PublicMirrors/you-get
1
0
Fork 0
mirror of https://github.com/soimort/you-get.git synced 2025-03-13 03:17:44 +03:00
Code Issues Actions Packages Projects Releases 1 Wiki Activity

[util.strings] use 4-space indentation

Browse Source
This commit is contained in:
Mort Yao 2016-03-26 16:49:03 +01:00
parent 0cabb980b6
commit bca2c3eed3
1 changed files with 17 additions and 17 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

34
src/you_get/util/strings.py
View File

@ -1,25 +1,25 @@
try:
# py 3.4
from html import unescape as unescape_html
# py 3.4
from html import unescape as unescape_html
except ImportError:
import re
from html.entities import entitydefs
import re
from html.entities import entitydefs
def unescape_html(string):
'''HTML entity decode'''
string = re.sub(r'&#[^;]+;', _sharp2uni, string)
string = re.sub(r'&[^;]+;', lambda m: entitydefs[m.group(0)[1:-1]], string)
return string
def unescape_html(string):
'''HTML entity decode'''
string = re.sub(r'&#[^;]+;', _sharp2uni, string)
string = re.sub(r'&[^;]+;', lambda m: entitydefs[m.group(0)[1:-1]], string)
return string
def _sharp2uni(m):
'''&#...; ==> unicode'''
s = m.group(0)[2:].rstrip(';；')
if s.startswith('x'):
return chr(int('0'+s, 16))
else:
return chr(int(s))
def _sharp2uni(m):
'''&#...; ==> unicode'''
s = m.group(0)[2:].rstrip(';；')
if s.startswith('x'):
return chr(int('0'+s, 16))
else:
return chr(int(s))
from .fs import legitimize
def get_filename(htmlstring):
return legitimize(unescape_html(htmlstring))
return legitimize(unescape_html(htmlstring))