Documentation

Here goes a high-level description of the code structure.

Core modules

Descriptors (stateinterpreter.descriptors)

compute_descriptors(traj[, descriptors])	Compute descriptors from trajectory: - Dihedral angles - CA distances - Hydrogen bonds distances - Hydrogen bonds contacts - Disulfide bonds dihedrals
load_descriptors(descriptors[, start, stop, ...])

Metastable (stateinterpreter.metastable)

identify_metastable_states(colvar, ...[, ...])	Label configurations based on free energy
approximate_FES(colvar, bandwidth[, ...])	Approximate Free Energy Surface (FES) in the space of selected_cvs through Gaussian Kernel Density Estimation

ML (stateinterpreter.ml)

prepare_training_dataset(descriptors, ...[, ...])	Sample points from trajectory
Classifier(dataset, features_names[, ...])

Utilities

Input/Output (stateinterpreter.utils.io)

load_dataframe(data[, start, stop, stride])	Load dataframe from object or from file.
load_trajectory(traj_dict[, start, stop, stride])	Load trajectory with mdtraj.

Visualize (stateinterpreter.utils.visualize)

visualize_features(trajectory, ...[, state, ...])	Visualize snapshots of each state highlighting the relevant features for a given state.
compute_residue_score(classifier, reg, ...)	Compute a residue score by aggregating all the features relevances by residues.
visualize_residue_score(trajectory, ...[, ...])	Visualize snapshots of each state coloring the residues with the score per each state.

Plot (stateinterpreter.utils.plot)

plot_states(colvar, state_labels, selected_cvs)
plot_regularization_path(classifier, reg)
plot_classifier_complexity_vs_accuracy(...)
plot_combination_states_features(colvar, ...)
plot_states_features(cv_x, cv_y, ...[, ...])
plot_histogram_features(descriptors, ...[, ...])