Terminal based IDE

Description

This project contains the code necessary to build the docker image of linux terminal based IDE. It is built on NeoVim, Tmux and inlcudes most popular plugins. It supports GUI based matplotlib visualization using JupyterLab. This can be a preferable python IDE for developing deep learning models. The autobuilt docker image is available here.

Getting Started

Dependencies

• docker (verified with 20.10.6)
• docker-compose (verified in 1.29.2)

Running the IDE

1. Build & Run the required docker image by specifying any service like terminal-ide-18-py3.7-service mentioned in the docker-compose-file.

docker-compose up -d terminal-ide-18-py3.7-service

Note: -d is needed to connect interactively to the container

2. Attach to the running container

docker attach terminal-ide-18-py3.7-containerr

3. Run . /ide_start.sh to start the terminal based IDE (pre-defined tmux template).
4. To start jupyter lab server (useful for visualization),

jupyter lab --ip 0.0.0.0 --port 8899 --allow-root

Custom IDEs

You can add your own service to satisfy your requirements.

1. Edit the existing docker-compose.yml file at the project root folder by following the below format with the required inputs

version: "3.9"
services:
.
.
.
  terminal-ide-service: # change as you wish
    image: terminal-ide:yourversion # change as you wish
    build:
      context: .
      dockerfile: Dockerfile
      args:
        - baseimage=baseimage  # any image from dockerhub/local registry
        - uid=1000  # uid of the host
        - pythonversion=python3.x  # change as you wish (but python 2.x requires changes int he Docker file also)
        - user=${USER}  # change as you wish
    ports:
      - 8899:8899  # jupyterlab
    volumes:
      - /tmp/.X11-unix:/tmp/.X11-unix  # sync host clipboard
    environment:
      - DISPLAY  # sync host clipboard
    hostname: laptop-docker  # change it as you wish
    container_name: terminal-ide-container # change it as you wish
    stdin_open: true
    tty: true
    runtime: "nvidia" # needed only if you need CUDA support inside docker
    network_mode: "bridge"