fix(rag): distinguish query and document embeddings

scripts/openvino-embeddings-server.py

@@ -67,22 +67,29 @@ class EmbeddingService:
         cfg.batch_size = 1
         self.pipeline = ovg.TextEmbeddingPipeline(self.model_dir, self.device, cfg)
 
-    def embed_one(self, text: str) -> dict[str, Any]:
+    def embed_one(self, text: str, *, purpose: str = "query") -> dict[str, Any]:
         text = str(text or "")
         if not text.strip():
             raise ValueError("embedding input text is empty")
+        if purpose not in {"query", "document"}:
+            raise ValueError("embedding purpose must be 'query' or 'document'")
         before = npu_busy_time_us()
         started = time.perf_counter()
         # TextEmbeddingPipeline is a native object; serialize calls until proven
         # safe under concurrent NPU use. Tiny silicon clown-car avoidance clause.
         with self.lock:
-            vec = self.pipeline.embed_query(text)
+            if purpose == "document":
+                # batch_size=1 means embed_documents must receive exactly one doc.
+                vec = self.pipeline.embed_documents([text])[0]
+            else:
+                vec = self.pipeline.embed_query(text)
         after = npu_busy_time_us()
         vector = [float(x) for x in vec]
         self.embedding_dim = len(vector)
         return {
             "embedding": vector,
             "dim": len(vector),
+            "purpose": purpose,
             "duration_ms": round((time.perf_counter() - started) * 1000, 3),
             "npu_busy_delta_us": None if before is None or after is None else after - before,
         }
@@ -136,17 +143,19 @@ class Handler(BaseHTTPRequestHandler):
             payload = self.read_json()
             if path == "/api/embed":
                 texts = normalize_input(payload.get("input"))
-                results = [self.svc.embed_one(text) for text in texts]
+                purpose = str(payload.get("purpose") or payload.get("task") or "document")
+                results = [self.svc.embed_one(text, purpose=purpose) for text in texts]
                 self.write_json({
                     "model": payload.get("model") or self.svc.model_name,
                     "embeddings": [item["embedding"] for item in results],
                     "embedding_dim": results[0]["dim"] if results else None,
+                    "purpose": purpose,
                     "npu_busy_delta_us": sum((item.get("npu_busy_delta_us") or 0) for item in results),
                     "durations_ms": [item["duration_ms"] for item in results],
                 })
             elif path == "/api/embeddings":
                 text = payload.get("prompt") or payload.get("input")
-                result = self.svc.embed_one(str(text or ""))
+                result = self.svc.embed_one(str(text or ""), purpose="query")
                 self.write_json({
                     "model": payload.get("model") or self.svc.model_name,
                     "embedding": result["embedding"],
@@ -156,7 +165,8 @@ class Handler(BaseHTTPRequestHandler):
                 })
             elif path == "/v1/embeddings":
                 texts = normalize_input(payload.get("input"))
-                results = [self.svc.embed_one(text) for text in texts]
+                purpose = str(payload.get("purpose") or payload.get("task") or "query")
+                results = [self.svc.embed_one(text, purpose=purpose) for text in texts]
                 self.write_json({
                     "object": "list",
                     "model": payload.get("model") or self.svc.model_name,
@@ -166,6 +176,7 @@ class Handler(BaseHTTPRequestHandler):
                     ],
                     "usage": {"prompt_tokens": 0, "total_tokens": 0},
                     "embedding_dim": results[0]["dim"] if results else None,
+                    "purpose": purpose,
                     "npu_busy_delta_us": sum((item.get("npu_busy_delta_us") or 0) for item in results),
                     "durations_ms": [item["duration_ms"] for item in results],
                 })