config

tit.analyzer.config

Configuration dataclass for the analyzer runner.

Pure Python -- no SimNIBS, numpy, nibabel, or other heavy dependencies. Mirrors the tit.opt.config / tit.sim.config pattern: a thin, typed wrapper around the dict keys :mod:tit.analyzer.__main__ already reads (both its "single" and "group" dispatch branches), so :mod:tit.config_io can generate a JSON Schema for it and the desktop UI can validate a form against that schema before submitting the job.

This module does not change tit.analyzer.__main__ in any way -- it is a read of what that entry point already accepts (see its docstring and the _run_single / _run_group dict lookups), not a new contract the runner must adopt. See AnalyzerConfig's "Notes" for the two runner behaviors this dataclass cannot yet express.

See Also

tit.analyzer.main : Reads the exact dict keys this dataclass mirrors. tit.analyzer.group.run_group_analysis : Consumed by the "group" branch. tit.analyzer.analyzer.Analyzer : Consumed by the "single" branch.

AnalysisMode

Bases: StrEnum

Which entry-point branch handles this config.

Attributes

SINGLE : str One subject, via :class:tit.analyzer.Analyzer. GROUP : str Multiple subjects, via :func:tit.analyzer.run_group_analysis.

AnalyzerSpace

Bases: StrEnum

Where the field is read from.

Attributes

MESH : str SimNIBS surface/volume mesh. VOXEL : str MNI-space NIfTI volume.

AnalysisType

Bases: StrEnum

Region-of-interest shape.

Attributes

SPHERICAL : str A sphere at center with radius radius (implemented today). CORTICAL : str An atlas region on atlas (implemented today). SUBCORTICAL : str A volumetric atlas region. Not implemented by the runner yet -- see :class:AnalyzerConfig's Notes.

AnalyzerCoordinateSpace

Bases: StrEnum

Space of center for a spherical ROI.

Attributes

SUBJECT : str Native subject-space coordinates (mm). MNI : str MNI coordinates, transformed to subject space before analysis.

AnalyzerConfig dataclass

AnalyzerConfig(mode: AnalysisMode = SINGLE, subject_id: str | None = None, subject_ids: list[str] = field(default_factory=list), simulation: str = '', space: AnalyzerSpace = MESH, tissue_type: str = 'GM', analysis_type: AnalysisType = SPHERICAL, field: str | None = None, center: list[float] | None = None, radius: float | None = None, coordinate_space: AnalyzerCoordinateSpace = SUBJECT, atlas: str | None = None, region: str | list[str] | None = None, output_dir: str | None = None, visualize: bool = True, mask_path: str | None = None)

Configuration for one analyzer run (single subject or group).

Attributes

mode : AnalysisMode "single" (one subject, subject_id) or "group" (multiple subjects, subject_ids). subject_id : str or None Subject identifier. Required when mode is "single". subject_ids : list of str Subject identifiers. Required (non-empty) when mode is "group". simulation : str Simulation (montage) folder name. space : AnalyzerSpace "mesh" or "voxel". tissue_type : str "GM", "WM", or "both" (voxel space only). analysis_type : AnalysisType ROI shape; see :class:AnalysisType. field : str or None Field to analyze (a name from constants.FIELD_REGISTRY). None resolves the TI_max/mTI_max envelope automatically. center : list of float or None [x, y, z] sphere center in mm. Required for analysis_type="spherical". radius : float or None Sphere radius in mm. Required for analysis_type="spherical". coordinate_space : AnalyzerCoordinateSpace Space of center (spherical only). atlas : str or None Atlas name. Required for analysis_type="cortical". region : str or list of str or None Region name(s) within atlas. Required for analysis_type="cortical". mask_path : str or None NIfTI mask; positive voxels select the ROI in coordinate_space. output_dir : str or None Override output directory. None derives it from PathManager. visualize : bool Generate visualization artifacts.

Raises

ValueError If mode is "single" without subject_id, "group" without subject_ids, if analysis_type is "spherical" without both center and radius, or "cortical" without atlas.

Notes

Two things this dataclass models but the runner does not (yet) honor, reported rather than silently patched into tit/analyzer/__main__.py (owned by agent B4):

• analysis_type="subcortical" is accepted here (and by :class:AnalysisType) but tit.analyzer.__main__._run_single only branches on "spherical" and "cortical" -- a subcortical request is silently a no-op today.
• region is a single canonical field here; the runner reads data.get("regions") or data.get("region") (two keys, "regions" preferred). Serializing this dataclass writes only "region", which the runner already falls back to, so single-region submissions work unchanged; only the plural alias is not reproduced.

See Also

tit.analyzer.main : The entry point whose accepted keys this dataclass mirrors.