API Reference¶

The public API exposes four groups of functions: the core workflow, system-function builders, the reference-state search helpers used internally by run_ref_extraction_by_mcs(), and the active-acquisition rule that chooses which unknown sample it resolves next.

Core workflow¶

`rsr.rsr.run_ref_extraction_by_mcs`	Extract boundary reference states via Monte Carlo search.
`rsr.rsr.get_comp_cond_sys_prob`	P(system state \| given component states).
`rsr.rsr.get_comp_cond_sys_prob_multi`	Estimate P(system state = s \| given component states) for multi-state systems by Monte Carlo.

System-function builders¶

`rsr.igraph_sfun.make_igraph_sfun_global_conn`	Create an igraph-accelerated sfun for global connectivity.
`rsr.igraph_sfun.make_igraph_sfun_conn`	Create an igraph-accelerated sfun for 1OD connectivity.
`rsr.igraph_sfun.eval_global_conn_igraph`	igraph-accelerated version of eval_global_conn_k.
`rsr.igraph_sfun.eval_1od_connectivity_igraph`	igraph-accelerated version of eval_1od_connectivity.
`rsr.igraph_sfun.nx_to_igraph`	Convert a NetworkX graph to igraph, preserving node names and edge IDs.

Reference-state search¶

`rsr.rsr.minimise_upper_states_random`	Random greedy reduction of component states (binary-search variant).
`rsr.rsr.minimise_lower_states_random`	Random greedy reduction of component states (binary-search variant).
`rsr.rsr.update_refs`	Insert a new reference into the store, removing any it dominates.
`rsr.rsr.sample_new_comp_st_to_test`	Search for a component-state vector not yet covered by `refs_mat`.
`rsr.rsr.classify_samples_with_indices`	Classify samples as upper, lower, or unknown using subset checks, and return indices for each class.

Active acquisition¶

Deciding which unknown sample to resolve next, used by run_ref_extraction_by_mcs() when active_ref_search is enabled.

`rsr.rsr.select_refs_by_acquisition`	Pick the unresolved candidates with the highest acquisition score.
`rsr.rsr.compute_deficits`	Smallest componentwise violation of each candidate against a reference set.
`rsr.rsr.acquisition_score`	Acquisition score `(d+ + d-) - gamma * \|d+ - d-\|`.
`rsr.rsr.refs_mat_to_states`	Convert a binary reference tensor to integer component-state coordinates.
`rsr.rsr.validate_ref_consistency`	Raise if any lower reference state dominates an upper reference state.