maintenance: replace langcodes by babel

courlan/filters.py

@@ -9,7 +9,7 @@
 from typing import Any, Optional, Tuple
 from urllib.parse import urlsplit
 
-from langcodes import Language, tag_is_valid
+from babel import Locale, UnknownLocaleError  # type: ignore
 
 from .langinfo import COUNTRY_CODES, LANGUAGE_CODES
 
@@ -181,23 +181,19 @@ def extension_filter(urlpath: str) -> bool:
 
 
 def langcodes_score(language: str, segment: str, score: int) -> int:
-    """Use langcodes on selected URL segments and integrate
-    them into a score."""
-    # see also: https://babel.pocoo.org/en/latest/locale.html
+    """Use language codes or locale parser on selected URL segments and
+    integrate them into a score."""
     # test if the code looks like a country or a language
-    if segment[:2] not in COUNTRY_CODES and segment[:2] not in LANGUAGE_CODES:
-        return score
-    # test if tag is valid (caution: private codes are)
-    if tag_is_valid(segment):
-        # try to identify language code
-        identified = Language.get(segment).language
-        # see if it matches
-        if identified is not None:
-            LOGGER.debug("langcode %s found in URL segment %s", identified, segment)
-            if identified != language:
-                score -= 1
-            else:
+    beginning = segment[:2]
+    if beginning in LANGUAGE_CODES or beginning in COUNTRY_CODES:
+        # use locale parser
+        try:
+            if Locale.parse(segment).language == language:
                 score += 1
+            else:
+                score -= 1
+        except UnknownLocaleError:
+            pass
     return score
 
 

setup.py

@@ -106,7 +106,7 @@ def get_long_description():
     include_package_data=True,
     python_requires=">=3.6",
     install_requires=[
-        "langcodes >= 3.3.0",
+        "babel >= 2.11.0",
         "tld == 0.12.6; python_version < '3.7'",
         "tld >= 0.13; python_version >= '3.7'",
         "urllib3 >= 1.26, < 2; python_version < '3.7'",

tests/unit_tests.py

@@ -37,7 +37,13 @@
     lang_filter,
 )
 from courlan.core import filter_links
-from courlan.filters import domain_filter, extension_filter, path_filter, type_filter
+from courlan.filters import (
+    domain_filter,
+    extension_filter,
+    langcodes_score,
+    path_filter,
+    type_filter,
+)
 from courlan.meta import clear_caches
 from courlan.urlutils import _parse, get_tldinfo, is_known_link
 
@@ -440,6 +446,10 @@ def test_lang_filter():
         lang_filter("http://bz.berlin1.de/kino/050513/fans.html", "de", strict=True)
         is False
     )
+    assert langcodes_score("en", "en_HK", 0) == 1
+    assert langcodes_score("en", "en_XY", 0) == 0
+    assert langcodes_score("en", "de_DE", 0) == -1
+
     # assert lang_filter('http://www.verfassungen.de/ch/basel/verf03.htm'. 'de') is True
     # assert lang_filter('http://www.uni-stuttgart.de/hi/fnz/lehrveranst.html', 'de') is True
     # http://www.wildwechsel.de/ww/front_content.php?idcatart=177&lang=4&client=6&a=view&eintrag=100&a=view&eintrag=0&a=view&eintrag=20&a=view&eintrag=80&a=view&eintrag=20