languages: sync with CLDR

- add missing definitions present in CLDR - use CLDR as plural rules instead of default
WeblateOrg · Jan 29, 2024 · d4ada97 · d4ada97
1 parent 3ad3735
commit d4ada97
Show file tree

Hide file tree

Showing 122 changed files with 2,489 additions and 134 deletions.
diff --git a/PLURALS_DIFF.md b/PLURALS_DIFF.md
diff --git a/languages.csv b/languages.csv
@@ -133,7 +133,7 @@ den,Slave (Athapascan),2,n != 1
 dgr,Dogrib,2,n != 1
 din,Dinka,2,n != 1
 dnk,Dengka,2,n != 1
-doi,Dogri,2,n != 1
+doi,Dogri,2,n > 1
 dry,Darai,2,n != 1
 dsb,Lower Sorbian,4,(n % 100 == 1) ? 0 : ((n % 100 == 2) ? 1 : ((n % 100 == 3 || n % 100 == 4) ? 2 : 3))
 dua,Duala,2,n != 1
@@ -258,6 +258,7 @@ hil,Hiligaynon,2,n != 1
 hit,Hittite,2,n != 1
 hmn,Hmong,2,n != 1
 hne,Chhattisgarhi,2,n != 1
+hnj,Hmong Njua,1,0
 ho,Hiri Motu,2,n != 1
 hr,Croatian,3,n%10==1 && n%100!=11 ? 0 : n%10>=2 && n%10<=4 && (n%100<10 || n%100>=20) ? 1 : 2
 hrx,Hunsrik,2,n != 1
@@ -344,6 +345,7 @@ lez,Lezghian,2,n != 1
 lfn,Lingua Franca Nova,2,n != 1
 lg,Luganda,2,n != 1
 li,Limburgish,2,n != 1
+lij,Ligurian,2,n != 1
 lki,Laki,2,n != 1
 lkt,Lakota,1,0
 ln,Lingala,2,n > 1
@@ -453,6 +455,7 @@ pam,Pampanga,2,n != 1
 pap,Papiamento,2,n != 1
 pau,Palauan,2,n != 1
 pbb,Páez,2,n != 1
+pcm,Pidgin (Nigeria),2,n > 1
 peo,Persian (Old),2,n != 1
 phn,Phoenician,2,n != 1
 pi,Pali,2,n != 1
@@ -613,7 +616,7 @@ uz,Uzbek,2,n != 1
 uz_Latn,Uzbek (latin),2,n != 1
 vai,Vai,2,n != 1
 ve,Venda,2,n != 1
-vec,Venetian,2,n != 1
+vec,Venetian,3,(n == 1) ? 0 : ((n != 0 && n % 1000000 == 0) ? 1 : 2)
 vi,Vietnamese,1,0
 vls,Flemish (West),2,n != 1
 vo,Volapük,2,n != 1

diff --git a/scripts/lint b/scripts/lint
@@ -23,6 +23,8 @@ def parse_csv(name):
 
 languages = parse_csv("languages.csv")
 aliases = parse_csv("aliases.csv")
+cldr = parse_csv("cldr.csv")
+default_countries = parse_csv("default_countries.csv")
 
 for alias in aliases:
     if not alias.islower():
@@ -35,3 +37,24 @@ if missing:
 overlap = set(languages.keys()) & set(aliases.keys())
 if overlap:
     raise ValueError(f"Overlaping languages and aliases: {overlap}")
+
+missing = set(cldr.keys()) - set(languages.keys())
+# Remove aliases (these use lower case)
+missing -= {miss for miss in missing if miss.lower() in aliases}
+# Remove default countries (these use lower case)
+missing -= {miss for miss in missing if miss.lower() in default_countries}
+# Remove unwanted languages
+missing -= {"ar_001"}
+if missing:
+    raise ValueError(f"Missing from CLDR: {missing}")
+
+# Validate CLDR plural rules match
+exceptions = {"es", "it", "ca", "es_MX", "es_419"}
+matching = (set(cldr.keys()) & set(languages.keys())) - exceptions
+for match in matching:
+    plural_our = languages[match][3]
+    plural_cldr = cldr[match][3]
+    if plural_our == "n != 1" and plural_cldr != "n != 1":
+        raise ValueError(
+            f"Mismatching plural form for {match}: {plural_our!r} != {plural_cldr!r}"
+        )
diff --git a/weblate_language_data/language_codes.py b/weblate_language_data/language_codes.py
@@ -413,6 +413,7 @@
     "hmn",
     "hmo",
     "hne",
+    "hnj",
     "ho",
     "hr",
     "hrv",
@@ -556,6 +557,7 @@
     "lfn",
     "lg",
     "li",
+    "lij",
     "lim",
     "lin",
     "lit",
@@ -717,6 +719,7 @@
     "pap",
     "pau",
     "pbb",
+    "pcm",
     "peo",
     "per",
     "phn",

diff --git a/weblate_language_data/languages.py b/weblate_language_data/languages.py
@@ -1230,7 +1230,7 @@
         # or other variant.
         _("Dogri"),
         2,
-        "n != 1",
+        "n > 1",
     ),
     (
         "dry",
@@ -2348,6 +2348,15 @@
         2,
         "n != 1",
     ),
+    (
+        "hnj",
+        # Translators: Language name for ISO code "hnj". The parenthesis clarifies
+        # variant of the language. It could contain a region, age (Old, Middle, ...)
+        # or other variant.
+        _("Hmong Njua"),
+        1,
+        "0",
+    ),
     (
         "ho",
         # Translators: Language name for ISO code "ho". The parenthesis clarifies
@@ -3122,6 +3131,15 @@
         2,
         "n != 1",
     ),
+    (
+        "lij",
+        # Translators: Language name for ISO code "lij". The parenthesis clarifies
+        # variant of the language. It could contain a region, age (Old, Middle, ...)
+        # or other variant.
+        _("Ligurian"),
+        2,
+        "n != 1",
+    ),
     (
         "lki",
         # Translators: Language name for ISO code "lki". The parenthesis clarifies
@@ -4103,6 +4121,15 @@
         2,
         "n != 1",
     ),
+    (
+        "pcm",
+        # Translators: Language name for ISO code "pcm". The parenthesis clarifies
+        # variant of the language. It could contain a region, age (Old, Middle, ...)
+        # or other variant.
+        _("Pidgin (Nigeria)"),
+        2,
+        "n > 1",
+    ),
     (
         "peo",
         # Translators: Language name for ISO code "peo". The parenthesis clarifies
@@ -5549,8 +5576,8 @@
         # variant of the language. It could contain a region, age (Old, Middle, ...)
         # or other variant.
         _("Venetian"),
-        2,
-        "n != 1",
+        3,
+        "(n == 1) ? 0 : ((n != 0 && n % 1000000 == 0) ? 1 : 2)",
     ),
     (
         "vi",

diff --git a/weblate_language_data/locale/ab/LC_MESSAGES/django.po b/weblate_language_data/locale/ab/LC_MESSAGES/django.po
@@ -7,7 +7,7 @@ msgid ""
 msgstr ""
 "Project-Id-Version: Weblate Language Data\n"
 "Report-Msgid-Bugs-To: https://github.com/WeblateOrg/language-data/issues/\n"
-"POT-Creation-Date: 2024-01-29 12:57+0100\n"
+"POT-Creation-Date: 2024-01-29 13:13+0100\n"
 "PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"
 "Last-Translator: Automatically generated\n"
 "Language-Team: none\n"
@@ -1570,6 +1570,12 @@ msgstr ""
 msgid "Chhattisgarhi"
 msgstr ""
 
+#. Translators: Language name for ISO code "hnj". The parenthesis clarifies
+#. variant of the language. It could contain a region, age (Old, Middle, ...)
+#. or other variant.
+msgid "Hmong Njua"
+msgstr ""
+
 #. Translators: Language name for ISO code "ho". The parenthesis clarifies
 #. variant of the language. It could contain a region, age (Old, Middle, ...)
 #. or other variant.
@@ -2086,6 +2092,12 @@ msgstr ""
 msgid "Limburgish"
 msgstr ""
 
+#. Translators: Language name for ISO code "lij". The parenthesis clarifies
+#. variant of the language. It could contain a region, age (Old, Middle, ...)
+#. or other variant.
+msgid "Ligurian"
+msgstr ""
+
 #. Translators: Language name for ISO code "lki". The parenthesis clarifies
 #. variant of the language. It could contain a region, age (Old, Middle, ...)
 #. or other variant.
@@ -2740,6 +2752,12 @@ msgstr ""
 msgid "Páez"
 msgstr ""
 
+#. Translators: Language name for ISO code "pcm". The parenthesis clarifies
+#. variant of the language. It could contain a region, age (Old, Middle, ...)
+#. or other variant.
+msgid "Pidgin (Nigeria)"
+msgstr ""
+
 #. Translators: Language name for ISO code "peo". The parenthesis clarifies
 #. variant of the language. It could contain a region, age (Old, Middle, ...)
 #. or other variant.

diff --git a/weblate_language_data/locale/af/LC_MESSAGES/django.po b/weblate_language_data/locale/af/LC_MESSAGES/django.po
@@ -6,7 +6,7 @@ msgid ""
 msgstr ""
 "Project-Id-Version: Weblate Language Data\n"
 "Report-Msgid-Bugs-To: https://github.com/WeblateOrg/language-data/issues/\n"
-"POT-Creation-Date: 2024-01-29 12:57+0100\n"
+"POT-Creation-Date: 2024-01-29 13:13+0100\n"
 "PO-Revision-Date: 2023-07-12 09:53+0000\n"
 "Last-Translator: Gideon Wentink <[email protected]>\n"
 "Language-Team: Afrikaans <https://hosted.weblate.org/projects/weblate/"
@@ -1623,6 +1623,12 @@ msgstr "Hmong"
 msgid "Chhattisgarhi"
 msgstr ""
 
+#. Translators: Language name for ISO code "hnj". The parenthesis clarifies
+#. variant of the language. It could contain a region, age (Old, Middle, ...)
+#. or other variant.
+msgid "Hmong Njua"
+msgstr ""
+
 #. Translators: Language name for ISO code "ho". The parenthesis clarifies
 #. variant of the language. It could contain a region, age (Old, Middle, ...)
 #. or other variant.
@@ -2169,6 +2175,13 @@ msgstr ""
 msgid "Limburgish"
 msgstr "Limburgs"
 
+#. Translators: Language name for ISO code "lij". The parenthesis clarifies
+#. variant of the language. It could contain a region, age (Old, Middle, ...)
+#. or other variant.
+#, fuzzy
+msgid "Ligurian"
+msgstr "Litaus"
+
 #. Translators: Language name for ISO code "lki". The parenthesis clarifies
 #. variant of the language. It could contain a region, age (Old, Middle, ...)
 #. or other variant.
@@ -2863,6 +2876,12 @@ msgstr "Palauaans"
 msgid "Páez"
 msgstr ""
 
+#. Translators: Language name for ISO code "pcm". The parenthesis clarifies
+#. variant of the language. It could contain a region, age (Old, Middle, ...)
+#. or other variant.
+msgid "Pidgin (Nigeria)"
+msgstr ""
+
 #. Translators: Language name for ISO code "peo". The parenthesis clarifies
 #. variant of the language. It could contain a region, age (Old, Middle, ...)
 #. or other variant.

diff --git a/weblate_language_data/locale/afh/LC_MESSAGES/django.po b/weblate_language_data/locale/afh/LC_MESSAGES/django.po
@@ -7,7 +7,7 @@ msgid ""
 msgstr ""
 "Project-Id-Version: Weblate Language Data\n"
 "Report-Msgid-Bugs-To: https://github.com/WeblateOrg/language-data/issues/\n"
-"POT-Creation-Date: 2024-01-29 12:57+0100\n"
+"POT-Creation-Date: 2024-01-29 13:13+0100\n"
 "PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"
 "Last-Translator: Automatically generated\n"
 "Language-Team: none\n"
@@ -1570,6 +1570,12 @@ msgstr ""
 msgid "Chhattisgarhi"
 msgstr ""
 
+#. Translators: Language name for ISO code "hnj". The parenthesis clarifies
+#. variant of the language. It could contain a region, age (Old, Middle, ...)
+#. or other variant.
+msgid "Hmong Njua"
+msgstr ""
+
 #. Translators: Language name for ISO code "ho". The parenthesis clarifies
 #. variant of the language. It could contain a region, age (Old, Middle, ...)
 #. or other variant.
@@ -2086,6 +2092,12 @@ msgstr ""
 msgid "Limburgish"
 msgstr ""
 
+#. Translators: Language name for ISO code "lij". The parenthesis clarifies
+#. variant of the language. It could contain a region, age (Old, Middle, ...)
+#. or other variant.
+msgid "Ligurian"
+msgstr ""
+
 #. Translators: Language name for ISO code "lki". The parenthesis clarifies
 #. variant of the language. It could contain a region, age (Old, Middle, ...)
 #. or other variant.
@@ -2740,6 +2752,12 @@ msgstr ""
 msgid "Páez"
 msgstr ""
 
+#. Translators: Language name for ISO code "pcm". The parenthesis clarifies
+#. variant of the language. It could contain a region, age (Old, Middle, ...)
+#. or other variant.
+msgid "Pidgin (Nigeria)"
+msgstr ""
+
 #. Translators: Language name for ISO code "peo". The parenthesis clarifies
 #. variant of the language. It could contain a region, age (Old, Middle, ...)
 #. or other variant.

diff --git a/weblate_language_data/locale/ang/LC_MESSAGES/django.po b/weblate_language_data/locale/ang/LC_MESSAGES/django.po
@@ -7,7 +7,7 @@ msgid ""
 msgstr ""
 "Project-Id-Version: Weblate Language Data\n"
 "Report-Msgid-Bugs-To: https://github.com/WeblateOrg/language-data/issues/\n"
-"POT-Creation-Date: 2024-01-29 12:57+0100\n"
+"POT-Creation-Date: 2024-01-29 13:13+0100\n"
 "PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"
 "Last-Translator: Automatically generated\n"
 "Language-Team: none\n"
@@ -1570,6 +1570,12 @@ msgstr ""
 msgid "Chhattisgarhi"
 msgstr ""
 
+#. Translators: Language name for ISO code "hnj". The parenthesis clarifies
+#. variant of the language. It could contain a region, age (Old, Middle, ...)
+#. or other variant.
+msgid "Hmong Njua"
+msgstr ""
+
 #. Translators: Language name for ISO code "ho". The parenthesis clarifies
 #. variant of the language. It could contain a region, age (Old, Middle, ...)
 #. or other variant.
@@ -2086,6 +2092,12 @@ msgstr ""
 msgid "Limburgish"
 msgstr ""
 
+#. Translators: Language name for ISO code "lij". The parenthesis clarifies
+#. variant of the language. It could contain a region, age (Old, Middle, ...)
+#. or other variant.
+msgid "Ligurian"
+msgstr ""
+
 #. Translators: Language name for ISO code "lki". The parenthesis clarifies
 #. variant of the language. It could contain a region, age (Old, Middle, ...)
 #. or other variant.
@@ -2740,6 +2752,12 @@ msgstr ""
 msgid "Páez"
 msgstr ""
 
+#. Translators: Language name for ISO code "pcm". The parenthesis clarifies
+#. variant of the language. It could contain a region, age (Old, Middle, ...)
+#. or other variant.
+msgid "Pidgin (Nigeria)"
+msgstr ""
+
 #. Translators: Language name for ISO code "peo". The parenthesis clarifies
 #. variant of the language. It could contain a region, age (Old, Middle, ...)
 #. or other variant.