Skip to content
Fetching contributors…
Cannot retrieve contributors at this time
executable file 120 lines (96 sloc) 4.17 KB
#!/usr/bin/env python
"""A small wrapper file for parsing AsciiDoc files at Github."""
__author__ = "Devin Weaver"
__copyright__ = "Copyright (C) 2009 Devin Weaver"
__license__ = "Public Domain"
__version__ = "0.1"
"""
github_asciidoc.py
------------------
This is a wrapper file for parsing AsciiDoc files at github. It wraps the
current AsciiDoc API.
AsciiDoc specifications suggest using the file extension of `.txt` however this
causes conflict because there is no way to determine if a text file is an
AsciiDoc or not without pre-processing the file. This gives us two simple
options:
1. **Parse all text files**. We could have all files ending in `.txt` or
``README.txt`` be parsed through AsciiDoc. It will print pretty text fine
even if it isn't formatted as such. However this could be *not what the user
expects*.
2. **Pick a unique extension**. We could pick a unique extension (i.e.
`.asciidoc`) to prevent clashing. Although not directly suggested by the
author of AsciiDoc there is no standard or practice to the contrary.
Option two is recommended by myself.
Requirements
~~~~~~~~~~~~
The AsciiDoc API comes in two parts. The first is the system installation of
AsciiDoc which has a simple install_. The second part is the API script. You
can either copy this to the current directory or the application's lib folder.
There is more information on the `API page`_
The `re` package is imported here for the purpose to accomplish E-Mail address
cloaking. AsciiDoc does not offer it's own cloaking algorithm like docutils
does. So I made a simple one here to do the same. **If the expense of regex's
is too high it can be easily commented out.**
.. tip::
AsciiDoc by default runs in *safe mode* which means it will not include
external files that are **not** in the same directory as the `infile`.
However since we use a StringIO through the API it should be based on the
current working directory.
.. _install: http://www.methods.co.nz/asciidoc/userguide.html
.. _API page: http://www.methods.co.nz/asciidoc/asciidocapi.html
"""
try:
import locale
locale.setlocale(locale.LC_ALL, '')
except:
pass
import sys
import cStringIO # faster then StringIO
from asciidocapi import AsciiDocAPI
from asciidocapi import AsciiDocError
import re # only needed to simulate cloak_email_addresses
def main():
"""
Parses the given AsciiDoc file or the redirected string input and returns
the HTML body.
Usage: asciidoc2html < README.rst
asciidoc2html README.rst
"""
try:
text = open(sys.argv[1], 'r').read()
except IOError: # given filename could not be found
return ''
except IndexError: # no filename given
text = sys.stdin.read()
infile = cStringIO.StringIO(text)
outfile = cStringIO.StringIO()
asciidoc = AsciiDocAPI()
asciidoc.options('-s')
try:
asciidoc.execute(infile, outfile, 'xhtml11')
except AsciiDocError, strerror:
str = "%s" % (strerror)
str = str.replace("&", "&amp;") # Must be done first
str = str.replace("<", "%lt;")
str = str.replace(">", "%gt;")
outfile.write ("<blockquote><strong>AsciiDoc ERROR: %s</strong></blockquote>" % (str))
"""
Cloak email addresses
AsciiDoc API does not have a `cloak_email_addresses` option. We can do the
same with a set of regex but that can be expensive. Keep section commented
to disable. So ``abc@mail.example.com`` becomes:
-----------
<a class="reference" href="mailto:abc&#37;&#52;&#48;mail&#46;example&#46;org">
abc<span>&#64;</span>mail<span>&#46;</span>example<span>&#46;</span>org</a>
-----------
"""
def mangleEmail(matches):
email1 = "%s&#37;&#52;&#48;%s" % (matches.group(1), matches.group(2))
email1 = email1.replace(".", "&#46;")
email2 = "%s<span>&#64;</span>%s" % (matches.group(1), matches.group(2))
email2 = email2.replace(".", "<span>&#46;</span>")
return "<a class=\"reference\" href=\"mailto:%s\">%s</a>" % (email1, email2)
return re.sub(r'<a href="mailto:([^@]+)@([^@]+)">([^@]+)@([^@]+)</a>', mangleEmail, outfile.getvalue())
#return outfile.getvalue()
if __name__ == '__main__':
print main()
Something went wrong with that request. Please try again.