Embed images (data URI) into HTML document

embedder.py

#!/usr/bin/env python
# -*- coding: utf-8 -*-
"""
Embed images into HTML document
"""

import base64
import mimetypes
import os
import re
import io
import sys

## from https://gist.github.com/jsocol/1089733
class FileNotFoundError(Exception):
    pass

def img_to_data(path):
    """Convert a file (specified by a path) into a data URI."""
    if not os.path.exists(path):
        raise FileNotFoundError
    mime, _ = mimetypes.guess_type(path)
    with open(path, 'rb') as fp:
        data = fp.read()
        data64 = u''.join(base64.encodestring(data).splitlines())
        return u'data:%s;base64,%s' % (mime, data64)
    
def usage(argv):
    print('Embed images into HTML document')
    print('Output file is postfixed _embedded')
    print('Usage: {:s} <path-to-file>'.format(argv[0]))

if __name__ == '__main__':
    try:
        infile = sys.argv[1]
    except IndexError:
        usage(sys.argv)
        sys.exit(1)

    outfile = re.sub(r'(.+)\.(.+)$', r'\1_embedded.\2', infile);
    print("Reading: {:s}\nSaving to: {:s}".format(infile, outfile))       

    fin = io.open(infile, 'r', encoding="utf-8")
    fout = io.open(outfile, 'w', encoding="utf-8")
    
    din = fin.read()
    fin.close()
    
    found = re.findall("<img[^>]*src=\"([^\"]+)", din)
    for img in found:
        try:
            uri = img_to_data(img)
        except FileNotFoundError:
            print("File {:s} is not found".format(img))
            sys.exit(2)
            
        print ("Replace {:s} with {:s}...".format(img, uri[0:50]))
        din = din.replace(img, uri)
    
    fout.write(din)
    fout.close()