Creole to Markdown

creole_md.py

import functools
import re
import sys


def header_repl(matchobj):
    return "{} {}".format(len(matchobj.group(1)) * '#', matchobj.group(2))


def list_repl(matchobj, ordered=True):
    level = len(matchobj.group(1))
    tmpl = '{}{}'.format('  ' * (level - 1), '1.' if ordered else '*')
    return tmpl

patterns = [
    (re.compile('\{\{\{(.*?)\}\}\}'), r'`\1`'),
    (re.compile('\{\{\{\s*.*?[!#]{2}', re.MULTILINE), r'```'),
    (re.compile('^[ ]*(#+)', re.MULTILINE), list_repl),
    (re.compile('^[ ]*(\*{2,})', re.MULTILINE), functools.partial(list_repl, ordered=False)),
    (re.compile('\{\{\{', re.MULTILINE), r'```'),
    (re.compile('\}\}\}'), r'```'),
    (re.compile('^[ ]*(=+)\s*([^=]+?)\s*=+', re.MULTILINE), header_repl),
    (re.compile('//(.*)//'), r'*\1*'),
]


def creole_to_markdown(data):
    """
    Convert creole markup (http://en.wikipedia.org/wiki/Creole_(markup)) to markdown.
    Does not support tables.

    >>> creole_to_markdown("some //italics//")
    'some *italics*'

    >>> print creole_to_markdown('''# A
    ... # B
    ...  ## C
    ... ### D
    ...   # E''')
    1. A
    1. B
      1. C
        1. D
    1. E

    >>> print creole_to_markdown('''* A
    ... * B
    ...  ** C
    ... *** D
    ... * E''')
    * A
    * B
      * C
        * D
    * E

    >>> print creole_to_markdown('''{{{
    ... !#xml
    ... <some><xml></xml></some>
    ... }}}
    ...
    ... text
    ...
    ... {{{
    ... random code block without syntax formatting
    ... }}}
    ... {{{
    ... #!xml
    ... <morexml></morexml>
    ... }}}''')
    ```xml
    <some><xml></xml></some>
    ```
    <BLANKLINE>
    text
    <BLANKLINE>
    ```
    random code block without syntax formatting
    ```
    ```xml
    <morexml></morexml>
    ```
    >>> print creole_to_markdown('''text
    ...
    ... {{{
    ... random code block without syntax formatting
    ... }}}''')
    text
    <BLANKLINE>
    ```
    random code block without syntax formatting
    ```
    >>> creole_to_markdown("* {{{inline code}}}")
    '* `inline code`'

    >>> print creole_to_markdown('''= Header 1 =
    ... text
    ... == Header 2==
    ...  === Header 3 ===''')
    # Header 1
    text
    ## Header 2
    ### Header 3

    >>> creole_to_markdown('{{{valid}}} and {{{expires}}}')
    '`valid` and `expires`'
    """
    for pattern, repl in patterns:
        data = pattern.sub(repl, data)
    return data

if __name__ == '__main__':
    file_name = sys.argv[1]
    with open(file_name) as f:
        data = f.read()
        f.close()

    sys.stdout.write(creole_to_markdown(data))

More robust version here but no markdown support yet: https://bitbucket.org/thesheep/wikicreole/wiki/Home