wordpress_to_jekyll_postprocess.py

import os, sys, re

patterns = [
    # <img src='http://s.wordpress.com/latex.php?latex=%5Cmathcal%7BO%7D%28n%29&#038;bg=T&#038;fg=000000&#038;s=0' alt='\mathcal{O}(n)' title='\mathcal{O}(n)' class='latex' />
    (r'<img.*?title=\'(.*?)\' class=\'latex\' />',
     '$$ \\1 $$ ',
     0),

    # [<img class=" size-full wp-image-1722 aligncenter" src="http://erikbern.com/wp-content/uploads/2016/01/avg.png" alt="avg" width="512" height="512" />](http://erikbern.com/wp-content/uploads/2016/01/avg.png)
    (r'\[(<img.*?)\]\(.*?\)',
     '\\1',
     0),

    # <img class="alignnone size-full wp-image-600" title="RNN 1" src="http://45.55.214.245/wp-content/uploads/2014/06/RNN-1.png" alt="" width="741" height="187" />
    (r'<img.*?src="(.*?)".*?/>',
     '![image](\\1)',
     0),

    # <figcaption class="wp-caption-text">blah blah</figcaption></figure>
    (r'<figcaption.*?>(.*?)</figcaption></figure>',
     '*\\1*',
     re.DOTALL),

    # gists
    (r'<div class="oembed-gist">.*?href="https://gist.github.com/(.*?)".*?</div>',
     '{% gist \\1 %}',
     0),

    # encoding issue
    (r'\xc2\xa0',
     ' ',
     0),

    # italics
    (r' _([^_]*?) _',
     ' _\\1_ ',
     0),
    
    # emphasis
    (r' \*\*([^*]*?) \*\*',
     ' **\\1** ',
     0),

    # links
    (r'http://45.55.214.245',
     '',
     0),

    # pipes (for mathjax
    (r'\|',
     '\mid',
     0)
]

for root, subfolders, files in os.walk(sys.argv[1]):
    for fn in files:
        fn = os.path.join(root, fn)
        output = []
        with open(fn) as f:
            s = f.read()
            for pattern, replacement, flags in patterns:
                s2 = re.sub(pattern, replacement, s, 0, flags)
                if s2 != s:
                    print fn, pattern
                s = s2

        f = open(fn, 'w')
        f.write(s)
        f.close()