pdstools.explanations.Schema
============================

.. py:module:: pdstools.explanations.Schema

.. autoapi-nested-parse::

   Expected schema of the pre-aggregated explanations parquet files.



Attributes
----------

.. autoapisummary::

   pdstools.explanations.Schema.AGGREGATE_SCHEMA


Functions
---------

.. autoapisummary::

   pdstools.explanations.Schema.apply_schema


Module Contents
---------------

.. py:data:: AGGREGATE_SCHEMA
   :type:  dict[str, polars.DataType]

   Columns read from ``OVERVIEW.parquet`` / ``BY_CONTEXT.parquet``, and the
   dtype each is cast to. Casting on read means every downstream aggregation
   sees the same types regardless of how the exporting Pega version wrote the
   file.

.. py:function:: apply_schema(lf: polars.LazyFrame) -> polars.LazyFrame

   Select the expected columns from *lf* and cast them to their dtypes.

   :param lf: Scan over one of the aggregated parquet files.
   :type lf: pl.LazyFrame

   :returns: *lf* narrowed to :data:`AGGREGATE_SCHEMA` and cast to its dtypes.
   :rtype: pl.LazyFrame

   :raises ValueError: If *lf* is missing any of the expected columns.