LLM-based Argumentation Services

• Make sure to set the environment variable OPENAI_API_KEY to your OpenAI API key.
• Use poetry install in the base directory to install the project's dependencies.

Argument Mining

• Start the gRPC Server with poetry run python -m mining.server.
• Start the gRPC Client with poetry run python -m mining.client to run a test case.

This project contains one folder for each of the argument extraction pipeline steps: entailment (argument relation classification), adu extraction, major claim prediction and graph construction. In each of these folders, a gRPC Servicer is defined, which handles the gRPC types, i.e., mapping a gRPC request to a gRPC response. Internally, it calls the respective model.py, which uses the OpenAI API to implement the actual processing.

Quality Assessment

• Start the gRPC Server with poetry run python -m quality.server.

The service currently only supports quality explanation.

Argument Ranking

• Start the gRPC Server with poetry run python -m ranking.server.

The service currently only supports fine-granular clustering.