My own fastapi postgres boilerplate
This project only works with Python 3.8, 3.9 and 3.10. If you want to use a virtual environment first create the environment:
python3 -m venv .venv
source venv/bin/activate
You can install the required libraries with pip. The following command will install all the required libraries for the project. Check out the different files under requirements to more specifically see which library is used and for what reason.
pip install -r ./requirements/all.txt
A PostgreSQL user and two databases are required ('boilerplate' is the password used by default).
createuser -sP boilerplate
createdb boilerplate -O boilerplate
createdb boilerplate-test -O boilerplate
Now you should be able to start a hot reloading, api server:
PYTHONPATH=. uvicorn server.main:app --reload --port 8080
Or run a threaded server and auto-apply migrations on launch:
/bin/server
PYTHONPATH=. pytest tests/unit_tests
All configuration is done via ENV vars.
export SESSION_SECRET="SUPER_DUPER_SECRET"
export TESTING=False
Note: FastAPI will detect and automatically load an existing
.env
file.
There is a script in the main folder that creates a initial user with superuser rights:
PYTHONPATH=. python server/create_initial_user.py
Note: you can influence the default pass from the settings / env vars.
The database schema is maintained by migrations (see /migrations
for the
definitions). Pending migrations are automatically applied when starting the
server.
There are 2 migration branches that move independently of one another. The data branch which contains all needed data (e.g. examples etc.) and the Schema branch.
To create a new schema migration:
PYTHONPATH=. alembic revision --autogenerate -m "New schema" --head=schema@head
This opens a new migration in /migrations/versions/
The initial scheme was created with:
PYTHONPATH=. alembic revision --autogenerate -m "Initial scheme" --head=schema@head --version-path=migrations/versions/schema
To create a data migration do the following:
PYTHONPATH=. alembic revision --message "Name of the migration" --head=general@head
This will also create a new revision file where normal SQL can be written like so:
conn = op.get_bind()
res = conn.execute("INSERT INTO products VALUES ('x', 'y', 'z')")
Activate a python env with SAM installed, fire up Docker if it's not already running and run:
sam validate
sam build --use-container --debug
sam package --s3-bucket YOUR_S3_BUCKET \
--output-template-file out.yml --region eu-central-1
And then deploy it with:
sam deploy --template-file out.yml \
--stack-name fastapi-postgres-boilerplate \
--region eu-central-1 --no-fail-on-empty-changeset \
--capabilities CAPABILITY_IAM
An more detailed explanation about the deployment on Amazon lambda can be found on: renedohmen.nl/deploy-fastapi-on-amazon-serverless
- deliver the tests for DB model:
maps
- get it running on AWS yourself
- fix db default and product save() problem