datawhalechina/hello-agents · error · Error

生成失败

Error message

生成失败

What it means

'生成失败' (generation failed) is the default message assigned in StockAnalysis.vue's stream loop when an NDJSON frame arrives with type:'error' but carries neither message nor content fields. The frame signals a server-side error mid-stream; the thrown Error surfaces it after the read loop exits. It only shows the literal '生成失败' when the server sent a bare error event with no detail.

Source

Thrown at Co-creation-projects/lcyting-StockSage-agent/frontend/src/views/StockAnalysis.vue:1085

        buf = buf.slice(nl + 1)
        if (!line) continue
        let obj
        try {
          obj = JSON.parse(line)
        } catch {
          continue
        }
        const deltaPiece = obj.text ?? obj.content
        if (obj.type === 'delta' && deltaPiece) {
          buffettStreamRaw.value += deltaPiece
        } else if (obj.type === 'error') {
          streamError = obj.message || obj.content || '生成失败'
        } else if (obj.type === 'done') {
          streamFinished = true
        }
      }

      if (streamError) throw new Error(streamError)
      if (streamFinished || readerDone) break
    }

    const tail = buf.trim()
    if (tail) {
      try {
        const obj = JSON.parse(tail)
        if (obj.type === 'error') throw new Error(obj.message || obj.content || '生成失败')
        const tailDelta = obj.text ?? obj.content
        if (obj.type === 'delta' && tailDelta) buffettStreamRaw.value += tailDelta
      } catch (e) {
        if (!(e instanceof SyntaxError)) throw e
      }
    }

    const md = buffettStreamRaw.value.trim()
    buffettAiMarkdown.value = md
    buffettStreamRaw.value = ''

View on GitHub (pinned to 606a07d341)

Solutions

1. Check the backend's error-frame emitter — make it always include the exception message so this default never fires blind.
2. Read backend logs at the timestamp of the failed stream for the real exception.
3. If it's provider rate-limiting/quota, fix keys/limits upstream and retry the analysis.
4. For timeout-class failures, chunk the analysis request or increase server-side LLM timeouts.

function isErrorFrame(obj: unknown): obj is { type: 'error'; message?: string; content?: string } {
  return typeof obj === 'object' && obj !== null && (obj as { type?: unknown }).type === 'error';
}

try {
  await streamBuffettAnalysis();
} catch (err) {
  const m = (err as Error).message;
  if (m === '生成失败') {
    // bare error frame — check server logs; likely LLM provider fault
    showError('生成失败：上游模型异常，请稍后重试');
  } else {
    showError(m);
  }
}

When it happens

Trigger: The analysis stream opens OK, deltas flow, then the server emits {type:'error'} (optionally with message/content) — LLM provider failure mid-generation, upstream timeout, token limit hit, or internal exception in the analysis pipeline. The default text appears specifically when the error frame is empty.

Common situations: LLM API rate-limit/key errors surfacing mid-stream; long analyses hitting provider timeouts; backend catching an exception and emitting a bare error frame without details.

Related errors

• msg
• 无法读取响应流
• 研究请求失败，状态码：${response.status}
• stream request failed: ${resp.status} ${message}
• 智能体流式请求失败: ${resp.status}

AI-assisted analysis of datawhalechina/hello-agents@606a07d341 (2026-08-14). Data as JSON: /api/errors/e3440af8552bb448. Report an issue: GitHub.