v0.2.2

What's new

Added

• ValidationResult.referenced_columns — populated on successful spec validation; a dict[str, list[str]] mapping each spec field to the unqualified column names it references. None when the spec is invalid.

  Enables downstream consumers who hold a warehouse connection to verify that every column referenced in a spec is present in the source table before calling compute().

  result = metric_spec.validate()
  if result.valid:
      print(result.referenced_columns)
      # {'numerator': ['revenue'], 'denominator': ['impressions'], 'timestamp_col': ['created_at']}

  Fields extracted per spec type:

  • Metric: "numerator", "denominator" (if set), "timestamp_col", "entities" (if set)
  • Slice leaf: "values[i].where" (one key per value)
  • Slice wildcard: "where"
  • Slice composite: empty dict {}

See the changelog and column introspection docs for details.