ivanmkc / forecasting-live-demo-backend

Built on Vertex AI: Backend

This is the backend code for "Built on Vertex AI".

It uses the following packages:

• FastAPI
• Google Vertex AI SDK for Python
• Google BigQuery SDK for Python

Available Commands

In the project directory, you can run:

Deployment

Here are a few options on how to deploy the backend to your own Google Cloud project. Run these commands from the same folder as this README.md file.

1. Build and deploy to Google Cloud Run in one step

gcloud beta run deploy built-on-vertex-ai-backend --source . --region=us-central1 --platform=managed --cpu=2 --memory=8G --timeout=3600 --allow-unauthenticated --session-affinity --min-instances=1

Package the backend app into a container using the Dockerfile, save it in the Google Container Registry and deploy it on Google Cloud Run.

2. Build a container using Google Cloud Build

gcloud builds submit --tag gcr.io/your-project-name/built-on-vertex-ai-backend

This uses the Dockerfile to build the container and save it in the Google Container Registry. This can then be deployed to Cloud Run or other platforms.

See Deploy a Python service to Cloud Run for more information.

Development

Start local development server

python -m uvicorn main:app --reload --port=8000

Files

Some important files and folders.

TODO