Build with Python language and FastAPI web framework. It runs in ASGI web server Gunicorn

Set correct environment variables by using configuration templates from subfolder. Then run following commands:

docker-compose build
docker-compose up

On first run migrations have to be executed:

docker-compose exec api bash
$ /entrypoint.sh
docker-compose exec site_api bash
$ /entrypoint.sh

Then, app has to be authorized with Azure ML:

docker-compose exec api bash
(docker)$ az login --use-device-code

• REST API launches on localhost:8080.
• Docstring auto-generated OpenAPI docs available at: /docs.

Database should be accessible on URL set by environment variable DB_URL.

Create python virtual environment in root project directory and install alembic

virtualenv .venv
source .venv/bin/activate
pip install -r api/config/requirements.txt
export DB_URL="postgresql+asyncpg:https://user:password@localhost/neurai"
alembic -c api/config/alembic.ini revision --autogenerate -m "message"

In case of site api, its requirements have to be installed and you have to change URL to sqlite file (e.g. https://docs.sqlalchemy.org/en/14/core/engines.html#sqlite)

pip install -r site_api/config/requirements.txt
export DB_URL="sqlite+aiosqlite:https:///site_api/sqlite/neurai.db"
alembic -c site_api/config/alembic.ini revision --autogenerate -m "message"

Execute into container and run alembic. Environment variable should already be set.

$ docker-compose exec api bash
(docker)$ alembic -c api/config/alembic.ini upgrade head

https://alembic.sqlalchemy.org/en/latest/cookbook.html In order to seed test data to database run next command instead. if you already updated schema to newest version you have to downgrade in order to force data seeding

(docker)$ alembic -c api/config/alembic.ini downgrade base
(docker)$ alembic -c api/config/alembic.ini -x data=true upgrade head

To recreate sqlite databse run this command in sqlite folder:

sqlite3 neurai.db ""

Encryption key and sigma is generated and set as environment variable

import os  
import base64

b_key = os.urandom(32)
key = base64.b64encode(b_key)

b_sig = os.urandom(16)
sig = base64.b64encode(b_sig)

print('key: ', key.decode())
print('sig: ', sig.decode())