Python Semantic Matching Service

This is a proof-of-concept implementation of a semantic matching service using Python. The idea behind this service is to create a uniform interface for exchanging semantic matches, that are stored decentralized across multiple semantic matching services.

Warning

This project is not about finding semantic matches or semantic similarities, it rather requires these as input.

Structure

The project is structured as follows:

• algorithm.py implements the decentralized match aggregation algorithm
• service.py offers the service implementation using the FastAPI framework. Note that you can find a detailed interface description (OpenAPI) by running the server locally and navigating to <endpoint>/docs.

How to Use

There are two main options to run:

Run service.py

You need a working Python installation (3.11 or higher).

• Clone this repository
• Create a new virtual environment: python3 -m venv venv (Note that the service is only tested on Linux systems)
• Activate the virtual environment: source venv/bin/activate
• Install the package: pip install .
• Copy the config.ini.default to config.ini and adapt the fields as necessary
• Run the service: python3 semantic_matcher/service.py

Run via Docker

To run via docker, you obviously need a working docker installation.

• Clone the repository
• Copy the config.ini.default to config.ini and adapt the fields as necessary
• Choose one of the two options below:

Option 1: Build and run manually

In the project root directory:

docker build -t semantic_matching_service .

docker run -d -p 8000:8000 semantic_matching_service

Note

You may have to change the port in the run command if you changed it in your config.ini.

Option 2: Use the docker compose file

In the project root directory:

docker compose up

Note

If you changed your config.ini, you might need to adapt the respective fields in the compose.yaml.