@@ -24,11 +24,11 @@ pip install /path/to/pydra-fsl/
24
24
` FSLConverter ` class (from ` tools/converter.py ` ) requires three parts of information:
25
25
26
26
- Nipype spec: converter loads nipype interface and reads ` _cmd ` , ` input_spec ` and ` output_spec `
27
- - yml file with additional spec: ` specs/fsl_conv_params .yml ` contains additional spec that are written based
27
+ - yml file with additional spec: ` specs/fsl_{module_name}_params .yml ` contains additional spec that are written based
28
28
on additional functions from nipype (including ` list_outputs ` ), each interface can have the following fields:
29
29
- inputs_metadata: additional metadata for fields from input_spec
30
30
(it will be included in ` metadata ` in pydra spec),
31
- e.g., used for ` FAST ` to set default value for ` number_classes `
31
+ e.g., used in ` specs/fsl_preprocess_params.yml ` for ` FAST ` to set default value for ` number_classes `
32
32
(it's not part of nipype's spec, but it's set in ` list_output ` )
33
33
34
34
- output_requirements: providing required fields for the output to be created,
@@ -61,7 +61,10 @@ The converter can be used by running:
61
61
62
62
python tools/converter.py --interface_name <name of teh interface> --module_name <module_name>
63
63
64
- The pydra task will be created in saved in ` pydra/tasks/fsl/{module_name}/{interface_name}.py `
64
+ The pydra task will be created and saved in ` pydra/tasks/fsl/{module_name}/{interface_name}.py ` .
65
+ Note, that the spec file has to be present for the specific module name in order to run the converter.
66
+ If no ` interface_name ` is provided, the default value ` all ` will be used
67
+ and the converter will be run for all interfaces from the spec file.
65
68
66
69
Tests are written based on the fields from the yml file:
67
70
` tests_inputs ` and ` tests_outputs ` (the lengths should be the same).
0 commit comments