feat(ui): explicit chat state machine — typing indicator, streaming progress, timeout and error recovery

app/api/chat.py

@@ -34,7 +34,7 @@ from app.rag.llm import EmbeddingError, LLMClient, LLMError
 from app.rag.prompts import build_deflect_prompt, build_high_prompt
 from app.rag.retriever import RetrievedChunk, retrieve, select_documents, weak_hit_titles
 from app.rag.suggestions import derive_suggestions
-from app.schemas import ChatDoneEvent, ChatRequest, SourceRef
+from app.schemas import ChatDoneEvent, ChatErrorEvent, ChatRequest, SourceRef
 
 logger = logging.getLogger("app.chat")
 router = APIRouter(tags=["chat"])
@@ -122,12 +122,19 @@ async def chat(
         try:
             question_vec = await llm.embed_one(request.message)
         except EmbeddingError as e:
-            logger.error("chat: embedding failed question=%r — %s", request.message, e)
+            embed_ms = int((time.monotonic() - t0) * 1000)
+            total_ms = int((time.monotonic() - started) * 1000)
+            logger.error(
+                "chat: question=%r embed_ms=%d total_ms=%d — embedding failed: %s",
+                request.message,
+                embed_ms,
+                total_ms,
+                e,
+            )
             yield sse_event(
-                {
-                    "type": "error",
-                    "detail": "I couldn't reach the embedding model — please try again.",
-                }
+                ChatErrorEvent(
+                    detail="I couldn't reach the embedding model — please try again."
+                ).model_dump()
             )
             return
         embed_ms = int((time.monotonic() - t0) * 1000)
@@ -139,12 +146,15 @@ async def chat(
             chunks = retrieve(db, question_vec)
             plan = plan_turn(chunks, settings)
         except Exception:  # noqa: BLE001 — DB failure mid-turn
-            logger.exception("chat: retrieval failed question=%r", request.message)
+            logger.exception(
+                "chat: retrieval failed question=%r total_ms=%d",
+                request.message,
+                int((time.monotonic() - started) * 1000),
+            )
             yield sse_event(
-                {
-                    "type": "error",
-                    "detail": "The knowledge base went offline mid-question — is Postgres up?",
-                }
+                ChatErrorEvent(
+                    detail="The knowledge base went offline mid-question — is Postgres up?"
+                ).model_dump()
             )
             return
         source_paths = [f"{d.source}/{d.path}" for d in plan.docs]
@@ -158,9 +168,16 @@ async def chat(
             async for piece in llm.chat_stream(messages):
                 yield sse_event({"type": "delta", "text": piece})
         except LLMError as e:
-            logger.error("chat: LLM stream failed question=%r — %s", request.message, e)
+            logger.error(
+                "chat: LLM stream failed question=%r total_ms=%d — %s",
+                request.message,
+                int((time.monotonic() - started) * 1000),
+                e,
+            )
             yield sse_event(
-                {"type": "error", "detail": "The chat model dropped the connection — try again?"}
+                ChatErrorEvent(
+                    detail="The chat model dropped the connection — try again?"
+                ).model_dump()
             )
             return