Skip to content

NumPy Adapter - Counting Helpers

Stateless counting helpers used by the NumPy adapter. They were extracted into this sibling module so that causaliq_data.numpy stays within the module size limits described in the development policy.

The adapter passes all state to these functions as arguments, so the NumPy.MAX_BINCOUNT class attribute remains the single place that selects the fast packing path over the slower NumPy unique path.

Combination Counting

combinations_from_binning 

combinations_from_binning(
    sample_cols: ndarray, num_vals: ndarray, n_reqd: int, minlength: int
) -> Tuple[ndarray, ndarray]

Count unique combinations of values by packing them into integers.

Parameters:

  • sample_cols

    (ndarray) –

    Sample values for the required columns.

  • num_vals

    (ndarray) –

    Number of values in each of those columns.

  • n_reqd

    (int) –

    Number of columns required.

  • minlength

    (int) –

    Maximum possible number of value combinations.

Returns:

  • Tuple[ndarray, ndarray] –

    Tuple of unique value combinations and their counts.

crosstab_counts 

crosstab_counts(
    combos: ndarray, combo_counts: ndarray, n_child_values: int
) -> Tuple[ndarray, ndarray]

Return a counts matrix of child values against parental combos.

Parameters:

  • combos

    (ndarray) –

    Unique value combinations, with child value first.

  • combo_counts

    (ndarray) –

    Count of each unique value combination.

  • n_child_values

    (int) –

    Number of distinct child values.

Returns:

  • Tuple[ndarray, ndarray] –

    Tuple of counts matrix and unique parental combinations.

Categorical Counts

categorical_value_counts 

categorical_value_counts(
    sample: ndarray,
    categories: ndarray,
    nodes: Tuple[str, ...],
    orig_to_ext: Dict[str, str],
) -> Dict[str, Dict[str, int]]

Return value counts for each categorical column of a sample.

Parameters:

  • sample

    (ndarray) –

    Sample data array.

  • categories

    (ndarray) –

    Category values for each column.

  • nodes

    (Tuple[str, ...]) –

    Internal (i.e. original) node names.

  • orig_to_ext

    (Dict[str, str]) –

    Map from original to external names.

Returns:

  • Dict[str, Dict[str, int]] –

    Value counts for each categorical node.