davebb/linear-coding-agent
1
0
Fork 0
Code Issues Pull Requests 1 Actions Packages Projects Releases Wiki Activity
115 Commits 2 Branches 0 Tags
7707429d878374c6dd8fb43708986ebd884b4d82
Commit Graph

4 Commits

Author SHA1 Message Date
David Blanc Brioir
4400d25a58 fix: MCP tools migrated from StateVector V1 to StateTensor V2 
- identity_tools.py: rewritten to read StateTensor (8x1024 named vectors)
  instead of StateVector (single 1024-dim). Uses CATEGORY_TO_DIMENSION mapping.
- mcp_server.py: get_state_vector renamed to get_state_tensor
- __init__.py: updated exports

Now returns S(30) with architecture v2_tensor instead of S(2) from V1.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
 2026-02-06 19:42:04 +01:00
David Blanc Brioir
c633ab5aff Fix unified tools to use near_vector instead of near_text 
The Weaviate collections have vectorizer: "none", so near_text
searches fail silently. Changed all search handlers to:
- Import get_embedder from embedding_service
- Generate query vectors manually
- Use near_vector for semantic search

Affected handlers:
- search_memories_handler
- trace_concept_evolution_handler
- check_consistency_handler

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
 2026-01-31 21:46:33 +01:00
David Blanc Brioir
9e657cbf29 Add unified memory tools: search_memories, trace_concept_evolution, check_consistency, update_thought_evolution_stage 
- Create memory/mcp/unified_tools.py with 4 new handlers:
  - search_memories: unified search across Thoughts and Conversations
  - trace_concept_evolution: track concept development over time
  - check_consistency: verify statement alignment with past content
  - update_thought_evolution_stage: update thought maturity stage
- Export new tools from memory/mcp/__init__.py
- Register new tools in mcp_server.py with full docstrings

These tools complete the Ikario memory toolset to match memoryTools.js expectations.

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
 2026-01-30 17:24:30 +01:00
David Blanc Brioir
2f34125ef6 feat: Add Memory system with Weaviate integration and MCP tools 
MEMORY SYSTEM ARCHITECTURE:
- Weaviate-based memory storage (Thought, Message, Conversation collections)
- GPU embeddings with BAAI/bge-m3 (1024-dim, RTX 4070)
- 9 MCP tools for Claude Desktop integration

CORE MODULES (memory/):
- core/embedding_service.py: GPU embedder singleton with PyTorch
- schemas/memory_schemas.py: Weaviate schema definitions
- mcp/thought_tools.py: add_thought, search_thoughts, get_thought
- mcp/message_tools.py: add_message, get_messages, search_messages
- mcp/conversation_tools.py: get_conversation, search_conversations, list_conversations

FLASK TEMPLATES:
- conversation_view.html: Display single conversation with messages
- conversations.html: List all conversations with search
- memories.html: Browse and search thoughts

FEATURES:
- Semantic search across thoughts, messages, conversations
- Privacy levels (private, shared, public)
- Thought types (reflection, question, intuition, observation)
- Conversation categories with filtering
- Message ordering and role-based display

DATA (as of 2026-01-08):
- 102 Thoughts
- 377 Messages
- 12 Conversations

DOCUMENTATION:
- memory/README_MCP_TOOLS.md: Complete API reference and usage examples

All MCP tools tested and validated (see test_memory_mcp_tools.py in archive).

Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
 2026-01-08 18:08:13 +01:00