sct_tutorial_data

This dataset contains image files required for the Spinal Cord Toolbox tutorials. It also contains two scripts with all of the tutorial commands, to demonstrate how commands can be linked together to form a pipeline.

Using this dataset

There are three ways to use this dataset:

1. Download the entire sct_tutorial_data dataset (Releases -> "Source code (.zip)"), then follow along with each tutorial in the documentation from start to finish. (The tutorials link together.)
2. Download the entire dataset, but use the included batch scripts (batch_single_subject.sh and process_data.sh) to execute all of the commands together as a pipeline.
3. Download a single dataset for a specific tutorial (Releases -> e.g. "data_spinalcord-segmentation.zip"), and complete just that tutorial.

Making changes to this dataset

If you've written or modified a tutorial for SCT, and your tutorial relies on certain files, updating this dataset requires 4 things:

1. Update the batch scripts (batch_single_subject.sh and process_data.sh) with any new or modified commands.
2. Update the file tutorial-datasets.csv for your tutorial.
   • All of the files you specify will be automatically packaged into a .zip download. You can then link to this download at the start of your tutorial.
   • Be sure to specify each file that needs to be present in order to complete your tutorial.
   • Try to re-use the intermediate files generated by earlier tutorial commands. (If you can't, you may commit new files to this repo.)
3. Create a new release by clicking "Run workflow" on this GitHub Actions page.
   • First, the workflow executes the batch_single_subject.sh script to generate all of the necessary files.
   • Then, it packages up sets of data files into .zip downloads.
   • Finally, it creates a new release with the tutorial-specific .zip files attached.
4. In the SCT pull request for your tutorial changes, go to the conf.py file and update extlinks to point to your new release tag.
   • Note: The reason we don't use an automatically updating "latest" link is to prevent any future changes from silently breaking old documentation pages.