false-mirror

Automated TV show downloader for AniLibria and LostFilm. Monitors tracked shows for new episodes, downloads them via qBittorrent, and moves completed files to a Synology NAS. Includes a Telegram bot for managing the watch list.

How it works

1. You send a link to the Telegram bot (/download <url>)
2. Bot saves the show and immediately checks for available episodes
3. Episodes are downloaded via qBittorrent to the NAS
4. A scheduler rechecks all shows hourly (each show at a random time to spread load)
5. Post-download tracker verifies files landed on NAS and alerts via Telegram

Quick Start

git clone https://github.com/icyberdeveloper/false-mirror.git
cd false-mirror/deploy

Edit compose.yml and set environment variables:

Variable	Required	Description
LF_SESSION	Yes	LostFilm session cookie
TG_BOT_TOKEN	Yes	Telegram bot token from @BotFather
HEALTHCHECK_TG_CHAT_ID	Yes	Your Telegram chat ID for alerts
QB_USERNAME	No	qBittorrent username (default: admin)
QB_PASSWORD	No	qBittorrent password (default: adminadmin)

docker-compose build
docker-compose up -d

Telegram Bot Commands

• /download <url> - Add a show or movie. Accepts:
  • lostfilm.*/series/... — TV series (tracked, periodic checks)
  • lostfilm.*/movies/... — movies (fire-and-forget, downloads immediately)
  • anilibria.*/release/... — anime series (tracked, periodic checks)
• /list - Show all tracked series
• /start - Help

Project Structure

false-mirror/
├── app/                    # Python application
│   ├── scheduler.py        # Periodic checks (entry point)
│   ├── bot.py              # Telegram bot (entry point)
│   ├── worker.py           # Isolated per-series check functions
│   ├── config.py           # Config parsing
│   ├── clients/            # Data providers
│   │   ├── anilibria.py    # AniLibria REST API client
│   │   └── lostfilm.py     # LostFilm HTML scraper
│   └── services/           # Shared services
│       ├── database.py     # TinyDB storage
│       ├── library.py      # Filesystem scanner (duplicate prevention)
│       ├── qbittorrent.py  # qBittorrent API wrapper
│       ├── tracker.py      # Post-download verification + alerts
│       ├── renamer.py      # File renaming (Show - s01e01.mkv)
│       └── network.py      # HTTP client with retry
├── deploy/                 # Infrastructure
│   ├── compose.yml         # Docker Compose
│   ├── Dockerfile          # Scheduler container
│   ├── Dockerfile.bot      # Bot container
│   ├── bootstrap.sh        # Full server recovery script
│   └── backup-to-nas.sh    # Daily backup to NAS
├── config.yaml             # Application config
└── requirements.txt

Architecture

Two triggers:

• Immediate - Bot adds a show and instantly checks for episodes in a background thread
• Periodic - Scheduler assigns each show a random time slot within a 1-hour cycle, spreading load evenly

Error isolation: Each show is processed independently. If one show's source is down, others continue working normally.

Duplicate prevention (three layers):

1. Database check - is this episode tracked?
2. Filesystem scan - is it already on NAS?
3. qBittorrent queue check - is it already downloading?

Post-download verification: Tracker monitors qBittorrent state per torrent. On completion, verifies the file exists on NAS. Alerts via Telegram on problems (stalled downloads, missing files, errors).

Docker Images

All images are hosted on Docker Hub and built by CI on every push to main:

Image	Description
icyberdeveloper/false-mirror	Scheduler (periodic episode checks)
icyberdeveloper/nocron	Telegram bot
icyberdeveloper/qbittorrent	Mirror of linuxserver/qbittorrent (pinned version)

Images are tagged with latest and git SHA for version tracking.

cd deploy
docker-compose pull && docker-compose up -d   # Update from registry
docker-compose up -d --build                   # Build locally (development)

Configuration

Application config is in config.yaml. Secrets are set via environment variables in deploy/compose.yml.

Providers

AniLibria - REST API. Selects best quality torrent (prefers non-HEVC). Downloads full seasons.

LostFilm - HTML scraper. Follows multi-step redirects to get torrent URLs. Requires LF_SESSION cookie (get it from browser after logging in to lostfilm.tv). Prefers 1080p, falls back to 720p/HDTVRip. Downloads individual episodes.

Disaster Recovery

All configs and data are backed up daily to NAS. To rebuild the server from scratch:

apt update && apt install -y git nfs-common
mkdir -p /mnt/backups && mount -t nfs4 192.168.1.150:/volume1/backups /mnt/backups
git clone https://github.com/icyberdeveloper/false-mirror /app/false-mirror
bash /app/false-mirror/deploy/bootstrap.sh

See CLAUDE.md for full infrastructure documentation.