Micro chunk HLS checker

hls-checker.py

# -*- coding: utf8 -*-
from __future__ import absolute_import, division, print_function

import logging
import sys

import requests

MIN_CHUNK_SIZE = 1.0

logging.basicConfig(
    level=logging.DEBUG,
    format='%(asctime)s - %(name)s - %(levelname)s - %(message)s'
)


def main():
    fname = sys.argv[1]
    list_file = open(fname)
    for raw_line in list_file:
        line = raw_line[:-1]
        line_list = line.split('\t')
        logging.info("line_list = %s", line_list)
        check_bad_pl(
            *line_list
        )


def check_bad_pl(uuid, url):
    """

    :param str uuid:
    :param str url:
    :return:
    """
    chunk_pl_list = parse_master_pl(url)
    for chunk_pl in chunk_pl_list:
        is_bad_chunk = parse_chunk_pl(chunk_pl)
        if is_bad_chunk:
            logging.error("bad = %s, %s", uuid, url)
            print('bad = ', uuid, url)
            return True
    return False


def parse_master_pl(url):
    """

    :param str url:
    :return:
    """

    url = url.strip()
    response = requests.get(url)
    content = response.content
    content_list = content.split('\n')

    url_parts = url.split('/')
    url_base = '/'.join(url_parts[:-1])

    for content_item in content_list:
        if 'm3u8' in content_item:
            chunk_pl = '{}/{}'.format(url_base, content_item)
            yield chunk_pl


def parse_chunk_pl(url):
    """

    :param str url:
    :return:
    """
    url = url.strip()
    response = requests.get(url)
    content = response.content
    content_list = content.split('\n')
    content_seq = filter_extinf(content_list)
    content_list = list(content_seq)
    for content_item in content_list[:-1]:
        content_item = content_item.replace('#EXTINF:', '')
        content_item = content_item.replace(',', '')
        chunk_size = float(content_item)
        if chunk_size < MIN_CHUNK_SIZE:
            return True
    return False


def filter_extinf(content_list):
    """

    :param list[str] content_list:
    :return:
    """
    for content_item in content_list:
        if '#EXTINF:' in content_item:
            yield content_item


if __name__ == '__main__':
    main()