Is it possible to access the input column names in onnx converter functions?

[paranjapeved15]

In the below code, I am appending the vehicleType to the prefix to get which column name to use from my input Dataframe. So for example if vehicleType = 'car' then I would return 'features_car' feature value.
Data:
df = pandas.DataFrame([["car",0.1,0.0],["car",0.2,0.0],["suv",0.0,0.2]],columns=['vehicleType','features_car','features_suv'])
Custom Transformer:

class GetScore(BaseEstimator, TransformerMixin):  # type: ignore
    """Apply binarize transform for matching values to filter_value."""

    def __init__(self, prefix: str):
        """Initialize transformer with expected columns."""
        self.prefix = prefix
        pass

    def dot_product(self, x) -> float:
        """Return 1.0 if input == filter_value, else 0."""
        print("type of x:")
        print(type(x))
        return x[self.prefix+x.vehicleType]


    def fit(self, X, y=None):  # type: ignore
        """Fit the transformer."""
        return self

    def transform(self, X: pandas.DataFrame | numpy.ndarray, y: None = None) -> numpy.ndarray:
        """Transform the given data."""
        if type(X) == pandas.DataFrame:
            x = X.apply(lambda x: self.dot_product(x), axis=1)
            return x.values.reshape((-1, 1))
        # elif type(X) == numpy.ndarray:
        #     vector_func = numpy.vectorize(self.dot_product)
        #     x = vector_func(X)
        #     return x.reshape((-1, 1))

    def get_feature_names_out(self) -> None:
        """Return feature names. Required for onnx conversion."""
        pass

sklearn pipeline:

preprocessor = ColumnTransformer(
        transformers=[
            #("",make_pipeline(OneHotEncoder(categories=[["car", "suv"]], sparse_output=False)), ['vehicleType','features_car','features_suv']),
            ("features_computed",GetScore("features_"), ['vehicleType','features_car','features_suv']),
            ],
    #remainder="passthrough",
    verbose_feature_names_out=False,
)

To write a custom converter for my GetScore, I would need to be able to access the input by the column name. Is that accessible in the converter inputs? Or would I have to come up with another approach?

[xadupre]

The library was started before scikit-learn implemented this feature and this information is not used right now. We could probably make a code change to improve the mapping between onnx name and scikit-learn name. I'm not sure it is possible to guarantee an exact mapping in particular if scikit-learn allows name to be reused but it is not a very complicated changed. Do you think it should be the exact name, or extended with a prefix or suffix?