unslothai/unsloth · error · HTTPException

Unstructured seed support not available (missing data_design

Error message

Unstructured seed support not available (missing data_designer_unstructured_seed)

What it means

HTTP 500 raised by _read_preview_rows_from_unstructured_file when the optional module data_designer_unstructured_seed could not be imported (the functions resolve_chunking / build_unstructured_preview_rows are None). The unstructured-seed feature is an optional dependency, so the route degrades to a 500 when it is absent.

Source

Thrown at studio/backend/routes/data_recipe/seed.py:255

        raise
    except (ValueError, OSError) as exc:
        raise log_and_http_error(
            exc,
            422,
            "seed inspect failed",
            event = "data_recipe.seed.local_preview_failed",
            log = logger,
        ) from exc

    rows = df.to_dict(orient = "records")
    return _serialize_preview_rows(rows)


def _read_preview_rows_from_unstructured_file(
    *, path: Path, preview_size: int, chunk_size: int | None, chunk_overlap: int | None
) -> list[dict[str, Any]]:
    if resolve_chunking is None or build_unstructured_preview_rows is None:
        raise HTTPException(
            500,
            "Unstructured seed support not available (missing data_designer_unstructured_seed)",
        )
    size, overlap = resolve_chunking(chunk_size, chunk_overlap)
    try:
        rows = build_unstructured_preview_rows(
            source_path = path,
            preview_size = preview_size,
            chunk_size = size,
            chunk_overlap = overlap,
        )
    except (FileNotFoundError, RuntimeError, ValueError, OSError) as exc:
        raise log_and_http_error(
            exc,
            422,
            "seed inspect failed",
            event = "data_recipe.seed.unstructured_preview_failed",
            log = logger,

View on GitHub (pinned to 203007d190)

Solutions

1. Install the optional unstructured-seed extra for studio backend (e.g. pip install 'gretel-studio[unstructured]' or add data_designer_unstructured_seed to the environment).
2. Check backend startup logs for the ImportError that zeroed these symbols and fix the underlying missing package.
3. Until installed, avoid the unstructured preview path — use CSV/JSON seeds.

Example fix

# before
pip install studio-backend

# after
pip install 'studio-backend[unstructured-seed]'

// Probe capability once at app start
const capabilities = await fetch('/data_recipe/capabilities').then(r => r.json());
if (!capabilities.unstructuredSeed) disableUnstructuredUI();

Catch HTTPException 500 with 'Unstructured seed support not available' and degrade gracefully: hide unstructured options and show an install hint, since retrying cannot help.

When it happens

Trigger: POST /seed/inspect-upload with unstructured files (pdf/docx/txt/md) on a backend installation where the data_designer_unstructured_seed extra was never installed or its import failed.

Common situations: Slim install without optional extras; partial virtualenv; dependency renamed or moved in a version bump; module raises on import due to its own missing deps (e.g. pdf parser) leaving the names None.

Related errors

• NeMo Data Designer Hugging Face integration is not installed
• GPU selection is unavailable on this host: {exc}
• Unsloth: MLX inference requires unsloth-zoo with the MLX mod
• Speech-to-text needs PyTorch, Transformers, and PyAV. Run `u
• Failed to read the local dataset cache.

AI-assisted analysis of unslothai/unsloth@203007d190 (2026-08-15). Data as JSON: /api/errors/bc924bda5a490642. Report an issue: GitHub.