dicomgenerator

Generate pydicom datasets and data elements for use in testing


License
MIT
Install
pip install dicomgenerator==0.8.1

Documentation

dicomgenerator

CI PyPI PyPI - Python Version Code style: black Checked with mypy

Generate pydicom datasets and data elements for use in testing.

  • Free software: MIT license
  • Status: Alpha. Tests run but there are loose ends

Features

  • Extends factory-boy factories to produce pydicom Datasets and DicomElements
  • Generate valid DICOM values for person name, time, date, and UID
  • Create json-based editable templates from any dicom file
  • quick_dataset(): single-line pydicom dataset init

Installation

Install with pip::

pip install dicomgenerator

Usage

Quick dataset

I have found this quite useful in testing:

    from dicomgenerator.generators import quick_dataset
    ds = quick_dataset(PatientName='Jane', StudyDescription='Test')

    # >>> ds.PatientName -> 'Jane'     
    # >>> ds.StudyDescription -> 'Test'

Generating a dataset

Generate a realistic CT dataset

    from dicomgenerator.factory import CTDatasetFactory

    # Generate from template
    >>> CTDatasetFactory().PatientName -> 'van Haarlem^Anouk'  #  generated random name
    >>> CTDatasetFactory().PatientName -> 'Loreal^Casper'      #  generated random name

    # Overwrite arbitrary DICOM elements
    ds.CTDatasetFactory(PatientSex='M', PatientName='Smith^Harry')
    >>> ds.PatientName -> 'Smith^Harry'
    >>> ds.PatientSex  -> 'M'

    # generated UIDs and dates are valid DICOM
    >>> CTDatasetFactory().StudyTime        -> '130624.929'
    >>> CTDatasetFactory().StudyDate        -> '20110508'
    >>> CTDatasetFactory().StudyInstanceUID -> '1.2.826.0.1.3680'

Generating a data element

    # import
    from dicomgenerator.factory import DataElementFactory

    # Creating a DICOM data element by name will give a realistic value and correct VR
    >>> DataElementFactory(tag='PatientName').value -> "van Ooyen^Fiene"
    >>> DataElementFactory(tag='PatientName').VR -> 'PN'

    # You can also give DICOM tags as hex
    >>> DataElementFactory(tag=0x00100010).value -> "Weil^Jack"

    # Dates, times and UIDs all work.
    >>> DataElementFactory(tag="AcquisitionTime").value   -> '184146.928'
    >>> DataElementFactory(tag="PatientBirthDate").value  -> '20120511'
    >>> DataElementFactory(tag="SeriesInstanceUID").value -> '1.2.826.0.1.3680'

In reproducible tests

You can set the random seed in factory-boy like this:

    from factory import random

    def test_one:
        """The random patient name in this test will always be the same"""
        random.reseed_random('any string you want')
        assert element = DataElementFactory(tag='PatientName').value == "van Ooyen^Fiene"

Command Line Interface

You can convert a DICOM file to AnnotatedDataset via the commandline. by default this will write an annotated dataset to the same folder, appending_template.json

$ dicomgen convert to-json /tmp/dicom_file
$ ls
dicom_file  dicom_file_template.json

For options, use

$ dicomgen convert to-json --help

Credits

This package was originally created with Cookiecutter and the audreyr/cookiecutter-pypackage project template.