instancespace.data.model

Apply the shared MATLAB/Python closed-boundary TRACE membership policy.

Holds initial dataset from metadata and processed data after operations.

Retain full instance data before density-based subsampling.

Stores processed and raw features/performance, derived good/bad labels, winners, and instance metadata so SIFTED can re-filter instances using its selected features.

Contains preliminary output metrics calculated from the data.

Results of the sifting process in the data analysis pipeline.

Results of the Pilot process in the data analysis pipeline.

Results of the Cloister process in the data analysis pipeline.

Results of the Pythia process in the data analysis pipeline.

A class to represent a footprint with geometric and statistical properties.

Attributes

• polygon (Polygon | MultiPolygon | TetrahedralMesh | None): The geometric shape of the footprint.
• area (float): The area (2D) or volume (3D), retained under the legacy field name.
• elements (int): The number of data points within the footprint.
• good_elements (int): The number of "good" data points within the footprint.
• density (float): The density of points within the footprint.
• purity (float): The purity of "good" elements in relation to all elements in the footprint.
• dimension (int): The geometry coordinate count. Defaults to two for legacy constructors.

Results of the Trace process in the data analysis pipeline.

Results of applying trained model to test data.

This class encapsulates all outputs from the explore() pipeline, which applies a trained instance space model to new test data.

Attributes

• dataset_id (str): Identifier for this test dataset.
• timestamp (datetime): When the explore operation was performed.
• x (NDArray[np.double]): Processed features after PRELIM/SIFTED transformations. Shape: (n_instances, n_selected_features).
• z (NDArray[np.double]): Projected coordinates from PILOT. Shape: (n_instances, n_projection_dimensions), where the projection has two or three coordinates. TRACE3 supports full 3D exploration.
• y_hat (NDArray[np.bool_] | None): Binary algorithm predictions from PYTHIA SVMs. Shape: (n_instances, n_algorithms). None if PYTHIA not applied. n_algorithms includes a trailing column (always False) per test-set-only algorithm absent from training when ground truth is present - see algo_labels.
• pr0_hat (NDArray[np.double] | None): Probability predictions from PYTHIA SVMs. Shape: (n_instances, n_algorithms). None if PYTHIA not applied. Same new-algorithm-column convention as y_hat (trailing 0.0).
• selection0 (NDArray[np.int_] | None): Recommended algorithm index for each instance. Shape: (n_instances,). None if PYTHIA not applied. Never points at a new-algorithm column (see y_hat) - those are given zero selection precision, so they can never be the recommended algorithm.
• in_good (NDArray[np.bool_] | None): Whether each instance falls in "good" footprint for each algorithm. Shape: (n_instances, n_algorithms). None if TRACE not applied. Same new-algorithm-column convention as y_hat (trailing False - no trained footprint exists to test membership against).
• in_best (NDArray[np.bool_] | None): Whether each instance falls in "best" footprint for each algorithm. Shape: (n_instances, n_algorithms). None if TRACE not applied. Same new-algorithm-column convention as in_good.
• inst_labels (pd.Series): Instance labels/identifiers from the test metadata.
• y_actual (NDArray[np.bool_] | None): Ground-truth binary good/bad label per algorithm, derived from test_metadata's own algo_* performance columns using the trained PerformanceOptions. Shape: (n_instances, n_algorithms). None when the test set carries no algorithm performance columns at all (the feature-only case - existing callers see no change).
• y_best_actual (NDArray[np.double] | None): Best observed ground-truth performance per instance. Shape: (n_instances,). None under the same condition as y_actual.
• p_actual (NDArray[np.int_] | None): 1-based index of the best-performing algorithm per instance, ties broken with MATLAB-compatible seeded random selection. Shape: (n_instances,). None under the same condition as y_actual.
• beta_actual (NDArray[np.bool_] | None): Whether each instance clears the beta threshold's fraction of good algorithms. Shape: (n_instances,). None under the same condition as y_actual.
• accuracy_actual (NDArray[np.double] | None): Per-algorithm accuracy of PYTHIA's trained classifiers against y_actual. Shape: (n_algorithms,). NaN for an algorithm absent from the test set's ground truth, or a test-set-only algorithm with no trained classifier to score (both "no CV model", matching MATLAB's convention). None under the same condition as y_actual.
• precision_actual (NDArray[np.double] | None): Per-algorithm precision, same shape/NaN convention as accuracy_actual.
• recall_actual (NDArray[np.double] | None): Per-algorithm recall, same shape/NaN convention as accuracy_actual.
• cvcmat_actual (NDArray[np.double] | None): Per-algorithm confusion matrix [tn, fp, fn, tp] against y_actual. Shape: (n_algorithms, 4). NaN row under the same conditions as accuracy_actual. None under the same condition as y_actual.
• algo_labels (list[str] | None): The full algorithm order (trained algorithms, then any test-set-only algorithm absent from training, in first-encountered order) that y_hat/pr0_hat/in_good/in_best/accuracy_actual/etc.'s per-algorithm columns index into (including new test-set algorithms) - needed to interpret those columns once new algorithms widen them beyond Model.data.algo_labels' own length. None under the same condition as y_actual (when there's no ground truth, column order is implicitly Model.data.algo_labels and never widened).
• trace_out (TraceOut | None): Trained TRACE geometry rescored against this test set's ground truth. Geometry and area remain unchanged; evidence metrics are updated. It is None for feature-only explore data.

Holds indices for feature selection.