Skip to content

Latest commit

 

History

History
140 lines (104 loc) · 2.71 KB

README.md

File metadata and controls

140 lines (104 loc) · 2.71 KB
Raw

Data Recommendation API

Recommendation API with FastAPI + Psql + VertexAI endpoints

See /docs for API endpoints

Folders structure

+-- src
| + huggy
|  +-- core
|    +-- endpoint - configuration of VertexAi endpoint for retrieval and ranking
|    +-- model_engine - orchestration of the scoring pipeline
|    +-- model_selection - model selection based on the model_endpoint parameter
|    +-- scorer - retrieve the base of recommendable offers and score them
|  +-- crud - functions for reading/writing to the database
|  +-- database - logic for handling db connexion
|  +-- models - model of the database tables - 1 file per table
|  +-- schemas - definition of object properties
|  +-- utils
|  +-- views - router views
|
+-- main.py
|
+-- tests - tests files

How to DEV

Set-up a dev-env:

Install pyenv and install `python 3.9``

pyenv install 3.9

Create a virtual env

pyenv virtualenv 3.9 reco_fastapi
pyenv shell reco_fastapi
cd apps/recommendation/api/
pip3 install -r requirements.txt

Run tests

pip install "psycopg[binary,pool]"

Set up a testdb :

cd apps/recommendation/api/
export DATA_GCP_TEST_POSTGRES_PORT=5432
export DB_NAME="db"
docker-compose up -d testdb
cd apps/recommendation/api
pyenv shell reco_fastapi
export API_LOCAL=1
pytest

In case you have some troubles to run tests

Connect to db, create db database, add postgis extension

docker exec -it <docker_id> /bin/sh
psql -U postgres

create database db;
\c db;
create extension postgis;

Run it locally

cd apps/recommendation/api/src
export API_LOCAL=1
uvicorn main:app --reload

How to PROD

CI-CD

github-actions handle de CI/CD deployment

  • we have a main (testing), staging and production branches that deploy for each environment. Everything is done automatically.

Deploy by hand

In case of emergency you still can deploy the api by hand:

1: Build

cd apps/recommendation/api
gcloud builds submit \
  --tag eu.gcr.io/<PROJECT-ID>/data-gcp/<IMAGE-NAME>
  • PROJECT-ID : (passculture-data-<env>)
  • IMAGE-NAME : (api-recommendation-<env>)

2: Deploy

cd apps/recommendation/api

gcloud run deploy <SERVICE> \
--image <IMAGE>:latest \
--region europe-west1 \
--allow-unauthenticated \
--platform managed
  • SERVICE : Service Name (api-recommendation-<env>)
  • IMAGE : Docker image (eu.gcr.io/passculture-data-<env>/data-gcp/api-recommendation)

Staging

gcloud builds submit --tag eu.gcr.io/passculture-data-ehp/data-gcp/apireco-stg

gcloud run deploy apireco-stg \
--image eu.gcr.io/passculture-data-ehp/data-gcp/apireco-stg:latest \
--region europe-west1 \
--allow-unauthenticated \
--platform managed