The backend is written in Python. We use a simple Flask server, and it is deployed with Google Cloud App Engine. This is a serverless service that automatically scales our backend code. The model inference is done with our custom model that will be fetched from Google Cloud Storage.
In order to run the backend locally you would need to set up the google-cloud-sdk so that you can access the services there. Make sure you have accepted the google cloud invite before beginning.
Follow this guide for installing the CLI for GCP.
Follow the instructions. Make sure to enter project ID p4ds-team-2 as project, leaving other values as default.
Now, restart your terminal. Then you can continue
Then you also need to configure docker to be able to authenticate
gcloud auth configure-dockerFinally, you need to add the service file to allow your docker container to
run google cloud apis. To do this you need to download a service-account file from the slack and put it in
root/backend/secret.json
Then, make sure to add the following line to your .env file in the backend folder
GOOGLE_APPLICATION_CREDENTIALS="secret.json"You can now move on to running the backend with docker shown below.
You can run the backend locally with docker using either
make start-backend
running the docker compose commands yourself.
The app will then expose our apis on localhost:8080.
If you want to reload the app without rebuilding every image run
make reload-backendYou can also run it directly on your local machine. Make sure to use python3.8 and
install the requirements in requirements.txt. Also make sure to update
this file whenever new requirements are added.
NB! our Google Cloud Platorm account is not yet set up, so do not deploy anything yet (you can't either haha) Deployments should also be handled in our pipeline once everything is set up.
If you want to play around with it: (should not be needed at all unless you want to learn more) Follow install instructions here:
Simply run the command
gcloud app deploywhile inside the root/backend directory.
Our Backend API endpoints are documented and can be tested under: http://localhost:8080/apidocs/
To update or add to the API, see root/backend/main.py and make sure to update/add documentation in the api_specs.py file.