Awesome

Repost FastAPI

This is the FastAPI implementation of the Repost API.

Installation

Python 3 must be installed and accessible through the use of a terminal and the keyword python or python3. Below are the steps for a proper setup using VENV (Python Virtual Environment).

Clone the repository

git clone https://github.com/pckv/repost-fastapi.git

Navigate to the repost-fastapi directory and create a new VENV

cd repost-fastapi
python -m venv venv

3 (Linux). Activate the venv (alternatively: run all commands after this step prefixed with venv/bin/)

source venv/bin/activate

3 (Windows). Activate the venv (alternatively: run all commands after this step prefixed with venv\Scripts\)

venv\Scripts\activate

Install the required packages

pip install -r requirements.txt

Configurations

Configurations are set by environment variables. Follow the instructions below to run the server once and a file config.env will be created in the root directory. Otherwise, the following settings can also be set using exported environment variables.

REPOST_CLIENT_ID - The OAuth2 client_id. Default is repost
REPOST_JWT_SECRET - The secret key used for JSON Web Tokens
REPOST_JWT_ALGORIGHTM - The algorithm used for the key above
REPOST_DATABASE_URL - An SQLAlchemy database url. See Engine Configuration
REPOST_ORIGINS - A list of CORS URLs separated by ;

Running the API with uvicorn

Uvicorn is a single-threaded ASGI server designed around uvloop to run fast. It is included in the requirements and should be used to run the API.

uvicorn repost:app

The default host and port is localhost and 8000. They can be changed with the --host and --port arguments. To run the server publically, set the host to 0.0.0.0 like so.

uvicorn repost:app --host 0.0.0.0

Running the API with gunicorn

Gunicorn is a WSGI server that can manage multiple workers. Uvicorn has a worker for Gunicorn that can be used to run multiple Uvicorn workers. Since Repost is a stateless API, this works perfectly and will allow utilizing more processing power.

gunicorn repost:app -w 17 -k uvicorn.workers.UvicornWorker

The example above uses 17 workers for a system with 8 CPUs (16 threads + 1 workers). This value can be tweaked to your setup. You can also set the host and port in gunicorn with the -b argument, which includes both host and port in the same argument.

gunicorn repost:app -b 0.0.0.0:8000 -w 17 -k uvicorn.workers.UvicornWorker

Documentation

Documentation for the API is available after deployment at the /api/swagger and /api/docs endpoints.