API (Pipe)

AAanalysis exposes the same analysis through two interfaces:

import aaanalysis as aa            # explicit interface (the building blocks)
import aaanalysis.pipe as aap      # implicit interface (the golden pipelines)

The building blocks (aa) are the explicit objects and functions you compose for full control over every step. The golden pipelines (aap) chain the standard load to CPP to model to explain to plot workflow into a single call, much as pyplot sits over Matplotlib’s Axes and Figure. They are stateless wrappers whose defaults match the explicit path, and they live in their own module under a separate alias (import aaanalysis.pipe as aap). Reach for aa when you want control over each step, and aap when you want a sensible default workflow in one call.

Each pipeline below documents its inputs, outputs, and the building blocks it composes:

obtain_samples(df_seq[, pos_col, strategy, ...])	Obtain a balanced training set from a described sampling situation in one call.
find_features(labels, df_seq[, search, ...])	Identify discriminating features in one call via a staged, interpretable CPP AutoML search.
predict_samples(list_df_feat, df_seq, labels)	Train and compare predictors across feature sets and models in one call.
plot_eval(df_eval[, metric, score_col, figsize])	Decompose a find_features() sweep into a set of publication-ready evaluation figures.
explain_features(df_feat, df_seq, labels[, ...])	Explain a feature set in one call: compute per-sample SHAP impact and draw the SHAP feature map.