miniQC

---

miniQC is a BIDS App for performing minimal quality control beyond validation. Its goal is to rapidly detect data corruption, as opposed to dataset coherence or image quality.

Table of Contents

• Installation
• Usage
• Testing
• License

Installation

pip install miniqc

Usage

miniqc is a BIDS-App, so can be run as follows:

miniqc /path/to/dataset /output participant

Because miniqc does not currently generate any outputs or modify its behavior based on analysis level, output directory and analysis level are optional.

This tool is meant as a CLI application and does not provide a public API.

Example

$ miniqc tests/data/bids_dataset
[
  [
    "sub-01/anat/sub-01_acq-truncated_T2w.nii.gz",
    "FailedCheck",
    "Expected 477 bytes; found 352"
  ],
  [
    "sub-01/anat/sub-01_acq-dangling_T2w.nii.gz",
    "FileNotFoundError",
    "[Errno 2] No such file or directory: 'tests/data/bids_dataset/sub-01/anat/sub-01_acq-dangling_T2w.nii.gz'"
  ]
]

Single file usage is also permitted:

$ miniqc tests/data/bids_dataset/sub-01/anat/sub-01_acq-truncated_T2w.nii.gz
[
  [
    "/git/miniqc/tests/data/bids_dataset/sub-01/anat/sub-01_acq-truncated_T2w.nii.gz",
    "FailedCheck",
    "Expected 477 bytes; found 352"
  ]
]

Outputs

The output of this tool is a JSON array of arrays, each of length 3. Each sub-array contains the failed file (relative to dataset root), the type of error, and a message with more detail.

Testing

With the hatch project management tool installed:

hatch run test:cov

Alternately, just run pytest, although you will need the dependencies installed.

License

miniqc is distributed under the terms of the MIT license.