ignis621/watch-batch-transcode repository overview

⁠Watch & Batch Transcode

A simple Docker container that watches a specified input directory for video files, automatically transcodes them using FFmpeg with configurable settings (specifically for single-pass CRF), moves files to success/failure directories, logs processing details to a CSV file, and can send notifications via ntfy.sh⁠.

It is built on the linuxserver.io FFmpeg base image⁠, inheriting its user/group ID handling and S6-overlay init system.

⁠Features

• Directory watch: Continuously scans the input directory for new files
• Automatic Processing: Automatically picks up the first file found and starts processing
• FFmpeg Transcoding: Transcodes the file with FFmpeg
• Configurable Settings: Easy configuration of video codec, CRF, encode preset, audio codec, and audio bitrate via environment variables
• Retry Mechanism: Retries multiple times before giving up
• Success/failure classification: Moves successfully processed files to the completed directory and failed files to the failed directory. When processing starts, original files are moved to the sources_processed directory
• CSV Logging: Appends processing details (filename, input size, output size, ratio, time taken) to a CSV file in the completed directory
• NTFY Notifications: Optional notifications via ntfy.sh⁠ on success and failure. Supports custom NTFY server
• Source Deletion: Can be configured to automatically delete the original source file after success

⁠Getting Started

⁠Prerequisites

• Docker⁠ installed
• Docker Compose⁠ installed

⁠Setup

1. Create a docker-compose.yaml file using the template from the repo:

   curl -O https://raw.githubusercontent.com/ignis621/watch-batch-transcode/main/docker-compose.yaml
   

   Copy

2. Configure it to suit your needs - volumes, environment variables, etc.

⁠Running with Docker Compose

1. Run using the prebuilt Docker image:

   docker-compose up -d
   

   Copy

   This pulls the latest image from Docker Hub and starts the container in detached mode.

2. Check logs:

   docker-compose logs -f
   

   Copy

3. Stop the container:

   docker-compose down
   

   Copy

⁠Configuration

Configuration is done via environment variables set in the docker-compose.yaml file.

⁠Directory Structure

The container uses the following directory structure for file processing, mapped via Docker volumes:

• /sources: Input Directory. Place video files here for processing. The script will pick up the first file it finds.
• /sources_processed: Working Directory. Files are moved here from /sources before processing begins. They remain here by default after successful processing unless DELETE_AFTER is set.
• /completed: Output Directory. Successfully transcoded files (in the format specified by FINAL_VIDEO_CONTAINER) are placed here. The .processed_files.csv log is also stored in this directory.
• /failed: Error Directory. Original files that failed all attempts are moved here.
• /tmp: Temporary Directory. Used for intermediate files during transcoding. Explicitly mapping this is a good idea for extremely big files, or if /tmp on the host is on a faster drive than your other data volumes.

⁠How it Works

1. Script scans the /sources directory.
2. If a file is found, it makes sure its not still being written or moved. If not ready, it waits and rechecks.
3. Once ready, the file is moved to /sources_processed.
4. FFmpeg attempts to transcode the file from /sources_processed to an MKV file in /tmp using parameters from environment variables (OVERRIDE_ENCODE_ARGS or VIDEO_CODEC, VIDEO_CRF, etc.)
5. If encoding is successful, FFmpeg tries to remux the MKV to the final output file (eg. MP4) in /completed using remuxing parameters (OVERRIDE_REMUX_ARGS or default copy options).
6. If both encoding and remuxing are successful:
   • The temporary MKV file is deleted
   • Processing time and file size details are calculated
   • An entry is appended to the .processed_files.csv file in /completed
   • If ntfy is configured, a success notification is sent
   • If DELETE_AFTER is set, the original file is deleted from /sources_processed
7. If either encoding or remuxing fails after MAX_RETRIES attempts:
   • Any temp files are cleaned up
   • The original file is moved from /sources_processed to /failed
   • If ntfy is configured, a failure notification is sent
8. Rinse Wait and repeat

⁠CSV Log File (.processed_files.csv)

A CSV file named .processed_files.csv is created in the /completed dir if it doesn't exist. Each time a file is successfully processed, a new row is appended.

⁠NTFY Notifications

To receive notifications on your phone or other devices via ntfy.sh⁠:

1. Get a topic name (either from the public instance or your own server).
2. Set the NTFY_TOPIC environment variable in your docker-compose.yaml to this topic name.
3. (Optional) If using a custom NTFY server, set the NTFY_SERVER environment variable to its address.

Success and fail messages will be sent to your specified topic.

⁠Building the Container

If you want to build the image yourself instead of pulling from Docker Hub:

1. Clone the repository:

   git clone https://github.com/ignis621/watch-batch-transcode
   cd watch-batch-transcode
   

   Copy

2. Build and run using the provided build config:

   docker-compose -f docker-compose.build.yaml up -d --build
   

   Copy

3. Stop the container:

   docker-compose -f docker-compose.build.yaml down
   

   Copy

⁠Contributing

Contributions are welcome! Feel free to open issues for bugs or feature requests, or submit PRs

⁠License

This project is licensed under the GNU General Public License v3 (GPL-3). Check tl;drLegal⁠ for an explanation or LICENSE⁠ if you're rlly bored.

⁠Acknowledgements

• Uses the FFmpeg⁠ multimedia framework
• Built on linuxserver.io FFmpeg Docker base image⁠
• Uses ntfy.sh⁠ for simple push notifications