Upload latest GT on OpenNeuro

[jcohenadad]

After discussing with @rohanbanerjee I realized that the current GT on OpenNeuro are the original ones provided by the sites. Since then, this GTs have been revisited. We should upload them on OpenNeuro (replace the current ones to avoid confusion).

Related to #30

[rohanbanerjee]

The latest GTs have been uploaded to Openneuro and is the v1.1.0 version of the dataset.

More details:

• List of subjects that have been updated can be found in issues Training and inference discussion for active learning round 1 #35, Training and inference discussion for active learning round 2 #38 and Training and inference discussion for active learning round 3 #40 under the filename finetuning.yml
• Script used for updating the data on Openneuro can be found at https://github.com/sct-pipeline/fmri-segmentation/blob/main/scripts/openneuro_data_update.py
• Logic flow of the script

1. Reads subjects from the respective yml files (to know which subjects were manually corrected)
2. Copies the manually segmented files from the nnUNet datasets (that were used for training the active learning rounds) to the openneuro dataset.
3. Updates/Appends the json sidecar files with the below information
{	
      'Name': 'Manual',                                        
      'Author': 'Rohan Banerjee and Merve Kaptan (Manually corrected after initial segmentation done by EPISeg model 
      (https://github.com/sct-pipeline/fmri-segmentation/releases/tag/v0.2))',                                         
      'Date': time.strftime('%Y-%m-%d %H:%M:%S')
}

The above json file updation follows this comment

[rohanbanerjee]

Steps/Commands followed to download and update the dataset on Openneuro:

1. Download the data:
openneuro upload --dataset ds005143 -i openneuro_data

2. Updating the dataset using the script mentioned in the above comment

3. Upload the new snapshot of the dataset: 
openneuro upload --dataset ds005143 -i openneuro_data

[rohanbanerjee]

{
'Name': 'Manual',
'Author': 'Rohan Banerjee and Merve Kaptan (Manually corrected after initial segmentation done by EPISeg model
(https://github.com/sct-pipeline/fmri-segmentation/releases/tag/v0.2))',
'Date': time.strftime('%Y-%m-%d %H:%M:%S')
}

There should be three values. For eg in the above json, there should be, for our case:

• Initial author
• AI Model
• Manual rater

[rohanbanerjee]

The mentioned update has been done in the new version of the dataset --> https://openneuro.org/datasets/ds005143/versions/1.2.0