23032: Adds new parameter to react methods, "feature_post_process_code_map". MINOR (#370)

Adds new parameter, "feature_post_process_code_map" to all react*
method. This parameter takes a mapping feature name to custom code
string that is used to derive an updated feature value based on whatever
code is defined. At the time of execution, the code will have access to
all context features and previously generated action features (including
the feature itself).

Co-authored-by: Andrew Bassett <[email protected]>

Author: cademack

howso/client/base.py

@@ -1463,6 +1463,7 @@ def react(  # noqa: C901
         details: t.Optional[Mapping] = None,
         exclude_novel_nominals_from_uniqueness_check: bool = False,
         feature_bounds_map: t.Optional[Mapping] = None,
+        feature_post_process_code_map: t.Optional[Mapping] = None,
         generate_new_cases: GenerateNewCases = "no",
         goal_features_map: t.Optional[Mapping] = None,
         initial_batch_size: t.Optional[int] = None,
@@ -1932,6 +1933,15 @@ def react(  # noqa: C901
                     "feature_c": {"max": 1}
                 }
 
+        feature_post_process_code_map : dict of str, optional
+            A mapping of feature name to custom code strings that will be
+            evaluated to update the value of the feature they are mapped from.
+            The custom code is evaluated just after a feature value is predicted
+            or synthesized to update the value of the feature, meaning that the
+            resulting value will be used as part of the context for following
+            action features. The custom code will have access to all context
+            feature values and previously generated action feature values.
+
         generate_new_cases : {"always", "attempt", "no"}, default "no"
             (Optional) Whether to generate new cases.
 
@@ -2158,6 +2168,7 @@ def react(  # noqa: C901
                 "action_features": action_features,
                 "derived_context_features": derived_context_features,
                 "derived_action_features": derived_action_features,
+                "feature_post_process_code_map": feature_post_process_code_map,
                 "goal_features_map": goal_features_map,
                 "post_process_features": post_process_features,
                 "post_process_values": post_process_values,
@@ -2199,6 +2210,7 @@ def react(  # noqa: C901
                 "action_features": action_features,
                 "derived_context_features": derived_context_features,
                 "derived_action_features": derived_action_features,
+                "feature_post_process_code_map": feature_post_process_code_map,
                 "post_process_features": post_process_features,
                 "post_process_values": post_process_values,
                 "use_regional_residuals": use_regional_residuals,
@@ -2595,6 +2607,7 @@ def react_series(  # noqa: C901
         details: t.Optional[Mapping] = None,
         exclude_novel_nominals_from_uniqueness_check: bool = False,
         feature_bounds_map: t.Optional[Mapping[str, Mapping[str, t.Any]]] = None,
+        feature_post_process_code_map: t.Optional[Mapping] = None,
         final_time_steps: t.Optional[list[t.Any]] = None,
         generate_new_cases: GenerateNewCases = "no",
         goal_features_map: t.Optional[Mapping] = None,
@@ -2695,6 +2708,16 @@ def react_series(  # noqa: C901
             If True, will exclude features which have a subtype defined in their feature
             attributes from the uniqueness check that happens when ``generate_new_cases``
             is True. Only applies to generative reacts.
+        feature_post_process_code_map : dict of str, optional
+            A mapping of feature name to custom code strings that will be
+            evaluated to update the value of the feature they are mapped from.
+            The custom code is evaluated just after a feature value is predicted
+            or synthesized to update the value of the feature, meaning that the
+            resulting value will be used as part of the context for following
+            action features. The custom code will have access to all context
+            feature values and previously generated action feature values of
+            the timestep being generated, as well as the feature values of all
+            previously generated timesteps.
         series_context_features : iterable of str, optional
             List of context features corresponding to ``series_context_values``.
         series_context_values : list of list of list of object or list of DataFrame, optional
@@ -2864,6 +2887,7 @@ def react_series(  # noqa: C901
             react_params = {
                 "action_features": action_features,
                 "continue_series": continue_series,
+                "feature_post_process_code_map": feature_post_process_code_map,
                 "final_time_steps": final_time_steps,
                 "init_time_steps": init_time_steps,
                 "series_stop_maps": series_stop_maps,
@@ -2913,6 +2937,7 @@ def react_series(  # noqa: C901
                 "num_series_to_generate": num_series_to_generate,
                 "action_features": action_features,
                 "continue_series": continue_series,
+                "feature_post_process_code_map": feature_post_process_code_map,
                 "final_time_steps": final_time_steps,
                 "init_time_steps": init_time_steps,
                 "series_stop_maps": series_stop_maps,

howso/engine/trainee.py

@@ -1179,6 +1179,7 @@ def react(
         details: t.Optional[Mapping[str, t.Any]] = None,
         exclude_novel_nominals_from_uniqueness_check: bool = False,
         feature_bounds_map: t.Optional[Mapping[str, Mapping[str, t.Any]]] = None,
+        feature_post_process_code_map: t.Optional[Mapping] = None,
         generate_new_cases: GenerateNewCases = "no",
         goal_features_map: t.Optional[Mapping] = None,
         initial_batch_size: t.Optional[int] = None,
@@ -1599,6 +1600,15 @@ def react(
                     "feature_c": {"max": 1}
                 }
 
+        feature_post_process_code_map : dict of str, optional
+            A mapping of feature name to custom code strings that will be
+            evaluated to update the value of the feature they are mapped from.
+            The custom code is evaluated just after a feature value is predicted
+            or synthesized to update the value of the feature, meaning that the
+            resulting value will be used as part of the context for following
+            action features. The custom code will have access to all context
+            feature values and previously generated action feature values.
+
         generate_new_cases : {"always", "attempt", "no"}, default "no"
             This parameter takes in a string that may be one of the following:
 
@@ -1721,6 +1731,7 @@ def react(
             details=details,
             exclude_novel_nominals_from_uniqueness_check=exclude_novel_nominals_from_uniqueness_check,
             feature_bounds_map=feature_bounds_map,
+            feature_post_process_code_map=feature_post_process_code_map,
             generate_new_cases=generate_new_cases,
             goal_features_map=goal_features_map,
             initial_batch_size=initial_batch_size,
@@ -1753,6 +1764,7 @@ def react_series(
         details: t.Optional[Mapping[str, t.Any]] = None,
         exclude_novel_nominals_from_uniqueness_check: bool = False,
         feature_bounds_map: t.Optional[Mapping[str, Mapping[str, t.Any]]] = None,
+        feature_post_process_code_map: t.Optional[Mapping] = None,
         final_time_steps: t.Optional[list[t.Any]] = None,
         generate_new_cases: GenerateNewCases = "no",
         goal_features_map: t.Optional[Mapping] = None,
@@ -1820,6 +1832,16 @@ def react_series(
             is True. Only applies to generative reacts.
         feature_bounds_map : map of str -> map of str -> object, optional
             See parameter ``feature_bounds_map`` in :meth:`react`.
+        feature_post_process_code_map : dict of str, optional
+            A mapping of feature name to custom code strings that will be
+            evaluated to update the value of the feature they are mapped from.
+            The custom code is evaluated just after a feature value is predicted
+            or synthesized to update the value of the feature, meaning that the
+            resulting value will be used as part of the context for following
+            action features. The custom code will have access to all context
+            feature values and previously generated action feature values of
+            the timestep being generated, as well as the feature values of all
+            previously generated timesteps.
         final_time_steps: list of object, optional
             The time steps at which to end synthesis. Time-series only.
             Time-series only. Must provide either one for all series, or
@@ -1938,6 +1960,7 @@ def react_series(
                 details=details,
                 exclude_novel_nominals_from_uniqueness_check=exclude_novel_nominals_from_uniqueness_check,
                 feature_bounds_map=feature_bounds_map,
+                feature_post_process_code_map=feature_post_process_code_map,
                 final_time_steps=final_time_steps,
                 generate_new_cases=generate_new_cases,
                 goal_features_map=goal_features_map,