pdstools.adm.Aggregates¶

Classes¶

Aggregates

Module Contents¶

class Aggregates(datamart: pdstools.adm.ADMDatamart.ADMDatamart)¶

Parameters:: datamart (pdstools.adm.ADMDatamart.ADMDatamart)

datamart¶

cdh_guidelines¶

last(*, data: polars.LazyFrame | None = None, table: Literal['model_data', 'predictor_data', 'combined_data'] = 'model_data')¶

Gets the last snapshot of the given table

Parameters:

data (Optional[pl.LazyFrame], optional) – If provided, subsets to just that dataframe, by default None
table (Literal['model_data', 'predictor_data', 'combined_data'], optional) – If provided, specifies the table to get data from, by default “model_data”

Returns:

_description_

Return type:

_type_

_combine_data(model_df: polars.LazyFrame | None, predictor_df: polars.LazyFrame | None) → polars.LazyFrame | None¶

Combines the model and predictor tables to the combined_data attribute

Parameters:

model_df (pl.LazyFrame) – The model snapshots table
predictor_df (pl.LazyFrame) – The predictor binning snapshots table

Returns:

The resulting data, joined on the ModelID column

Return type:

pl.LazyFrame

predictor_performance_pivot(*, query: pdstools.utils.types.QUERY | None = None, active_only: bool = False, by='Name', top_predictors: int | None = None, top_groups: int | None = None) → polars.LazyFrame¶

Creates a pivot table of the predictor performance per ‘group’

Parameters:

query (Optional[QUERY], optional) – A query to apply to the data before creating the pivot, by default None
by (str, optional) – A group by which to ‘facet’, by default “Name”. If, for instance, the ‘by’ argument is set to ‘Configuration’, each row will be a distinct configuration
top_predictors (Optional[int], optional) – Specify the maximum number of predictors, by default None
top_groups (Optional[int], optional) – Specify the maximum number of ‘groups’ specified in the ‘by’ argument, by default None
active_only (bool)

Returns:

A LazyFrame with a column for each predictor, and a row for each ‘group’. The values represent the weighted performance for that predictor

Return type:

pl.LazyFrame

model_summary(by: str = 'Name', query: pdstools.utils.types.QUERY | None = None) → polars.LazyFrame¶

Generate a summary of statistic for each model (based on model ID)

If you want to generate statistics at a model name or treatment level, specify this in the ‘by’ column.

Parameters:

by (str, optional) – The column to define the ‘counts’ for, by default “ModelID” Must be part of the context keys in the ADMDatamart class
query (Optional[QUERY], optional) – A query to apply to the data before summarization, by default None

Returns:

A LazyFrame, with one row for each context key combination

Return type:

pl.LazyFrame

predictor_counts(*, facet: str = 'Configuration', by: str = 'Type', query: pdstools.utils.types.QUERY | None = None) → polars.LazyFrame¶

Returns the count of each predictor grouped by a certain column

Parameters:

by (str, optional) – The column to group the data by, by default “Type”
query (Optional[QUERY], optional) – A query to apply to the data, by default None
facet (str)

Returns:

A LazyFrame, with one row per predictor and ‘by’ combo

Return type:

pl.LazyFrame

static _top_n(df: polars.DataFrame, top_n: int, metric: str = 'PredictorPerformance', facets: list | None = None)¶

Subsets DataFrame to contain only top_n predictors.

Parameters:

df (pl.DataFrame) – Table to subset
top_n (int) – Number of top predictors
metric (str) – Metric to use for comparing predictors
facets (list) – Subsets top_n predictors over facets. Seperate top predictors for each facet

Returns:

Subsetted dataframe

Return type:

pl.DataFrame

summary_by_channel(custom_channels: Dict[str, str] | None = None, by_period: str | None = None, keep_lists: bool = False) → polars.LazyFrame¶

Summarize ADM models per channel

Parameters:

custom_channels (Dict[str, str], optional) – Optional list with custom channel/direction name mappings. Defaults to None.
by_period (str, optional) – Optional grouping by time period. Format string as in polars.Expr.dt.truncate (https://docs.pola.rs/api/python/stable/reference/expressions/api/polars.Expr.dt.truncate.html), for example “1mo”, “1w”, “1d” for calendar month, week day. If provided, creates a new Period column with the truncated date/time. Defaults to None.
keep_lists (bool, optional) – Internal flag to keep some columns (action and treatment names etc) as full lists.

Returns:

Dataframe with summary per channel (and optionally a period)

Return type:

pl.LazyFrame

summary_by_configuration() → polars.DataFrame¶

Generates a summary of the ADM model configurations.

Returns:: A Polars DataFrame containing the configuration summary.
Return type:: pl.DataFrame

predictors_overview() → polars.DataFrame | None¶

Generate a summary of the last snapshot of predictor data.

This method creates a summary of predictor data by joining the last snapshots of predictor_data and model_data, then performing various aggregations and calculations. It excludes the “Classifier” predictor from the analysis.

Returns:: A Polars DataFrame containing the predictor summary if successful, None if the required data is not available.
Return type:: pl.DataFrame or None

overall_summary(custom_channels: Dict[str, str] = None, by_period: str = None) → polars.LazyFrame¶

Overall ADM models summary. Only valid data is included.

Parameters:

custom_channels (Dict[str, str], optional) – Optional list with custom channel/direction name mappings. Defaults to None.
by_period (str, optional) – Optional grouping by time period. Format string as in polars.Expr.dt.truncate (https://docs.pola.rs/api/python/stable/reference/expressions/api/polars.Expr.dt.truncate.html), for example “1mo”, “1w”, “1d” for calendar month, week day. If provided, creates a new Period column with the truncated date/time. Defaults to None.

Returns:

Summary across all valid ADM models as a dataframe

Return type:

pl.LazyFrame