DeepEvalEx
Copyright 2025 Steven Holdsworth (@holsee)

This project is a port/derivative work of DeepEval, originally developed
by Confident AI.

================================================================================

DeepEval
Copyright 2023-2024 Confident AI

Licensed under the Apache License, Version 2.0
https://github.com/confident-ai/deepeval

================================================================================

This Elixir implementation (DeepEvalEx) is based on the algorithms, metrics,
and evaluation patterns from the original DeepEval Python library. While the
code has been rewritten in Elixir to be idiomatic to the BEAM ecosystem, the
core evaluation logic, metric definitions, and prompt templates are derived
from the original work.

Key components derived from DeepEval:
- Evaluation metric algorithms (G-Eval, Faithfulness, Hallucination, etc.)
- Prompt templates for LLM-as-judge evaluation
- Test case and result data structures
- Metric scoring and aggregation logic

================================================================================

Third-Party Notices:

This project may include or reference algorithms and techniques from academic
papers, including but not limited to:

- G-Eval: NLG Evaluation using GPT-4 with Better Human Alignment
  https://arxiv.org/abs/2303.16634

- RAGAS: Automated Evaluation of Retrieval Augmented Generation
  https://arxiv.org/abs/2309.15217