Show Wikipedia ZIMs

list-zim.py

#!/usr/bin/env python3

import os
import re
import argparse

def download():
    os.system("wget https://dumps.wikimedia.org/kiwix/zim/wikipedia/ -O zim.html")


def extract(lines:list[str]) -> tuple[str,str,int]:
    pat = re.compile(r"<a [^>]+>([a-zA-Z0-9_.-]+?)</a>\s+([0-9]{2}-[A-Za-z]{3}-[0-9]{4} [0-9:]{5})\s+([0-9]+)")
    kept = []

    for _line in lines:
        if m := re.match(pat, _line):
            kept.append( (m.group(1), m.group(2), int(m.group(3)) ) )
    return kept


def parseargs():
    parser = argparse.ArgumentParser()
    parser.add_argument("--update", "-u", action="store_true")
    parser.add_argument("filters", nargs="*")
    return parser.parse_args()


def ithas(name, targets):
    if len(targets) == 0:
        return True
    return all([f"_{t}_" in name for t in targets])


def main():
    args = parseargs()

    if args.update:
        download()

    with open("zim.html") as fh:
        refs = extract(fh.readlines())

    results = [(name, date, as_size(bytesize)) for name, date, bytesize in refs if ithas(name, args.filters)]

    [print(f"{size.ljust(10)}\t{date}\t{name}") for name,date,size in results]


def as_size(bytesize):
    units = ["B", "KB", "MB", "GB", "TB"]
    value = bytesize
    N = 0

    while N < len(units)+1:
        if value <= 1024:
            return f"{value:.2f} {units[N]}"

        if value > 1024:
            N += 1
            value = bytesize / (1024 ** N)

main()