scdlbot/__main__.py
#!/usr/bin/env python
import asyncio
import concurrent.futures
import datetime
import logging
import os
import pathlib
import platform
import random
import shutil
import tempfile
import threading
import time
import traceback
from importlib import resources
from logging.handlers import SysLogHandler
from multiprocessing import get_context
from subprocess import PIPE, TimeoutExpired # skipcq: BAN-B404
from urllib.parse import urljoin
from uuid import uuid4
import ffmpeg
import prometheus_client
import requests
import sdnotify
from mutagen.id3 import ID3, ID3v1SaveOptions
from mutagen.mp3 import EasyMP3 as MP3
from telegram import Bot, Chat, ChatMember, InlineKeyboardButton, InlineKeyboardMarkup, MessageEntity, Update
from telegram.constants import ChatAction
from telegram.error import BadRequest, ChatMigrated, Forbidden, NetworkError, TelegramError, TimedOut
from telegram.ext import Application, ApplicationBuilder, CallbackQueryHandler, CommandHandler, ContextTypes, MessageHandler, PicklePersistence, filters
from telegram.helpers import escape_markdown
from telegram.request import HTTPXRequest
# import gc
# from boltons.urlutils import find_all_links
# from fake_useragent import UserAgent
# from telegram_handler import TelegramHandler
# Support different old versions just in case:
# https://github.com/yt-dlp/yt-dlp/wiki/Forks
try:
import yt_dlp as ydl
except ImportError:
try:
import youtube_dl as ydl
except ImportError:
import youtube_dlc as ydl
from boltons.urlutils import URL
from plumbum import ProcessExecutionError, local
TG_BOT_TOKEN = os.environ["TG_BOT_TOKEN"]
TG_BOT_API = os.getenv("TG_BOT_API", "https://api.telegram.org")
# https://github.com/python-telegram-bot/python-telegram-bot/wiki/Local-Bot-API-Server
# https://github.com/tdlib/telegram-bot-api#usage
TG_BOT_API_LOCAL_MODE = False
if "TG_BOT_API_LOCAL_MODE" in os.environ:
TG_BOT_API_LOCAL_MODE = bool(int(os.getenv("TG_BOT_API_LOCAL_MODE", "0")))
elif "127.0.0.1" in TG_BOT_API or "localhost" in TG_BOT_API:
TG_BOT_API_LOCAL_MODE = True
HTTP_VERSION = "2"
if TG_BOT_API_LOCAL_MODE:
HTTP_VERSION = "1.1"
TG_BOT_OWNER_CHAT_ID = int(os.getenv("TG_BOT_OWNER_CHAT_ID", "0"))
CHAT_STORAGE = os.path.expanduser(os.getenv("CHAT_STORAGE", "/tmp/scdlbot.pickle"))
DL_DIR = os.path.expanduser(os.getenv("DL_DIR", "/tmp/scdlbot"))
BIN_PATH = os.getenv("BIN_PATH", "")
scdl_bin = local[os.path.join(BIN_PATH, "scdl")]
bcdl_bin = local[os.path.join(BIN_PATH, "bandcamp-dl")]
BCDL_ENABLE = False
WORKERS = int(os.getenv("WORKERS", 2))
# TODO try to change to spawn or forkserver to save RAM
mp_method = "fork"
if platform.system() == "Windows":
mp_method = "spawn"
# https://stackoverflow.com/a/66113051
# https://superfastpython.com/processpoolexecutor-multiprocessing-context/
# https://docs.python.org/3/library/multiprocessing.html#contexts-and-start-methods
# https://docs.python.org/3/library/concurrent.futures.html#concurrent.futures.ProcessPoolExecutor
EXECUTOR = concurrent.futures.ProcessPoolExecutor(max_workers=WORKERS, mp_context=get_context(method=mp_method))
DL_TIMEOUT = int(os.getenv("DL_TIMEOUT", 300))
CHECK_URL_TIMEOUT = int(os.getenv("CHECK_URL_TIMEOUT", 30))
# Timeouts: https://www.python-httpx.org/advanced/
COMMON_CONNECTION_TIMEOUT = int(os.getenv("COMMON_CONNECTION_TIMEOUT", 10))
MAX_TG_FILE_SIZE = int(os.getenv("MAX_TG_FILE_SIZE", "45_000_000"))
MAX_CONVERT_FILE_SIZE = int(os.getenv("MAX_CONVERT_FILE_SIZE", "80_000_000"))
NO_FLOOD_CHAT_IDS = list(map(int, os.getenv("NO_FLOOD_CHAT_IDS", "0").split(",")))
COOKIES_FILE = os.getenv("COOKIES_FILE", None)
PROXIES = []
if "PROXIES" in os.environ:
PROXIES = [None if x == "direct" else x for x in os.getenv("PROXIES").split(",")]
SOURCE_IPS = []
if "SOURCE_IPS" in os.environ:
SOURCE_IPS = os.getenv("SOURCE_IPS").split(",")
BLACKLIST_TELEGRAM_DOMAINS = [
"telegram.org",
"telegram.me",
"t.me",
"telegram.dog",
"telegra.ph",
"te.legra.ph",
"graph.org",
"tdesktop.com",
"desktop.telegram.org",
"telesco.pe",
"contest.com",
"contest.dev",
]
WHITELIST_DOMAINS = {}
if "WHITELIST_DOMAINS" in os.environ:
WHITELIST_DOMAINS = set(x for x in os.getenv("WHITELIST_DOMAINS").split(","))
BLACKLIST_DOMAINS = {}
if "BLACKLIST_DOMAINS" in os.environ:
BLACKLIST_DOMAINS = set(x for x in os.getenv("BLACKLIST_DOMAINS").split(","))
WHITELIST_CHATS = []
if "WHITELIST_CHATS" in os.environ:
try:
WHITELIST_CHATS = set(int(x) for x in os.getenv("WHITELIST_CHATS").split(","))
except ValueError:
raise ValueError("Your whitelisted chats list does not contain valid integers.")
BLACKLIST_CHATS = []
if "BLACKLIST_CHATS" in os.environ:
try:
BLACKLIST_CHATS = set(int(x) for x in os.getenv("BLACKLIST_CHATS").split(","))
except ValueError:
raise ValueError("Your blacklisted chats list does not contain valid integers.")
# Webhook:
WEBHOOK_ENABLE = bool(int(os.getenv("WEBHOOK_ENABLE", "0")))
WEBHOOK_HOST = os.getenv("HOST", "127.0.0.1")
WEBHOOK_PORT = int(os.getenv("PORT", "5000"))
WEBHOOK_APP_URL_ROOT = os.getenv("WEBHOOK_APP_URL_ROOT", "")
WEBHOOK_APP_URL_PATH = os.getenv("WEBHOOK_APP_URL_PATH", TG_BOT_TOKEN.replace(":", ""))
WEBHOOK_CERT_FILE = os.getenv("WEBHOOK_CERT_FILE", None)
WEBHOOK_KEY_FILE = os.getenv("WEBHOOK_KEY_FILE", None)
WEBHOOK_SECRET_TOKEN = os.getenv("WEBHOOK_SECRET_TOKEN", None)
# Prometheus metrics:
METRICS_HOST = os.getenv("METRICS_HOST", "127.0.0.1")
METRICS_PORT = int(os.getenv("METRICS_PORT", "8000"))
REGISTRY = prometheus_client.CollectorRegistry()
EXECUTOR_TASKS_REMAINING = prometheus_client.Gauge(
"executor_tasks_remaining",
"Value: executor_tasks_remaining",
registry=REGISTRY,
)
BOT_REQUESTS = prometheus_client.Counter(
"bot_requests_total",
"Value: bot_requests_total",
labelnames=["type", "chat_type", "mode"],
registry=REGISTRY,
)
# Logging:
logging_handlers = []
LOGLEVEL = os.getenv("LOGLEVEL", "INFO").upper()
HOSTNAME = os.getenv("HOSTNAME", "scdlbot-host")
console_formatter = logging.Formatter("[%(name)s] %(levelname)s: %(message)s")
console_handler = logging.StreamHandler()
console_handler.setFormatter(console_formatter)
console_handler.setLevel(LOGLEVEL)
logging_handlers.append(console_handler)
SYSLOG_ADDRESS = os.getenv("SYSLOG_ADDRESS", None)
if SYSLOG_ADDRESS:
syslog_formatter = logging.Formatter("%(asctime)s " + HOSTNAME + " %(name)s: %(message)s", datefmt="%b %d %H:%M:%S")
syslog_host, syslog_udp_port = SYSLOG_ADDRESS.split(":")
syslog_handler = SysLogHandler(address=(syslog_host, int(syslog_udp_port)))
syslog_handler.setFormatter(syslog_formatter)
syslog_handler.setLevel(LOGLEVEL)
logging_handlers.append(syslog_handler)
# telegram_handler = TelegramHandler(token=TG_BOT_TOKEN, chat_id=str(TG_BOT_OWNER_CHAT_ID))
# telegram_handler.setLevel(logging.WARNING)
# logging_handlers.append(telegram_handler)
logging.basicConfig(
format="%(asctime)s - %(name)s - %(levelname)s - %(message)s",
datefmt="%Y-%m-%d %H:%M:%S",
level=LOGLEVEL,
handlers=logging_handlers,
)
logger = logging.getLogger(__name__)
# Systemd watchdog monitoring:
SYSTEMD_NOTIFIER = sdnotify.SystemdNotifier()
# TODO randomize User-Agent
# UA = UserAgent()
# UA.update()
# Text constants from resources:
def get_response_text(file_name):
# https://stackoverflow.com/a/20885799/2490759
# https://docs.python.org/3/library/importlib.resources.html
return resources.files("scdlbot").joinpath("texts").joinpath(file_name).read_text(encoding="UTF-8")
HELP_TEXT = get_response_text("help.tg.md")
SETTINGS_TEXT = get_response_text("settings.tg.md")
DL_TIMEOUT_TEXT = get_response_text("dl_timeout.txt").format(DL_TIMEOUT // 60)
WAIT_BIT_TEXT = [get_response_text("wait_bit.txt"), get_response_text("wait_beat.txt"), get_response_text("wait_beet.txt")]
NO_URLS_TEXT = get_response_text("no_urls.txt")
FAILED_TEXT = get_response_text("failed.txt")
REGION_RESTRICTION_TEXT = get_response_text("region_restriction.txt")
DIRECT_RESTRICTION_TEXT = get_response_text("direct_restriction.txt")
LIVE_RESTRICTION_TEXT = get_response_text("live_restriction.txt")
OLD_MSG_TEXT = get_response_text("old_msg.txt")
# RANT_TEXT_PRIVATE = "Read /help to learn how to use me"
# RANT_TEXT_PUBLIC = f"[Start me in PM to read help and learn how to use me](t.me/{TG_BOT_USERNAME}?start=1)"
# Known and supported site domains:
# support soundcloud.com and soundcloud.app.goo.gl links:
DOMAIN_SC = "soundcloud"
DOMAIN_SC_API = "api.soundcloud.com"
DOMAIN_BC = "bandcamp.com"
DOMAIN_YT = "youtube.com"
DOMAIN_YT_BE = "youtu.be"
DOMAIN_YMR = "music.yandex.ru"
DOMAIN_YMC = "music.yandex.com"
DOMAIN_TT = "tiktok.com"
DOMAIN_IG = "instagram.com"
DOMAIN_TW = "twitter.com"
DOMAIN_TWX = "x.com"
DOMAINS = [DOMAIN_SC, DOMAIN_SC_API, DOMAIN_BC, DOMAIN_YT, DOMAIN_YT_BE, DOMAIN_YMR, DOMAIN_YMC, DOMAIN_TT, DOMAIN_IG, DOMAIN_TW, DOMAIN_TWX]
# TODO get rid of these dumb exceptions:
class FileNotSupportedError(Exception):
def __init__(self, file_format):
self.file_format = file_format
class FileTooLargeError(Exception):
def __init__(self, file_size):
self.file_size = file_size
class FileSplittedPartiallyError(Exception):
def __init__(self, file_parts):
self.file_parts = file_parts
class FileNotConvertedError(Exception):
def __init__(self):
pass
class FileSentPartiallyError(Exception):
def __init__(self, sent_audio_ids):
self.sent_audio_ids = sent_audio_ids
def get_random_wait_text():
return random.choice(WAIT_BIT_TEXT)
def get_link_text(urls):
link_text = ""
for i, url in enumerate(urls):
link_text += "[Source Link #{}]({}) | `{}`\n".format(str(i + 1), url, URL(url).host)
# TODO long link message split in many
direct_urls = urls[url].splitlines()[:3]
for idx, direct_url in enumerate(direct_urls):
if direct_url.startswith("http"):
content_type = ""
if "googlevideo" in direct_url:
if "audio" in direct_url:
content_type = "Audio"
else:
content_type = "Video"
link_text += "• {} #{} [Direct Link]({})\n".format(content_type, str(idx + 1), direct_url)
link_text += "\n*Note:* Final download URLs are only guaranteed to work on the same machine/IP where extracted"
return link_text
def get_settings_inline_keyboard(chat_data):
mode = chat_data["settings"]["mode"]
flood = chat_data["settings"]["flood"]
allow_unknown_sites = chat_data["settings"]["allow_unknown_sites"]
emoji_radio_selected = "🟢"
emoji_radio_unselected = "🟡"
emoji_toggle_enabled = "✅"
emoji_toggle_disabled = "❌"
emoji_close = "❌"
button_dl = InlineKeyboardButton(text=" ".join([emoji_radio_selected if mode == "dl" else emoji_radio_unselected, "Download"]), callback_data=" ".join(["settings", "dl"]))
button_link = InlineKeyboardButton(text=" ".join([emoji_radio_selected if mode == "link" else emoji_radio_unselected, "Links"]), callback_data=" ".join(["settings", "link"]))
button_ask = InlineKeyboardButton(text=" ".join([emoji_radio_selected if mode == "ask" else emoji_radio_unselected, "Ask"]), callback_data=" ".join(["settings", "ask"]))
button_flood = InlineKeyboardButton(text=" ".join([emoji_toggle_enabled if flood else emoji_toggle_disabled, "Captions"]), callback_data=" ".join(["settings", "flood"]))
button_allow_unknown_sites = InlineKeyboardButton(
text=" ".join([emoji_toggle_enabled if allow_unknown_sites else emoji_toggle_disabled, "Unknown sites"]), callback_data=" ".join(["settings", "allow_unknown_sites"])
)
button_close = InlineKeyboardButton(text=" ".join([emoji_close, "Close settings"]), callback_data=" ".join(["settings", "close"]))
inline_keyboard = InlineKeyboardMarkup([[button_dl, button_link, button_ask], [button_allow_unknown_sites, button_flood], [button_close]])
return inline_keyboard
def chat_allowed(chat_id):
if WHITELIST_CHATS:
if chat_id not in WHITELIST_CHATS:
return False
if BLACKLIST_CHATS:
if chat_id in BLACKLIST_CHATS:
return False
return True
def url_valid_and_allowed(url, allow_unknown_sites=False):
host = url.host
if host in BLACKLIST_TELEGRAM_DOMAINS:
return False
if WHITELIST_DOMAINS:
if host not in WHITELIST_DOMAINS:
return False
if BLACKLIST_DOMAINS:
if host in BLACKLIST_DOMAINS:
return False
if allow_unknown_sites:
return True
if any((site in host for site in DOMAINS)):
return True
else:
return False
async def start_help_commands_callback(update: Update, context: ContextTypes.DEFAULT_TYPE):
message = None
if update.channel_post:
message = update.channel_post
elif update.message:
message = update.message
chat_id = update.effective_chat.id
chat_type = update.effective_chat.type
command_name = "help"
# Determine the original command:
entities = message.parse_entities(types=[MessageEntity.BOT_COMMAND])
for entity_value in entities.values():
command_name = entity_value.replace("/", "").replace(f"@{context.bot.username}", "")
break
logger.debug(command_name)
BOT_REQUESTS.labels(type=command_name, chat_type=chat_type, mode="None").inc()
await context.bot.send_message(chat_id=chat_id, text=HELP_TEXT, parse_mode="Markdown", disable_web_page_preview=True)
async def settings_command_callback(update: Update, context: ContextTypes.DEFAULT_TYPE):
command_name = "settings"
chat_id = update.effective_chat.id
chat_type = update.effective_chat.type
logger.debug(command_name)
BOT_REQUESTS.labels(type=command_name, chat_type=chat_type, mode="None").inc()
init_chat_data(
chat_data=context.chat_data,
mode=("dl" if chat_type == Chat.PRIVATE else "ask"),
flood=(chat_id not in NO_FLOOD_CHAT_IDS),
)
await context.bot.send_message(chat_id=chat_id, parse_mode="Markdown", reply_markup=get_settings_inline_keyboard(context.chat_data), text=SETTINGS_TEXT)
async def dl_link_commands_and_messages_callback(update: Update, context: ContextTypes.DEFAULT_TYPE):
message = None
if update.channel_post:
message = update.channel_post
elif update.message:
message = update.message
chat_id = update.effective_chat.id
chat_type = update.effective_chat.type
if not chat_allowed(chat_id):
await context.bot.send_message(chat_id=chat_id, text="This command isn't allowed in this chat.")
return
init_chat_data(
chat_data=context.chat_data,
mode=("dl" if chat_type == Chat.PRIVATE else "ask"),
flood=(chat_id not in NO_FLOOD_CHAT_IDS),
)
# Determine the original command:
command_entities = message.parse_entities(types=[MessageEntity.BOT_COMMAND])
allow_unknown_sites = context.chat_data["settings"]["allow_unknown_sites"]
mode = context.chat_data["settings"]["mode"]
command_passed = False
action = None
if command_entities:
command_passed = True
# Try to determine action from command:
action = None
for entity_value in command_entities.values():
action = entity_value.replace("/", "").replace("@{}".format(context.bot.username), "")
break
# If no command then it is just a message and use message action from settings:
if not action:
action = mode
if action == "silent":
return
if command_passed and not context.args:
# TODO rant for empty commands?
# rant_text = RANT_TEXT_PRIVATE if chat_type == Chat.PRIVATE else RANT_TEXT_PUBLIC
# rant_text += "\nYou can simply send message with links (to download) OR command as `/{} <links>`.".format(mode)
# rant_and_cleanup(context.bot, chat_id, rant_text, reply_to_message_id=reply_to_message_id)
return
command_name = f"{action}_cmd" if command_passed else f"{action}_msg"
logger.debug(command_name)
BOT_REQUESTS.labels(type=command_name, chat_type=chat_type, mode=mode).inc()
apologize = False
# Apologize for fails: always in PM; only when it was explicit command in non-PM:
if chat_type == Chat.PRIVATE or command_passed:
apologize = True
reply_to_message_id = message.message_id
source_ip = None
if SOURCE_IPS:
source_ip = random.choice(SOURCE_IPS)
proxy = None
if PROXIES:
proxy = random.choice(PROXIES)
wait_message_id = None
if action in ["dl", "link"]:
await context.bot.send_chat_action(chat_id=chat_id, action=ChatAction.TYPING)
wait_message = await context.bot.send_message(chat_id=chat_id, reply_to_message_id=reply_to_message_id, parse_mode="Markdown", text=f"_{get_random_wait_text()}_")
wait_message_id = wait_message.message_id
urls_dict = {}
# Get our main running asyncio loop:
loop_main = asyncio.get_event_loop()
# a) Run heavy task blocking the main running asyncio loop.
# Needs to have timeout signals in function, but they are bad.
# urls_dict = get_direct_urls_dict(message, action, proxy, source_ip, allow_unknown_sites)
# b) Run heavy task in separate process or thread, but without blocking the main running asyncio loop.
# Function will continue working till the end: https://stackoverflow.com/a/34457515/2490759
# You may add timeout signals in function, but they are bad.
# If ThreadPoolExecutor - no signals.
# We monitor EXECUTOR process pool task queue, so we use it.
# FIXME maybe don't use wait_for here because it includes pool queue waiting
# pool = concurrent.futures.ThreadPoolExecutor()
try:
# https://docs.python.org/3/library/asyncio-eventloop.html#asyncio.loop.run_in_executor
# https://docs.python.org/3/library/asyncio-task.html#asyncio.wait_for
urls_dict = await asyncio.wait_for(
loop_main.run_in_executor(EXECUTOR, get_direct_urls_dict, message, action, proxy, source_ip, allow_unknown_sites),
timeout=CHECK_URL_TIMEOUT * 10,
)
except asyncio.TimeoutError:
logger.debug("get_direct_urls_dict took too much time and was dropped (but still running)")
except Exception:
logger.debug("get_direct_urls_dict failed for some unhandled reason")
# pool.shutdown(wait=False, cancel_futures=True)
logger.debug(f"prepare_urls: urls dict: {urls_dict}")
urls_values = " ".join(urls_dict.values())
# Continue only if any good direct url status exist (or if we deal with trusted urls):
if action == "dl":
if not urls_dict:
if apologize:
await context.bot.send_message(chat_id=chat_id, reply_to_message_id=reply_to_message_id, text=NO_URLS_TEXT, parse_mode="Markdown")
await context.bot.delete_message(chat_id=chat_id, message_id=wait_message_id)
else:
await context.bot.send_chat_action(chat_id=chat_id, action=ChatAction.TYPING)
for url in urls_dict:
direct_urls_status = urls_dict[url]
if direct_urls_status in ["failed", "restrict_direct", "restrict_region", "restrict_live", "timeout"]:
if direct_urls_status == "failed":
await context.bot.send_message(chat_id=chat_id, reply_to_message_id=reply_to_message_id, text=FAILED_TEXT, parse_mode="Markdown")
elif direct_urls_status == "timeout":
await context.bot.send_message(chat_id=chat_id, reply_to_message_id=reply_to_message_id, text=DL_TIMEOUT_TEXT, parse_mode="Markdown")
elif direct_urls_status == "restrict_direct":
await context.bot.send_message(chat_id=chat_id, reply_to_message_id=reply_to_message_id, text=DIRECT_RESTRICTION_TEXT, parse_mode="Markdown")
elif direct_urls_status == "restrict_region":
await context.bot.send_message(chat_id=chat_id, reply_to_message_id=reply_to_message_id, text=REGION_RESTRICTION_TEXT, parse_mode="Markdown")
elif direct_urls_status == "restrict_live":
await context.bot.send_message(chat_id=chat_id, reply_to_message_id=reply_to_message_id, text=LIVE_RESTRICTION_TEXT, parse_mode="Markdown")
else:
kwargs = {
"bot_options": {
"token": context.bot.token,
"base_url": context.bot.base_url.split("/bot")[0] + "/bot",
"base_file_url": context.bot.base_file_url.split("/file/bot")[0] + "/file/bot",
"local_mode": context.bot.local_mode,
},
"chat_id": chat_id,
"url": url,
"flood": context.chat_data["settings"]["flood"],
"reply_to_message_id": reply_to_message_id,
"wait_message_id": wait_message_id,
"cookies_file": COOKIES_FILE,
"source_ip": source_ip,
"proxy": proxy,
}
await context.bot.send_chat_action(chat_id=chat_id, action=ChatAction.RECORD_VOICE)
# Run heavy task in separate process, "fire and forget":
EXECUTOR.submit(download_url_and_send, **kwargs)
elif action == "link":
if "http" not in urls_values:
if apologize:
await context.bot.send_message(chat_id=chat_id, reply_to_message_id=reply_to_message_id, text=NO_URLS_TEXT, parse_mode="Markdown")
else:
await context.bot.send_chat_action(chat_id=chat_id, action=ChatAction.TYPING)
await context.bot.send_message(
chat_id=chat_id, reply_to_message_id=reply_to_message_id, parse_mode="Markdown", disable_web_page_preview=True, text=get_link_text(urls_dict)
)
await context.bot.delete_message(chat_id=chat_id, message_id=wait_message_id)
elif action == "ask":
if "http" not in urls_values:
if apologize:
await context.bot.send_message(chat_id=chat_id, reply_to_message_id=reply_to_message_id, text=NO_URLS_TEXT, parse_mode="Markdown")
else:
url_message_id = str(reply_to_message_id)
context.chat_data[url_message_id] = {"urls": urls_dict, "source_ip": source_ip, "proxy": proxy}
question = "🎶 links found, what to do?"
button_dl = InlineKeyboardButton(text="⬇️ Download", callback_data=" ".join([url_message_id, "dl"]))
button_link = InlineKeyboardButton(text="🔗️ Get links", callback_data=" ".join([url_message_id, "link"]))
button_cancel = InlineKeyboardButton(text="❌", callback_data=" ".join([url_message_id, "cancel"]))
inline_keyboard = InlineKeyboardMarkup([[button_dl, button_link, button_cancel]])
await context.bot.send_message(chat_id=chat_id, reply_to_message_id=reply_to_message_id, reply_markup=inline_keyboard, text=question)
async def button_press_callback(update: Update, context: ContextTypes.DEFAULT_TYPE):
button_message = update.callback_query.message
button_message_id = button_message.message_id
user_id = update.callback_query.from_user.id
chat = update.effective_chat
chat_id = update.effective_chat.id
chat_type = update.effective_chat.type
# get message id and action from button data:
# TODO create separate callbacks by callback query data pattern
url_message_id, button_action = update.callback_query.data.split()
if not chat_allowed(chat_id):
await update.callback_query.answer(text="This command isn't allowed in this chat.")
return
if url_message_id == "settings":
# button on settings message:
if chat_type != Chat.PRIVATE:
chat_member = await chat.get_member(user_id)
# logger.debug(chat_member.status)
if chat_member.status not in [ChatMember.OWNER, ChatMember.ADMINISTRATOR] and user_id != TG_BOT_OWNER_CHAT_ID:
logger.debug("settings_fail")
await update.callback_query.answer(text="You're not chat admin.")
return
command_name = f"settings_{button_action}"
logger.debug(command_name)
BOT_REQUESTS.labels(type=command_name, chat_type=chat_type, mode="None").inc()
if button_action == "close":
await context.bot.delete_message(chat_id, button_message_id)
else:
setting_changed = False
if button_action in ["dl", "link", "ask"]:
# Radio buttons:
current_setting = context.chat_data["settings"]["mode"]
if button_action != current_setting:
setting_changed = True
context.chat_data["settings"]["mode"] = button_action
elif button_action in ["flood", "allow_unknown_sites"]:
# Toggles:
# TODO support multiple settings windows
current_setting = context.chat_data["settings"][button_action]
context.chat_data["settings"][button_action] = not current_setting
setting_changed = True
if setting_changed:
await update.callback_query.answer(text="Settings changed")
await update.callback_query.edit_message_reply_markup(reply_markup=get_settings_inline_keyboard(context.chat_data))
else:
await update.callback_query.answer(text="Settings not changed")
elif url_message_id in context.chat_data:
# mode is ask, we got data from button on asking message.
# if it asked, then we were in prepare_urls:
url_message_data = context.chat_data.pop(url_message_id)
urls_dict = url_message_data["urls"]
command_name = f"{button_action}_msg"
logger.debug(command_name)
BOT_REQUESTS.labels(type=command_name, chat_type=chat_type, mode="ask").inc()
if button_action == "dl":
await update.callback_query.answer(text=get_random_wait_text())
wait_message = await update.callback_query.edit_message_text(parse_mode="Markdown", text=f"_{get_random_wait_text()}_")
for url in urls_dict:
kwargs = {
"bot_options": {
"token": context.bot.token,
"base_url": context.bot.base_url.split("/bot")[0] + "/bot",
"base_file_url": context.bot.base_file_url.split("/file/bot")[0] + "/file/bot",
"local_mode": context.bot.local_mode,
},
"chat_id": chat_id,
"url": url,
"flood": context.chat_data["settings"]["flood"],
"reply_to_message_id": url_message_id,
"wait_message_id": wait_message.message_id,
"cookies_file": COOKIES_FILE,
"source_ip": url_message_data["source_ip"],
"proxy": url_message_data["proxy"],
}
await context.bot.send_chat_action(chat_id=chat_id, action=ChatAction.RECORD_VOICE)
# Run heavy task in separate process, "fire and forget":
EXECUTOR.submit(download_url_and_send, **kwargs)
elif button_action == "link":
await context.bot.send_message(chat_id=chat_id, reply_to_message_id=url_message_id, parse_mode="Markdown", disable_web_page_preview=True, text=get_link_text(urls_dict))
await context.bot.delete_message(chat_id=chat_id, message_id=button_message_id)
elif button_action == "cancel":
await context.bot.delete_message(chat_id=chat_id, message_id=button_message_id)
else:
await update.callback_query.answer(text=OLD_MSG_TEXT)
await context.bot.delete_message(chat_id=chat_id, message_id=button_message_id)
async def blacklist_whitelist_callback(update: Update, context: ContextTypes.DEFAULT_TYPE):
chat_id = update.effective_chat.id
if not chat_allowed(chat_id):
await context.bot.leave_chat(chat_id)
async def unknown_command_callback(update: Update, context: ContextTypes.DEFAULT_TYPE):
return
async def error_callback(update: Update, context: ContextTypes.DEFAULT_TYPE): # skipcq: PYL-R0201
try:
raise context.error
except Forbidden:
# remove update.message.chat_id from conversation list
logger.debug(f"Update {update} caused Forbidden error: {context.error}")
except BadRequest:
# handle malformed requests - read more below!
logger.debug(f"Update {update} caused BadRequest error: {context.error}")
except TimedOut:
# handle slow connection problems
logger.debug(f"Update {update} caused TimedOut error: {context.error}")
except NetworkError:
# handle other connection problems
logger.debug(f"Update {update} caused NetworkError error: {context.error}")
except ChatMigrated as e:
# the chat_id of a group has changed, use e.new_chat_id instead
logger.debug(f"Update {update} caused ChatMigrated error: {context.error}")
except TelegramError:
# handle all other telegram related errors
logger.debug(f"Update {update} caused TelegramError error: {context.error}")
def init_chat_data(chat_data, mode="dl", flood=True):
if "settings" not in chat_data:
chat_data["settings"] = {}
if "mode" not in chat_data["settings"]:
chat_data["settings"]["mode"] = mode
if "flood" not in chat_data["settings"]:
chat_data["settings"]["flood"] = flood
if "allow_unknown_sites" not in chat_data["settings"]:
chat_data["settings"]["allow_unknown_sites"] = False
def get_direct_urls_dict(message, mode, proxy, source_ip, allow_unknown_sites):
# If telegram message passed:
urls = []
url_entities = message.parse_entities(types=[MessageEntity.URL])
url_caption_entities = message.parse_caption_entities(types=[MessageEntity.URL])
url_entities.update(url_caption_entities)
for entity in url_entities:
url_str = url_entities[entity]
if "://" not in url_str:
url_str = "http://" + url_str
# TODO try except
url = URL(url_str)
if url_valid_and_allowed(url, allow_unknown_sites=allow_unknown_sites):
logger.info("Entity URL parsed: %s", url)
urls.append(url)
else:
logger.info("Entry URL is not valid or blacklisted: %s", url_str)
text_link_entities = message.parse_entities(types=[MessageEntity.TEXT_LINK])
text_link_caption_entities = message.parse_caption_entities(types=[MessageEntity.TEXT_LINK])
text_link_entities.update(text_link_caption_entities)
for entity in text_link_entities:
url = URL(entity.url)
if url_valid_and_allowed(url, allow_unknown_sites=allow_unknown_sites):
logger.info("Entity Text Link parsed: %s", url)
urls.append(url)
else:
logger.info("Entry Text Link is not valid or blacklisted: %s", url)
# If message just some text passed (not isinstance(message, Message)):
# all_links = find_all_links(message, default_scheme="http")
# urls = [link for link in all_links if url_valid_and_allowed(link)]
logger.info(f"prepare_urls: urls list: {urls}")
urls_dict = {}
for url_item in urls:
# FIXME better check domain in hostname
unknown_site = not any((site in url_item.host for site in DOMAINS))
# unshorten soundcloud.app.goo.gl and unknown sites links
# example: https://soundcloud.app.goo.gl/mBMvG
if unknown_site or DOMAIN_SC in url_item.host:
proxy_arg = None
if proxy:
proxy_arg = {"http": proxy, "https": proxy}
try:
url = URL(
requests.head(
url_item.to_text(full_quote=True),
allow_redirects=True,
timeout=2,
proxies=proxy_arg,
# headers={"User-Agent": UA.random},
headers={"User-Agent": "Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:105.0) Gecko/20100101 Firefox/105.0"},
).url
)
except:
url = url_item
else:
url = url_item
unknown_site = not any((site in url.host for site in DOMAINS))
url_text = url.to_text(full_quote=True)
logger.debug(f"unshortened link: {url_text}")
# url_text = url_text.replace("m.soundcloud.com", "soundcloud.com")
url_parts_num = len([part for part in url.path_parts if part])
if mode == "link" or unknown_site:
# We run it if it was explicitly requested as per "link" mode.
# We run it for links from unknown sites (if they were allowed).
# If it's known site, we need to check it more thoroughly below.
urls_dict[url_text] = ydl_get_direct_urls(url_text, COOKIES_FILE, source_ip, proxy)
elif DOMAIN_SC in url.host and (2 <= url_parts_num <= 4 or DOMAIN_SC_API in url.host) and (not "you" in url.path_parts):
# SoundCloud: tracks, sets and widget pages, no /you/ pages
# TODO support private sets URLs that have 5 parts
# We know for sure these links can be downloaded, so we just skip running ydl_get_direct_urls
urls_dict[url_text] = "http"
elif DOMAIN_BC in url.host and (2 <= url_parts_num <= 2):
# Bandcamp: tracks and albums
# We know for sure these links can be downloaded, so we just skip running ydl_get_direct_urls
urls_dict[url_text] = "http"
elif (DOMAIN_YT in url.host or DOMAIN_YT_BE in url.host) and (DOMAIN_YT_BE in url.host or "watch" in url.path or "playlist" in url.path):
# YouTube: videos and playlists
# We still run it for checking YouTube region restriction to avoid fake asking:
urls_dict[url_text] = ydl_get_direct_urls(url_text, COOKIES_FILE, source_ip, proxy)
elif DOMAIN_YMR in url.host or DOMAIN_YMC in url.host:
# YM: tracks. Note that the domain includes x.com..
# We know for sure these links can be downloaded, so we just skip running ydl_get_direct_urls
urls_dict[url_text] = "http"
elif DOMAIN_TT in url.host:
# TikTok: videos
# We know for sure these links can be downloaded, so we just skip running ydl_get_direct_urls
urls_dict[url_text] = "http"
elif DOMAIN_IG in url.host:
# Instagram: videos, reels
# TODO We run it for checking Instagram ban to avoid fake asking:
urls_dict[url_text] = ydl_get_direct_urls(url_text, COOKIES_FILE, source_ip, proxy)
elif (DOMAIN_TW in url.host or DOMAIN_TWX in url.host) and (DOMAIN_YMC not in url.host) and (3 <= url_parts_num <= 3):
# Twitter: videos
# We know for sure these links can be downloaded, so we just skip running ydl_get_direct_urls
urls_dict[url_text] = "http"
return urls_dict
def ydl_get_direct_urls(url, cookies_file=None, source_ip=None, proxy=None):
# TODO transform into unified ydl function and deduplicate
logger.debug("Entering: ydl_get_direct_urls: %s", url)
status = ""
cmd_name = "ydl_get_direct_urls"
ydl_opts = {
"format": "bestaudio/best",
"noplaylist": True,
"skip_download": True,
# "forceprint": {"before_dl":}
}
if proxy:
ydl_opts["proxy"] = proxy
if source_ip:
ydl_opts["source_address"] = source_ip
cookies_download_file = None
if cookies_file:
cookies_download_file = tempfile.NamedTemporaryFile(mode="wb", delete=False)
cookies_download_file_path = pathlib.Path(cookies_download_file.name)
if cookies_file.startswith("http"):
# URL for downloading cookie file:
try:
r = requests.get(cookies_file, allow_redirects=True, timeout=5)
cookies_download_file.write(r.content)
cookies_download_file.close()
ydl_opts["cookiefile"] = str(cookies_download_file_path)
except:
logger.debug("download_url_and_send could not download cookies file")
pass
elif cookies_file.startswith("firefox:"):
# TODO better handling of env var
cookies_file_components = cookies_file.split(":", maxsplit=2)
if len(cookies_file_components) == 3:
cookies_sqlite_file = cookies_file_components[2]
cookies_download_sqlite_path = pathlib.Path.home() / ".mozilla" / "firefox" / cookies_file_components[1] / "cookies.sqlite"
# URL for downloading cookie sqlite file:
try:
r = requests.get(cookies_sqlite_file, allow_redirects=True, timeout=5)
with open(cookies_download_sqlite_path, "wb") as cfile:
cfile.write(r.content)
ydl_opts["cookiesfrombrowser"] = ("firefox", cookies_file_components[1], None, None)
logger.debug("download_url_and_send downloaded cookies.sqlite file")
except:
logger.debug("download_url_and_send could not download cookies.sqlite file")
pass
else:
ydl_opts["cookiesfrombrowser"] = ("firefox", cookies_file_components[1], None, None)
else:
# cookie file local path:
cookies_download_file.write(open(cookies_file, "rb").read())
cookies_download_file.close()
ydl_opts["cookiefile"] = str(cookies_download_file_path)
# FIXME apply CHECK_URL_TIMEOUT by using a process (as we did before). Maybe use https://github.com/noxdafox/pebble
logger.debug("%s starts: %s", cmd_name, url)
try:
info_dict = ydl.YoutubeDL(ydl_opts).extract_info(url, download=False)
# TODO actualize checks, fix for youtube playlists
if "url" in info_dict:
direct_url = info_dict["url"]
elif "entries" in info_dict:
direct_url = "\n".join([x["url"] for x in info_dict["entries"] if "url" in x])
else:
raise Exception()
if "yt_live_broadcast" in direct_url:
status = "restrict_live"
elif "returning it as such" in direct_url:
status = "restrict_direct"
elif "proxy server" in direct_url:
status = "restrict_region"
# end actualize checks
else:
status = direct_url
logger.debug("%s succeeded: %s", cmd_name, url)
except Exception:
logger.debug("%s failed: %s", cmd_name, url)
logger.debug(traceback.format_exc())
status = "failed"
if cookies_file:
cookies_download_file.close()
os.unlink(cookies_download_file.name)
return status
def download_url_and_send(
bot_options,
chat_id,
url,
flood=False,
reply_to_message_id=None,
wait_message_id=None,
cookies_file=None,
source_ip=None,
proxy=None,
):
logger.debug("Entering: download_url_and_send")
# loop_main = asyncio.get_event_loop()
# We use ProcessPoolExecutor that runs "fork".
# It has really useful fire-and-forget ProcessPoolExecutor.submit() that only accepts sync functions.
# Also, there is no point in using asyncio event loop/async functions inside it.
# Hence, download_url_and_send() is sync function, not async.
# But we still want to use async Bot functions from ptb framework here.
# We can't use loop_main.run_in_executor(None) because it doesn't return the result.
# We can't use loop_main.run_until_complete() because loop is already running in our forked process.
# We can't use asyncio.run_coroutine_threadsafe(coro, loop_main) because it doesn't work (interferes with framework?).
# So we run additional loop in additional thread and just use it:
loop_additional = asyncio.new_event_loop()
thread_additional = threading.Thread(target=loop_additional.run_forever, name="Async Runner", daemon=True)
def run_async(coro):
if not thread_additional.is_alive():
thread_additional.start()
future = asyncio.run_coroutine_threadsafe(coro, loop_additional)
return future.result()
# We must not pass context/bot here, because they need to get serialized/pickled on "fork" (and they cannot be).
# https://docs.python-telegram-bot.org/en/v20.1/telegram.bot.html
# So we create and use new Bot object:
bot = Bot(
token=bot_options["token"],
base_url=bot_options["base_url"],
base_file_url=bot_options["base_file_url"],
local_mode=bot_options["local_mode"],
request=HTTPXRequest(http_version=HTTP_VERSION),
get_updates_request=HTTPXRequest(http_version=HTTP_VERSION),
)
run_async(bot.initialize())
logger.debug(bot.token)
download_dir = os.path.join(DL_DIR, str(uuid4()))
shutil.rmtree(download_dir, ignore_errors=True)
os.makedirs(download_dir)
url_obj = URL(url)
host = url_obj.host
download_video = False
status = "initial"
add_description = ""
cmd = None
cmd_name = ""
cmd_args = ()
cmd_input = None
if (DOMAIN_SC in host and DOMAIN_SC_API not in host) or (DOMAIN_BC in host and BCDL_ENABLE):
# If link is sc/bc, we try scdl/bcdl first:
if DOMAIN_SC in host and DOMAIN_SC_API not in host:
cmd = scdl_bin
cmd_name = str(cmd)
cmd_args = (
"-l",
url, # URL of track/playlist/user
"-c", # Continue if a music already exist
"--path",
download_dir, # Download the music to a custom path
"--onlymp3", # Download only the mp3 file even if the track is Downloadable
"--addtofile", # Add the artist name to the filename if it isn't in the filename already
"--addtimestamp",
# Adds the timestamp of the creation of the track to the title (useful to sort chronologically)
"--no-playlist-folder",
# Download playlist tracks into directory, instead of making a playlist subfolder
"--extract-artist", # Set artist tag from title instead of username
)
cmd_input = None
elif DOMAIN_BC in host and BCDL_ENABLE:
cmd = bcdl_bin
cmd_name = str(cmd)
cmd_args = (
"--base-dir",
download_dir, # Base location of which all files are downloaded
"--template",
"%{track} - %{artist} - %{title} [%{album}]", # Output filename template
"--overwrite", # Overwrite tracks that already exist
"--group", # Use album/track Label as iTunes grouping
# "--embed-art", # Embed album art (if available)
"--no-slugify", # Disable slugification of track, album, and artist names
url, # URL of album/track
)
cmd_input = "yes"
env = None
if proxy:
env = {"http_proxy": proxy, "https_proxy": proxy}
logger.debug("%s starts: %s", cmd_name, url)
cmd_proc = cmd[cmd_args].popen(env=env, stdin=PIPE, stdout=PIPE, stderr=PIPE, universal_newlines=True)
try:
cmd_stdout, cmd_stderr = cmd_proc.communicate(input=cmd_input, timeout=DL_TIMEOUT)
cmd_retcode = cmd_proc.returncode
# listed are common scdl problems for one track with 0 retcode, all its output is always in stderr:
if cmd_retcode or (any(err in cmd_stderr for err in ["Error resolving url", "is not streamable", "Failed to get item"]) and ".mp3" not in cmd_stderr):
raise ProcessExecutionError(cmd_args, cmd_retcode, cmd_stdout, cmd_stderr)
logger.debug("%s succeeded: %s", cmd_name, url)
status = "success"
except TimeoutExpired:
cmd_proc.kill()
logger.debug("%s took too much time and dropped: %s", cmd_name, url)
except ProcessExecutionError:
logger.debug("%s failed: %s", cmd_name, url)
logger.debug(traceback.format_exc())
if status == "initial":
# If link is not sc/bc or scdl/bcdl just failed, we use ydl
cmd_name = "ydl_download"
# https://github.com/yt-dlp/yt-dlp/blob/master/yt_dlp/YoutubeDL.py#L172
# https://github.com/yt-dlp/yt-dlp/blob/master/yt_dlp/utils/_utils.py
ydl_opts = {
# https://github.com/yt-dlp/yt-dlp#output-template
# Default outtmpl is "%(title)s [%(id)s].%(ext)s"
# Take first 16 symbols of title:
"outtmpl": os.path.join(download_dir, "%(title).16s [%(id)s].%(ext)s"),
"restrictfilenames": True,
"windowsfilenames": True,
# TODO Support ffmpeg_location parameter or just use BIN_PATH here:
# "ffmpeg_location": "/home/gpchelkin/.local/bin/",
# "trim_file_name": 32,
}
if DOMAIN_TT in host:
download_video = True
ydl_opts["format"] = "mp4"
elif (DOMAIN_TW in host or DOMAIN_TWX in host) and (DOMAIN_YMC not in host):
download_video = True
ydl_opts["format"] = "mp4"
elif DOMAIN_IG in host:
download_video = True
ydl_opts.update(
{
"format": "mp4",
"postprocessors": [
# Instagram gives VP9 cideo codec (when downloading with cookies) and it doesn't play in Telegram iOS client.
# We need AVC (x264/h264) or HEVC (x265/h265) video (+ AAC audio) from Instagram videos.
# "FFmpegVideoConvertor" doesn't work here since the original file is already mp4.
# We don't touch audio and just copy it here since it's probably OK in original. But we can change copy to aac.
# --use-postprocessor FFmpegCopyStream --ppa copystream:"-codec:v libx264 -crf 27 -preset veryfast -codec:a copy -f mp4"
# https://github.com/yt-dlp/yt-dlp/issues/7607
# https://github.com/yt-dlp/yt-dlp/issues/5859
# https://github.com/yt-dlp/yt-dlp/issues/8904
# {"key": "FFmpegVideoConvertor", "preferedformat": "mp4"},
{"key": "FFmpegCopyStream"},
],
"postprocessor_args": {
"copystream": ["-codec:v", "libx264", "-crf", "35", "-preset", "ultrafast", "-codec:a", "copy", "-f", "mp4"],
},
}
)
else:
ydl_opts.update(
{
"format": "bestaudio/best",
"postprocessors": [
{"key": "FFmpegExtractAudio", "preferredcodec": "mp3", "preferredquality": "320"},
{"key": "FFmpegMetadata"},
# {"key": "EmbedThumbnail"},
],
"noplaylist": True,
}
)
if proxy:
ydl_opts["proxy"] = proxy
if source_ip:
ydl_opts["source_address"] = source_ip
cookies_download_file = None
if cookies_file:
cookies_download_file = tempfile.NamedTemporaryFile(mode="wb", delete=False)
cookies_download_file_path = pathlib.Path(cookies_download_file.name)
if cookies_file.startswith("http"):
# URL for downloading cookie file:
try:
r = requests.get(cookies_file, allow_redirects=True, timeout=5)
cookies_download_file.write(r.content)
cookies_download_file.close()
ydl_opts["cookiefile"] = str(cookies_download_file_path)
except:
logger.debug("download_url_and_send could not download cookies file")
pass
elif cookies_file.startswith("firefox:"):
# TODO better handling of env var
cookies_file_components = cookies_file.split(":", maxsplit=2)
if len(cookies_file_components) == 3:
cookies_sqlite_file = cookies_file_components[2]
cookies_download_sqlite_path = pathlib.Path.home() / ".mozilla" / "firefox" / cookies_file_components[1] / "cookies.sqlite"
# URL for downloading cookie sqlite file:
try:
r = requests.get(cookies_sqlite_file, allow_redirects=True, timeout=5)
with open(cookies_download_sqlite_path, "wb") as cfile:
cfile.write(r.content)
ydl_opts["cookiesfrombrowser"] = ("firefox", cookies_file_components[1], None, None)
logger.debug("download_url_and_send downloaded cookies.sqlite file")
except:
logger.debug("download_url_and_send could not download cookies.sqlite file")
pass
else:
ydl_opts["cookiesfrombrowser"] = ("firefox", cookies_file_components[1], None, None)
else:
# cookie file local path:
cookies_download_file.write(open(cookies_file, "rb").read())
cookies_download_file.close()
ydl_opts["cookiefile"] = str(cookies_download_file_path)
# FIXME apply DL_TIMEOUT by using a process (as we did before). Maybe use https://github.com/noxdafox/pebble
logger.debug("%s starts: %s", cmd_name, url)
try:
# TODO check result
info_dict = ydl.YoutubeDL(ydl_opts).download([url])
logger.debug("%s succeeded: %s", cmd_name, url)
status = "success"
if download_video:
info_dict = ydl.YoutubeDL(ydl_opts).extract_info(url, download=False)
if "description" in info_dict and info_dict["description"]:
# TODO better handle right-to-left hashtags https://www.instagram.com/reel/CtZbNhtrJv3/
add_description = escape_markdown(info_dict["description"][:800], version=1)
except Exception as exc:
print(exc)
logger.debug("%s failed: %s", cmd_name, url)
logger.debug(traceback.format_exc())
status = "failed"
if cookies_file:
cookies_download_file.close()
os.unlink(cookies_download_file.name)
# gc.collect()
if status == "failed":
run_async(bot.send_message(chat_id=chat_id, reply_to_message_id=reply_to_message_id, text=FAILED_TEXT, parse_mode="Markdown"))
elif status == "timeout":
run_async(bot.send_message(chat_id=chat_id, reply_to_message_id=reply_to_message_id, text=DL_TIMEOUT_TEXT, parse_mode="Markdown"))
elif status == "success":
file_list = []
for d, dirs, files in os.walk(download_dir):
for file in files:
file_list.append(os.path.join(d, file))
if not file_list:
logger.debug("No files in dir: %s", download_dir)
run_async(
bot.send_message(
chat_id=chat_id, reply_to_message_id=reply_to_message_id, text="*Sorry*, I couldn't download any files from some of the provided links", parse_mode="Markdown"
)
)
else:
for file in sorted(file_list):
file_name = os.path.split(file)[-1]
file_parts = []
try:
file_root, file_ext = os.path.splitext(file)
file_format = file_ext.replace(".", "").lower()
file_size = os.path.getsize(file)
if file_format not in ["mp3", "m4a", "mp4", "webm"]:
raise FileNotSupportedError(file_format)
# We don't convert videos (from tiktok or instagram or twitter):
if file_format in ["m4a", "mp4", "webm"] and not download_video:
if file_size > MAX_CONVERT_FILE_SIZE:
raise FileTooLargeError(file_size)
logger.debug("Converting: %s", file)
try:
file_converted = file.replace(file_ext, ".mp3")
ffinput = ffmpeg.input(file)
# https://kkroening.github.io/ffmpeg-python/#ffmpeg.output
# We could set audio_bitrate="320k", but we don't need it now
ffmpeg.output(ffinput, file_converted, vn=None).run()
file = file_converted
file_root, file_ext = os.path.splitext(file)
file_format = file_ext.replace(".", "").lower()
file_size = os.path.getsize(file)
except Exception:
raise FileNotConvertedError
file_parts = []
if file_size <= MAX_TG_FILE_SIZE:
file_parts.append(file)
else:
logger.debug("Splitting: %s", file)
id3 = None
try:
id3 = ID3(file, translate=False)
except:
pass
parts_number = file_size // MAX_TG_FILE_SIZE + 1
# https://github.com/c0decracker/video-splitter
# https://superuser.com/a/1354956/464797
try:
# file_duration = float(ffmpeg.probe(file)['format']['duration'])
part_size = file_size // parts_number
cur_position = 0
for i in range(parts_number):
file_part = file.replace(file_ext, ".part{}{}".format(str(i + 1), file_ext))
ffinput = ffmpeg.input(file)
if i == (parts_number - 1):
ffmpeg.output(ffinput, file_part, codec="copy", vn=None, ss=cur_position).run()
else:
ffmpeg.output(ffinput, file_part, codec="copy", vn=None, ss=cur_position, fs=part_size).run()
part_duration = float(ffmpeg.probe(file_part)["format"]["duration"])
cur_position += part_duration
if id3:
try:
id3.save(file_part, v1=ID3v1SaveOptions.CREATE, v2_version=4)
except:
pass
file_parts.append(file_part)
except Exception:
raise FileSplittedPartiallyError(file_parts)
except FileNotSupportedError as exc:
# If format is not some extra garbage from downloaders:
if not (exc.file_format in ["m3u", "jpg", "jpeg", "png", "finished", "tmp"]):
logger.debug("Unsupported file format: %s", file_name)
run_async(
bot.send_message(
chat_id=chat_id,
reply_to_message_id=reply_to_message_id,
text="*Sorry*, downloaded file `{}` is in format I could not yet convert or send".format(file_name),
parse_mode="Markdown",
)
)
except FileTooLargeError as exc:
logger.debug("Large file for convert: %s", file_name)
run_async(
bot.send_message(
chat_id=chat_id,
reply_to_message_id=reply_to_message_id,
text="*Sorry*, downloaded file `{}` is `{}` MB and it is larger than I could convert (`{} MB`)".format(
file_name, exc.file_size // 1000000, MAX_CONVERT_FILE_SIZE // 1000000
),
parse_mode="Markdown",
)
)
except FileSplittedPartiallyError as exc:
file_parts = exc.file_parts
logger.debug("Splitting failed: %s", file_name)
run_async(
bot.send_message(
chat_id=chat_id,
reply_to_message_id=reply_to_message_id,
text="*Sorry*, not enough memory to convert file `{}`..".format(file_name),
parse_mode="Markdown",
)
)
except FileNotConvertedError as exc:
logger.debug("Splitting failed: %s", file_name)
run_async(
bot.send_message(
chat_id=chat_id,
reply_to_message_id=reply_to_message_id,
text="*Sorry*, not enough memory to convert file `{}`..".format(file_name),
parse_mode="Markdown",
)
)
caption = None
reply_to_message_id_send = None
if flood:
addition = ""
if DOMAIN_YT in host or DOMAIN_YT_BE in host:
source = "YouTube"
file_root, file_ext = os.path.splitext(file_name)
file_title = file_root.replace(file_ext, "")
addition = ": " + file_title
elif DOMAIN_SC in host:
source = "SoundCloud"
elif DOMAIN_BC in host:
source = "Bandcamp"
else:
source = url_obj.host.replace(".com", "").replace(".ru", "").replace("www.", "").replace("m.", "")
# TODO fix youtube id in []
caption = "@{} _got it from_ [{}]({}){}".format(bot.username.replace("_", r"\_"), source, url, addition.replace("_", r"\_"))
if add_description:
caption += "\n\n" + add_description
# logger.debug(caption)
reply_to_message_id_send = reply_to_message_id
sent_audio_ids = []
for index, file_part in enumerate(file_parts):
path = pathlib.Path(file_part)
file_name = os.path.split(file_part)[-1]
# file_name = translit(file_name, 'ru', reversed=True)
logger.debug("Sending: %s", file_name)
run_async(bot.send_chat_action(chat_id=chat_id, action=ChatAction.UPLOAD_VOICE))
caption_part = None
if len(file_parts) > 1:
caption_part = "Part {} of {}".format(str(index + 1), str(len(file_parts)))
if caption:
if caption_part:
caption_full = caption_part + " | " + caption
else:
caption_full = caption
else:
if caption_part:
caption_full = caption_part
else:
caption_full = ""
# caption_full = textwrap.shorten(caption_full, width=190, placeholder="..")
retries = 3
for i in range(retries):
try:
logger.debug(f"Trying {i+1} time to send file part: {file_part}")
if file_part.endswith(".mp3"):
mp3 = MP3(file_part)
duration = round(mp3.info.length)
performer = None
title = None
try:
performer = ", ".join(mp3["artist"])
title = ", ".join(mp3["title"])
except:
pass
if TG_BOT_API_LOCAL_MODE:
audio = path.absolute().as_uri()
logger.debug(audio)
else:
audio = open(file_part, "rb")
# Bot.send_audio() has connection troubles when running async in parallel:
# Works bad on my computer with official API (good with high timeout)
# Works good on server with local API.
audio_msg = run_async(
bot.send_audio(
chat_id=chat_id,
reply_to_message_id=reply_to_message_id_send,
audio=audio,
duration=duration,
performer=performer,
title=title,
caption=caption_full,
parse_mode="Markdown",
read_timeout=COMMON_CONNECTION_TIMEOUT,
write_timeout=COMMON_CONNECTION_TIMEOUT,
connect_timeout=COMMON_CONNECTION_TIMEOUT,
pool_timeout=COMMON_CONNECTION_TIMEOUT,
),
)
sent_audio_ids.append(audio_msg.audio.file_id)
logger.debug("Sending audio succeeded: %s", file_name)
break
elif download_video:
video = open(file_part, "rb")
duration = int(float(ffmpeg.probe(file_part)["format"]["duration"]))
videostream = next(item for item in ffmpeg.probe(file_part)["streams"] if item["codec_type"] == "video")
width = int(videostream["width"])
height = int(videostream["height"])
video_msg = run_async(
bot.send_video(
chat_id=chat_id,
reply_to_message_id=reply_to_message_id_send,
video=video,
supports_streaming=True,
duration=duration,
width=width,
height=height,
caption=caption_full,
parse_mode="Markdown",
read_timeout=COMMON_CONNECTION_TIMEOUT,
write_timeout=COMMON_CONNECTION_TIMEOUT,
connect_timeout=COMMON_CONNECTION_TIMEOUT,
pool_timeout=COMMON_CONNECTION_TIMEOUT,
),
)
sent_audio_ids.append(video_msg.video.file_id)
logger.debug("Sending video succeeded: %s", file_name)
break
except TelegramError:
### ??? print(traceback.format_exc())
if i == retries - 1:
logger.debug("Sending failed because of TelegramError: %s", file_name)
else:
time.sleep(5)
if len(sent_audio_ids) != len(file_parts):
run_async(
bot.send_message(
chat_id=chat_id,
reply_to_message_id=reply_to_message_id,
text="*Sorry*, could not send file `{}` or some of it's parts..".format(file_name),
parse_mode="Markdown",
)
)
logger.debug("Sending some parts failed: %s", file_name)
shutil.rmtree(download_dir, ignore_errors=True)
if wait_message_id:
try:
run_async(
bot.delete_message(
chat_id=chat_id,
message_id=wait_message_id,
),
)
except:
pass
run_async(bot.shutdown())
async def post_shutdown(application: Application) -> None:
EXECUTOR.shutdown(wait=False, cancel_futures=True)
async def post_init(application: Application) -> None:
SYSTEMD_NOTIFIER.notify("READY=1")
SYSTEMD_NOTIFIER.notify(f"STATUS=Application initialized")
async def callback_watchdog(context: ContextTypes.DEFAULT_TYPE):
SYSTEMD_NOTIFIER.notify("WATCHDOG=1")
SYSTEMD_NOTIFIER.notify(f"STATUS=Watchdog was sent {datetime.datetime.now()}")
async def callback_monitor(context: ContextTypes.DEFAULT_TYPE):
logger.debug(f"EXECUTOR pending work items: {len(EXECUTOR._pending_work_items)} tasks remain")
EXECUTOR_TASKS_REMAINING.set(len(EXECUTOR._pending_work_items))
def main():
# Start exposing Prometheus/OpenMetrics metrics:
prometheus_client.start_http_server(addr=METRICS_HOST, port=METRICS_PORT, registry=REGISTRY)
# Maybe we can use token again if we will buy SoundCloud Go+
# https://github.com/flyingrub/scdl/issues/429
# if sc_auth_token:
# config = configparser.ConfigParser()
# config['scdl'] = {}
# config['scdl']['path'] = DL_DIR
# config['scdl']['auth_token'] = sc_auth_token
# config_dir = os.path.join(os.path.expanduser('~'), '.config', 'scdl')
# config_path = os.path.join(config_dir, 'scdl.cfg')
# os.makedirs(config_dir, exist_ok=True)
# with open(config_path, 'w') as config_file:
# config.write(config_file)
persistence = PicklePersistence(filepath=CHAT_STORAGE)
# https://docs.python-telegram-bot.org/en/v20.1/telegram.ext.applicationbuilder.html#telegram.ext.ApplicationBuilder
# We use concurrent_updates with limit instead of unlimited create_task.
# https://github.com/python-telegram-bot/python-telegram-bot/wiki/Concurrency#applicationconcurrent_updates
# https://github.com/python-telegram-bot/python-telegram-bot/issues/3509
application = (
ApplicationBuilder()
.token(TG_BOT_TOKEN)
.local_mode(TG_BOT_API_LOCAL_MODE)
# https://github.com/python-telegram-bot/python-telegram-bot/issues/3556
.http_version(HTTP_VERSION)
.get_updates_http_version(HTTP_VERSION)
.base_url(f"{TG_BOT_API}/bot")
.base_file_url(f"{TG_BOT_API}/file/bot")
.persistence(persistence)
.post_init(post_init)
.post_shutdown(post_shutdown)
.concurrent_updates(256)
.connection_pool_size(512)
.connect_timeout(COMMON_CONNECTION_TIMEOUT)
.read_timeout(COMMON_CONNECTION_TIMEOUT)
.write_timeout(COMMON_CONNECTION_TIMEOUT)
.pool_timeout(COMMON_CONNECTION_TIMEOUT)
.build()
)
bot_username = requests.get(f"{TG_BOT_API}/bot{TG_BOT_TOKEN}/getMe").json()["result"]["username"]
blacklist_whitelist_handler = MessageHandler(filters.StatusUpdate.NEW_CHAT_MEMBERS, blacklist_whitelist_callback)
start_command_handler = CommandHandler("start", start_help_commands_callback)
help_command_handler = CommandHandler("help", start_help_commands_callback)
settings_command_handler = CommandHandler("settings", settings_command_callback)
dl_command_handler = CommandHandler("dl", dl_link_commands_and_messages_callback, filters=~filters.UpdateType.EDITED_MESSAGE & ~filters.FORWARDED)
link_command_handler = CommandHandler("link", dl_link_commands_and_messages_callback, filters=~filters.UpdateType.EDITED_MESSAGE & ~filters.FORWARDED)
message_with_links_handler = MessageHandler(
~filters.UpdateType.EDITED_MESSAGE
& ~filters.ForwardedFrom(username=bot_username)
& ~filters.COMMAND
& (
(filters.TEXT & (filters.Entity(MessageEntity.URL) | filters.Entity(MessageEntity.TEXT_LINK)))
| (filters.CAPTION & (filters.CaptionEntity(MessageEntity.URL) | filters.CaptionEntity(MessageEntity.TEXT_LINK)))
),
dl_link_commands_and_messages_callback,
)
button_query_handler = CallbackQueryHandler(button_press_callback)
unknown_handler = MessageHandler(filters.COMMAND, unknown_command_callback)
application.add_handler(blacklist_whitelist_handler)
application.add_handler(start_command_handler)
application.add_handler(help_command_handler)
application.add_handler(settings_command_handler)
application.add_handler(dl_command_handler)
application.add_handler(link_command_handler)
application.add_handler(message_with_links_handler)
application.add_handler(button_query_handler)
application.add_handler(unknown_handler)
application.add_error_handler(error_callback)
job_queue = application.job_queue
job_watchdog = job_queue.run_repeating(callback_watchdog, interval=60, first=10)
job_monitor = job_queue.run_repeating(callback_monitor, interval=5, first=5)
if WEBHOOK_ENABLE:
application.run_webhook(
drop_pending_updates=True,
listen=WEBHOOK_HOST,
port=WEBHOOK_PORT,
url_path=WEBHOOK_APP_URL_PATH,
webhook_url=urljoin(WEBHOOK_APP_URL_ROOT, WEBHOOK_APP_URL_PATH),
secret_token=WEBHOOK_SECRET_TOKEN,
max_connections=1024,
cert=WEBHOOK_CERT_FILE,
key=WEBHOOK_KEY_FILE,
)
else:
# TODO await it somehow
application.bot.delete_webhook()
application.run_polling(
drop_pending_updates=True,
)
if __name__ == "__main__":
main()