Add multi-turn evaluators Python sample

[JoseCSantos]

Adds a new Python sample under samples/python/multi-turn-evaluators showing how to evaluate pre-recorded multi-turn conversations using the four Azure AI Foundry multi-turn evaluators (Coherence, Task Completion, Customer Satisfaction, Groundedness).

What's included

• multi_turn_evaluators.ipynb: end-to-end notebook (load traces, convert to Foundry schema, create per-evaluator evals, submit runs, retrieve and analyze results).
• data/: 15 small labeled conversation traces covering all four evaluators.
• multi_turn_traces_foundry.jsonl: pre-built JSONL in Foundry format that can also be uploaded manually via the Foundry portal UI.
• README.md, requirements.txt, .env.example (with documented value formats).

Notes

• A GPT-5.x judge model is recommended/expected.
• The notebook groups traces by their target evaluator and creates one eval per evaluator to avoid noisy cross-evaluation.

[github-actions]

👋 Thanks for your interest in contributing, @JoseCSantos!

This repository does not accept pull requests directly. If you'd like to report a bug, suggest an improvement, or propose a new sample, please open an issue instead.

See CONTRIBUTING.md for more details.