binary-husky/gpt_academic · error · RuntimeError

用户取消了程序。

Error message

用户取消了程序。

What it means

RuntimeError ('用户取消了程序。') from the Cohere bridge's watchdog in predict_no_ui_long_connection. While consuming Cohere's SSE events (text-generation frames appended to result), the code checks observe_window[1] — a timestamp the UI thread refreshes — against watch_dog_patience (5 s). If new text arrives but the timestamp is stale, the loop aborts on the assumption the user cancelled or the UI died. Standard cooperative-cancellation mechanism replicated across all bridges.

Source

Thrown at request_llms/bridge_cohere.py:122

    while True:
        try: chunk = next(stream_response)
        except StopIteration:
            break
        except requests.exceptions.ConnectionError:
            chunk = next(stream_response) # 失败了，重试一次？再失败就没办法了。
        chunk_decoded, chunkjson, has_choices, choice_valid, has_content, has_role = decode_chunk(chunk)
        if chunkjson['event_type'] == 'stream-start': continue
        if chunkjson['event_type'] == 'text-generation':
            result += chunkjson["text"]
            if not console_silence: print(chunkjson["text"], end='')
            if observe_window is not None:
                # 观测窗，把已经获取的数据显示出去
                if len(observe_window) >= 1:
                    observe_window[0] += chunkjson["text"]
                # 看门狗，如果超过期限没有喂狗，则终止
                if len(observe_window) >= 2:
                    if (time.time()-observe_window[1]) > watch_dog_patience:
                        raise RuntimeError("用户取消了程序。")
        if chunkjson['event_type'] == 'stream-end': break
    return result


def predict(inputs:str, llm_kwargs:dict, plugin_kwargs:dict, chatbot:ChatBotWithCookies,
            history:list=[], system_prompt:str='', stream:bool=True, additional_fn:str=None):
    """
    发送至chatGPT，流式获取输出。
    用于基础的对话功能。
    inputs 是本次问询的输入
    top_p, temperature是chatGPT的内部调优参数
    history 是之前的对话列表（注意无论是inputs还是history，内容太长了都会触发token数量溢出的错误）
    chatbot 为WebUI中显示的对话列表，修改它，然后yield出去，可以直接修改对话界面内容
    additional_fn代表点击的哪个按钮，按钮见functional.py
    """
    # if is_any_api_key(inputs):
    #     chatbot._cookies['api_key'] = inputs
    #     chatbot.append(("输入已识别为Cohere的api_key", what_keys(inputs)))

View on GitHub (pinned to d6bde0fa54)

Solutions

1. If deliberate, re-send the query after stopping.
2. Keep the UI tab in the foreground during generation.
3. Custom callers: update observe_window[1] = time.time() at ~1 Hz for the duration of the stream.
4. Raise watch_dog_patience in request_llms/bridge_cohere.py for slow setups.

try:
    result = predict_no_ui_long_connection(inputs, llm_kwargs, history, sys_prompt, observe_window=win)
except RuntimeError as e:
    if '用户取消了程序' in str(e):
        handle_cancel_or_stale_watchdog()
    else:
        raise

When it happens

Trigger: User stops generation in the Gradio UI during a Cohere response; UI polling thread stops updating observe_window[1] (suspended tab, stalled Gradio queue); custom integration passes observe_window without a feeding thread while text-generation events stream in.

Common situations: Intentional stops mid-answer; slow Cohere API or relay causing UI thread starvation; backgrounded browser tabs freezing the watchdog feed.

Related errors

• 用户取消了程序。
• 用户取消了程序。
• 用户取消了程序。
• 程序终止。
• 用户取消了程序。

AI-assisted analysis of binary-husky/gpt_academic@d6bde0fa54 (2026-08-14). Data as JSON: /api/errors/674da743c1e3869f. Report an issue: GitHub.