WOWHoneypotのaccess_logよりcsvを作成するpython3スクリプトです。

access_log2csv.py

# -*- coding: utf-8 -*-
#WOWHoneypotのaccess_logよりcsvを作成するpython3スクリプトです。
#2019-04-13 mah-nyan
import sys
import os
import re
import csv

args = sys.argv

def create_csv(arg1):    
    try:
        csvdata = [["date", "time", "src_ip", "method", "path"]]
        log_file = open("./" + arg1)
        lines = log_file.readlines()
        log_file.close

        for line in lines:
            date = re.search ("\d{4}-\d{2}-\d{2}", line)
            time = re.search ("\d{2}:\d{2}:\d{2}", line)
            srcip = re.search ("]\s\d{1,3}.\d{1,3}.\d{1,3}.\d{1,3}", line)
            method = re.search ("(?<=\s\")\w+", line)
            uri = re.search ("\s/\s|\s(/[\w\$\&\?\(\)\.\+\-\\\|\[\]\^=~/%#:;',<>@{}]+)(?=\sHTTP/\d\.\d\")", line)
            uri2 = re.search ("\s(\w+[\w\$\&\?\(\)\.\+\-\\\|\[\]\^=~/%#:;',<>@{}]+)(?=\sHTTP/\d\.\d\")", line)
            uri3 = re.search ("\s\s(?=HTTP/\d\.\d\")", line)
            
            if uri is not None:
                rawuri = str(uri.group()).lstrip( )
                rawsrcip = str(srcip.group())
                csvdata += [[date.group(), time.group(), rawsrcip.replace("] ", ""), method.group(), rawuri]]
            
            if uri2 is not None:
                rawuri2 = str(uri2.group()).lstrip( )
                rawsrcip = str(srcip.group())
                csvdata += [[date.group(), time.group(), rawsrcip.replace("] ", ""), method.group(), rawuri2]]
            
            if uri3 is not None:
                rawuri3 = str(uri3.group()).lstrip( )
                rawsrcip = str(srcip.group())
                csvdata += [[date.group(), time.group(), rawsrcip.replace("] ", ""), method.group(), rawuri3]]
            
        
        if not os.path.exists("./log/"):
            os.mkdir("./log/")            
        
        with open('./log/' + arg1 + ".csv", "w") as file:
            writer = csv.writer(file, lineterminator='\n')
            writer.writerows(csvdata)

        return csvdata

    except Exception as e:
        csvdata += [["e", "rr", "o", "r", str(e)]]
        with open('./log/' + arg1 + ".csv", "w") as file:
            writer = csv.writer(file, lineterminator='\n')
            writer.writerows(csvdata)
        print (str(e))
        return csvdata


if __name__ == '__main__':
    if len(args)  <= 1:
        print("引数に対象のaccess_logを指定して下さい。")
    else:
        print(args[1] + "のdataを読み込みます。\n")
        csvdata = create_csv(args[1])
        print('./log/' + args[1] + ".csvを作成しました。")

2019-04-11 pathが無い場合も取得できるように対応しました。