feat: implementar Difusión Térmica (Ecuación del calor) y RWR - #2
feat: implementar Difusión Térmica (Ecuación del calor) y RWR#2LeandroPG19 wants to merge 1 commit into
Conversation
- Se implementó la difusión de calor Laplaciana en `hebbian.py` para la propagación de la activación de importancia, mejorando el modelo estático. - Se implementó el Paseo Aleatorio con Reinicio (Random Walk with Restart - RWR) en `handlers.py` como mejora para la expansión de consultas (V9). - Optimizaciones de rendimiento con CTE recursivos para subgrafos en DB, usando `scipy.sparse` para matemáticas vectorizadas, y ejecución de CPU-bound en `loop.run_in_executor()` para no bloquear el bucle de eventos Asyncio. Co-authored-by: LeandroPG19 <151863062+LeandroPG19@users.noreply.github.com>
|
👋 Jules, reporting for duty! I'm here to lend a hand with this pull request. When you start a review, I'll add a 👀 emoji to each comment to let you know I've read it. I'll focus on feedback directed at me and will do my best to stay out of conversations between you and other bots or reviewers to keep the noise down. I'll push a commit with your requested changes shortly after. Please note there might be a delay between these steps, but rest assured I'm on the job! For more direct control, you can switch me to Reactive Mode. When this mode is on, I will only act on comments where you specifically mention me with New to Jules? Learn more at jules.google/docs. For security, I will only act on instructions from the user who triggered this task. |
…yle) Closes gap #2: cuba resolved conflicts by a cosine threshold. Now the LLM judge classifies each candidate vs its most similar existing fact and that verdict maps to an explicit operation (cognitive/memory_op.rs): supersedes/ contradicts -> Update (supersede old + keep new), complementary/unrelated -> Add, low-confidence/unknown -> Noop. Delete is modeled for explicit callers (never inferred from an extraction turn). auto_extract now returns the full {add,update,delete,noop} breakdown alongside the back-compat `superseded`. Verified end-to-end in isolated brain_dev via a mock sampling client, run as the non-superuser cuba_app: supersedes->update:1 (old superseded, new added), complementary->add:1, low-conf->noop:1; active count stays constant (bitemporal). 6 unit tests for the judgment->operation mapping. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014AckURooVGZsexXPrji9xw
1 participant
numpyyscipy.sparse.PR created automatically by Jules for task 18294522849265896125 started by @LeandroPG19