Inspecting hydrogen inventory#
Separating indexed H atoms from virtual hydrogen annotations.
molsysmt.physchem.get_hydrogen_inventory() reads a selected chemical state
from any supported form with the necessary assignments. It never changes the
source, fills missing counts, predicts protonation or requires RDKit.
API documentation
Added in version 1.0.0.
For each selected atom, indexed_hydrogen_counts counts its covalent indexed H
neighbors. missing_hydrogen_counts is the sum of stored n_implicit_hydrogens
and n_explicit_hydrogens: both are virtual annotations, not indexed atoms.
total_hydrogen_counts adds the two kinds. Unknown stored counts and unknown
totals are -1, not zero. All counts are dimensionless int64 arrays aligned to
atom_indices. parent_hydrogen_pairs has integer shape (n_pairs, 2), parent
first, and retains only pairs whose parent is selected. A selected indexed H is
listed separately in explicit_hydrogen_atom_indices.
The full graph is inspected before output selection. Integer atom selections
are deduplicated and returned in sorted order; empty results have defined types
and shapes. chemical_state='structure' resolves the single selected frame’s
association. Structure indices are zero-based indices, not structure IDs.
The detached molsysmt.hydrogen_inventory@1 report can be available,
unassessed, conflict or empty. Read its indexed issues and remaining
unassessed_checks. Available certifies that this inventory is readable;
valence, protonation and coordinate quality are not independently validated.
A graph with existing H but missing virtual counts remains unassessed rather than
being declared complete. Do not use this inventory alone as a hydrogen-addition
or chemical-readiness certification.
See also
Getting chemical readiness — inspecting additional stored fields.
Adding hydrogens to a fixed chemical state — materializing a validated fixed inventory.