Territory languages

babel/core.py

@@ -43,6 +43,24 @@ def get_global(key):
     >>> get_global('zone_territories')['Europe/Berlin']
     u'DE'
 
+    The keys available are:
+
+    - ``currency_fractions``
+    - ``language_aliases``
+    - ``likely_subtags``
+    - ``parent_exceptions``
+    - ``script_aliases``
+    - ``territory_aliases``
+    - ``territory_currencies``
+    - ``territory_languages``
+    - ``territory_zones``
+    - ``variant_aliases``
+    - ``win_mapping``
+    - ``zone_aliases``
+    - ``zone_territories``
+
+    .. note:: The internal structure of the data may change between versions.
+
     .. versionadded:: 0.9
 
     :param key: the data key

babel/languages.py

@@ -0,0 +1,72 @@
+# -- encoding: UTF-8 --
+from babel.core import get_global
+
+
+def get_official_languages(territory, regional=False, de_facto=False):
+    """
+    Get the official language(s) for the given territory.
+
+    The language codes, if any are known, are returned in order of descending popularity.
+
+    If the `regional` flag is set, then languages which are regionally official are also returned.
+
+    If the `de_facto` flag is set, then languages which are "de facto" official are also returned.
+
+    .. warning:: Note that the data is as up to date as the current version of the CLDR used
+                 by Babel.  If you need scientifically accurate information, use another source!
+
+    :param territory: Territory code
+    :type territory: str
+    :param regional: Whether to return regionally official languages too
+    :type regional: bool
+    :param de_facto: Whether to return de-facto official languages too
+    :type de_facto: bool
+    :return: Tuple of language codes
+    :rtype: tuple[str]
+    """
+
+    territory = str(territory).upper()
+    allowed_stati = set(("official",))
+    if regional:
+        allowed_stati.add("official_regional")
+    if de_facto:
+        allowed_stati.add("de_facto_official")
+
+    languages = get_global("territory_languages").get(territory, {})
+    pairs = [
+        (info['population_percent'], language)
+        for language, info in languages.items()
+        if info.get('official_status') in allowed_stati
+    ]
+    pairs.sort(reverse=True)
+    return tuple(lang for _, lang in pairs)
+
+
+
+def get_territory_language_info(territory):
+    """
+    Get a dictionary of language information for a territory.
+
+    The dictionary is keyed by language code; the values are dicts with more information.
+
+    The following keys are currently known for the values:
+
+    * `population_percent`: The percentage of the territory's population speaking the
+                            language.
+    * `official_status`: An optional string describing the officiality status of the language.
+                         Known values are "official", "official_regional" and "de_facto_official".
+
+    .. warning:: Note that the data is as up to date as the current version of the CLDR used
+                 by Babel.  If you need scientifically accurate information, use another source!
+
+    .. note:: Note that the format of the dict returned may change between Babel versions.
+
+    See http://www.unicode.org/cldr/charts/latest/supplemental/territory_language_information.html
+
+    :param territory: Territory code
+    :type territory: str
+    :return: Language information dictionary
+    :rtype: dict[str, dict]
+    """
+    territory = str(territory).upper()
+    return get_global("territory_languages").get(territory, {}).copy()

docs/api/index.rst

@@ -10,6 +10,7 @@ public API of Babel.
    core
    dates
    lists
+   languages
    messages/index
    numbers
    plural

docs/api/languages.rst

@@ -0,0 +1,14 @@
+Languages
+=========
+
+.. module:: babel.languages
+
+The languages module provides functionality to access data about
+languages that is not bound to a given locale.
+
+Official Languages
+------------------
+
+.. autofunction:: get_official_languages
+
+.. autofunction:: get_territory_language_info

scripts/import_cldr.py

@@ -183,6 +183,7 @@ def main():
         territory_currencies = global_data.setdefault('territory_currencies', {})
         parent_exceptions = global_data.setdefault('parent_exceptions', {})
         currency_fractions = global_data.setdefault('currency_fractions', {})
+        territory_languages = global_data.setdefault('territory_languages', {})
 
         # create auxiliary zone->territory map from the windows zones (we don't set
         # the 'zones_territories' map directly here, because there are some zones
@@ -276,6 +277,16 @@ def main():
             cur_crounding = int(fraction.attrib.get('cashRounding', cur_rounding))
             currency_fractions[cur_code] = (cur_digits, cur_rounding, cur_cdigits, cur_crounding)
 
+        # Languages in territories
+        for territory in sup.findall('.//territoryInfo/territory'):
+            languages = {}
+            for language in territory.findall('./languagePopulation'):
+                languages[language.attrib['type']] = {
+                    'population_percent': float(language.attrib['populationPercent']),
+                    'official_status': language.attrib.get('officialStatus'),
+                }
+            territory_languages[territory.attrib['type']] = languages
+
         write_datafile(global_path, global_data, dump_json=dump_json)
 
     # build a territory containment mapping for inheritance

tests/test_languages.py

@@ -0,0 +1,14 @@
+# -- encoding: UTF-8 --
+from babel.languages import get_official_languages, get_territory_language_info
+
+
+def test_official_languages():
+    assert get_official_languages("FI") == ("fi", "sv")
+    assert get_official_languages("SE") == ("sv",)
+    assert get_official_languages("CH") == ("de", "fr", "it")
+    assert get_official_languages("CH", de_facto=True) == ("de", "gsw", "fr", "it")
+    assert get_official_languages("CH", regional=True) == ("de", "fr", "it", "rm")
+
+
+def test_get_language_info():
+    assert set(get_territory_language_info("HU").keys()) == set(("hu", "en", "de", "ro", "hr", "sk", "sl"))