Poe: handle "Response timed out" exception

[johnd0e]

At the moment this exception is thrown:

capybara : Traceback (most recent call last):
  File "D:\AI\poe-api-wrapper\poe_api_wrapper\api.py", line 481, in send_message
    message = self.message_queues[human_message_id].get(timeout=timeout)
  File "C:\Users\a\AppData\Local\Programs\Python\Python310\lib\queue.py", line 179, in get
    raise Empty
_queue.Empty

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "D:\AI\poe-api-wrapper\example.py", line 3, in <module>
    Poe.chat_with_bot(token)
  File "D:\AI\poe-api-wrapper\poe_api_wrapper\api.py", line 935, in chat_with_bot
    for chunk in client.send_message(bot, message, chatId, suggest_replies=True, file_path=file_urls):
  File "D:\AI\poe-api-wrapper\poe_api_wrapper\api.py", line 485, in send_message
    raise RuntimeError("Response timed out.")
RuntimeError: Response timed out.

I suppose Poe should handle this, as it is pretty common case.

[snowby666]

I think it tends to be a problem with the poe.com website itself. Sometimes, even when I manually submit the question in the browser, the answer is not generated.

[johnd0e]

Sure it is poe.com problem.
And the subject of this issue is proper handling of this case, as it seems to be pretty common.

[snowby666]

I tried to reproduce this and caused the bot to enter a 'did not respond' loop, but I was only able to make it generate again by clearing the context and resending the message (kinda awkward). I suppose you could implement a try-except for that.

[7trail]

Hello,

An issue I've been having is the POE backend not responding, but the api not catching the timeout and instead halting the program indefinitely. Do you think you might know what is causing this error? More details can be provided if needed.

[snowby666]

Hello,

An issue I've been having is the POE backend not responding, but the api not catching the timeout and instead halting the program indefinitely. Do you think you might know what is causing this error? More details can be provided if needed.

If your program is just halting but not in an inf loop then this may help:
You can inspect the api.py and will find that send_message has a timeout arg. You can set a lower value for this arg:

for chunk in client.send_message(bot, message, timeout=2):
      print(chunk['response'], end='', flush=True)

[johnd0e]

I suppose you could implement a try-except for that.

Do you mean something like this?

def main():
    try:
        Poe.chat_with_bot(token)
    except RuntimeError:
        main()

main()

This seems not to be a good solution, as then I must manually re-choose bot and thread (do I ever remember those?).
It would be much more convenient if Poe would recover from this error without user intervention, as it already knows both bot's name an thread's code.

[snowby666]

As I've mentioned above, I tried to reproduce this error, but my attempts to cancel and resend the message still results in "did not respond" loop. Currently, I could only clear the context and start over. Maybe this error has something to do with poe.com itself

[johnd0e]

Currently, I could only clear the context and start over.

Even if you are doing that on poe.com?

Currently, I could only clear the context and start over.

I understand. I this case instead of exception I would prefer to see such message:

Poe.com error. Please !clear context.

[snowby666]

Currently, I could only clear the context and start over.

Even if you are doing that on poe.com?

Currently, I could only clear the context and start over.

I understand. I this case instead of exception I would prefer to see such message:

Poe.com error. Please !clear context.

Yeah. I think I will let the user to choose whether or not to clear the context if encounter this error. For now, if I can find another solution, I will implement it later

[johnd0e]

Thank you!

[snowby666]

This is temp fixed in the new version. The stability is improved but there may be some unexpected errors in the future.