Module candidates

Structs§

AssignmentResult
CandidateOpts: Options for the simple candidate enumeration. Rule = RT overlap (seconds) AND group eligibility (mz ∩ isolation AND scans ∩ program).
FragmentGroupIndex
FragmentIndex
FragmentMetadata: Pre-computed per-MS2 metadata for fast filtering and lookup.
FragmentMetadataEntry: A single entry from FragmentMetadata.
FragmentQueryOpts: Axis-aligned query knobs for fragment candidates.
FragmentStorage: Storage for MS2 clusters with ID-based lookup.
PairFeatures: Compact feature bundle per pair for traceability.
PseudoBuildResult: Full result of the DIA → pseudo-DDA pipeline.
ScoreOpts: Scoring knobs. Defaults are conservative and width-aware but won’t punish pairs that lack good fits (we use shape_neutral when shape data is missing).
ThinPrecursor

best_ms1_for_each_ms2_any: Switch between geometric and XIC assignment.
best_ms1_for_each_ms2_geom: Geometric scoring (width / Jaccard / apex / overlap etc.).
best_ms1_for_each_ms2_xic: XIC-based assignment: score via RT/IM traces + intensity ratio, choose best MS1 per MS2, return Vec<Option<ms1_idx>> indexed by ms2_idx.
build_pseudo_spectra_all_pairs: NON-COMPETITIVE builder: mainly for debugging / exploration.
build_pseudo_spectra_end_to_end: End-to-end, geometric competitive builder: DIA index + MS1 clusters + MS2 clusters (+ optional features) → candidates → geometric scoring/assignment → pseudo-MS/MS spectra.
build_pseudo_spectra_end_to_end_xic: End-to-end, XIC-based competitive builder: DIA index + MS1 clusters + MS2 clusters (+ optional features) → candidates → XIC scoring/assignment → pseudo-MS/MS spectra.
fragment_from_cluster: Convert an MS2 ClusterResult1D into a PseudoFragment. Return None if the cluster is unusable (no m/z).