fix: review learned context before persona writes

[EterUltimate]

Summary

• stop realtime expression learning from immediately injecting learned examples into persona begin_dialogs
• route memory replay and multidimensional runtime context into existing WebUI persona review records before applying
• add regression coverage for review-before-persona-write behavior

Tests

• $env:TMP=(Resolve-Path '.tmp\\pytest-tmp').Path; $env:TEMP=$env:TMP; $env:PYTEST_ADDOPTS='-p no:cacheprovider'; python -m pytest tests/unit/test_learning_chain_regressions.py tests/unit/test_persona_review_service.py tests/unit/test_config.py tests/unit/test_config_service.py -q

Fixes #187

[sourcery-ai]

Sorry @EterUltimate, you have reached your weekly rate limit of 500000 diff characters.

Please try again later or upgrade to continue using Sourcery