Second Program 2

import re
from pprint import pprint
from datetime import datetime
f = open("raw.txt","r").readlines() ## ganti dengan path log file
filter_url = "dapi.ds-intel.net" ## filter url yang mau dicari

final = []
for data in f[1:]:
    data = data.strip()
    result = []
    matches = re.finditer(r'(\[[^"]*\])+|[^" ]+|("[^"]*")', data.strip())
    for matchNum, match in enumerate(matches):
        matchNum = matchNum + 1
        result.append(match.group())
    final.append(result)

hasil = []
for x in final:
    url = x[5].split(" ")[1]
    if url.find(filter_url) >= 0:
        date, time = x[0].replace("[","").replace("]","").split(':', maxsplit=1)
        date = datetime.strptime(date, '%d/%b/%Y').strftime('%Y-%m-%d')
        hasil.append({
            "date": date,
            "time": time[:8],
            "dest_ip": x[2],
            "src_ip": x[3],
            "status_code": x[4],
            "url": url
        })
if len(hasil) > 0:
    # simpan ke csv file
    simpan = open("hasil.csv","w")
    # add header
    simpan.write("{0},{1},{2},{3}\n".format("Date", "Time", "Destip","Srcip"))
    for x in hasil:
        # tulis detail
        simpan.write("{0},{1},{2},{3}\n".format(str(x['date']), str(x['time']), str(x['dest_ip']), str(x['src_ip'])))
    simpan.close()
else:
    print("Data tidak tersedia")