openclaw - 💡(How to fix) Fix Telegram typing indicator should stay active during tool execution, not just message composition [1 comments, 2 participants]

TL;DR

Implement a mechanism to continuously send typing indicators during tool execution phases, utilizing existing callbacks like tool_progress_callback and step_callback.

Guidance

• Investigate how the Hermes agent implements callback-driven progress indicators to keep the typing status alive during tool execution, as it may serve as a reference model.
• Modify the OpenClaw gateway to send typing indicators at regular intervals during tool execution, potentially using the existing tool_progress_callback and step_callback to trigger these indicators.
• Consider adding a timer or a scheduling mechanism to ensure the typing indicator is updated periodically, even if the tool execution progress callbacks are not frequently triggered.
• Review the Telegram API documentation to ensure compliance with any specific requirements or limitations for sending typing indicators.

Example

No specific code example can be provided without more context on the existing implementation, but the idea would be to integrate the typing indicator updates within the callbacks, such as:

def tool_progress_callback(progress):
    # Update typing indicator here
    send_typing_indicator()
    # Existing progress handling code

def step_callback(step):
    # Update typing indicator here
    send_typing_indicator()
    # Existing step handling code

Notes

The exact implementation details may vary based on the OpenClaw architecture and the Telegram API specifics. Ensuring that the typing indicator does not spam the user or violate any platform policies is crucial.

Recommendation

Apply a workaround by integrating the typing indicator updates with the existing callbacks, as this approach leverages existing functionality and does not require waiting for a potential future version fix.