dmoonfire/dosage
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Activity

Match before and after a tag.

Browse source
This commit is contained in:
Bastian Kleineidam 2012-10-12 21:11:44 +02:00
parent 6fbf2e5cc8
commit 9c032c9006
1 changed files with 7 additions and 2 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

9
dosagelib/util.py
View file

@ -23,7 +23,7 @@ has_curses = has_module("curses")
MAX_FILESIZE = 1024*1024*1 # 1MB MAX_FILESIZE = 1024*1024*1 # 1MB
def tagre(tag, attribute, value, quote='"'): def tagre(tag, attribute, value, quote='"', before="", after=""):
"""Return a regular expression matching the given HTML tag, attribute """Return a regular expression matching the given HTML tag, attribute
and value. It matches the tag and attribute names case insensitive, and value. It matches the tag and attribute names case insensitive,
and skips arbitrary whitespace and leading HTML attributes. The "<>" at and skips arbitrary whitespace and leading HTML attributes. The "<>" at
@ -36,6 +36,9 @@ def tagre(tag, attribute, value, quote='"'):
@ptype value: string @ptype value: string
@param quote: the attribute quote (default ") @param quote: the attribute quote (default ")
@ptype quote: string @ptype quote: string
@param after: match after attribute value but before end
@ptype after: string
@return: the generated regular expression suitable for re.compile() @return: the generated regular expression suitable for re.compile()
@rtype: string @rtype: string
""" """
@ -44,8 +47,10 @@ def tagre(tag, attribute, value, quote='"'):
attribute=case_insensitive_re(attribute), attribute=case_insensitive_re(attribute),
value=value, value=value,
quote=quote, quote=quote,
before=before,
after=after,
) )
return r'<\s*%(tag)s\s+(?:[^>]*\s+)?%(attribute)s\s*=\s*%(quote)s%(value)s%(quote)s[^>]*>' % attrs return r'<\s*%(tag)s\s+(?:[^>]*%(before)s[^>]*\s+)?%(attribute)s\s*=\s*%(quote)s%(value)s%(quote)s[^>]*%(after)s[^>]*>' % attrs
def case_insensitive_re(name): def case_insensitive_re(name):