feat(client): add get/set_evaluation_context methods

[balgaly]

Summary

Closes#500

Adds set_evaluation_context() and get_evaluation_context() to OpenFeatureClient per the OpenFeature spec requirement 3.2.2.

The client-level context is already merged correctly during flag evaluation (global → transaction → client → per-call). These methods expose a public setter/getter so callers can update the context after construction — matching what the Go, Java, and .NET SDKs already provide.

Changes

• openfeature/client.py: add get_evaluation_context() and set_evaluation_context() with type validation
• tests/test_client.py: 4 new tests covering set/get, type error, and integration with flag evaluation

Tests

test_client_set_evaluation_context_updates_context
test_client_get_evaluation_context_returns_current_context
test_client_set_evaluation_context_invalid_type_raises
test_client_set_evaluation_context_is_used_in_flag_evaluation

[gemini-code-assist]

Code Review

This pull request introduces get_evaluation_context and set_evaluation_context methods to the OpenFeatureClient class, allowing for client-level evaluation context management as per the OpenFeature specification. The implementation includes type validation for the evaluation context and is supported by new unit tests covering context updates, retrieval, error handling for invalid types, and integration in flag evaluation. I have no feedback to provide.

[balgaly]

Picked up on @gruebel's note in #565 that the actual fix for #500 is the missing E2E tests, not the setter. Going to close this PR and open a new one targeting those tests instead.

[balgaly]

Superseded by #593, which implements the E2E tests @gruebel requested instead of adding a new setter.