YAML Configuration Reference

Auto-generated from the Pydantic config models in osipy.cli.config and runtime registries. This page is rebuilt on every documentation build, so it always reflects the current code.

Top-level fields

Every configuration file has these top-level keys:

PipelineConfig

Top-level pipeline configuration from YAML.

Field	Type	Default	Valid values
modality	string	required	dce, dsc, asl, ivim
pipeline	mapping	{}
data	DataConfig	(see DataConfig)
output	OutputConfig	(see OutputConfig)
backend	BackendConfig	(see BackendConfig)
logging	LoggingConfig	(see LoggingConfig)

Shared sections

These sections are common to all modalities.

data: (DataConfig)

Data loading configuration.

Field	Type	Default	Valid values
format	Literal	"auto"	auto, nifti, dicom, bids
mask	string or null	null
t1_map	string or null	null
aif_file	string or null	null
m0_data	string or null	null
b_values	list of numbers or null	null
b_values_file	string or null	null
subject	string or null	null
session	string or null	null

output: (OutputConfig)

Output configuration.

Field	Type	Default	Valid values
format	Literal	"nifti"	nifti

backend: (BackendConfig)

GPU/CPU backend configuration.

Field	Type	Default	Valid values
force_cpu	boolean	false

logging: (LoggingConfig)

Logging configuration.

Field	Type	Default	Valid values
level	Literal	"INFO"	DEBUG, INFO, WARNING

DCE Pipeline

Set modality: dce and configure these under pipeline:.

pipeline: (DCEPipelineYAML)

DCE pipeline settings from YAML.

Field	Type	Default	Valid values
model	TwoCUMConfig	TwoCXMConfig	ExtendedToftsConfig
t1_mapping_method	LookLockerConfig	VFAConfig	required
concentration	LinearConcentrationConfig	SPGRConcentrationConfig	required
aif_source	Literal	"population"	population, detect, manual
population_aif	FritzHansenAIFConfig	GeorgiouAIFConfig	McgrathAIFConfig
save_intermediate	boolean	false
acquisition	DCEAcquisitionYAML	(see DCEAcquisitionYAML)
fitting	DCEFittingConfig	(see DCEFittingConfig)

pipeline.acquisition: (DCEAcquisitionYAML)

DCE acquisition parameters from YAML.

Field	Type	Default	Valid values
baseline_frames	integer	5
relaxivity	number	4.5
tr	number or null	null
flip_angles	list of numbers or null	null
t1_assumed	number or null	null

pipeline.fitting: (DCEFittingConfig)

DCE model fitting configuration from YAML.

Field	Type	Default	Valid values
fitter	Literal	"lm"	bayesian, cSVD, lm, oSVD, sSVD, tikhonov
max_iterations	integer	100
tolerance	number	1e-06
r2_threshold	number	0.5
fit_delay	boolean	false
bounds	mapping or null	null
initial_guess	mapping or null	null

pipeline.model with method: extended_tofts

Extended Tofts model with plasma term (OSIPI: M.IC1.005).

Field	Type	Default	Valid values
method	Literal	"extended_tofts"

pipeline.model with method: patlak

Patlak model (OSIPI: M.IC1.006).

Field	Type	Default	Valid values
method	Literal	"patlak"

pipeline.model with method: tofts

Standard Tofts model (OSIPI: M.IC1.004).

Field	Type	Default	Valid values
method	Literal	"tofts"

pipeline.model with method: 2cum

Two-compartment uptake model (2CUM).

Field	Type	Default	Valid values
method	Literal	"2cum"

pipeline.model with method: 2cxm

Two-compartment exchange model (OSIPI: M.IC1.009).

Field	Type	Default	Valid values
method	Literal	"2cxm"

pipeline.t1_mapping_method with method: look_locker

Look-Locker inversion-recovery T1 mapping (OSIPI: P.NR2.004).

Field	Type	Default	Valid values
method	Literal	"look_locker"

pipeline.t1_mapping_method with method: vfa

Variable Flip Angle T1 mapping (OSIPI: P.NR2.002).

Field	Type	Default	Valid values
method	Literal	"vfa"
fit_method	Literal	"linear"

pipeline.concentration with method: linear

Linear signal-to-concentration approximation (small enhancement).

Field	Type	Default	Valid values
method	Literal	"linear"

pipeline.concentration with method: spgr

Spoiled gradient-echo signal-to-concentration model (OSIPI: P.SC1.001).

Field	Type	Default	Valid values
method	Literal	"spgr"

pipeline.population_aif with name: fritz_hansen

Population AIF: fritz_hansen.

Field	Type	Default	Valid values
name	Literal	"fritz_hansen"

pipeline.population_aif with name: georgiou

Population AIF: georgiou.

Field	Type	Default	Valid values
name	Literal	"georgiou"

pipeline.population_aif with name: mcgrath

Population AIF: mcgrath.

Field	Type	Default	Valid values
name	Literal	"mcgrath"

pipeline.population_aif with name: parker

Population AIF: parker.

Field	Type	Default	Valid values
name	Literal	"parker"

pipeline.population_aif with name: weinmann

Population AIF: weinmann.

Field	Type	Default	Valid values
name	Literal	"weinmann"

DSC Pipeline

Set modality: dsc and configure these under pipeline:.

pipeline: (DSCPipelineYAML)

DSC pipeline settings from YAML.

Field	Type	Default	Valid values
te	number	30.0
baseline_frames	integer	10
hematocrit_ratio	number	0.73
apply_leakage_correction	boolean	true
deconvolution	CSVDConfig	OSVDConfig	SSVDConfig

pipeline.deconvolution with method: cSVD

Block-circulant SVD deconvolution (delay-insensitive).

Field	Type	Default	Valid values
method	Literal	"cSVD"
threshold	number	0.2

pipeline.deconvolution with method: oSVD

Oscillation-index SVD deconvolution (per-voxel adaptive threshold).

Field	Type	Default	Valid values
method	Literal	"oSVD"
oscillation_index	number	0.035
default_threshold	number	0.2

pipeline.deconvolution with method: sSVD

Standard truncated-SVD deconvolution (single global threshold).

Field	Type	Default	Valid values
method	Literal	"sSVD"
threshold	number	0.2

ASL Pipeline

Set modality: asl and configure these under pipeline:.

pipeline: (ASLPipelineYAML)

ASL pipeline settings from YAML.

Field	Type	Default	Valid values
labeling_scheme	Literal	"pcasl"	pasl, casl, pcasl
pld	number	1800.0
label_duration	number	1800.0
t1_blood	number	1650.0
t1_tissue	number	1330.0
labeling_efficiency	number	0.85
partition_coefficient	number	0.9
m0	ReferenceRegionM0Config	SingleM0Config	VoxelwiseM0Config
difference	MeanDifferenceConfig	PairwiseDifferenceConfig	SurroundDifferenceConfig
quantification	MultiPLDConfig	SinglePLDConfig	required
label_control_order	Literal	"label_first"	label_first, control_first

pipeline.m0 with method: reference_region

Reference-region M0 calibration (OSIPI ASL Lexicon).

Field	Type	Default	Valid values
method	Literal	"reference_region"
reference_region	Literal	"csf"
t1_tissue	number	1330.0
tr_m0	number	6000.0
te_m0	number	13.0

pipeline.m0 with method: single

Single (mean) M0 calibration (OSIPI ASL Lexicon).

Field	Type	Default	Valid values
method	Literal	"single"
t1_tissue	number	1330.0
tr_m0	number	6000.0
te_m0	number	13.0

pipeline.m0 with method: voxelwise

Voxel-by-voxel M0 calibration (OSIPI ASL Lexicon).

Field	Type	Default	Valid values
method	Literal	"voxelwise"
t1_tissue	number	1330.0
tr_m0	number	6000.0
te_m0	number	13.0

pipeline.difference with method: mean

Mean subtraction (average controls and labels separately).

Field	Type	Default	Valid values
method	Literal	"mean"

pipeline.difference with method: pairwise

Pair-wise control-label subtraction, then average.

Field	Type	Default	Valid values
method	Literal	"pairwise"

pipeline.difference with method: surround

Surround subtraction (average adjacent controls per label).

Field	Type	Default	Valid values
method	Literal	"surround"

pipeline.quantification with mode: multi_pld

Multi-PLD CBF + ATT estimation via the Buxton general kinetic model.

Field	Type	Default	Valid values
mode	Literal	"multi_pld"
plds	list of numbers	required
att_model	string	"buxton"

pipeline.quantification with mode: single_pld

Single-PLD CBF quantification (one delay per voxel).

Field	Type	Default	Valid values
mode	Literal	"single_pld"

IVIM Pipeline

Set modality: ivim and configure these under pipeline:.

pipeline: (IVIMPipelineYAML)

IVIM pipeline settings from YAML.

Field	Type	Default	Valid values
fitting	BayesianFittingConfig	FullFittingConfig	SegmentedFittingConfig
model	BiexponentialModelConfig	SimplifiedModelConfig	required
normalize_signal	boolean	true

pipeline.fitting with method: bayesian

Two-stage Bayesian MAP IVIM fitting with empirical priors.

Field	Type	Default	Valid values
max_iterations	integer	500
tolerance	number	1e-06
bounds	mapping or null	null
initial_guess	mapping or null	null
method	Literal	"bayesian"
b_threshold	number	200.0
prior_scale	number	1.5
noise_std	number or null	null
compute_uncertainty	boolean	true

pipeline.fitting with method: full

Full bi-exponential IVIM fitting (all b-values jointly).

Field	Type	Default	Valid values
max_iterations	integer	500
tolerance	number	1e-06
bounds	mapping or null	null
initial_guess	mapping or null	null
method	Literal	"full"

pipeline.fitting with method: segmented

Two-step (segmented) IVIM fitting.

Field	Type	Default	Valid values
max_iterations	integer	500
tolerance	number	1e-06
bounds	mapping or null	null
initial_guess	mapping or null	null
method	Literal	"segmented"
b_threshold	number	200.0

pipeline.model with model: biexponential

IVIM bi-exponential signal model: S0, D, D*, f.

Field	Type	Default	Valid values
model	Literal	"biexponential"

pipeline.model with model: simplified

IVIM simplified signal model: S0, D, f (assumes D* >> D).

Field	Type	Default	Valid values
model	Literal	"simplified"
b_threshold	number	200.0