[Specification Shanoir Uploader] Mass import from a folder

Massive import in Shanoir Uploader

Mass import from a PACS is now available from Shanoir-Uploader using a CSV.

We want to be able to do the same from a folder with multiple datasets and subjects.

The idea is to have a program that scrolls through files and folders to prepare a list of available subjects/datasets.

Either:

• The main folder has to be organized with configuration files for every subject to configure them.
• The user configures for every group of dataset the subject / examination / etc.. as a normal import
• The user associates with such folder a similar CSV file as configured in PACS mass import that contains the configuration

We can imagine integrating a BIDS import that takes into account a full BIDS folder, iterates over it and upload it, as is, in Shanoir. This implies to be able to import nifti directly. Every files considered in the BIDs folder should be kept too. How to adapt such structure to the Shanoir structure ?

Not from BIDS

In that case, we will iterate over a folder to import data from dicom and import them iteratively into Shanoir.

In general we are asking for the folder to follow such organization:

• Subject 1
• • dicoms[]
• • associated files
• Subject 2
• • dicoms[]
• • associated files

Iterate simply over the folder to retrieve data and import using usual uploader path.

The dicoms are analyzed from present dicomDIRs and displayed in the arborescence of the uploader on the left. Then the user will be able to continue the import with the usual steps.

Use a CSV to import data from a folder

CSV with following column:

• Folder name
• Study ID
• StudyCard
• Shanoir subject name
• Examination comment

Given a folder in entry, Shanoir uploader will iterate over the parent folder, load all data from this folder and import automatically into Shanoir with given elements in the CSV file.

Having a unique, simple configuration to realize the import.

We can use this if the user wants to import ALL the subjects, in only one study, one study card. The subject name can be automatically generated with an index added to a suffix. Then we'll iterate over the folder to find every subfolder and import the study according to this.

Import from BIDS

BIDS has become one of the major standard in neuro-imaging for storing data. Its principal component is based around nifti.

BidsDataset

• Extra-data files.
• Nifti files -> DatasetFiles

Iterative import

Iterate over all folders to display all subjects -> select all subjects by default, and allow selection. Result files should be attached to the study directly ? (allow a directory foldder for every file ?)

Bids structure should be conserved.

[to be continued]