Containerized Web Server

NIH Containerization Project - Example Template

Web Server

Purpose:

Overall: Build a web server that is responsive to HTTP requests in order to run bioinformatics software.
Example: Create an example web app that runs a bioinformatics command-line executable code and produces 3D visualizations of input and output files. This project will then be made into a Docker container so other researchers can use it more easily, and it can be used as an example for containerization of bioinformatics software.

Goals:

  • Containerize bioinformatics software
  • Create a release system to push well-scripted software to the cloud and make the software callable by web requests
    • This would make the system easily scalable
  • Use Google Cloud Run to run server and compute
    • Takes this responsibility away from the authoring lab

Install and Run (Locally)

You must have node.js running on your machine. Once you have cloned this project you can run npm install to install all the packages for this project. Then running npm run dev will run the dev version of this code, which will run this project with nodemon. Nodemon auto-restarts the node server every time you make a change to a file. Very helpful when you are writing and testing code.

After running the dev version of the code, go to "localhost:8000" in your browser to see the frontend of the project.

Testing and Running ()

You can download the container and test on Google Cloud or on docker locally. Follow docker-container tutorial for these steps.

Useful Resources

Overview of pm2 and useful commands:

  • https://pm2.keymetrics.io/docs/usage/pm2-doc-single-page/

Explanation of exec/spawn to execute shell commands:

  • https://stackabuse.com/executing-shell-commands-with-node-js
  • https://nodejs.org/api/child_process.html

Explanation of multer (middleware to upload files):

  • https://www.npmjs.com/package/multer
  • https://www.bacancytechnology.com/blog/file-upload-using-multer-with-nodejs-and-express

Docker resources: Docker is a containerization software. To create a Docker container, all you have to do is create a Dockerfile in your root directory of your project. A Dockerfile is a simple text file with commands for how Docker will create an image for your container. A Docker image is essentially a blueprint for your container. The image becomes a container when it is activated.

  • https://www.docker.com/
  • make an account, download Docker Desktop (or do the tutorial on the cloud) and follow the getting-started tutorial, which I believe is prompted when you download Docker Desktop: https://docs.microsoft.com/en-us/visualstudio/docker/tutorials/docker-tutorial

To Dockerize your project: overview: https://blog.iron.io/how-to-create-a-docker-container/#1

  1. Create a Dockerfile
  • https://nodejs.org/en/docs/guides/nodejs-docker-webapp/
  • https://kapeli.com/cheat_sheets/Dockerfile.docset/Contents/Resources/Documents/index
  1. Use "docker build" to use your Dockerfile to create a Docker image
  • https://www.techrepublic.com/article/how-to-create-your-own-docker-image/
  • https://docs.docker.com/engine/reference/commandline/build/
  1. If applicable, push your image to the Docker Hub so it can be shared with others.
  • https://techtutorialsite.com/docker-push-images-to-dockerhub/
  1. Activate your Docker image to create a running container
  • https://blog.iron.io/how-to-create-a-docker-container/

OTHER POSSIBLY USEFUL LINKS https://codeburst.io/dockerize-deploy-1cfc4f0e7c72
https://github.com/nodejs/docker-node/blob/main/docs/BestPractices.md
https://medium.com/codebase/using-aws-s3-buckets-in-a-nodejs-app-74da2fc547a6
https://docs.docker.com/engine/reference/builder/ \

webgl-utils.js

This file contains functions every webgl program will need a version of one way or another.

Instead of setting up a context manually it is recommended to use. This will check for success or failure. On failure it will attempt to present an approriate message to the user.

  gl = WebGLUtils.setupWebGL(canvas);

For animated WebGL apps use of setTimeout or setInterval are discouraged. It is recommended you structure your rendering loop like this.

  function render() {
    window.requestAnimFrame(render, canvas);

    // do rendering
    ...
  }
  render();

This will call your rendering function up to the refresh rate of your display but will stop rendering if your app is not visible.

Source: