Upgrade to CLDR 45 (#1077)

* Upgrade to CLDR 45

* Handle 'localeRules="nonlikelyScript"' for parent locales

Locales of the form 'lang_Script' where 'Script' is not the
likely script for 'lang' should have 'root' as their parent
locale. For example, the parent of 'az_Arab' should not be
computed as 'az' by truncating from the end, but should be
'root' instead as 'Arab' is not the likely script for 'az'.

The list of such languages was previously specified using
an explicit 'locales' attribute. It is now handled dynamically
using the new 'localeRules' attribute.

babel/localedata.py

@@ -95,6 +95,27 @@ def locale_identifiers() -> list[str]:
     ]
 
 
+def _is_non_likely_script(name: str) -> bool:
+    """Return whether the locale is of the form ``lang_Script``,
+    and the script is not the likely script for the language.
+
+    This implements the behavior of the ``nonlikelyScript`` value of the
+    ``localRules`` attribute for parent locales added in CLDR 45.
+    """
+    from babel.core import get_global, parse_locale
+
+    try:
+        lang, territory, script, variant, *rest = parse_locale(name)
+    except ValueError:
+        return False
+
+    if lang and script and not territory and not variant and not rest:
+        likely_subtag = get_global('likely_subtags').get(lang)
+        _, _, likely_script, *_ = parse_locale(likely_subtag)
+        return script != likely_script
+    return False
+
+
 def load(name: os.PathLike[str] | str, merge_inherited: bool = True) -> dict[str, Any]:
     """Load the locale data for the given locale.
 
@@ -132,8 +153,11 @@ def load(name: os.PathLike[str] | str, merge_inherited: bool = True) -> dict[str
                 from babel.core import get_global
                 parent = get_global('parent_exceptions').get(name)
                 if not parent:
-                    parts = name.split('_')
-                    parent = "root" if len(parts) == 1 else "_".join(parts[:-1])
+                    if _is_non_likely_script(name):
+                        parent = 'root'
+                    else:
+                        parts = name.split('_')
+                        parent = "root" if len(parts) == 1 else "_".join(parts[:-1])
                 data = load(parent).copy()
             filename = resolve_locale_filename(name)
             with open(filename, 'rb') as fileobj:

scripts/download_import_cldr.py

@@ -9,10 +9,10 @@
 import zipfile
 from urllib.request import urlretrieve
 
-URL = 'https://unicode.org/Public/cldr/44/cldr-common-44.0.zip'
-FILENAME = 'cldr-common-44.0.zip'
-# Via https://unicode.org/Public/cldr/44/hashes/SHASUM512
-FILESUM = 'f2cd8733948caf308d6e39eae21724da7f29f528f8969d456514e1e84ecd5f1e6936d0460414a968888bb1b597bc1ee723950ea47df5cba21a02bb14f96d18b6'
+URL = 'https://unicode.org/Public/cldr/45/cldr-common-45.0.zip'
+FILENAME = 'cldr-common-45.0.zip'
+# Via https://unicode.org/Public/cldr/45/hashes/SHASUM512.txt
+FILESUM = '638123882bd29911fc9492ec152926572fec48eb6c1f5dd706aee3e59cad8be4963a334bb7a09a645dbedc3356f60ef7ac2ef7ab4ccf2c8926b547782175603c'
 BLKSIZE = 131072
 
 

scripts/import_cldr.py

@@ -315,6 +315,11 @@ def parse_global(srcdir, sup):
 
         for paternity in parentBlock.findall('./parentLocale'):
             parent = paternity.attrib['parent']
+            if parent == 'root':
+                # Since CLDR-45, the 'root' parent locale uses 'localeRules="nonlikelyScript"' instead of
+                # 'locales'. This special case is handled in babel when loading locale data
+                # (https://cldr.unicode.org/index/downloads/cldr-45#h.5rbkhkncdqi9)
+                continue
             for child in paternity.attrib['locales'].split():
                 parent_exceptions[child] = parent
 

tests/test_localedata.py

@@ -63,6 +63,21 @@ def test_load():
     assert localedata.load('en_US') is localedata.load('en_US')
 
 
+def test_load_inheritance(monkeypatch):
+    from babel.localedata import _cache
+
+    _cache.clear()
+    localedata.load('hi_Latn')
+    # Must not be ['root', 'hi_Latn'] even though 'hi_Latn' matches the 'lang_Script'
+    # form used by 'nonLikelyScripts'. This is because 'hi_Latn' has an explicit parent locale 'en_IN'.
+    assert list(_cache.keys()) == ['root', 'en', 'en_001', 'en_IN', 'hi_Latn']
+
+    _cache.clear()
+    localedata.load('az_Arab')
+    # Must not include 'az' as 'Arab' is not a likely script for 'az'.
+    assert list(_cache.keys()) == ['root', 'az_Arab']
+
+
 def test_merge():
     d = {1: 'foo', 3: 'baz'}
     localedata.merge(d, {1: 'Foo', 2: 'Bar'})

tests/test_numbers.py

@@ -250,7 +250,7 @@ def test_list_currencies():
 
     assert list_currencies(locale='pa_Arab') == {'PKR', 'INR', 'EUR'}
 
-    assert len(list_currencies()) == 305
+    assert len(list_currencies()) == 306
 
 
 def test_validate_currency():