mirror of
https://github.com/djohnlewis/stackdump
synced 2025-01-23 07:01:41 +00:00
993bee4fc1
Also rewrote part of the HTML rewriting code so it doesn't introduce an additional wrapping element in the output which was added due to a html5lib requirements on input.
70 lines
2.2 KiB
Python
70 lines
2.2 KiB
Python
#!/usr/bin/env python
|
|
|
|
"""
|
|
HTML Tidy Extension for Python-Markdown
|
|
=======================================
|
|
|
|
Runs [HTML Tidy][] on the output of Python-Markdown using the [uTidylib][]
|
|
Python wrapper. Both libtidy and uTidylib must be installed on your system.
|
|
|
|
Note than any Tidy [options][] can be passed in as extension configs. So,
|
|
for example, to output HTML rather than XHTML, set ``output_xhtml=0``. To
|
|
indent the output, set ``indent=auto`` and to have Tidy wrap the output in
|
|
``<html>`` and ``<body>`` tags, set ``show_body_only=0``.
|
|
|
|
[HTML Tidy]: http://tidy.sourceforge.net/
|
|
[uTidylib]: http://utidylib.berlios.de/
|
|
[options]: http://tidy.sourceforge.net/docs/quickref.html
|
|
|
|
Copyright (c)2008 [Waylan Limberg](http://achinghead.com)
|
|
|
|
License: [BSD](http://www.opensource.org/licenses/bsd-license.php)
|
|
|
|
Dependencies:
|
|
* [Python2.3+](http://python.org)
|
|
* [Markdown 2.0+](http://packages.python.org/Markdown/)
|
|
* [HTML Tidy](http://utidylib.berlios.de/)
|
|
* [uTidylib](http://utidylib.berlios.de/)
|
|
|
|
"""
|
|
|
|
import markdown
|
|
try:
|
|
import tidy
|
|
except ImportError:
|
|
tidy = None
|
|
|
|
class TidyExtension(markdown.Extension):
|
|
|
|
def __init__(self, configs):
|
|
# Set defaults to match typical markdown behavior.
|
|
self.config = dict(output_xhtml=1,
|
|
show_body_only=1,
|
|
char_encoding='utf8'
|
|
)
|
|
# Merge in user defined configs overriding any present if nessecary.
|
|
for c in configs:
|
|
self.config[c[0]] = c[1]
|
|
|
|
def extendMarkdown(self, md, md_globals):
|
|
# Save options to markdown instance
|
|
md.tidy_options = self.config
|
|
# Add TidyProcessor to postprocessors
|
|
if tidy:
|
|
md.postprocessors['tidy'] = TidyProcessor(md)
|
|
|
|
|
|
class TidyProcessor(markdown.postprocessors.Postprocessor):
|
|
|
|
def run(self, text):
|
|
# Pass text to Tidy. As Tidy does not accept unicode we need to encode
|
|
# it and decode its return value.
|
|
enc = self.markdown.tidy_options.get('char_encoding', 'utf8')
|
|
return unicode(tidy.parseString(text.encode(enc),
|
|
**self.markdown.tidy_options),
|
|
encoding=enc)
|
|
|
|
|
|
def makeExtension(configs=None):
|
|
return TidyExtension(configs=configs)
|