/ledfx-docker

An attempt at running LedFX inside a docker container

Primary LanguagePythonMIT LicenseMIT

ledfx-docker

An attempt at running LedFX inside a docker container

How these files work

Dockerfile:

This file uses the python:3.7-slim image:

  • It first creates a python virtual environment /ledfx/venv
  • "activate" the python venv by adding it to the PATH
  • Install compile and run dependencies with apt-get
  • Install ledfx-dev from pip
  • Purge compile dependencies with apt-get, clean and autoremove
  • Add ledfx user and create a home folder
  • Set WORKDIR to ledfx home directory and change USER to ledfx
  • EXPOSE ports 8888 and 5353
  • Launch ledfx

How to use: docker build -t ledfx .

Dockerfile-2stage-1:

This file uses the python:3.7-slim image:

  1. Create a "compile-image"

    • Set WORKDIR to /ledfx
    • Install compile dependencies with apt-get
    • Clone ledfx dev branch into /ledfx using git clone
    • cd to /ledfx/frontend
    • Install yarn and build the frontend
    • cd to /ledfx
    • Create a python venv /ledfx/venv and "activate" it by adding it to the PATH
    • Install ledfx-dev from source using pip

  2. Create a "build-image"

    • COPY /ledfx/venv from "compile-image" to current image
    • Activate the venv by adding it to the PATH
    • Install run dependencies with apt-get
    • Add ledfx user and create a home folder
    • Set WORKDIR to ledfx home directory and change USER to ledfx
    • EXPOSE ports 8888 and 5353
    • Launch ledfx

How to use: docker build -t ledfx-2stage -f Dockerfile-2stage-1 .

Dockerfile-4stage:

This file uses docker's multi-stage build technique to speed up the build process using the cache and cut down the overall size of the final image.

While all 4 images will be built initially, we can significantly cut down the build times by using the cached images that won't change regularly. Docker will use the cache up until the point in the Dockerfile that it detects a change. Any lines after that change will be build from scratch.

By creating the venv and compile images in the beggining of the Dockerfile, we can be fairly certain that these dependencies won't change regularly. This cuts down the time required to build by skipping the dependency installation process.

  1. Create "venv-image"

    • Create a python venv /ledfx/venv and "activate" it by adding it to the PATH
    • Install compile and run dependencies with apt-get

  2. Create "compile-image" from "venv-image"

    • Install build dependencies with apt-get

  3. Create "build-image" from "compile-image"

    • COPY /ledfx/venv from "venv-image" to current image
    • Activate the venv by adding it to the PATH
    • Set WORKDIR to /ledfx-git
    • Clone ledfx dev branch into /ledfx-git using git clone
    • cd to /ledfx/frontend
    • Install yarn and build the frontend
    • cd to /ledfx
    • Create a python venv /ledfx/venv and "activate" it by adding it to the PATH
    • Install ledfx-dev from source using pip

  4. Create "dist-image" from "venv-image"

    • COPY /ledfx/venv from "build-image" to current image
    • Activate the venv by adding it to the PATH
    • Remove all lists from /var/lib/apt/lists/
    • Add ledfx user and create a home folder
    • Set WORKDIR to ledfx home directory and change USER to ledfx
    • EXPOSE ports 8888 and 5353
    • Launch ledfx

How to use: docker build -t ledfx-4stage -f Dockerfile-4stage .

Dockerfile.venv, Dockerfile.compile, Dockerfile.build:

These files take a similar approach to Dockerfile-4stage. They differ in the fact that they use 3 independent Dockerfile's to achieve the same results. I'm not sure if this is significant, but I was learning how to create Dockerfile's when I wrote these so I decided to try it out.

  1. Dockerfile.venv

    • Create "venv-image" from python:3.7-slim
    • Create a python venv /ledfx/venv and "activate" it by adding it to the PATH
    • Install run dependencies with apt-get

How to use: docker build -t thatdonfc/ledfx-venv -f Dockerfile.venv .

  1. Dockerfile.compile

    • Create "compile-image" from "venv-image" named thatdonfc/ledfx-venv
    • Activate the venv by adding it to the PATH
    • Install build dependencies with apt-get
    • Install ledfx-dev with pip

How to use: docker build -t thatdonfc/ledfx-compile -f Dockerfile.compile .

  1. Dockerfile.build

    • Create build image from "venv-image" named thatdonfc/ledfx-venv
    • COPY /ledfx/venv from "compile-image" named thatdonfc/ledfx-compile
    • Remove all lists from /var/lib/apt/lists/
    • Add ledfx user and create a home folder
    • Set WORKDIR to ledfx home directory and change USER to ledfx
    • EXPOSE ports 8888 and 5353
    • Launch ledfx

How to use: docker build -t thatdonfc/ledfx-build -f Dockerfile.build .