= FFMPEG API

A web service for converting audio/video/image files using FFMPEG.

Based on:

* https://github.com/surebert/docker-ffmpeg-service
* https://github.com/jrottenberg/ffmpeg 
* https://github.com/fluent-ffmpeg/node-fluent-ffmpeg

FFMPEG API is provided as Docker image for easy consumption.

== Endpoints

* `GET /` - API Readme.
* `GET /endpoints` - Service endpoints as JSON.
* `POST /convert/audio/to/mp3` - Convert audio file in request body to mp3. Returns mp3-file.
* `POST /convert/audio/to/wav` - Convert audio file in request body to wav. Returns wav-file.
* `POST /convert/video/to/mp4` - Convert video file in request body to mp4. Returns mp4-file.
* `POST /convert/image/to/jpg` - Convert image file to jpg. Returns jpg-file.
* `POST /video/extract/audio` - Extract audio track from POSTed video file. Returns audio track as 1-channel wav-file.
** Query param: `mono=no` - Returns audio track, all channels.
* `POST /video/extract/images` - Extract images from POSTed video file as PNG. Default FPS is 1. Returns JSON that includes download links to extracted images.
** Query param: `compress=zip|gzip` - Returns extracted images as _zip_ or _tar.gz_ (gzip).
** Query param: `fps=2` - Extract images using specified FPS. 
* `GET /video/extract/download/:filename` - Downloads extracted image file and deletes it from server.
** Query param: `delete=no` - does not delete file.
* `POST /probe` - Probe media file, return JSON metadata.

== Docker image

=== Build your own

* Clone this repository.
* Build Docker image:
** `docker build -t ffmpeg-api .`
* Run image in foreground:
** `docker run -it --rm --name ffmpeg-api -p 3000:3000 ffmpeg-api`
* Run image in background:
** `docker run -d -name ffmpeg-api -p 3000:3000 ffmpeg-api`

=== Use existing

* Run image in foreground:
** `docker run -it --rm --name ffmpeg-api -p 3000:3000 kazhar/ffmpeg-api`
* Run image in background:
** `docker run -d --name ffmpeg-api -p 3000:3000 kazhar/ffmpeg-api`

=== Environment variables

* Default log level is _info_. Set log level using environment variable, _LOG_LEVEL_.
** Set log level to debug:
** `docker run -it --rm -p 3000:3000 -e LOG_LEVEL=debug kazhar/ffmpeg-api`
* Default maximum file size of uploaded files is 512MB. Use environment variable _FILE_SIZE_LIMIT_BYTES_ to change it:
** Set max file size to 1MB:
** `docker run -it --rm -p 3000:3000 -e FILE_SIZE_LIMIT_BYTES=1048576 kazhar/ffmpeg-api`
* All uploaded and converted files are deleted when they've been downloaded. Use environment variable _KEEP_ALL_FILES_ to keep all files inside the container /tmp-directory:
** `docker run -it --rm -p 3000:3000 -e KEEP_ALL_FILES=true kazhar/ffmpeg-api`


== Usage

Input file to FFMPEG API can be anything that ffmpeg supports. See https://www.ffmpeg.org/general.html#Supported-File-Formats_002c-Codecs-or-Features[ffmpeg docs for supported formats].

=== Convert

Convert audio/video/image files using the API.

* `curl -F "file=@input.wav" 127.0.0.1:3000/convert/audio/to/mp3  > output.mp3`
* `curl -F "file=@input.m4a" 127.0.0.1:3000/convert/audio/to/wav  > output.wav`
* `curl -F "file=@input.mov" 127.0.0.1:3000/convert/video/to/mp4  > output.mp4`
* `curl -F "file=@input.mp4" 127.0.0.1:3000/convert/videp/to/mp4  > output.mp4`
* `curl -F "file=@input.tiff" 127.0.0.1:3000/convert/image/to/jpg  > output.jpg`
* `curl -F "file=@input.png" 127.0.0.1:3000/convert/image/to/jpg  > output.jpg`

=== Extract images

Extract images from video using the API.

* `curl -F "file=@input.mov" 127.0.0.1:3000/video/extract/images`
** Returns JSON that lists image download URLs for each extracted image.
** Default FPS is 1.
** Images are in PNG-format.
** See sample: link:./samples/extracted_images.json[extracted_images.json].
* `curl 127.0.0.1:3000/video/extract/download/ba0f565c-0001.png`
** Downloads exracted image and deletes it from server.
* `curl 127.0.0.1:3000/video/extract/download/ba0f565c-0001.png?delete=no`
** Downloads exracted image but does not deletes it from server.
* `curl -F "file=@input.mov" 127.0.0.1:3000/video/extract/images?compress=zip > images.zip`
** Returns ZIP package of all extracted images.
* `curl -F "file=@input.mov" 127.0.0.1:3000/video/extract/images?compress=gzip > images.tar.gz`
** Returns GZIP (tar.gz) package of all extracted images.
* `curl -F "file=@input.mov" 127.0.0.1:3000/video/extract/images?fps=0.5`
** Sets FPS to extract images. FPS=0.5 is every two seconds, FPS=4 is four images per seconds, etc.

=== Extract audio

Extract audio track from video using the API.

* `curl -F "file=@input.mov" 127.0.0.1:3000/video/extract/audio`
** Returns 1-channel WAV-file of video's audio track.
* `curl -F "file=@input.mov" 127.0.0.1:3000/video/extract/audio?mono=no`
** Returns WAV-file of video's audio track, with all the channels as in input video.

=== Probe

Probe audio/video/image files using the API.

* `curl -F "file=@input.mov" 127.0.0.1:3000/probe`
** Returns JSON metadata of media file.
** The same JSON metadata as in ffprobe command: `ffprobe -of json -show_streams -show_format input.mov`.
** See sample of MOV-video metadata: link:./samples/probe_metadata.json[probe_metadata.json].


== Background

Originally developed by https://github.com/surebert[Paul Visco].                  

Changes include new functionality, updated Node.js version, Docker image based on Alpine, logging and other major refactoring.