opendatalab/MinerU · error · HTTPException

{detail}

Error message

{detail}

What it means

HTTPException that re-raises the upstream's non-200 status and body verbatim when fetching a task result. The upstream response body (or reason phrase if empty) becomes the error detail, and the failure is recorded on the task record.

Source

Thrown at mineru/cli/router.py:1320

                "GET",
                result_url,
                timeout=build_result_download_timeout(),
            ),
            stream=True,
        )
    except httpx.HTTPError as exc:
        await request.app.state.router_task_registry.increment_upstream_error(task.task_id, str(exc))
        raise HTTPException(status_code=502, detail=str(exc)) from exc

    if upstream_response.status_code != 200:
        body = await upstream_response.aread()
        await upstream_response.aclose()
        detail = body.decode("utf-8", errors="replace").strip() or upstream_response.reason_phrase
        await request.app.state.router_task_registry.increment_upstream_error(
            task.task_id,
            f"{upstream_response.status_code} {detail}",
        )
        raise HTTPException(
            status_code=upstream_response.status_code,
            detail=detail,
        )

    content_type = upstream_response.headers.get("content-type", "")
    if "application/json" in content_type:
        body = await upstream_response.aread()
        await upstream_response.aclose()
        return Response(
            content=body,
            status_code=200,
            media_type="application/json",
        )

    headers: dict[str, str] = {}
    content_disposition = upstream_response.headers.get("content-disposition")
    if content_disposition:
        headers["content-disposition"] = content_disposition

View on GitHub (pinned to 4fe4bde114)

Solutions

1. Read the status code and detail — they come straight from the upstream result endpoint
2. If 404 upstream: the result no longer exists there (restart/eviction); re-submit the task
3. Check upstream logs for the matching request; fix upstream result retention/storage if results vanish

r = client.get(f"{router}/tasks/{task_id}/result")
if r.status_code == 404:
    resubmit(files)  # upstream lost the result; nothing to fetch
elif r.status_code >= 500:
    retry_with_backoff(lambda: client.get(f"{router}/tasks/{task_id}/result"))

When it happens

Trigger: GET {upstream}/tasks/{id}/result returns e.g. 404 (result evicted upstream), 403, or 500; the router reads the body, closes the stream, and returns the same status/detail to the client.

Common situations: Upstream restarted and lost the finished result; result retention on the upstream expired; upstream error during result assembly. The status code is upstream's, so it identifies the true cause.

Related errors

• {exc.detail}
• No healthy upstream MinerU API servers are available
• MinerU upstream returned an invalid submit payload
• Invalid submit payload: {exc}
• Unknown process_mode: {process_mode}

AI-assisted analysis of opendatalab/MinerU@4fe4bde114 (2026-08-14). Data as JSON: /api/errors/4bc928acbebbf6fd. Report an issue: GitHub.