binary-husky/gpt_academic · error · ConnectionAbortedError

OpenAI拒绝了请求:{error_msg}

Error message

OpenAI拒绝了请求:{error_msg}

What it means

ConnectionAbortedError raised in bridge_openrouter.predict_no_ui_long_connection when a non-'data:' stream frame is received and the drained error text contains 'reduce the length' - the upstream's token-limit rejection. The bridge maps that substring to this exception so callers can treat it as context-overflow rather than a generic failure.

Source

Thrown at request_llms/bridge_openrouter.py:177

        chunkjson = json.loads(response.content.decode())
        gpt_replying_buffer = chunkjson['choices'][0]["message"]["content"]
        return gpt_replying_buffer

    stream_response = response.iter_lines()
    result = ''
    json_data = None
    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 len(chunk_decoded)==0 or chunk_decoded.startswith(':'): continue
        if not chunk_decoded.startswith('data:'):
            error_msg = get_full_error(chunk, stream_response).decode()
            if "reduce the length" in error_msg:
                raise ConnectionAbortedError("OpenAI拒绝了请求:" + error_msg)
            elif """type":"upstream_error","param":"307""" in error_msg:
                raise ConnectionAbortedError("正常结束，但显示Token不足，导致输出不完整，请削减单次输入的文本量。")
            else:
                raise RuntimeError("OpenAI拒绝了请求：" + error_msg)
        if ('data: [DONE]' in chunk_decoded): break # api2d 正常完成
        # 提前读取一些信息 （用于判断异常）
        json_data = chunkjson['choices'][0]
        delta = json_data["delta"]
        if len(delta) == 0: break
        if (not has_content) and has_role: continue
        if (not has_content) and (not has_role): continue # raise RuntimeError("发现不标准的第三方接口："+delta)
        if has_content: # has_role = True/False
            result += delta["content"]
            if not console_silence: print(delta["content"], 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] += delta["content"]

View on GitHub (pinned to d6bde0fa54)

Solutions

1. Trim history or shorten inputs so total tokens fit the model's context window
2. Switch llm_kwargs['llm_model'] to a model with a larger context window
3. Use the project's history-compression / split-input plugin instead of one giant prompt
4. Catch ConnectionAbortedError and show a 'reduce input' message to the user instead of retrying (retrying the same payload will fail identically)

Example fix

# before
history += [long_doc, reply]  # keeps growing

# after
MAX_TURNS = 8
history = history[-MAX_TURNS*2:]  # bound the context

# estimate token count before sending
approx_tokens = len(inputs) // 2 + sum(len(a) + len(b) for a, b in history) // 2
max_ctx = model_info[llm_kwargs['llm_model']].get('max_context', 8000)
assert approx_tokens < max_ctx * 0.9, f'input ~{approx_tokens} tokens exceeds safe budget {max_ctx}'

try:
    result = predict_no_ui_long_connection(inputs, llm_kwargs, history)
except ConnectionAbortedError as e:
    if 'reduce the length' in str(e):
        history = history[-6:]  # shrink context and retry once
        result = predict_no_ui_long_connection(inputs, llm_kwargs, history)
    else:
        raise

When it happens

Trigger: history plus inputs exceeding the model's context window so the API rejects the request; a long document plugin (e.g. paper reading) feeding too much text; a third-party OpenAI-compatible relay that forwards OpenAI's 'Please reduce the length of the messages' error verbatim.

Common situations: Summarizing very long PDFs; accumulated chat history growing past the model context; routing through OpenRouter to a model with a smaller window than expected.

Related errors

• OpenAI拒绝了请求:
• OpenAI拒绝了请求：{error_msg}
• 用户取消了程序。
• 程序终止。
• OpenAI拒绝了请求：

AI-assisted analysis of binary-husky/gpt_academic@d6bde0fa54 (2026-08-14). Data as JSON: /api/errors/ea43b08560371faa. Report an issue: GitHub.