Cutlery/calibre
1
0
Fork 0
mirror of https://github.com/kovidgoyal/calibre.git synced 2025-07-09 03:04:10 -04:00
Code Issues Packages Projects Releases Wiki Activity

ICU collator for searching ignoring punctuation

Browse Source
This commit is contained in:
Kovid Goyal 2022-04-23 12:14:04 +05:30
parent 4412d3f615
commit 303006cb11
No known key found for this signature in database
GPG Key ID: 06BC317B515ACE7C
2 changed files with 18 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

17
src/calibre/utils/icu.py
View File

@ -12,7 +12,8 @@ from calibre.utils.config_base import tweaks, prefs
from calibre_extensions import icu as _icu from calibre_extensions import icu as _icu
from polyglot.builtins import cmp from polyglot.builtins import cmp
_locale = _collator = _primary_collator = _sort_collator = _non_numeric_sort_collator = _numeric_collator = _case_sensitive_collator = None _locale = _collator = _primary_collator = _sort_collator = _non_numeric_sort_collator = _numeric_collator = None
_case_sensitive_collator = _primary_no_punc_collator = None
cmp cmp
_none = '' _none = ''
@ -66,7 +67,9 @@ def collator():
def change_locale(locale=None): def change_locale(locale=None):
global _locale, _collator, _primary_collator, _sort_collator, _numeric_collator, _case_sensitive_collator, _non_numeric_sort_collator global _locale, _collator, _primary_collator, _sort_collator, _numeric_collator, _case_sensitive_collator, _non_numeric_sort_collator
global _primary_no_punc_collator
_collator = _primary_collator = _sort_collator = _numeric_collator = _case_sensitive_collator = _non_numeric_sort_collator = None _collator = _primary_collator = _sort_collator = _numeric_collator = _case_sensitive_collator = _non_numeric_sort_collator = None
_primary_no_punc_collator = None
_locale = locale _locale = locale
@ -79,6 +82,16 @@ def primary_collator():
return _primary_collator return _primary_collator
def primary_collator_without_punctuation():
'Ignores case differences, accented characters and punctuation'
global _primary_no_punc_collator
if _primary_no_punc_collator is None:
_primary_no_punc_collator = collator().clone()
_primary_no_punc_collator.strength = _icu.UCOL_PRIMARY
_primary_no_punc_collator.set_attribute(_icu.UCOL_ALTERNATE_HANDLING, _icu.UCOL_SHIFTED)
return _primary_no_punc_collator
def sort_collator(): def sort_collator():
'Ignores case differences and recognizes numbers in strings (if the tweak is set)' 'Ignores case differences and recognizes numbers in strings (if the tweak is set)'
global _sort_collator global _sort_collator
@ -223,8 +236,10 @@ except AttributeError: # For people running from source
find = make_two_arg_func(collator, 'find') find = make_two_arg_func(collator, 'find')
primary_find = make_two_arg_func(primary_collator, 'find') primary_find = make_two_arg_func(primary_collator, 'find')
primary_no_punc_find = make_two_arg_func(primary_collator_without_punctuation, 'find')
contains = make_two_arg_func(collator, 'contains') contains = make_two_arg_func(collator, 'contains')
primary_contains = make_two_arg_func(primary_collator, 'contains') primary_contains = make_two_arg_func(primary_collator, 'contains')
primary_no_punc_contains = make_two_arg_func(primary_collator_without_punctuation, 'contains')
startswith = make_two_arg_func(collator, 'startswith') startswith = make_two_arg_func(collator, 'startswith')
primary_startswith = make_two_arg_func(primary_collator, 'startswith') primary_startswith = make_two_arg_func(primary_collator, 'startswith')
safe_chr = _icu.chr safe_chr = _icu.chr

2
src/calibre/utils/icu_test.py
View File

@ -114,6 +114,8 @@ class TestICU(unittest.TestCase):
self.assertTrue(icu.primary_contains('pena', 'peña')) self.assertTrue(icu.primary_contains('pena', 'peña'))
x = icu.primary_collator() x = icu.primary_collator()
self.ae(x.get_attribute(icu._icu.UCOL_STRENGTH), icu._icu.UCOL_PRIMARY), self.ae(x.get_attribute(icu._icu.UCOL_STRENGTH), icu._icu.UCOL_PRIMARY),
self.ae((0, 4), icu.primary_no_punc_find('pena"', 'peña'))
self.ae((0, 13), icu.primary_no_punc_find("typographers", 'typographer’s'))
def test_collation_order(self): def test_collation_order(self):
'Testing collation ordering' 'Testing collation ordering'