datawhalechina/hello-agents · warning · ValueError

议题不能为空

Error message

议题不能为空

What it means

run_historical_debate (the generator in debate_orchestrator.py) strips its topic argument and raises ValueError('议题不能为空') when the result is empty. This is deliberate input validation at the top of the debate pipeline: an empty topic would send meaningless prompts through four LLM phases (round1, digest, round2, synthesis), so the orchestrator refuses before creating an LLM client or emitting any events.

Solutions

1. Pass a non-empty topic: run_historical_debate("评价唐朝灭亡的原因") after stripping whitespace yourself.
2. If calling from the web layer, validate/strip req.topic in the endpoint (app.py already does this and returns HTTP 400) rather than letting the orchestrator raise.
3. In CLIs, argparse with required=True plus a check that topic.strip() is non-empty prevents silent empty strings.
4. Catch ValueError at the caller boundary and surface it as a user-facing '请输入议题' message instead of a stack trace.

Example fix

# before
run_historical_debate(topic=user_input)
# after
topic = (user_input or "").strip()
if not topic:
    raise SystemExit("请输入有效的议题")
run_historical_debate(topic=topic)

topic = (topic or "").strip()
if not topic:
    raise ValueError("议题不能为空 — provide a non-empty debate topic")
md = run_historical_debate(topic, ...)

def is_valid_topic(topic: object) -> bool:
    return isinstance(topic, str) and bool(topic.strip())

try:
    md = collect_debate(run_historical_debate(topic, ...))
except ValueError as e:
    if "议题不能为空" in str(e):
        # user-input problem — re-prompt, never retry the same input
        raise UserInputError(str(e)) from e
    raise

When it happens

Trigger: Calling run_historical_debate(""), run_historical_debate(None), or a topic of only whitespace/newlines; a UI passing an unvalidated textarea value straight through; a CLI arg like --topic '' ; upstream code trimming the topic to empty before forwarding it.

Common situations: Frontend form submitted with an empty topic field; programmatic callers forwarding user input without validation; trimming logic that reduces a punctuation-only topic to an empty string; API clients testing the endpoint with empty payloads.

Related errors

• 会议记录过短，请至少提供 20 个字符
• 议题不能为空
• 辩论未正常结束
• 缺少完整 Markdown
• site_id not found

AI-assisted analysis of datawhalechina/hello-agents@606a07d341 (2026-08-14). Data as JSON: /api/errors/a161b3af8ffb5e79. Report an issue: GitHub.

Appendix: source

Thrown at Co-creation-projects/meiguanxiHXX-historyReviewAgent/historical_review/debate_orchestrator.py:110

    llm_model: str | None = None,
    llm_max_tokens: int | None = 4096,
    llm_timeout: int | None = None,
) -> Iterator[dict[str, Any]]:
    """
    逐步产出辩论过程事件，供 SSE / 日志展示。

    事件类型
    --------
    - progress: step, total, message
    - round1_start / round1_end: role, content（end）
    - digest_start / digest_end: content（end）
    - round2_start / round2_end: role, content（end）
    - synthesis_start / synthesis_end: content（end）
    - complete: markdown（全文）
    """
    topic = (topic or "").strip()
    if not topic:
        raise ValueError("议题不能为空")

    if llm is None:
        llm = create_llm(
            api_key=llm_api_key,
            base_url=llm_base_url,
            model=llm_model,
            max_tokens=llm_max_tokens,
            timeout=llm_timeout,
            temperature=0.4,
        )

    step = 0

    yield _yield_progress(step, f"议题已接收：{topic[:80]}{'…' if len(topic) > 80 else ''}")
    step += 1

    evidence_block = ""
    if use_evidence_bundle:

View on GitHub (pinned to 606a07d341)