def reset(self, query):
#query is the list of keywords,eg:['yu','aoi']
query = [escape(word) for word in query]
words = '|'.join(query)
self.Q_PATTERN = re.compile(r'(?P<query>' + words + r')', re.I)
self.highlight = 0
BaseHTMLProcessor.reset(self)
def reset(self,query): #query is the list of keywords,eg:['yu','aoi'] query=[escape(word) for word in query] words='|'.join(query) self.Q_PATTERN=re.compile(r'(?P<query>' + words + r')',re.I) self.highlight=0 BaseHTMLProcessor.reset(self)
def end_a(self):
if self._start_a:
method = getattr( self, self._start_a.__name__.replace('start','end'), None )
if method:
method()
else:
BaseHTMLProcessor.unknown_endtag(self,"a")
def reset(self):
self._start_a = None
self.methodQueue = []
self.divClassStack = []
self.articleRec = []
BaseHTMLProcessor.reset(self)
def reset(self):
self._start_a = None
self.methodQueue = []
self.divClassStack = []
self.articleRec = []
BaseHTMLProcessor.reset(self)
def end_a(self):
if self._start_a:
method = getattr(self,
self._start_a.__name__.replace('start', 'end'),
None)
if method:
method()
else:
BaseHTMLProcessor.unknown_endtag(self, "a")
def unknown_endtag(self, tag):
self.flushcolor()
BaseHTMLProcessor.unknown_endtag(self, tag)
if self.needcolor:
self.colorindex = len(self.pieces)
def unknown_starttag(self, tag, attrs):
self.flushcolor()
BaseHTMLProcessor.unknown_starttag(self, tag, attrs)
if self.needcolor:
self.colorindex = len(self.pieces)
def __init__(self, basedir): BaseHTMLProcessor.__init__(self) self.basedir = basedir
def reset(self):
# extend (called from __init__ in ancestor)
# Reset all data attributes
self.verbatim = 0
BaseHTMLProcessor.reset(self)
def reset(self): '''extend (called from __init__ in ancestor)''' self.verbatim = 0 BaseHTMLProcessor.reset(self)
def unknown_starttag(self, tag, attrs): self.flushcolor() BaseHTMLProcessor.unknown_starttag(self, tag, attrs) if self.needcolor: self.colorindex = len(self.pieces)
def reset(self):
self.meta = {'data': ''}
self.div_stack = [] # keep track of div
self.methodStack = []
BaseHTMLProcessor.reset(self)
def __init__(self, basedir): BaseHTMLProcessor.__init__(self) self.basedir = basedir
def start_a(self,attrs):
if self._start_a:
self._start_a(attrs)
else:
BaseHTMLProcessor.unknown_starttag(self,"a",attrs)
def start_a(self, attrs):
if self._start_a:
self._start_a(attrs)
else:
BaseHTMLProcessor.unknown_starttag(self, "a", attrs)
def reset(self):
self.verbatim = 0
BaseHTMLProcessor.reset(self)
def reset(self): BaseHTMLProcessor.reset(self) self.colorindex = 0 self.needcolor = 0
def reset(self):
self.meta = {'data':''}
self.div_stack = [] # keep track of div
self.methodStack = []
BaseHTMLProcessor.reset(self)
#!/usr/bin/python
'''
htmlQuo.py
@author ffmmx
'''
from BaseHTMLProcessor import BaseHTMLProcessor
if __name__=='__main__':
htmlSource='''
<html>
<head>
<title>Test page</title>
</head>
<body>
<ul>
<li><a href=index.html>Home</a></li>
<li><a href=toc.html>Table of contents</a></li>
<li><a href=history.html>Revision history</a></li>
</body>
</html>
'''
parser=BaseHTMLProcessor()
print parser.feed(htmlSource)
def __init__(self, usefonts=0): BaseHTMLProcessor.__init__(self) self.usefonts = usefonts
#! /usr/bin/python
htmlSource = """
<html>
<head>
<title>Test page</title>
</head>
<body>
<ul>
<li><a href=index.html>Home</a></li>
<li><a href=toc.html>Table of contents</a></li>
<li><a href=history.html>Revision history</a></li>
</body>
</html>
"""
import sys
sys.path.append("../..")
from BaseHTMLProcessor import BaseHTMLProcessor
parser = BaseHTMLProcessor()
parser.feed(htmlSource)
print parser.output()
def flushcolor(self): if self.colorindex: buffer = "".join(self.pieces[self.colorindex:]) self.pieces = self.pieces[:self.colorindex] self.colorindex = 0 BaseHTMLProcessor.handle_data(self, self.HTMLfontify(buffer))
import urllib
from BaseHTMLProcessor import BaseHTMLProcessor
sock = urllib.urlopen("../Programmer/html/examples.html")
htmlSource = sock.read()
sock.close()
parser = BaseHTMLProcessor()
parser.feed(htmlSource)
f = open('../Programmer/html/examples1.html', 'w')
f.write(parser.output())
def unknown_endtag(self, tag): self.flushcolor() BaseHTMLProcessor.unknown_endtag(self, tag) if self.needcolor: self.colorindex = len(self.pieces)
def __init__(self, usefonts=0):
BaseHTMLProcessor.__init__(self)
self.usefonts = usefonts
def reset(self):
BaseHTMLProcessor.reset(self)
self.colorindex = 0
self.needcolor = 0
def flushcolor(self):
if self.colorindex:
buffer = "".join(self.pieces[self.colorindex:])
self.pieces = self.pieces[:self.colorindex]
self.colorindex = 0
BaseHTMLProcessor.handle_data(self, self.HTMLfontify(buffer))
def reset(self): # extend (called from __init__ in ancestor) # Reset all data attributes self.verbatim = 0 BaseHTMLProcessor.reset(self)
def reset(self):
'''extend (called from __init__ in ancestor)'''
self.verbatim = 0
BaseHTMLProcessor.reset(self)
