akadmin/Funk-lab
1
0
Fork 0
mirror of https://github.com/Alkatrazz24/Funk-lab.git synced 2026-07-08 11:04:43 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions

feat(stt): mémoire court-terme de session côté serveur (5a) (#8)

Browse source 
Serveur : SessionStore (historique borné + TTL, en mémoire), /v1/ask accepte session_id
et injecte l'historique dans l'appel LLM, /v1/reset l'efface. Dockerfile en 1 worker
(cohérence mémoire process).

Client : session_id généré par run (uuid), envoyé à chaque requête ; commande /reset
en mode texte.

Testé en process (TestClient) : historique croît 0→2→4, reset→0, sessions isolées,
sans session_id = sans état.

Mémoire long-terme Qdrant (5b) + réparation Qdrant (5c) à suivre.


Claude-Session: https://claude.ai/code/session_013FmcxGsyXZXogiAHQLjnZT

Co-authored-by: Claude <noreply@anthropic.com>
This commit is contained in:
ALI YESILKAYA 2026-06-17 16:15:57 +02:00 • committed by GitHub
parent bcdb4f8b2a
commit 815beaf37c
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
8 changed files with 107 additions and 17 deletions
Download patch file Download diff file Expand all files Collapse all files

18
stt/server/stt_server/app.py
View file

@ -14,13 +14,16 @@ from pydantic import BaseModel
from stt_server import __version__
from stt_server.brain import ask as brain_ask
from stt_server.config import settings
from stt_server.memory import SessionStore
app = FastAPI(title="STT-server", version=__version__)
sessions = SessionStore()
class AskRequest(BaseModel):
text: str
model: str | None = None # alias LiteLLM ; défaut serveur si absent
model: str | None = None # alias LiteLLM ; défaut serveur si absent
session_id: str | None = None # mémoire court-terme : fil de conversation
class AskReply(BaseModel):
@ -38,6 +41,13 @@ async def v1_models() -> dict:
return {"default": settings.model, "available": settings.allowed_models}
@app.post("/v1/reset")
async def v1_reset(req: AskRequest) -> dict:
if req.session_id:
sessions.reset(req.session_id)
return {"status": "reset"}
@app.post("/v1/ask", response_model=AskReply)
async def v1_ask(req: AskRequest) -> AskReply:
text = req.text.strip()
@ -49,10 +59,14 @@ async def v1_ask(req: AskRequest) -> AskReply:
status_code=400,
detail=f"modèle '{model}' non autorisé ; dispo : {settings.allowed_models}",
)
history = sessions.history(req.session_id) if req.session_id else None
try:
reply = await brain_ask(text, model)
reply = await brain_ask(text, model, history)
except httpx.HTTPError as e:
raise HTTPException(status_code=502, detail=f"upstream LiteLLM : {e}") from e
if req.session_id:
sessions.add(req.session_id, "user", text)
sessions.add(req.session_id, "assistant", reply)
return AskReply(reply=reply, model=model)