mirrors/hermes-agent
1
0
Fork 0
mirror of https://github.com/NousResearch/hermes-agent.git synced 2026-04-25 00:51:20 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

fix(cli, agent): improve streaming handling and state management

Browse source 
- Updated _stream_delta method in HermesCLI to handle None values, flushing the stream and resetting state for clean tool execution.
- Enhanced quiet mode handling in AIAgent to ensure proper display closure before tool execution, preventing display issues with intermediate streamed content.

These changes improve the robustness of the streaming functionality and ensure a smoother user experience during tool interactions.
This commit is contained in:
Test 2026-03-20 10:02:42 -07:00
parent 0e3b7b6a39
commit 76bc27199f
2 changed files with 23 additions and 3 deletions
Download patch file Download diff file Expand all files Collapse all files

16
run_agent.py
View file

@ -4838,7 +4838,7 @@ class AIAgent:
spinner.stop(cute_msg)
elif self.quiet_mode:
self._vprint(f" {cute_msg}")
elif self.quiet_mode and not self._has_stream_consumers():
elif self.quiet_mode:
face = random.choice(KawaiiSpinner.KAWAII_WAITING)
emoji = _get_tool_emoji(function_name)
preview = _build_tool_preview(function_name, function_args) or function_name
@ -6568,7 +6568,19 @@ class AIAgent:
self._vprint(f" ┊ 💬 {clean}")
messages.append(assistant_msg)
# Close any open streaming display (response box, reasoning
# box) before tool execution begins. Intermediate turns may
# have streamed early content that opened the response box;
# flushing here prevents it from wrapping tool feed lines.
# Only signal the display callback — TTS (_stream_callback)
# should NOT receive None (it uses None as end-of-stream).
if self.stream_delta_callback:
try:
self.stream_delta_callback(None)
except Exception:
pass
_msg_count_before_tools = len(messages)
self._execute_tool_calls(assistant_message, messages, effective_task_id, api_call_count)