title: API — the catalog description: pardesline.catalog: Catalog, Module, Method, Parameter with validation, rationale and validated_on.
Catalog¶
The typed catalog of modules, methods and parameters.
The catalog is generated from the same manifests that drive the PardesLine
platform (catalog/data/<module>.json). It is the single source of truth
for what an algorithm is called, which file types it accepts, which flags it
takes, their defaults and their ranges. Nothing here imports a native library.
>>> import pardesline as pl
>>> cat = pl.catalog()
>>> m = cat.method("point_cloud_processing", "voxel_downsample")
>>> m.validate({"voxel_size": 0.05})
{'voxel_size': '0.05'}
Catalog(modules: list[Module])
All modules, ordered as on the platform.
method(module_id: str, method_id: str | None = None) -> Method
method("mod", "meth") or method("mod.meth").
find(method_id: str, extension: str | None = None) -> Method
Resolve a bare method id ("voxel_downsample") to its method.
When several modules define the id, extension picks the one that
accepts the file; an ambiguous lookup raises :class:CatalogError.
methods_for(extension: str | None) -> list[Method]
Every method whose module accepts files with this extension.
dataclass
¶
Module(id: str, name: str, description: str, category: str, order: int, input_formats: tuple[str, ...], methods: dict[str, Method] = dict(), dual_input: bool = False, dual_input_label: str = '', theory: dict[str, Any] = dict(), videos: tuple[dict[str, str], ...] = ())
A family of methods sharing input formats.
accepts(extension: str | None) -> bool
True when the module reads files with this suffix. .nii.gz is
accepted wherever .nii is: the readers (SimpleITK, VTK) handle the
compression themselves.
dataclass
¶
Method(id: str, name: str, module_id: str, script: str, output_extension: str, dual_input: bool = False, parameters: tuple[Parameter, ...] = (), validated_on: str = '')
One algorithm: a script with a typed parameter list.
validate(params: dict[str, Any] | None) -> dict[str, str]
Check every supplied parameter and return the command-line strings.
Unknown names, wrong types and out-of-range values raise
:class:ParameterError with the accepted list.
dataclass
¶
Parameter(name: str, type: str, default: Any, description: str = '', min: float | None = None, max: float | None = None, options: tuple[str, ...] = (), advanced: bool = False, rationale: str = '', validated_on: str = '')
One tunable flag of a method, as declared in the catalog.
coerce(value: Any, *, where: str = '') -> str
Validate value against this parameter and return the string the
script receives on its command line. Raises :class:ParameterError.