This application is a simple REST API which provides one serviceto run a provided dockerfile.
These instructions will get you a copy of the project up and running on your local machine for development and testing purposes.
To run this project you'll need a Docker and/or Golang working environment.
Install Docker compose : https://docs.docker.com/compose/install/
Step-by-step installation guide Clone repo following go convention:
cd ~/go/src/github.com/alisstaki
git clone https://github.com/alisstaki/interview-exercise.git
cd interview-exercise
Start app
make start
The app should be up and running and see something like that in your terminal:
API ready to receive content...
- Third parties libraries used : https://github.com/gorilla/mux to manage services routing and https://github.com/moby/buildkit to validate dockerfile
For time optimization, I've decided to not unittest this api
You can upload your dockerfile using a POST to http://localhost:8080/run :
curl --request POST \
--url http://localhost:8080/run \
--header 'Content-Type: multipart/form-data; boundary=---011000010111000001101001' \
--form 'file=FROM ubuntu:16.04
# train machine learning model
```bash
cat <<EOF > Dockerfile.test
FROM ubuntu:16.04
ARG version
ENV VERSION=${version}
CMD echo '{"perf":0.99}' > "/data/perf-${VERSION}.json"
EOF
# save performances
CMD echo '\''{"perf":0.99}'\'' > /data/perf.json'
You can check job status and model performance using a GET to http://localhost:8080/status:
curl --request GET \
--url 'http://localhost:8080/status?id=62a86742-3faa-4212-a2f3-8f6ab0b6950a'
With given limited time, the focus was made to have a running project and a validity check on the dockerfile. To continue, I would focus on security and only allow a certain set of command in the provided dockerfile, and run it in a more secure environment. Also jwt token authentication would be added to allow only known client to upload a dockerfile. To manage the job I would use a Redis instead of in-memory map.
At last, additionnal time could be spent to have a cleaner code and to have a better error management.
This project is licensed under the MIT License - see the LICENSE.md file for details
- Thanks mostly to StackOverflow and Medium for the endless source of knowledge.
- And thanks for reading up until the end :) © 2021 GitHub, Inc. Terms Privacy Security Status Docs Contact GitHub Pricing API Training Blog About