Welcome to BikeRouter!

Description

This project combines the power of LLMs with location data and everyone's favourite computational problem: The Travelling Salesman problem!

• Enter the city you want to visit and you will get 10 must-visits in the city of choice!
• Pick the locations you want to visit
• Generate the shortest path between these locations

Requirements

This project uses both Mapbox and Replicate. You will need to get API keys for both services:

• Mapbox
• Replicate

Running the project

Set-up a virtual environment:

python -m venv .venv
source .venv/bin/activate

Install the necessary requirements

pip install -r requirements.txt

Create the API variables

export MAPBOX_TOKEN=
export REPLICATE_API_TOKEN=

Run the project

python -m streamlit run main.py

Deploying the project

You can deploy BikeRouter using a Docker container. For this, the Dockerfile in the project has been added. You can use AWS ECS or alternatives by Azure and GCP.

I have deployed BikeRouter on Railway: here

Next steps

• Add some context, i.e., who is doing the bike trip
• Clean up the code, for now all the code is in the same file. In the future, make things more modular
• Write tests to see if everything is working as expected. Here, a framework like PyTest can be leveraged