The idea is to use an LLM to decide whether a list of sentences are consistent.
For now, we have propositional sentences, and conjunctions and disjunctions at the search level.
In the future, we might want to be able to normalize and decrease the size of the constraint set. One idea is that the LLM could generate summaries.
First create a .env
file in the root directory containing your OpenAI API key:
OPENAI_API_KEY="sk-xxxx"
Also install the Python dependencies:
pip install -r requirements.txt
To use the ollama
server and model file
brew install ollama
ollama pull orca2
ollama create constraint_solver -f constraint_solver.model
Then start the server and you can communicate with the model via ollama run constraint_solver