This project serves as a demo and proof of concept for an API server built on the FastAPI architecture.
Thanks for stopping by, and make sure you skim the cliff notes below if you're interested in cloning and running this package.
This project was built using Python 3.11, and while it's very likely compatible with earlier versions, you will probably have some type hinting issues and may need to sort out some Enum/Flag and similar issues.
The other requirements are simply pip packages as per requirements.txt.
The server binds to localhost:80 by default, but can be changed easily by modifying the uvicorn flags for whichever launch mechanism you're using.
I'll go through a few ways you can run this project below, but you'll most likely want to interface with it as well. Before you go ahead and make a super-slick react app, you could start by loading the supplied Postman collection.
Once the server is running you can check out the Swagger docs. This particular link will only work at runtime, not necessarily right now as you're reading this on GitHub.
Open this entire repository as your workspace, and you'll find some reasonable settings along with some launch configurations.
Go to Run and Debug and select the API Server dropdown. You're all set — just hit the play button or F5.
Build the container with the supplied Dockerfile and run it.
If you're using a virtualized environment, you may have to inject some container bindings. For most scenarios I think something like this will serve you well: docker run --rm -it -p 80:80 <image>.
Your dynamic image name is very likely simplerestapi:latest if you didn't rename the repo.
Open up a terminal and do something along these lines:
# Linux
cd path/to/repo
python -m pip install --upgrade -r requirements.txt
python -m uvicorn simple_rest_api.main:app --reload --host 0.0.0.0 --port 80
# Windows
cd C:\path\to\repo
python.exe -m pip install --upgrade -r requirements.txt
python.exe -m uvicorn simple_rest_api.main:app --reload --host 0.0.0.0 --port 80Just make sure you're launching the correct Python interpreter. The system default isn't always the version you want — particularly not when there's likely a couple of version 3.11 dependencies in this package.
For the sake of this demo, the chosen authentication method is the reasonably straightforward API key scheme. However, I have chosen to expand upon this to require a set of paired keys; a secret key along with a known app name.
In an actual production environment, these keys will most definitely be stored in an external key vault. For our purposes here though, they are one-way hashed and stored in the repository itself.
To get started, you can use the default credentials:
app-name =
demo
api-key =z_IlZ7JiE2A4_MCdGK2SR5YbIwm3-64ePyZ3rGqKfZ
Or if you prefer, you can generate and validate API keys with the supplied VSCode launch settings. If that still doesn't cut the mustard, you can go straight to the standalone CLI via generate_api_key.py and validate_api_key.py.
The keys need to either be injected into the headers of a request, or passed as query parameters. The specific naming for this is controlled in config.py, but looks something like this:
> X-App-Name: demo
> X-Api-Key: z_IlZ7JiE2A4_MCdGK2SR5YbIwm3-64ePyZ3rGqKfZ
[GET] http://localhost/
[GET] http://localhost/?x_app_name=demo&x_api_key=z_IlZ7JiE2A4_MCdGK2SR5YbIwm3-64ePyZ3rGqKfZ
This demo uses a simple In-Memory SQLite instance, which will not persist once you stop the running process. Initial data population happens in init_db.py, which you can tweak to your specific requirements.
That said, if you're using this project as the basis for something real, you will need to use a proper backing database. That's if you want a database, which you may not.
PS: I've used relative imports throughout this package, making it easier to rename if someone wants any of the code. However, the name simple_rest_api is hardcoded in the Dockerfile, the CLI utils and the tests.