feat: implement memory compression and deduplication (issue #141)

[NTLx]

Summary

Implemented an intelligent memory optimization system to reduce redundancy and improve memory quality.

• Deduplication: Added exact match (hash-based) and semantic (vector similarity) deduplication.
• Compression: Implemented LLM-based memory summarization using greedy clustering of similar memories.
• Integration: Exposed the optimization interface via memory.optimize() in the core Memory class.

Addressing oceanbase/seekdb#123 (Track 7 - Memory Compression and Deduplication).

Test plan

• Added unit tests for MemoryOptimizer: tests/unit/intelligence/test_memory_optimizer.py
• Added integration tests for Memory.optimize: tests/unit/core/test_memory_optimize.py
• All tests passed locally using uv run pytest.

AI Contribution Details

• Analysis: Explored IntelligentMemoryManager and Memory core logic.
• Env: Set up a clean feature branch and confirmed dependency compatibility.
• Impl: Created modular MemoryOptimizer and logic-driven prompt templates.
• Validation: Verified with TDD-style unit tests and delegation tests.

🤖 Generated with Claude Code

[Teingi]

The test pipeline has failed.

[NTLx]

Hi @Teingi, the test failures have been fixed!

Issues resolved:

• Fixed in tests
• Changed test approach to use instead of instantiating Memory class
• This avoids vector store initialization that requires real provider configuration

Results:

• All 6 tests now pass ✅ (2 in test_memory_optimize.py + 4 in test_memory_optimizer.py)
• test_memory_optimize_delegation: PASSED
• test_memory_compress_delegation: PASSED
• test_deduplicate_exact_match: PASSED
• test_calculate_similarity: PASSED
• test_deduplicate_semantic: PASSED
• test_compress_logic: PASSED

Please re-trigger the CI pipeline to verify. Thanks!