GS 3G CN Downloader

import filetype
import os
import requests
import shutil
from threading import Thread
from urllib.parse import urlparse, parse_qs

MAX_TRIES = 50
POOL_SIZE = 10
START_ID = "aaaaaa"

headers = {
    "Accept": "text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8",
    "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:83.0) Gecko/20100101 Firefox/83.0",
    "Accept-Encoding": "gzip, deflate",
    "Connection": "keep-alive"
}


def get_image(url, num_try=0):
    if num_try > MAX_TRIES:
        return None
    t = requests.get(url, headers=headers)
    if t.url != url:
        parsed = urlparse(t.url)
        parsed_q = parse_qs(parsed.query)
        print(url)
        print(t.url)
        return parsed_q["u"][0]
    else:
        if t.text.startswith('{"state":"fail"'):
            return "FAIL"
        else:
            return get_image(url, num_try=num_try + 1)


ALLOWED_CHARS = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz"


def inc_char(c):
    i = ALLOWED_CHARS.index(c)
    if i < len(ALLOWED_CHARS) - 1:
        return ALLOWED_CHARS[i + 1]
    return ALLOWED_CHARS[0]


def get_id(prev_id):
    last = get_last_increment(prev_id)

    if last == -1:
        return ""

    unchanged = prev_id[:last]
    changed = prev_id[last:]
    first_inc = inc_char(changed[0])
    rest_count = len(changed) - 1

    return unchanged + first_inc + (ALLOWED_CHARS[0] * rest_count)


def get_last_increment(str):
    c = len(str) - 1
    while c >= 0:
        if str[c] != 'z':
            return c
        c -= 1

    return -1


def build_url(id):
    return f"http://gs.3g.cn/D/{id}/w"


def download_data(id):
    url = build_url(id)
    print(f"ID: {id} URL: {url}")
    try:
        img = get_image(url)
        if img is not None and img != "FAIL":
            img_data = requests.get(img, stream=True)
            img_data.decode_content = True
            with open(f"output/{id}.bin", "wb") as output_file:
                shutil.copyfileobj(img_data.raw, output_file)
            del img_data

            kind = filetype.guess(f"output/{id}.bin")
            if kind is None:
                print(f"Unknown file type for {id}")
            else:
                os.rename(f"output/{id}.bin", f"output/{id}.{kind.extension}")
                print(f"Saved file \"output/{id}.{kind.extension}\"")
    except:
        print(f"Skipped {id} due to many fails.")
        pass
    print()


thread_pool = []

id = get_id(START_ID)
while id != "":
    while len(thread_pool) < POOL_SIZE:
        t = Thread(target=download_data, args=(id,))
        t.start()
        thread_pool.append(t)
        id = get_id(id)

    to_remove = []
    for thread in thread_pool:
        if not thread.is_alive():
            to_remove.append(thread)

    for thread in to_remove:
        thread_pool.remove(thread)
        try:
            thread.join()
        except:
            pass