langchain-ai/langchain · error · ValueError

`score_threshold` is not specified with a float value(0~1) i

Error message

`score_threshold` is not specified with a float value(0~1) in `search_kwargs`.

What it means

Companion validation in the `VectorStoreRetriever` field validator: when `search_type='similarity_score_threshold'`, the retriever requires `search_kwargs['score_threshold']` to be present and to be a `float` between 0 and 1. Without a threshold the filtered search has no cutoff, so construction is rejected.

Source

Thrown at libs/core/langchain_core/vectorstores/base.py:1015

        Raises:
            ValueError: If `search_type` is not one of the allowed search types.
            ValueError: If `score_threshold` is not specified with a float value(`0~1`)
        """
        search_type = values.get("search_type", "similarity")
        if search_type not in cls.allowed_search_types:
            msg = (
                f"search_type of {search_type} not allowed. Valid values are: "
                f"{cls.allowed_search_types}"
            )
            raise ValueError(msg)
        if search_type == "similarity_score_threshold":
            score_threshold = values.get("search_kwargs", {}).get("score_threshold")
            if (score_threshold is None) or (not isinstance(score_threshold, float)):
                msg = (
                    "`score_threshold` is not specified with a float value(0~1) "
                    "in `search_kwargs`."
                )
                raise ValueError(msg)
        return values

    def _get_ls_params(self, **kwargs: Any) -> LangSmithRetrieverParams:
        """Get standard params for tracing."""
        kwargs_ = self.search_kwargs | kwargs

        ls_params = super()._get_ls_params(**kwargs_)

        ls_params["ls_vector_store_provider"] = self.vectorstore.__class__.__name__

        if self.vectorstore.embeddings:
            ls_params["ls_embedding_provider"] = (
                self.vectorstore.embeddings.__class__.__name__
            )
        elif hasattr(self.vectorstore, "embedding") and isinstance(
            self.vectorstore.embedding, Embeddings
        ):
            ls_params["ls_embedding_provider"] = (

View on GitHub (pinned to e32fa9a52e)

Solutions

1. Add a float threshold inside `search_kwargs`: `as_retriever(search_type='similarity_score_threshold', search_kwargs={'score_threshold': 0.5})`.
2. Cast config-loaded values: `float(config['score_threshold'])`.
3. Ensure the literal is a float (`0.75`, `1.0`), not an int or string.

Example fix

# before
retriever = store.as_retriever(search_type="similarity_score_threshold")

# after
retriever = store.as_retriever(
    search_type="similarity_score_threshold",
    search_kwargs={"score_threshold": 0.5},
)

def validate_retriever_config(search_type: str, search_kwargs: dict) -> None:
    if search_type == "similarity_score_threshold":
        st = search_kwargs.get("score_threshold")
        if st is None or not isinstance(st, float) or not 0.0 <= st <= 1.0:
            raise ValueError(
                "score_threshold must be a float in [0, 1] inside search_kwargs, "
                f"got {st!r}"
            )

def has_valid_score_threshold(search_kwargs: dict) -> bool:
    """True when score_threshold exists and is a float (not int/str)."""
    st = search_kwargs.get("score_threshold")
    return isinstance(st, float) and 0.0 <= st <= 1.0

from pydantic import ValidationError

try:
    retriever = store.as_retriever(
        search_type="similarity_score_threshold", search_kwargs=search_kwargs
    )
except ValidationError as e:
    if "score_threshold" in str(e):
        retriever = store.as_retriever(
            search_type="similarity_score_threshold",
            search_kwargs={**search_kwargs, "score_threshold": 0.5},
        )
    else:
        raise

When it happens

Trigger: `as_retriever(search_type='similarity_score_threshold')` with no `search_kwargs`, with `score_threshold` missing from `search_kwargs`, or with a non-float value such as the string `"0.5"` or the integer `1` (note `isinstance(1, float)` is `False`).

Common situations: Switching `search_type` to `'similarity_score_threshold'` but leaving `search_kwargs` empty; JSON/YAML config parsing thresholds as strings; passing `score_threshold=1` (int) instead of `1.0`; forgetting the key lives inside `search_kwargs`, not top-level.

Related errors

• search_type of {search_type} not allowed. Valid values are:
• maxsize must be greater than 0
• Could not resolve content_key {full_path!r}: expected a mapp
• Could not resolve content_key {full_path!r}: missing key {ke
• If multiple pydantic schemas are provided then args_only sho

AI-assisted analysis of langchain-ai/langchain@e32fa9a52e (2026-08-14). Data as JSON: /api/errors/a1a93d2f5e70d2fd. Report an issue: GitHub.