sub "У" "U"
+    sub "Ф" "F"
+    sub "Х" "H"
+    sub "Ц" "C"
+    sub "Ч" "Č"
+    sub "Џ" "Dž"
+    sub "Ш" "Š"
+    sub "а" "a"
+    sub "б" "b"
+    sub "в" "v"
+    sub "г" "g"
+    sub "д" "d"
+    sub "ѓ" "ǵ"
+    sub "е" "e"
+    sub "ж" "ž"
+    sub "з" "z"
+    sub "ѕ" "dz"
+    sub "и" "i"
+    sub "ј" "j"
+    sub "к" "k"
+    sub "л" "l"
+    sub "љ" "lj"
+    sub "м" "m"
+    sub "н" "n"
+    sub "њ" "nj"
+    sub "о" "o"
+    sub "п" "p"
+    sub "р" "r"
+    sub "с" "s"
+    sub "т" "t"
+    sub "ќ" "ḱ"
+    sub "у" "u"
+    sub "ф" "f"
+    sub "х" "h"
+    sub "ц" "c"
+    sub "ч" "č"
+    sub "џ" "dž"
+    sub "ш" "š"
+  }
+  # POSTRULES
+  # rewritten by hand to work properly
+  # DZ
+  sub {
+    from capture(maybe(upper)) + "Dz" + capture(upper)
+    to ref(1) + "DZ" + ref(2)
+  }
+  sub {
+    from capture(upper) + "Dz" + capture(maybe(upper))
+    to ref(1) + "DZ" + ref(2)
+  }
+
+  #LJ
+  sub {
+    from capture(upper) + "Lj" + capture(maybe(upper))
+    to ref(1) + "LJ" + ref(2)
+  }
+  sub {
+    from capture(maybe(upper)) + "Lj" + capture(upper)
+    to ref(1) + "LJ" + ref(2)
+  }
+
+  #NJ
+  sub {
+    from capture(upper) + "Nj" + capture(maybe(upper))
+    to ref(1) + "NJ" + ref(2)
+  }
+  sub {
+    from capture(maybe(upper)) + "Nj" + capture(upper)
+    to ref(1) + "NJ" + ref(2)
+  }
+
+  #DŽ
+  sub {
+    from capture(upper) + "Dž" + capture(maybe(upper))
+    to ref(1) + "DŽ" + ref(2)
+  }
+  sub {
+    from capture(maybe(upper)) + "Dž" + capture(upper)
+    to ref(1) + "DŽ" + ref(2)
+  }
+
+  compose
+
+}
+
+}