Python script to extract Markdown from a VoodooPad file used for static blog generation.

ExtractMarkdownFromVoodoopad.py

#!/usr/bin/env python

import argparse
import codecs
import glob
import os
import plistlib
import re
import sys

from datetime import datetime

def parse_args(argv):
    if argv is None:
        argv = sys.argv[1:]

    parser = argparse.ArgumentParser(description="")
    parser.add_argument('-v', '--verbose', default=False, action='store_true')
    parser.add_argument('vpdoc_file',
                        metavar='SRC_FILE',
                        help="input .vpdoc filename")
    parser.add_argument('output_dir',
                        metavar='DST_DIR',
                        help="output directory")
    return parser.parse_args(argv)

def generate_file_pairs(vp_root):
    subdirs = os.listdir(os.path.join(vp_root, 'pages'))

    for subdir in subdirs:
        subdir = os.path.join(vp_root, 'pages', subdir)
        for plist in glob.glob(os.path.join(subdir, '*.plist')):
            src_file = plist.replace('.plist', '')
            if os.path.exists(src_file):
                yield (src_file, plist)

def generate_metadata(file_pair_generator):
    for source_filename, plist_filename in file_pair_generator:
        plist = plistlib.readPlist(plist_filename)
        attributes = plist.get('attributes', {})

        pubdate = attributes.get('meta.pubdate', None)
        if pubdate:
            pubdate = datetime.strptime(pubdate, '%Y-%m-%d %H:%M:%S')

        ctime = plist.get('createdDate', None)
        publish_flag = attributes.get('meta.vpstatic.publish', False)

        if publish_flag and (pubdate or ctime):
            yield {
                'src_filename' : source_filename,
                'display_name' : plist.get('displayName', None),
                'date'         : pubdate or ctime,
                'attributes'   : attributes,
                'plist'        : plist,
            }

def generate_dst_filename(metadata_generator):
    for md in metadata_generator:
        # datestamp = md['date']
        disp_name = md['display_name']

        # filename = '%s_%s.md' % (datestamp.strftime('%Y-%m-%d'),
        #                       title_to_filename(disp_name))
        filename = '%s.md' % title_to_filename(disp_name)

        md['dst_filename'] = filename
        yield md

def make_slug(title):
    return re.sub(r'[ ,]', '_', title).lower()

def title_to_filename(title):
    return re.sub(r'[ ,-]', '_', title).lower()

def main(argv=None):
    options = parse_args(argv)

    if options.verbose:
        from pprint import pprint

    generator = generate_dst_filename(
                    generate_metadata(
                        generate_file_pairs(options.vpdoc_file)))
    for d in generator:
        src = d['src_filename']
        dst = os.path.join(options.output_dir, d['dst_filename'])

        if options.verbose:
            pprint(d['plist'])
            print "Copy %s" % src
            print "  to %s" % dst
            print

        with codecs.open(dst, 'w', 'utf-8') as out_fh:
            print >>out_fh, "title: ", d['display_name']
            print >>out_fh, "date:  ", d['date']
            print >>out_fh, "slug:  ", make_slug(d['display_name'])
            print >>out_fh

            with codecs.open(src, 'r', 'utf-8') as in_fh:
                for line in in_fh:
                    print >>out_fh, line,

if __name__ == "__main__":
    sys.exit(main())