added logging of RAG chunks for debugging

server.py

@@ -22,9 +22,16 @@ class ChatRequest(BaseModel):
     message: str
 
 
+class RetrievedChunk(BaseModel):
+    content: str
+    source: str
+    page: int | None
+
+
 class ChatResponse(BaseModel):
     answer: str
     error: str | None = None
+    retrieved: list[RetrievedChunk] | None = None
 
 
 @app.get("/", response_class=HTMLResponse)
@@ -42,8 +49,21 @@ def chat(request: ChatRequest):
     if not request.message or not request.message.strip():
         return ChatResponse(answer="", error="Message cannot be empty")
     try:
-        answer = rag.query(request.message.strip())
-        return ChatResponse(answer=answer)
+        result = rag.query(request.message.strip())
+        answer = result["answer"]
+        retrieved = result.get("retrieved", [])
+
+        # Server-side console trace: log retrieved chunks before LLM answer
+        if retrieved:
+            print(f"\n[RAG] Retrieved {len(retrieved)} chunk(s) for query: {request.message[:80]!r}")
+            for i, chunk in enumerate(retrieved):
+                content = chunk.get("content", "")
+                preview = (content[:1000] + "...") if len(content) > 1000 else content
+                print(f"  [{i + 1}] source={chunk.get('source', '')} page={chunk.get('page')} | {preview!r}")
+        else:
+            print(f"\n[RAG] Retrieved 0 chunks for query: {request.message[:80]!r}")
+
+        return ChatResponse(answer=answer, retrieved=retrieved)
     except Exception as e:
         return ChatResponse(answer="", error=str(e))