coot-coord-utils.cc — functions needing tests

18 coot::util functions defined in coot-utils/coot-coord-utils.cc that have no coverage in either test-coord-utils.cc (C++ unit tests) or python-tests/ (integration tests via the coot.* scripting API). (Was 19; get_residue_alt_confs removed — it’s #if 0‘d here and actually lives in geometry/mol-utils.cc. See row 16.)

Of the 18, 3 are BLOCKED on gemmi gaining graph matching (#1–#3), so the testable-now set is 15 (#4–#15, #17–#19).

CALLER-SCOPE PRIORITISATION (2026-06-20)

Only functions reached from the libcootapi/headless build (api/coot-utils/ideal/ligand/…) are in the current migration scope. Functions called ONLY from src/ (the GUI layer) are NOT compiled into coot_headless_api/libcootapi, so defer them (like #15 chain_atoms_segid, now being removed). Functions with NO external callers are internal/comparators or near-dead — investigate before any work. (current-state/ is a duplicate snapshot tree — ignored in this analysis.)

In scope — DO THESE: #5 close_residues (api, ideal), #12 chiral_4th_atom (api), #17 shift_to_origin (coot-utils, ligand), #18 translate_close_to_origin (coot-utils), #19 position_residue_by_internal_coordinates (coot-utils). [#13 nucleotide_is_DNA already DONE.]
src-only — DEFER: #7 mol_by_symmetry, #9 set_mol_cell, #10 rotate_atom_about, #11 standardize_peptide_C_N_distances.
No external callers (investigate / maybe dead): #4 compare_residues (comparator), #6 residues_near_residues_for_residues, #8 symmetry_move_atoms, #14 chains_in_atom_selection.

LESSON (user, 2026-06-20): check caller scope BEFORE picking a function to migrate/test — don’t prioritise src-only functions. chain_atoms_segid (#15) was such a trap (1 caller, src NCS code).

Excluded from this list: functions that shouldn’t be unit-tested directly (comparators, ctors), molecule copy/link plumbing that’s indirectly exercised, and the four already covered transitively by python-tests (cis_trans_conversion, cis_trans_convert, move_waters_around_protein, nucleotide_to_nucleotide).

These are all mmdb-typed and on the path to gemmi migration ([[project_mmdb_to_gemmi]]), so prefer writing them as mmdb-vs-gemmi differential tests (run both backends on the same structure, compare outputs) rather than hand-rolled expected values — see test-coord-utils-gemmi.cc for the pattern.

Header = coot-coord-utils.hh, def = coot-coord-utils.cc.

#	function	header decl	.cc def	notes
1	`graph_match`	`graph_match_info_t graph_match(mmdb::Residue res_moving, mmdb::Residue res_reference, bool apply_rtop_flag, ...)`	2023	BLOCKED — no gemmi equivalent. gemmi graph matching to be added separately/later ([[project_gemmi_graph_matching]]); decided 2026-06-15. Defer test until then.
2	`match_names`	`void match_names(mmdb::Residue *res_moving_names)`	2207	BLOCKED — graph-match family (see #1)
3	`matching_atoms`	`std::vector<mmdb::Atom > matching_atoms(mmdb::Residue residue)`	1686	BLOCKED — graph-match family (see #1)
4	`compare_residues`	`bool compare_residues(const std::pair<mmdb::Residue *, int> &a, ...)`	5432	predicate; borderline (used as comparator)
5	`close_residues`	`... close_residues(mmdb::Manager mol1, mmdb::Manager mol2, float dist)`	7319	spatial query across two mols
6	`residues_near_residues_for_residues`	`std::map<mmdb::Residue, std::set<mmdb::Residue>> residues_near_residues_for_residues(const std::map<...> &all_molecule_map, ...)`	601	neighbour map
7	`mol_by_symmetry`	`mmdb::Manager mol_by_symmetry(mmdb::Manager mol, ...)`	7119	returns a new molecule
8	`symmetry_move_atoms`	`... symmetry_move_atoms(const std::vector<clipper::Coord_orth> &protein_coords, ...)`	7654	clipper-typed in/out
9	`set_mol_cell`	`bool set_mol_cell(mmdb::Manager *mol, clipper::Cell cell)`	7791	mutates cell
10	`rotate_atom_about`	`void rotate_atom_about(const clipper::Coord_orth &direction, ...)`	6258	geometry edit
11	`standardize_peptide_C_N_distances`	`void standardize_peptide_C_N_distances(const std::vector<std::pair<mmdb::Atom, mmdb::Atom>> &C_N_pairs)`	6277	geometry edit
12	`chiral_4th_atom`	`mmdb::Atom chiral_4th_atom(mmdb::Residue residue_p, mmdb::Atom *at_centre, ...)`	8557	returns atom
13	`nucleotide_is_DNA`	`bool nucleotide_is_DNA(mmdb::Residue *r)`	5122	DONE — gemmi twin in coot-coord-utils-gemmi.{hh,cc} (unpadded “O2’”/”O2*”); test `test_nucleotide_is_DNA` PASS
14	`chains_in_atom_selection`	`std::vector<mmdb::Chain > chains_in_atom_selection(mmdb::Manager mol, int model_number, const std::string &atom_selection)`	1993	selection query
15	~~`chain_atoms_segid`~~	—	1105	REMOVED from scope. src-only (1 caller: src/molecule-class-info-ncs.cc:1423). User is deleting it and replacing with `chain_residues_segid()` (does not exist yet). My gemmi twin + test were reverted. The inconsistent-segID case never occurs in real data.
16	~~`get_residue_alt_confs`~~	—	—	NOT a coot-coord-utils.cc function. Parser false positive: the def at .cc:3191 is inside `#if 0` (“we already have one of these in geometry”). The LIVE function is `geometry/mol-utils.cc:31` (decl mol-utils.hh:36), used 12× in reduce.cc. If it needs a test, that belongs with mol-utils, not here. REMOVED from this list.
17	`shift_to_origin`	`clipper::Coord_frac shift_to_origin(mmdb::Manager *mol)`	7661/7917	mmdb overload here; the clipper overload was moved to coot-coord-utils-mmdb-free.cc
18	`translate_close_to_origin`	`void translate_close_to_origin(mmdb::Manager *mol)`	(mmdb overload)	clipper overload moved to coot-coord-utils-mmdb-free.cc
19	`position_residue_by_internal_coordinates`	`class position_residue_by_internal_coordinates { ... }` (ctor + `move_moving_residue()`)	8447/8481	a class, not a free fn; test via construct + move_moving_residue()

Suggested ordering (easy/scalar first)

13 nucleotide_is_DNA, 15 chain_atoms_segid, 14 chains_in_atom_selection, 17 shift_to_origin, 18 translate_close_to_origin → then the geometry edits (10, 11), spatial queries (5, 6), symmetry (7, 8, 9), match family (1, 2, 3, 4), 12 chiral_4th_atom, 19 position_residue_by_internal_coordinates. Resolve 16 get_residue_alt_confs (dead) separately.