UPDATE

rag/embedding.py

@@ -1,9 +1,12 @@
 """
 سرویس تعبیه‌سازی متن — از Adapter Pattern برای سوئیچ بین providers استفاده می‌کند
 """
+import logging
+import time
+
 from .api_provider import get_embedding_client
 from .config import RAGConfig, load_rag_config
-import logging
+from .observability import classify_exception, log_event, observe_operation, record_metric
 
 logger = logging.getLogger(__name__)
 
@@ -26,12 +29,13 @@ def embed_texts(
         لیست وکتورها
     """
     if not texts:
+        record_metric("rag.embedding.empty_input", operation="embed_texts")
         return []
 
     cfg = config or load_rag_config()
     client = get_embedding_client(cfg)
     model_name = model or cfg.embedding.model
-    logger.info(model_name)
+    provider = cfg.embedding.provider or "unknown"
     batch_size = cfg.embedding.batch_size
 
     all_embeddings: list[list[float]] = []
@@ -39,15 +43,44 @@ def embed_texts(
     if dimensions is not None:
         extra["dimensions"] = dimensions
 
-    for i in range(0, len(texts), batch_size):
-        batch = texts[i : i + batch_size]
-        resp = client.embeddings.create(
-            model=model_name,
-            input=batch,
-            **extra,
-        )
-        for item in sorted(resp.data, key=lambda x: x.index):
-            all_embeddings.append(item.embedding)
+    with observe_operation(source="rag.embedding", provider=provider, operation="embed_texts"):
+        for i in range(0, len(texts), batch_size):
+            batch = texts[i : i + batch_size]
+            started_at = time.monotonic()
+            try:
+                resp = client.embeddings.create(
+                    model=model_name,
+                    input=batch,
+                    **extra,
+                )
+            except Exception as exc:
+                failure = classify_exception(exc)
+                log_event(
+                    level=logging.ERROR,
+                    message="embedding batch request failed",
+                    source="rag.embedding",
+                    provider=provider,
+                    operation="embed_batch",
+                    result_status="error",
+                    duration_ms=(time.monotonic() - started_at) * 1000,
+                    error_code=failure.error_code,
+                    batch_size=len(batch),
+                    model=model_name,
+                )
+                raise
+            for item in sorted(resp.data, key=lambda x: x.index):
+                all_embeddings.append(item.embedding)
+            log_event(
+                level=logging.INFO,
+                message="embedding batch request completed",
+                source="rag.embedding",
+                provider=provider,
+                operation="embed_batch",
+                result_status="success",
+                duration_ms=(time.monotonic() - started_at) * 1000,
+                batch_size=len(batch),
+                model=model_name,
+            )
 
     return all_embeddings