jze9/code-memory
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
7 Commits 1 Branch 0 Tags
fa66a8098cf011a671dca1dc8126fc3dfaf0dacd
Commit Graph

7 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
jze9
fa66a8098c Rewrite extraction prompt and system prompt 
Extraction: system+user split with explicit SAVE/DON'T SAVE rules,
few-shot examples, stricter JSON-only format. Min fact length 15 chars.

System prompt: model now knows about its 3-level memory, how to use
recalled facts, and that user can control memory via chat commands.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-05-28 20:48:04 +05:00
jze9
9974ce8dd9 Add memory management: chat commands + REST API 
Chat commands (intercept before LLM):
  "запомни что X" / "remember X" — saves with importance 0.9
  "забудь про X" / "forget X"    — deletes by vector similarity

REST API:
  POST   /v1/memories             — add memory manually
  PATCH  /v1/memories/{id}        — update content or importance
  DELETE /v1/memories/{id}        — delete by id (existing)

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-05-28 20:41:53 +05:00
jze9
74c3fac2cc Add deduplication to store_memory 
Before inserting, check if a near-identical memory exists (cosine > 0.85).
If found, skip insert. Prevents flooding the DB with repeated facts
from the same conversation.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-05-28 20:39:28 +05:00
jze9
8e5149a7c3 Fix 404: map 'code-memory' model name to real Ollama model 
Open WebUI sends the proxy model name to our endpoint; passing it
directly to Ollama caused 404. Now stripped before forwarding.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-05-28 20:36:56 +05:00
jze9
3f392bffca Fix streaming TransferEncodingError 
DB saves were happening inside the generator after [DONE], causing
connection abort if any await failed. Moved to BackgroundTask.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-05-28 20:35:03 +05:00
jze9
7a0c819eca Add /v1/models endpoint for Open WebUI integration 
Open WebUI requires this endpoint to discover and list available models
from OpenAI-compatible connections.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-05-28 20:32:52 +05:00
jze9
3aa6ec8b3a Initial commit: code-memory service 
OpenAI-compatible FastAPI proxy with GraphRAG + persistent memory.
Includes 3-level Letta memory, CMA consolidation, AST-based code indexing.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-05-28 20:27:16 +05:00