P2.2+P2.3: NATS offload node-worker + router offload integration
Node Worker (services/node-worker/):
- NATS subscriber for node.{NODE_ID}.llm.request / vision.request
- Canonical JobRequest/JobResponse envelope (Pydantic)
- Idempotency cache (TTL 10min) with inflight dedup
- Deadline enforcement (DEADLINE_EXCEEDED on expired jobs)
- Concurrency limiter (semaphore, returns busy)
- Ollama + Swapper vision providers
Router offload (services/router/offload_client.py):
- NATS req/reply with configurable retries
- Circuit breaker per node+type (3 fails/60s → open 120s)
- Concurrency semaphore for remote requests
Model selection (services/router/model_select.py):
- exclude_nodes parameter for circuit-broken nodes
- force_local flag for fallback re-selection
- Integrated circuit breaker state awareness
Router /infer pipeline:
- Remote offload path when NCS selects remote node
- Automatic fallback: exclude failed node → force_local re-select
- Deadline propagation from router to node-worker
Tests: 17 unit tests (idempotency, deadline, circuit breaker)
Docs: ops/offload_routing.md (subjects, envelope, verification)
Made-with: Cursor
This commit is contained in:
42
services/node-worker/providers/swapper_vision.py
Normal file
42
services/node-worker/providers/swapper_vision.py
Normal file
@@ -0,0 +1,42 @@
|
||||
"""Swapper vision provider for node-worker."""
|
||||
import logging
|
||||
from typing import Any, Dict, List, Optional
|
||||
|
||||
import httpx
|
||||
|
||||
from config import SWAPPER_URL, DEFAULT_VISION
|
||||
|
||||
logger = logging.getLogger("provider.swapper_vision")
|
||||
|
||||
|
||||
async def infer(
|
||||
images: Optional[List[str]] = None,
|
||||
prompt: str = "",
|
||||
model: str = "",
|
||||
system: str = "",
|
||||
max_tokens: int = 1024,
|
||||
temperature: float = 0.2,
|
||||
timeout_s: float = 60.0,
|
||||
) -> Dict[str, Any]:
|
||||
model = model or DEFAULT_VISION
|
||||
|
||||
payload: Dict[str, Any] = {
|
||||
"model": model,
|
||||
"prompt": prompt or "Describe this image.",
|
||||
"max_tokens": max_tokens,
|
||||
"temperature": temperature,
|
||||
}
|
||||
if images:
|
||||
payload["images"] = images
|
||||
if system:
|
||||
payload["system"] = system
|
||||
|
||||
async with httpx.AsyncClient(timeout=timeout_s) as c:
|
||||
resp = await c.post(f"{SWAPPER_URL}/vision", json=payload)
|
||||
resp.raise_for_status()
|
||||
data = resp.json()
|
||||
return {
|
||||
"text": data.get("text", data.get("response", "")),
|
||||
"model": model,
|
||||
"provider": "swapper_vision",
|
||||
}
|
||||
Reference in New Issue
Block a user