Add File organization and DICOM to NIfTI conversion sections (#3)

* Add 'Data structure' section * Add note that only the `sourcedata` directory is required * Add participants.tsv example * Fix subject naming * Update comments * Add 'data_processed' and 'bids' folders * Store the first version of the script. This version interactively prompts the user to provide paths to the nifti images. * Add '-bids-folder' input arg; check whether the BIDS folder already exists, if so, warn the user and exit * Remove unused import; improve arg help * Add prints * t2_path --> path_t2w * dwi_path --> path_dwi * Refactor the script to allow selecting the images for further processing from the list of images * Update description * Update docstring * Update docstrings * Use os.path.expanduser * Allow to specify the contrasts to process using an input arg * Add dcm2niix requirement * Add example of the input and output file structure * Add TODOs * Add '-debug' flag to keep the temp folder (output of dcm2niix) * Move print into `copy_files_to_bids_folder` * Fix `-contrasts` to properly accept several contrasts * Add 'File organization' subsection * Add '3.2 DICOM to NIfTI conversion' subsection * Update TOC * Briefly describe the dependencies * typos * typo * update dimensions and pixel_size formatting * include also series description for dcm2niix * update df printing * use logging instead of print statements * Sort nii files based on the series number (the last number in the file name before the .nii.gz extension) * Switch from 'logging.eror' to 'logging.info("Warning: ...' * Include "Error" into logging.error print * Explicitly tell users the rows number to select from * Check if there are any NIfTI files in the folder, if not, print error message and exit * Write a new entry to the participant.tsv - add new args for sex and age * Add '2.3 Downloading this repository' section * Update script path * Add `-contrasts` arg to the example usage * Add example usage * Update README.md * Simplify TOC - add 'Getting Started' - move 'Dependencies' and 'Installation' under 'Getting Started' * Move printing * Use 'formatter_class=argparse.RawTextHelpFormatter' * Add an example on how to distinguish between two images of the same contrast with different orientation, e.g. 'acq-axial_T2w' and 'acq-sag_T2w' --------- Co-authored-by: Julien Cohen-Adad <[email protected]>
sct-pipeline · Sep 5, 2024 · d16c6f1 · d16c6f1
1 parent 168ee26
commit d16c6f1
Show file tree

Hide file tree

Showing 2 changed files with 558 additions and 16 deletions.
diff --git a/README.md b/README.md
@@ -1,31 +1,37 @@
 # balgrist-sci
 
-Repository containing pipeline for processing spinal cord injury (SCI) patients (both DCM and tSCI) using the [Spinal Cord Toolbox (SCT)](https://github.com/spinalcordtoolbox/spinalcordtoolbox).
+Repository containing code to process MRI data from spinal cord injury (SCI) patients (both DCM and tSCI) using the [Spinal Cord Toolbox (SCT)](https://github.com/spinalcordtoolbox/spinalcordtoolbox).
 
-Pipeline steps:
-1. DICOM to nii (BIDS) conversion
+Steps:
+1. DICOM to NIfTI (BIDS) conversion
 2. Processing (spinal cord and lesion segmentation, vertebral labeling)
 3. Quality control (QC) + manual compression level labeling
 4. Lesion metric computation
 
 ## Table of contents
-* [1. Dependencies](#1-dependencies)
-* [2. Installation](#2-installation)
-  * [2.1 SCT Installation](#21-sct-installation)
-  * [2.2 dcm2niix Installation](#22-dcm2niix-installation)
-* [3. Processing pipeline](#3-processing-pipeline)
+- [1. Getting Started](#1-getting-started)
+  - [1.1 Dependencies](#11-dependencies)
+  - [1.2 Installation](#12-installation)
+    - [SCT Installation](#sct-installation)
+    - [dcm2niix Installation](#dcm2niix-installation)
+    - [Downloading this repository](#downloading-this-repository)
+- [2. Data structure](#2-data-structure)
+  - [2.1 File organization](#21-file-organization)
+  - [2.2 DICOM to NIfTI conversion](#22-dicom-to-nifti-conversion)
 
-## 1. Dependencies
+## 1. Getting Started
 
-* [Spinal Cord Toolbox v6.4](https://github.com/spinalcordtoolbox/spinalcordtoolbox/releases/tag/6.4)
-* [dcm2niix >= v1.0.20220505](https://github.com/rordenlab/dcm2niix?tab=readme-ov-file#install)
+### 1.1 Dependencies
 
-## 2. Installation
+* [Spinal Cord Toolbox v6.4](https://github.com/spinalcordtoolbox/spinalcordtoolbox/releases/tag/6.4): toolbox for processing spinal cord MRI data
+* [dcm2niix >= v1.0.20220505](https://github.com/rordenlab/dcm2niix?tab=readme-ov-file#install): tool for converting DICOM images into the NIfTI format
+
+### 1.2 Installation
 
 > [!NOTE]
 > The installation process below is currently only supported on macOS.
 
-### 2.1 SCT Installation
+#### SCT Installation
 
 <details><summary>Click the triangle to expand/collapse the section</summary>
 
@@ -69,7 +75,7 @@ The expected output is `[OK]` for all dependencies.
 
 </details>
 
-### 2.2 dcm2niix Installation
+#### dcm2niix Installation
 
 <details><summary>Click the triangle to expand/collapse the section</summary>
 
@@ -99,6 +105,123 @@ The expected output is the version of `dcm2niix`.
 
 </details>
 
-## 3. Processing pipeline
+#### Downloading this repository
+
+<details><summary>Click the triangle to expand/collapse the section</summary>
+
+1. Open a new terminal (if you closed the previous one):
+
+Press <kbd>command</kbd> + <kbd>space</kbd> and type `Terminal` and press <kbd>return/enter</kbd>.
+
+2. Run the following commands in the terminal (you can copy-paste the whole block):
+
+```bash
+# Go to your home directory
+cd ~
+# Download (clone) the repository --> the repository will be downloaded to a directory named balgrist-sci
+git clone https://github.com/sct-pipeline/balgrist-sci.git balgrist-sci
+```
+
+TODO: Replace `git clone` by `wget` once we publish a repo release because Apple Developer Tools are needed for `git`.
+
+3. Check that the repository was downloaded correctly:
+
+```bash
+# Activate SCT conda environment
+source ./python/etc/profile.d/conda.sh
+conda activate venv_sct
+# Call the help of the file_loader.py script
+python ~/balgrist-sci/file_loader.py --help
+```
+
+</details>
 
-TODO
+## 2. Data structure
+
+### 2.1 File organization
+
+<details><summary>Click the triangle to expand/collapse the section</summary>
+
+A file organization according to the [BIDS](https://bids-specification.readthedocs.io/en/stable/) is shown below.
+
+Note that only the `sourcedata` directory containing folders with DICOM files for each subject is initially required. 
+The rest of the directories and files will be created during the processing; see the next section.
+
+```
+├── participants.tsv        --> file with participants information; see example below
+├── sourcedata              --> folder containing DICOM files for each subject
+│   ├── dir_20230711        --> folder with DICOM files for first subject and first session
+│   ├── dir_20230711        --> folder with DICOM files for second subject and first session
+│   ├── ... 
+│   ├── dir_20240815        --> folder with DICOM files for first subject and second session
+│   └── ... 
+├── bids                    --> folder with BIDS-compliant data
+│    ├── sub-001            --> folder containing NIfTI files for first subject
+│    │   ├── ses-01         --> first session
+│    │   │  ├── anat        --> folder with anatomical data
+│    │   │  │  ├── sub-001_ses-01_T1w.nii.gz
+│    │   │  │  ├── sub-001_ses-01_T2w.nii.gz
+│    │   │  │  ├── ...
+│    │   │  └── dwi         --> folder with diffusion data
+│    │   │     ├── sub-001_ses-01_dwi.nii.gz
+│    │   │     ├── sub-001_ses-01_dwi.bval
+│    │   │     ├── sub-001_ses-01_dwi.bvec
+│    │   └── ses-02         --> second session
+│    │      ├── ...
+│    ├── sub-002            --> folder containing NIfTI files for second subject
+│    │   ├── ...
+│    ├── ...
+├── data_processed          --> folder with processed data  
+│    ├── sub-001            --> folder with processed data for first subject
+│    │   ├── ses-01         --> first session
+│    │   │  ├── anat        --> folder with processed anatomical data
+│    │   │  │  ├── ...
+└── derivatives             --> folder to store visually checked and/or manually corrected data (for example, spinal cord segmentations)
+    └── labels
+        ├── sub-001         --> first subject
+        │   ├── ses-01      --> first session
+        │   │  ├── anat
+        │   │  │  ├── sub-001_ses-01_T2w_label-SC_seg.nii.gz              --> spinal cord (SC) binary segmentation 
+        │   │  │  ├── sub-001_ses-01_T2w_label-compression_label.nii.gz   --> binary compression labeling
+        │   │  │  ├── ...
+        │   │  └── dwi
+        │   │     ├── sub-001_ses-01_dwi_label-SC_seg.nii.gz
+        │   │     ├── ...
+        │   └── ses-02      --> second session
+        │      ├── ...
+        ├── sub-002 
+        └── ...
+```
+
+`participants.tsv` example:
+
+| participant_id | ses_id | source_id | age | sex |
+|----------------|--------|-----------|-----|-----|
+| sub-001        | ses-01 | dir_20230711 | 42  | M   |
+| sub-001        | ses-02 | dir_20240815 | 43  | M   |
+| sub-002        | ses-01 | dir_20230713 | 57  | F   |
+
+ℹ️ Notice that we use one row per session. This means that, for example, `sub-001` has two rows in the table because they have two sessions.
+
+</details>
+
+### 2.2 DICOM to NIfTI conversion
+
+A single subject DICOM images can be converted to NIfTI and organized according to the BIDS standard using 
+the `file_loader.py` script.
+
+Example usage:
+
+```bash
+# Activate SCT conda environment
+cd $SCT_DIR
+source ./python/etc/profile.d/conda.sh
+conda activate venv_sct
+# Run the script
+python ~/balgrist-sci/file_loader.py \
+  -dicom-folder ~/data/experiments/balgrist-sci/source_data/dir_20231010 \
+  -bids-folder ~/data/experiments/balgrist-sci/bids \ 
+  -participant sub-001 \
+  -session ses-01 \
+  -contrasts T2w dwi
+```