`file_load` support for organizing loaded object

[ekhahniii]

Hello,

Is it possible to provide a parent item for an object loaded via file_load? For example, it would be useful to create a markup directory and load a markup (generated by an external program) into that directory. I see slicerio.server.file_load accepts a properties kwarg, but the only documented use case is providing name.

Thanks!

[lassoan]

Can you write a bit more about your use case? How would you create and use the folders?

What would be an ideal syntax for the API? Would it be sufficient to expose these operations in Slicer?

• add/delete/rename a subject hierarchy folder
• move a node (or subject hierarchy folder) into a subject hierarchy folder

[ekhahniii]

Hi Andras, certainly.

My goal is to stage .mrml files with outputs of a pipeline for later review/correction in Slicer. The main issue now is organizing those outputs automatically in Slicer to expedite user interaction. Currently, this organization and then saving the .mrml must be performed manually by a user after running a script to load the pipeline outputs into Slicer.

Indeed, I believe these operations cover the data loading aspect of the use case:

add/delete/rename a subject hierarchy folder
move a node (or subject hierarchy folder) into a subject hierarchy folder

[ekhahniii]

@lassoan To satisfy the use case described previously, it might be more appropriate to do something like:
Slicer.exe --no-splash --no-main-window --python-script "/path/to/staging/script.py [args]" , where script.py performs the necessary data loading, organization, and saving.

Would you agree?

[lassoan]

You can certainly do a one-time setup by running a Python script in Slicer. You could also run any Python script at any time (not just at startup) in Slicer by using the /slicer/exec REST API endpoint. However, for commonly used operations, such as organizing data in a tree (or loading/saving the scene as you asked for in #5) it would probably make sense to have a convenient slicerio.server API.

The slicerio.server API may be a simplified abstraction of the Slicer scene. For example, in Slicer the folder structure is managed via a subject hierarchy node that contains folder and node items, but maybe we don't need to expose this on the slicerio.server API. The syntax for loading a node could be something like this:

slicerio.server.file_load("c:/tmp/Segmentation.nrrd", "SegmentationFile", folder="SomePatient/SomeStudy01")

WebServer module in Slicer could create the folder hierarchy based on the SomePatient/SomeStudy01 string.