Constants (consts.py)

This module defines various string constants used throughout the ABACUS library, particularly in configuration files, data structures, and function arguments. Using these constants helps ensure consistency and avoids typos.

Dictionary Keys

  • KEY_GRANULARITY = "granularity": Key used in configuration or data dictionaries to specify time granularity.

  • KEY_DATE_STRS = "date_strs": Key for date string arrays/lists.

  • KEY_METRICS = "metrics": Key related to metric data.

  • KEY_OBSERVATIONS = "observations": Key related to observation data.

Time Granularity

Constants representing the time granularity of the input data.

  • GRANULARITY_DAILY = "daily": Indicates daily time series data.

  • GRANULARITY_WEEKLY = "weekly": Indicates weekly time series data.

Model/Transformation Names (Legacy/Reference)

These constants might refer to specific model types or transformations, potentially from earlier versions or related libraries.

  • FIT_abacus_MMM_MODELNAME_HILL_ADSTOCK = "hill_adstock"

  • FIT_abacus_MMM_MODELNAME_ADSTOCK = "adstock"

  • FIT_abacus_MMM_MODELNAME_CARRYOVER = "carryover"

Outlier Removal Types

Constants specifying methods for handling outliers, likely used in preprocessing functions (e.g., abacus.prepro.outliers).

  • REMOVE_OUTLIERS_TYPE_REPLACE_WITH_TRIMMED_MEAN = "replace_with_trimmed_mean": Replace outliers with the trimmed mean of the non-outlier values.

  • REMOVE_OUTLIERS_TYPE_REPLACE_WITH_P10_VALUE = "replace_with_p10_value": Replace outliers with the 10th percentile value.

DataFrame Column Suffixes

Standard suffixes used when creating column names in pandas DataFrames generated by ABACUS functions (e.g., in DataToFit.to_data_frame).

  • DATA_FRAME_COST_SUFFIX = "cost": Suffix for columns representing cost data.

  • DATA_FRAME_IMPRESSIONS_SUFFIX = "impressions": Suffix for columns representing impression or spend data (often used interchangeably in MMM contexts).