AKERN Assistant

Have gcloud CLI configured (https://docs.cloud.google.com/sdk/docs/install-sdk?hl=it).
Setup the Google Cloud project:

Login

gcloud auth login

Set the correct Google project

gcloud config set core/project akqa-ita-ai-poc1

Build the image

Setup Docker first:

gcloud auth configure-docker \
    europe-west3-docker.pkg.dev

This allows Docker to use gcloud credentials to authenticate in the registry. Then to push the image:

docker buildx build \
    --platform linux/amd64 \
    -t europe-west3-docker.pkg.dev/akqa-ita-ai-poc1/akern-repository/akern_assistant:latest \
    --push \
    .

Deployment

gcloud run deploy akern-assistant \
      --image europe-west3-docker.pkg.dev/akqa-ita-ai-poc1/akern-repository/akern_assistant:latest \
      --platform managed \
      --region europe-west3 \
      --port 8080 \
      --allow-unauthenticated \
      --set-secrets "GOOGLE_APPLICATION_CREDENTIALS_JSON=google-credentials:latest,AUTH_USER=auth-user:latest,PASSWORD=auth-password:latest"

Run the project locally

1 - Install dependencies

uv sync

2 - Add the `credentials.json` to the project root

The file must be asked to a coworker (ask Matteo or Antonio).

3 - Run the database "migration"

The project currently uses a very minimal sqlite3 db. To initialize it, run:

uv run models/orm.py

This will create the example.db file.

After doing this, you will have to create at least one prompt record into the database. I suggest to use a SQLite viewer software for desktop called DB Browser for SQLite. Once you have downloaded the program, open the database that has just been created by the migration and add at least one row for one prompt.

You can use the prompt.md file to have a prompt example.

4 - Run the dev server

uv run fastapi dev app.py

1.9 KiB Raw Blame History