AND Export Guide

Export PRISM datasets to AND (Austrian NeuroCloud) compatible format.

Quick Start

Web Interface (Recommended)

Open PRISM Studio:
```
./prism-studio.py
```
Go to Projects page
Open your project
Scroll to Data Export / Share section
Enable AND Export checkbox
(Optional) Configure options:
- Convert to Git LFS: Only if AND requires it
- Include CI/CD Examples: Add validation workflows
- Edit Metadata: Click to customize README/CITATION details
Click Export for AND button
Find exported dataset in folder ending with _anc_export

Command Line

# Basic export (keeps DataLad compatible)
python -m src.converters.anc_export /path/to/my_dataset

# Export with Git LFS conversion (for AND submission)
python -m src.converters.anc_export /path/to/my_dataset --git-lfs

# Custom output location
python -m src.converters.anc_export /path/to/my_dataset -o /path/to/anc_ready_dataset

# With additional metadata
python -m src.converters.anc_export /path/to/my_dataset --metadata metadata.json

# Include CI/CD example files
python -m src.converters.anc_export /path/to/my_dataset --include-ci-examples

Metadata JSON Example

Create a metadata.json file with dataset information:

{
  "DATASET_NAME": "My Cognitive Study",
  "DATASET_DESCRIPTION": "fMRI study investigating working memory in healthy adults",
  "DATASET_CONTENTS": "30 participants, fMRI (T1w, BOLD), behavioral data",
  
  "INDEPENDENT_VARIABLES": "- Working memory load (low vs. high)\n- Task difficulty",
  "DEPENDENT_VARIABLES": "- Reaction time\n- BOLD signal\n- Accuracy",
  "CONTROL_VARIABLES": "- Age (20-35 years)\n- Education level\n- Handedness",
  
  "SUBJECT_DESCRIPTION": "30 healthy adults (15 female, 15 male) aged 20-35",
  "RECRUITMENT_INFO": "Recruited via university mailing lists",
  "INCLUSION_CRITERIA": "Right-handed, normal or corrected vision, no neurological conditions",
  "EXCLUSION_CRITERIA": "MRI contraindications, psychiatric disorders, left-handedness",
  
  "APPARATUS_DESCRIPTION": "3T Siemens Prisma scanner, 64-channel head coil",
  "TASK_ORGANIZATION": "Two runs per session, counterbalanced order",
  "TASK_DETAILS": "N-back task with 2-back and 3-back conditions",
  
  "CONTACT_NAME": "Dr. Jane Doe",
  "CONTACT_EMAIL": "jane.doe@university.edu",
  "CONTACT_ORCID": "https://orcid.org/0000-0001-2345-6789",
  
  "AUTHOR_GIVEN_NAME": "Jane",
  "AUTHOR_FAMILY_NAME": "Doe",
  "AUTHOR_AFFILIATION": "University of Example, Department of Psychology",
  
  "LICENSE": "CC-BY-4.0",
  "FUNDING": "Example Foundation Grant 12345",
  "ETHICS_APPROVALS": "University Ethics Committee Approval #2024-001"
}

What Gets Generated

The AND export creates:

Required Files

✅ README.md - ANC-structured dataset documentation
✅ CITATION.cff - Citation metadata for the dataset
✅ .bids-validator-config.json - BIDS validator configuration
✅ participants.tsv - Participant demographics (copied from source)
✅ dataset_description.json - BIDS dataset description (copied from source)

Optional Files

📄 GIT_LFS_SETUP.md - Instructions for Git LFS setup (if --git-lfs used)
📄 DATALAD_NOTE.md - DataLad usage notes (default)
📄 ANC_EXPORT_REPORT.json - Validation report
📄 .gitlab-ci.yml.example - CI/CD example for GitLab (if --include-ci-examples)
📄 .github/workflows/validate.yml.example - CI/CD example for GitHub (if --include-ci-examples)
📄 CI_SETUP.md - Instructions for CI/CD setup (if --include-ci-examples)

Dataset Structure

my_dataset_anc_export/
├── README.md                          # ANC-formatted documentation
├── CITATION.cff                       # Citation metadata
├── .bids-validator-config.json        # Validator config
├── .gitattributes                     # Git LFS config (if --git-lfs)
├── dataset_description.json           # BIDS description
├── participants.tsv                   # Participant data
├── participants.json                  # Participant metadata
├── sub-01/                            # Subject folders
│   └── ses-01/
│       ├── anat/
│       ├── func/
│       └── ...
└── code/                              # Analysis code

DataLad vs Git LFS

Default Export (DataLad-friendly)

python -m src.converters.anc_export /path/to/my_dataset

Keeps dataset DataLad compatible
No .gitattributes for Git LFS
Adds DATALAD_NOTE.md with instructions

Git LFS Export (For AND)

python -m src.converters.anc_export /path/to/my_dataset --git-lfs

Creates .gitattributes for Git LFS
Adds GIT_LFS_SETUP.md with instructions
Required if AND mandates Git LFS

Ask AND first! They may accept DataLad datasets.

Workflow

1. Prepare Your PRISM Dataset

Ensure your dataset is valid:

python prism.py /path/to/my_dataset

2. Create Metadata File

Create metadata.json with your dataset information (see example above).

3. Run Export

python -m src.converters.anc_export /path/to/my_dataset \
  --metadata metadata.json \
  --git-lfs  # Only if AND requires Git LFS

4. Review Generated Files

Check the exported dataset:

cd my_dataset_anc_export
cat README.md           # Review documentation
cat CITATION.cff        # Review citation
cat ANC_EXPORT_REPORT.json  # Check validation

5. Edit as Needed

Manually edit any generated files to add missing information:

README.md - Add detailed methods
CITATION.cff - Update author affiliations
.bids-validator-config.json - Adjust validation rules

6. Validate

Run BIDS validator:

bids-validator . --config .bids-validator-config.json

7. Submit to AND

Follow AND submission guidelines with your prepared dataset.

Integration with PRISM Studio (Web UI)

The AND export is integrated into the PRISM Studio web interface:

Projects → Data Export / Share section
Enable AND Export checkbox to show options
Configure export settings:
- Git LFS conversion (if required)
- CI/CD example files
- Dataset metadata (optional)
Click Export for AND to create submission-ready dataset
Review generated files in <project>_anc_export/ folder

Features

✅ Form-based metadata editing - No JSON required
✅ Visual progress tracking - See export status
✅ One-click export - Complete AND package in seconds
✅ Integrated validation - Checks requirements before export

See Web Interface Documentation for more details.

Programmatic Usage

from src.converters.anc_export import ANCExporter

# Create exporter
exporter = ANCExporter(
    dataset_path="/path/to/my_dataset",
    output_path="/path/to/anc_export"
)

# Define metadata
metadata = {
    "DATASET_NAME": "My Study",
    "CONTACT_EMAIL": "me@university.edu",
    # ... more fields
}

# Export
output_path = exporter.export(
    metadata=metadata,
    convert_to_git_lfs=False,  # Keep DataLad compatible
    copy_data=True
)

print(f"Exported to: {output_path}")

Troubleshooting

Missing Required Files

If export reports missing files, check your source dataset:

dataset_description.json → Required by BIDS
participants.tsv → Required by BIDS

README Placeholders Not Replaced

Create a complete metadata JSON file with all fields. See example above.

Git LFS Conversion Issues

Ensure source files are not symlinks (DataLad issue)
Use datalad get to retrieve actual files first
Then run export with --git-lfs

Validation Failures

Check ANC_EXPORT_REPORT.json for details on missing files or validation issues.

Support

For questions about AND export:

PRISM Issues: https://github.com/MRI-Lab-Graz/prism-studio/issues
AND Support: Contact AND directly for submission requirements