#!/usr/bin/env python
import ConfigParser, os, re
config_file = "/path/to/apache_util.config"


class ApacheUtil:
    """
    example config file:
    [main]
    apache_log = /var/log/apache2/access.log
    ban_file = /etc/deny.hosts
    ban_after = 5
    ban_level = ALL
    """
    def __init__(self):
        self.config = ConfigParser.ConfigParser()
        self.config.read(config_file)
        self.apache_log = self.config.get("main", "apache_log")
        self.ban_file = self.config.get("main", "ban_file")
        self.ban_after = int(self.config.get("main", "ban_after"))
        self.ban_level = self.config.get("main", "ban_level")
        self.parts = [
            r'(?P<host>\S+)',
            r'\S+',
            r'(?P<user>\S+)',
            r'\[(?P<time>.+)\]',
            r'"(?P<request>.+)"',
            r'(?P<status>[0-9]+)',
            r'(?P<size>\S+)',
            r'"(?P<referer>.*)"',
            r'"(?P<agent>.*)"',
        ]
        self.pattern = re.compile(r'\s+'.join(self.parts)+r'\s*\Z')
        self.matches = {
            "rfi":{}
        }
    def add_match(self, type, ip):
        if type not in self.matches:
            return "Invalid type"
        else:
            if ip in self.matches[type]:
                self.matches[type][ip] += 1
            else:
                self.matches[type][ip] = 1
    def ban_matches(self, type):
        file = open(self.ban_file, "r")
        contents = file.read()
        file.close()
        banned = ""
        total = 0
        for match in self.matches[type]:
            if self.matches[type][match] >= self.ban_after and contents.find(match) == -1:
                banned += "%s: %s \n" % match
                total += 1
        if banned != "":
            file = open(self.ban_file, "a")
            file.write(banned)
            file.close()
            print "banned %d %s's" % (total, type)

    def ban_rfi(self):
        for line in file(self.apache_log):
            m = self.pattern.match(line)
            request = m.groupdict()
            #print request
            pat = re.compile(r'(.*)(http|https|ftp):\/\/(.*)')
            if pat.match(request['request'].split()[1]):
                self.add_match("rfi", request['host'])
        self.ban_matches('rfi')


if __name__ == "__main__":
    ap = ApacheUtil()
    ap.ban_rfi()