twitch vod downloading and splitting tool, also with chat downloading abilities.

main.py

#!/bin/env python3

# note:      you will need TwitchDownloaderCLI for it to work.
# also note: dont expect me to change stuff here, i made it fully for myself
#            and throwing it here for sharing sake, and code is messy. happy using!

import subprocess
import datetime
import requests
import shutil
import json
import sys
import csv
import os

if not os.path.isdir("vods"):
    os.mkdir("vods")

if not os.path.isdir("temp"):
    os.mkdir("temp")

def get_size(path):
    fmt = subprocess.run(["ffprobe", "-i", path, "-show_format", "-v", "quiet"], capture_output=True).stdout.decode()
    for i in fmt.split("\n"):
        if i.startswith("duration="):
            return float(i[9:])
    return -1

def isint(string):
    try:
        int(string)
        return True
    except ValueError:
        return False

def fmt_time(seconds):
    s = round(seconds % 60, 2)
    m = seconds // 60 % 60
    h = seconds // 60 // 60
    return f"{int(h)//1:0>2}:{int(m):0>2}:{s:0>2}"

def print_help(prg="vodis"):
    print(f"usage: {prg} <vod id>")

error = lambda x: print("[!]", x) or sys.exit(1)
def log(*args, end="\n", sep=" "):
    print("[i]", *args, end=end, sep=sep)

def parse_args(args):
    program = args.pop(0)

    if len(args) == 0:
        print_help(program)
        error("expected vod id")

    return args.pop(0)

def norm_date(date):
    datetime_object = datetime.datetime.fromisoformat(date)
    return datetime_object.strftime("%d.%m.%Y")

def parse_time(secs):
    secs = int(secs)
    s = secs % 60
    m = secs // 60 % 60
    h = secs // (60**2)
    h = int(h)
    return f"{h:0>2}:{m:0>2}:{s:0>2}"

def get_info(vodid):
    request = requests.post("https://gql.twitch.tv/gql",
                            headers={"Client-ID": "kimne78kx3ncx6brgo4mv6wki5h1ko"},
                            json={"query": "query { video(id: " + str(vodid) + ") {createdAt, title}}"})
    data = json.loads(request.text)
    return norm_date(data["data"]["video"]["createdAt"]), data["data"]["video"]["title"]

def run(args, **kwargs):
    print("[c] $", " ".join(args))
    return subprocess.run(args, **kwargs)

#def run_in_thread(args, **kwargs):
#    print("[c] $", " ".join(args))
#    t = threading.Thread(target=dl, args=(args, **kwargs))
#    return subprocess.run(args, **kwargs)

def download_chat(vodid, path):
    run(["./ttvdl", "chatdownload", "-u", vodid, "-o", path, "--temp-path", "temp/"])

def get_chapters(chat_json):
    chat = json.load(open(chat_json, "r"))
    chapters = chat["video"]["chapters"]
    chapters_real = []
    for i in chapters:
        chapters_real.append({"name": i["description"], "start": int(i["startMilliseconds"]//1000), "length": int(i["lengthMilliseconds"]//1000)})
    return chapters_real

if __name__ == "__main__":
    vodid = parse_args(sys.argv)
    voddate, vodname = get_info(vodid)
    log(f"[{voddate}] \"{vodname}\"")

    if os.path.isfile(os.path.join("vods", f"{vodid}.mp4")):
        log("vod already downloaded, skipping")
    else:
        log("downloading vod")
        call = run(["./ttvdl", "videodownload", "-u", vodid, "-o", f"vods/{vodid}.mp4", "--temp-path", "temp/"])
        print()
        if call.returncode != 0:
            error("process exited with non-zero exitcode")

    log("splitting vod")

    video_path = os.path.join("vods", vodid + ".mp4")
    directory_path = os.path.join("vods", vodid)

    if os.path.isdir(directory_path):
        shutil.rmtree(directory_path)

    os.mkdir(directory_path)

    chunk_size = 1_900_000_000

    orig_size = get_size(video_path)
    chunk_cur = 0
    chunks = 1

    chunk_average = -1
    chunk_indicies = []

    while chunk_cur < orig_size:
        print(f"[i] #{chunks}, cur: {fmt_time(chunk_cur)}, full: {fmt_time(orig_size)}", end="")
        if chunk_average == -1: print()
        else: print(f", average chunk size: {fmt_time(chunk_average)}")
        call = run(["ffmpeg", "-ss", str(chunk_cur), "-i", video_path, "-fs", str(int(chunk_size)), "-c", "copy", os.path.join(directory_path, f"{chunks}.mp4")], capture_output=True)
        if call.returncode != 0:
            error("error: exited with non-zero code")
        this_chunk_size = get_size(os.path.join(directory_path, f"{chunks}.mp4"))

        chunk_indicies.append((chunks, chunk_cur, chunk_cur+this_chunk_size))
        chunk_cur += this_chunk_size

        if chunk_average == -1: chunk_average = this_chunk_size
        else: chunk_average = (chunk_average + this_chunk_size) / 2
        
        chunks += 1

    if os.path.isfile(os.path.join("vods", f"{vodid}.json")):
        log("chat already downloaded, skipping")
    else:
        log("downloading chat")
        download_chat(vodid, os.path.join("vods", f"{vodid}.json"))

    log("done!")
    
    chapters = get_chapters(os.path.join("vods", f"{vodid}.json"))
    chapter_sec = 0
    chapter_off = 0
    c = 0

    for k, s, e in chunk_indicies:
        print(f"\n[часть №{k}]")
        while chapter_sec + chapter_off < e:
            chapter = chapters[c] if 0 <= c < len(chapters) else None
            if chapter is None: break
            if chapter_sec + chapter["length"] > e:
                print(f"{parse_time((chapter_sec + chapter_off) - s)} - {parse_time(e - s)} ~ {chapter['name']}")
                chapter_off += e - (chapter_sec + chapter_off)
            else:
                print(f"{parse_time((chapter_sec + chapter_off) - s)} - {parse_time((chapter_sec + chapter['length']) - s)} ~ {chapter['name']}")
                chapter_sec += chapter['length']
                chapter_off = 0
                c += 1
    
    rm = lambda x: os.remove(x) if os.path.isfile(x) else None
    rm("COPYRIGHT.txt")
    rm("THIRD-PARTY-LICENSES.txt")
    
    # nvm this, use this to not write that every time i upload a vod to telegram lol
    print(f"\n[{voddate}] {vodname}\nв комментариях 👀")