llm_change_agent.evaluations package
Submodules
llm_change_agent.evaluations.evaluator module
Evaluation script for the LLM Change Agent.
- llm_change_agent.evaluations.evaluator.compare_changes(expected_dir, output_dir)
Compare the actual changes with the predicted changes.
- llm_change_agent.evaluations.evaluator.download_document(url, input_dir)
Download the document from the URL.
- llm_change_agent.evaluations.evaluator.generate_changes_via_llm(eval_dir, output_dir, provider, model)
Generate changes via the LLM Change Agent.
- llm_change_agent.evaluations.evaluator.get_comparison_metrics(expected_changes, output_changes)
Compare the expected changes with the output changes.
- llm_change_agent.evaluations.evaluator.prepare_eval_and_expected_yamls(input_dir)
Prepare the evaluation and expected YAMLs for the input documents.
- llm_change_agent.evaluations.evaluator.run_evaluate(model, provider)
Evaluate the LLM Change Agent.
- llm_change_agent.evaluations.evaluator.run_llm_change_agent(prompt, provider, model, docs=None)
Run the LLM Change Agent.
- Return type:
List
Module contents
Evaluate the LLM Change Agent.