import argparse
import csv
import gzip
import sys


def domains(path):
    with gzip.open(path, "rt", encoding="utf-8") as source:
        for line in source:
            domain = line.strip().lower().rstrip(".")
            if domain:
                yield domain

from collections import Counter

parser = argparse.ArgumentParser(description="Count topic signals in a domain list")
parser.add_argument("file")
parser.add_argument("--keywords", nargs="+", required=True)
args = parser.parse_args()
counts = Counter()
total = 0
for domain in domains(args.file):
    total += 1
    for word in args.keywords:
        if word.lower() in domain.rsplit(".", 1)[0]:
            counts[word] += 1
writer = csv.writer(sys.stdout)
writer.writerow(["input_file", "keyword", "matches", "total", "share_percent"])
for word in args.keywords:
    writer.writerow([args.file, word, counts[word], total,
                     round(100 * counts[word] / total, 4) if total else 0])