Parse the logs from apache access log and count the most acessing ip while the script it is running

apache2_access_log_analizer.py

#!/usr/bin/python3

import subprocess
from os import system
import re
from dataclasses import dataclass
from datetime import datetime
from collections import Counter

LOGFILE = "/var/log/apache2/live-access_log"
LOG = 10


@dataclass
class Access:
    address: str
    date: datetime
    method: str
    path: str

    def __post_init__(self):
        if isinstance(self.date, str):
            self.date = datetime.strptime(self.date, '%d/%b/%Y:%H:%M:%S')

    def __str__(self):
        address, date, method, path = self.address, self.date, self.method, self.path
        return f"{address=} {date=} {method=} {path=}"


def tailf(filename: str):
    command = f"tail -f {filename}"
    p = subprocess.Popen(
        command.split(), stdout=subprocess.PIPE, universal_newlines=True)
    for line in p.stdout:
        yield line


def parse_log_line(line: str):
    try:
        return Access(*re.match(r'^(\S+) .*\[(\S+) \S+\] "(\S+) (\S+) .*"', line).groups())
    except AttributeError:
        return None


if __name__ == "__main__":
    ip_hits = Counter()
    path_hits = Counter()
    ip_path_hits = {}

    for access in (parse_log_line(line) for line in tailf(LOGFILE)):
        if access is None:
            continue

        ip = access.address
        path_hits[access.path] += 1
        ip_hits[ip] += 1
        if ip not in ip_path_hits:
            ip_path_hits[ip] = Counter()
        ip_path_hits[ip][access.path] += 1
        ip_path_hits[ip] = Counter({k: v for k, v in ip_path_hits[ip].most_common(100)})

        # print 10 most acessing ips and their respective most accessed paths
        system("clear")
        print("CLIENTS")
        for ip, count in ip_hits.most_common(LOG):
            print(f"{ip}: {count=}: ", end="")
            for path, count in ip_path_hits[ip].most_common(1):
                print(f"{path}: {count}, ", end="")
            print()

        print(30 * "-")
        # Most accessed paths overall
        print("MOST ACESSED PATHS")
        for path, count in path_hits.most_common(5):
            print(f"{path} -> {count}")