Skip to content

FinceptTerminal is a modern finance application offering advanced market analytics, investment research, and economic data tools, designed for interactive exploration and data-driven decision-making in a user-friendly environment.

refactor(ai_chat): clear pending request context on stream completion and remove unused handle_response

Abhi Kakadiya57660e11857b
diff --git a/fincept-qt/src/screens/ai_chat/AiChatScreen.cpp b/fincept-qt/src/screens/ai_chat/AiChatScreen.cpp
index 5f8fd72e4..2f14c6a71 100644
--- a/fincept-qt/src/screens/ai_chat/AiChatScreen.cpp
+++ b/fincept-qt/src/screens/ai_chat/AiChatScreen.cpp
@@ -134,8 +134,11 @@ void AiChatScreen::refresh_theme() {
void AiChatScreen::showEvent(QShowEvent* e) {
QWidget::showEvent(e);
- connect(&ai_chat::LlmService::instance(), &ai_chat::LlmService::finished_streaming, this,
- &AiChatScreen::on_streaming_done, Qt::UniqueConnection);
+ // Connect LlmService finished_streaming to handle streaming responses using request‑time context
+ connect(&ai_chat::LlmService::instance(), &ai_chat::LlmService::finished_streaming,
+ this, &AiChatScreen::on_streaming_done, Qt::UniqueConnection);
+ // No additional code needed here.
+
connect(&ai_chat::LlmService::instance(), &ai_chat::LlmService::config_changed, this,
&AiChatScreen::on_provider_changed, Qt::UniqueConnection);
// Phase 7: rehydrate the linked symbol from SymbolContext on every
diff --git a/fincept-qt/src/screens/ai_chat/AiChatScreen.h b/fincept-qt/src/screens/ai_chat/AiChatScreen.h
index f77b1ab17..52cb96092 100644
--- a/fincept-qt/src/screens/ai_chat/AiChatScreen.h
+++ b/fincept-qt/src/screens/ai_chat/AiChatScreen.h
@@ -64,11 +64,6 @@ class AiChatScreen : public QWidget, public IStatefulScreen, public fincept::IGr
void on_rename_session();
void on_stream_chunk(const QString& chunk, bool done);
void on_streaming_done(ai_chat::LlmResponse response);
- // New helper that receives the LLM response together with the request snapshot
- void handle_response(const ai_chat::LlmResponse &response,
- const QString &req_session,
- const QString &req_provider,
- const QString &req_model);
void on_provider_changed();
void on_search_changed(const QString& text);
void on_typing_indicator_tick();
diff --git a/fincept-qt/src/screens/ai_chat/AiChatScreen_Messaging.cpp b/fincept-qt/src/screens/ai_chat/AiChatScreen_Messaging.cpp
index 5f4d650a4..6f9062356 100644
--- a/fincept-qt/src/screens/ai_chat/AiChatScreen_Messaging.cpp
+++ b/fincept-qt/src/screens/ai_chat/AiChatScreen_Messaging.cpp
@@ -215,8 +215,18 @@ void AiChatScreen::on_stream_chunk(const QString& chunk, bool done) {
void AiChatScreen::on_streaming_done(ai_chat::LlmResponse response) {
streaming_ = false;
+ // Clear pending request context early to avoid stale data on error paths
+ pending_req_session_.clear();
+ pending_req_provider_.clear();
+ pending_req_model_.clear();
show_typing(false);
-
+
+ // Ensure a bubble exists for any terminal state (success + content,
+ // failure with an error message, or success-but-empty). Without this,
+ // a kimi/openai response that arrives via the non-streaming fallback
+ // with no per-chunk emission, or fails before any chunk lands, would
+ // be silently dropped — the typing indicator hides and the input is
+ // re-enabled but nothing is shown.
// Ensure a bubble exists for any terminal state (success + content,
// failure with an error message, or success-but-empty). Without this,
// a kimi/openai response that arrives via the non-streaming fallback