CaptainCuddleCube/ap_location

AP location

This an an api that will start up, and expect results to be published to the /ap-location endpoint.

This is a FastAPI application that, including simple tests, and a Docker file to ensure the environment is correctly setup. The setup uses Gunicorn as the process manager and ASGI server (using uvicorn), with 2 cores set.

The current solution uses a simple in-process cache for it's simplicity. To improve on this design, it is recommended using Redis and it's Hash table to store the reply objects. This would allow the web server to go down and have the cache persist.

The caching algorithm uses the RSSI and the BSSID to determine whether a signal should be used to generate a hash. The RSSI threshold of -82dBm was chosen because this is approximately the 10th percentile.

Setup

You can build using the Dockerfile, or using a python venv, please note, this has only been tested with python 3.7 and 3.9

Docker setup

From this directory, please run:

docker build -t ap_location .

When this is complete, you have the docker environment setup.

Using Python venv

In this directory, run:

python3 -m venv venv

Then:

./venv/bin/pip install -r requirements.txt

Starting the webserver:

The webserver is an asynchronous web server, and has been setup to use Gunicorn as the process manager. The Gunicorn setup at been set to use 2 cores, since this allows for decent processor performance, and prevents too many cache misses.

It must be noted that caching is done per process - and not through a cache server like redis. It would be recommended to use a separate caching database.

The API_TOKEN and INDEX_THRESHOLDare configurable through environment variables.

Docker

To run the web server using Docker, run the following command. You will need to provide the API_TOKEN as an environment variable:

docker run -e API_TOKEN <the api token> -it --rm -p 8001:8000 ap_location

This will now start the webserver and have it listening on port 8001

Python venv

Note, you will need to pass the API_TOKEN in via an environment variable.

API_TOKEN=<api token> ./venv/bin/gunicorn ap_location.main:app -w 2 -k uvicorn.workers.UvicornWorker --bind 0.0.0.0:8001

This will start the web server listening on port 8001.

Running the tests

Running the test is super easy!

Running from Docker

docker run --rm ap_location pytest ap_location/test.py

Running from Python venv

./venv/bin/pytest ap_location/test.py