easylist2privoxy.py

#!/usr/bin/env python
# Takes converted for privoxy (wrongly*) easylist/easyprivacy action file
# and extracts list of adresses that privoxy should block
# See https://github.com/Andrwe/privoxy-blocklist
# and my
# (*) privoxy won't interpret '\.' as regex escape in hostnames
import re
import os

host_pattern = re.compile(
    # '^(\.((?!-)[A-Za-z0-9-]{1,63}(?<!-)\\\.)+[A-Za-z]{2,6})')
    '^(\.((?!-)[A-Za-z0-9-]{1,63}(?<!-)\\\.)+[A-Za-z]{2,6})(\/|\[)(.*)\n')

os.unlink("easyprivacy.action")
# touch file first if it doesn't exist
output = open("easyprivacy.action", "w")
output.write("# Block from AdBlock easylist\n")
output.write("# \n")

output.write("\n# This is privoxy statement!\n{+block{easylist.}}\n\n")

with open('easyprivacy.script.action', 'r') as f:
    for line in f.readlines():
        m = host_pattern.match(line)
        if m:
            if m.group(0) is not None:
                    # output.write(m.group(0)) # entire match
                    # output.write("host: " + m.group(1) + "\n") # host
                    h = re.sub(r"(\\\.)", ".", m.group(1))
                    output.write(h + m.group(3) + m.group(4) + "\n")
                    # output.write(m.group(2) + "\n") # some substring
                    # output.write("separator: " + m.group(3) + "\n") # / or [
                    # output.write("all the rest: " + m.group(4) + "\n") # rest


output.close()