What causes "SubQuery weight must be positive, got {self.weight}" in nexu-io/open-design?

Ensure the planner emits strictly positive weights (e.g. >= 0.1).

nexu-io/open-design · error · ValueError

SubQuery weight must be positive, got {self.weight}

Error message

SubQuery weight must be positive, got {self.weight}

What it means

Second invariant in SubQuery.__post_init__: weight must be > 0. Weight is used downstream for RRF-style fusion and ranking; a zero or negative weight would silently drop or invert the subquery's contribution, so construction fails loudly instead.

Source

Thrown at design-templates/last30days/scripts/lib/schema.py:55

    rerank_model: str
    x_search_backend: Literal["xai", "bird"] | None = None


@dataclass(frozen=True)
class SubQuery:
    """Planner-emitted retrieval unit."""

    label: str
    search_query: str
    ranking_query: str
    sources: list[str]
    weight: float = 1.0

    def __post_init__(self) -> None:
        if not self.sources:
            raise ValueError("SubQuery must have at least one source")
        if self.weight <= 0:
            raise ValueError(f"SubQuery weight must be positive, got {self.weight}")


@dataclass
class QueryPlan:
    """Planner output."""

    intent: str
    freshness_mode: str
    cluster_mode: str
    raw_topic: str
    subqueries: list[SubQuery]
    source_weights: dict[str, float]
    notes: list[str] = field(default_factory=list)


@dataclass
class SourceItem:
    """Generic normalized evidence item."""

View on GitHub (pinned to 5be4028344)

Solutions

Ensure the planner emits strictly positive weights (e.g. >= 0.1).
After any weight normalization, clamp each weight to a small positive floor before constructing SubQuery.
Omit the weight argument to inherit the default 1.0.

Example fix

# before
SubQuery(label='t', search_query='q', ranking_query='q', sources=['x'], weight=0.0)

# after
w = max(raw_weight, 0.1)
SubQuery(label='t', search_query='q', ranking_query='q', sources=['x'], weight=w)

Defensive patterns

Strategy: validation

Validate before calling

def build_subquery(label, search_query, ranking_query, sources, weight=1.0):
    if not sources:
        raise ValueError(f"subquery {label!r} needs at least one source")
    if weight <= 0:
        raise ValueError(f"subquery {label!r} weight must be > 0, got {weight}")
    return SubQuery(label, search_query, ranking_query, list(sources), weight=weight)

Type guard

def is_positive_weight(w) -> bool:
    return isinstance(w, (int, float)) and w > 0

Try / catch

null

Prevention

Clamp normalized weights to a small positive floor (e.g. max(w, 0.1)) before constructing SubQuery.
Validate the planner weight field is a positive number at parse time.
Default to 1.0 when the planner omits weight.

When it happens

Trigger: Constructing SubQuery(..., weight=0) or a negative weight. Reached when the planner emits a weight of 0, when a normalization step divides by a sum that produced 0, or when a default of 0 was used instead of the dataclass default 1.0.

Common situations: Planner JSON with 'weight': 0. Weight normalization that scales all weights to 0. Test fixture setting weight=0 by accident. Float parse of '0.0' string.

Related errors

AI-assisted analysis of nexu-io/open-design@5be4028344 (2026-08-12). Data as JSON: /api/errors/fd021b0a96b4524b. Report an issue: GitHub.