download alpine repo

README.md

generate urls

python3 alpine.py > urls

feed urls to aria2c

aria2c -d main/x86_64/ -i urls

alpine.py

#!/usr/bin/env python3
# -*- coding: utf-8 -*-

import os.path
import tarfile
import urllib.request
import logging


BASE_URL = 'https://mirror.us.leaseweb.net/alpine/v3.5/main/x86_64/'
INDEX_FILE = 'APKINDEX.tar.gz'
INDEX_URL = BASE_URL + INDEX_FILE
REPO_PATH = 'main/x86_64'
INDEX_PATH = os.path.join(REPO_PATH, INDEX_FILE)

FORBIDDEN = "linux-firmware gcc-gnat dbg doc src.rpm abiword gnome gtk xfce\
            firefox thunderbird dahdi email linux-vanilla-dev perl-dev\
            ffmpeg asterisk gimp gnumeric gparted guile sound valgrind icon-theme qt4\
            qt5 office vlc qt- qemu libvirt boost-dev tests xorg ntop clang-libs\
            django ghostscript country wireshark rt4 phonenumbers freeswitch\
            twisted graphviz pidgin xen- motif gfortran hydrogen ffmpeg xscreensaver\
            mesa- mpd- ipptool teeworld vala- thunar xfdesk".split()


logger = logging.getLogger()
logger.setLevel(logging.DEBUG)

formatter = logging.Formatter('%(asctime)s - %(levelname)s - %(message)s')

fh = logging.FileHandler('alpine.log')
fh.setLevel(logging.DEBUG)
fh.setFormatter(formatter)
logger.addHandler(fh)


def init_repo():
    os.makedirs(REPO_PATH, exist_ok=True)
    if not os.path.exists(INDEX_PATH):
        logger.info('Downloading %s' % INDEX_URL)
        urllib.request.urlretrieve(INDEX_URL, INDEX_PATH)
    else:
        logger.info('index ok')

def get_packages():
    index = read_index()
    packages = []
    name = ''
    for line in index:
        if line.startswith(b'P:'):
            name = line.strip(b'\n').split(b':')[1]
        elif line.startswith(b'V:'):
            packages.append((str(name, 'UTF-8'), str(line.strip(b'\n').split(b':')[1], 'UTF-8')))
    return packages


def read_index():
    if os.path.exists(INDEX_PATH):
        logger.info('reading file %s' % INDEX_PATH)
        return read_gz_index(INDEX_PATH)
    else:
        logger.error('index not found')
        raise Exception("Index file not found, check download process")


def read_gz_index(file_path):
    with tarfile.open(file_path, 'r:gz') as gz_file:
        return gz_file.extractfile('APKINDEX').readlines()


def bad_package(name):
    for bad in FORBIDDEN:
        if bad in name:
            return True
    return False
    

if __name__ == '__main__':
    init_repo()
    p = get_packages()
    for i, v in p: 
        if not bad_package(i):
            if not os.path.exists(os.path.join(REPO_PATH, "%s-%s.apk" % (i, v))):
                print("%s%s-%s.apk" % (BASE_URL, i, v))