train255 / ml_competition_api_server

API server to calculate metrics for machine learning competition platform

ml_competition_api_server

• Web API server to calculate competition metrics for this machine learning competition platform
• This code is written in Python and Flask web framework.

Overall Architecture

How to setup development environment

Setup environment variables.

cp .env.sample .env # Edit .env file according to your favorable environment
vim .env

Copy the following text to .env file. (In production environment, a strong key is recommended)

API_KEY_TOKEN=secret_key

About API_KEY_TOKEN

• API_KEY_TOKEN is used for authorization for API client.
• Therefore, in production environment, API_KEY_TOKEN value that cannot be easily guessed is recommended.
• When client calls API, correct API_KEY_TOKEN value should be specified as header value of X-Authorization-Key(sample code)

Start server

start server with following command

python main.py

Endpoint to access API server

http://localhost:8000

Default metrics

• AUC for classification (/metrics/AUC)
• mAP for detection (/metrics/mAP)

How to define original metric

• Still work in progress

License

MIT