Orientation par les compétences

Project structure

• model/: A python package containing the model which computes job and skill recommendations
• case-studies/: A streamlit app to showcase the model
• api/: An HTTP API interface for the model
• etl/: A python package containing Extract-Transform-Load pipelines to preprocess the data fed to the model
• frontend/: (obsolete) a javascript/react UI for the model
• docker/: Dockerfiles to bundle everything together

Running locally

The case-studies/ and api/ folders contain apps that can be run locally with the script up.sh in each folder.

Deployment in the cloud

build.sh builds a docker of the case-studies, the api and the frontend

release-production.sh sends the docker image, the file docker/docker-compose.yaml and the file .env-production to a remote server. The latter file must contain the following environment variables:

COMPOSE_PROJECT_NAME
COMPOSE_FILE
CORS_ALLOWED_ORIGINS
FRONTEND_PORT
API_PORT
API_ROOT_PATH
STREAMLIT_PORT
STREAMLIT_SERVER_BASE_URL_PATH
NEO4J_URI
NEO4J_USER
NEO4J_PASS

run-production.sh launches the services defined in docker/docker-compose.yaml