ffmpeg¶

To deploy an application using ffmpeg (in our case Python), we need to install this set of programs into the container itself, and not just the python library that is installed via the pip package manager.

Installation via Dockerfile¶

Using Docker to install FFMpeg.

Steps:

Create a Dockerfile in the project directory.
In the Dockerfile we specify the base Python image
```
FROM python:3.8-slim
```

Run the FFMpeg installation command:

RUN apt-get update && apt-get install -y ffmpeg

Copy the project to the image being built and assign WORKDIR:
```
COPY . /usr/src/app
WORKDIR /usr/src/app
```
Add a command to launch the application:
```
CMD ["python", "main.py"]
```

The resulting Dockerfile:

FROM python:3.9

RUN apt-get update && apt-get install -y ffmpeg

COPY . /usr/src/app
WORKDIR /usr/src/app
RUN pip install -r requirements.txt

COPY . /usr/src/app

CMD ["python", "bot.py"]

The created file must be placed in the root of the project.

Note

Instead of 3.8-slim you can specify any other Python version you need..
If you are using a different programming language, you will need to change the FFMpeg installation command to match your language..

Note:

Instead of 3.8-slim it is possible to specify another Python version that is required by the application.
If you use a different programming language, you must select a different base image with the desired language and install the library for that language.

Example¶

Let’s create a telegram bot in python using the telebot and ffmpeg-python libraries, which will accept user videos and add watermarks to them.

Note

The code is a demo example and we strongly recommend against storing bot tokens or similar data in the code. Use environment variables (secrets)!

import os

import telebot
import ffmpeg

bot = telebot.TeleBot('YOUR_BOT_TOKEN')

@bot.message_handler(commands=['start'])
def handle_start(message):
    bot.send_message(message.chat.id, "Hi! Send me the video and I will add watermark using Ffmpeg.")

@bot.message_handler(content_types=['video'])
def handle_video(message):
    video = message.video
    file_info = bot.get_file(video.file_id)
    downloaded_file = bot.download_file(file_info.file_path)

    with open('video.mp4', 'wb') as new_file:
        new_file.write(downloaded_file)

    in_file = ffmpeg.input('video.mp4')
    overlay_file = ffmpeg.input('watermark.png')
    (
        ffmpeg
        .concat(
            in_file.trim(start_frame=10, end_frame=20),
            in_file.trim(start_frame=30, end_frame=40),
        )
        .overlay(overlay_file.hflip())
        .drawbox(50, 50, 120, 120, color='red', thickness=5)
        .output('out.mp4')
        .run()
    )

    with open('out.mp4', 'rb') as video:
        bot.send_video(message.chat.id, video)

    os.remove('video.mp4')
    os.remove('out.mp4')  

bot.polling()

Let’s add the following files to the project root:

yaml configuration file:

meta:
  environment: docker
  toolchain:
    name: docker
    version: latest
build:
  dockerfile: Dockerfile
run:
  persistenceMount: /data
  containerPort: 80

Dockerfile:

FROM python:3.9

RUN apt-get update && apt-get install -y ffmpeg

COPY . /usr/src/app
WORKDIR /usr/src/app
RUN pip install -r requirements.txt

COPY . /usr/src/app

CMD ["python", "bot.ru"]

The file with dependencies requirements.txt. In our case, it is pyTelegramBotAPI (the official name of telebot) and ffmpeg-python.
```
pyTelegramBotAPI==4.15.4
ffmpeg-python==0.2.0
```

Possible errors¶

Python can’t find .ru file - check if you have changed the name of main.ru in the CMD parameter of Dockerfile or in the command parameter of amverum.yml to the name of your script or the path to it.
AttributeError: module 'ffmpeg' has no attribute 'input' - make sure you installed ffmpeg-python and not python-ffmpeg.