Hi, is it possible to define input hydration functions without attaching them to a custom type ? For example, when a "csv" parameter in the yaml file is specified

@input_hydration_config(Selector({"csv": Field(str)}))
def csv_input_hydration_config(context, selector) -> DataFrame:
    return read_csv(selector["csv"])

config.yaml

solids:
  my_solid:
    inputs:
      csv: 
        value: ...path_to_csv

without attaching them to a custom type

Nope, you’ll need to attach them to a custom type.

thanks alex

a heads up - we’re about to push out 0.7.0 which changes up a lot of the dagster type api surface. Also here is a link to the docs against master as a preview of whats to come - linked to a section on DataFrames you may find relevant https://dagster.readthedocs.io/en/latest/sections/learn/guides/dagster_pandas/dagster_pandas.html

exciting ! thanks for the heads up !