Release Notes

1.2.1

• Fixed Siemens CSA ASCCONV header parsing (#85).

1.2.0

• Changed license to MIT.

• Removed NiBabel from requirements.

• Made pandas an optional requirement.

• Added new Image attributes and properties:

  • affine()

  • image_orientation_patient()

  • image_position()

  • image_shape()

  • number

  • rotation_matrix()

  • slice_normal()

  • spatial_resolution()

• data() now also applies rescaling as (and if) defined by the Rescale Slope and Rescale Intercept header fields. See:

  • rescale_data()

  • fix_data()

• Fixed private tag exclusion in the Header class’ to_dict() method (see #59).

• Created the dicom_parser.utils.multi_frame.multi_frame.MultiFrame class to handle multi-frame images.

• Improved sequence detection logic and created a new bids module, meant to enable automated BIDS-compatible relative path generation based on the detected sequences.

• Fixed partial CSA reads to include both the general header and the nested ASCII-encoded (ASCCONV) header as parsed dictionaries.

• Fixed ValueError raised when calling the dicom_parser.image.Image class’ dicom_parser.image.Image.data() property with a missing ImageType header field.

• Improved CSA ASCCONV header parsing based on NiBabel’s implementation.

1.1.1

• Added Codacity and CodeFactor integrations to improve code quality and implemented suggestions.

• Added issue templates, code of conduct, and RTD configuration file.

• Added documentation building and test execution instructions in the project’s README.

1.1.0

• Updated project structure and packaging to conform with PEP517 and PEP518.

• Migrated CI to GitHub Actions.

• Fixed dcm generation by mime-type bug (#30) for nested directories.

• Improved tests and increased coverage.

1.0.1

• Fixed ImportError bug upon importing the python-magic library in a non-Linux OS.

1.0.0

• Improved test coverage and added more docstrings.

0.1.7

• Added the keyword_contains() method to easily look for data elements using a substring of the keyword.

• Increased tests coverage.

0.1.6

• Lots of documentation updates and linting fixes.

• Integrated isort.

• Changed invalid integer strings (“IS” data elements) to return None (see IntegerString).

• Added mime option to Series.

• Updated development status classifier to beta.

0.1.5

• Changed PrivateDataElement’s parse_value() method to try and call bytes.decode() on the raw value by default.

• Removed deprecated dicom_parser.parser.Parser class.

• Fixed some linting and docstring issues.

0.1.4

• Changed the return type of data elements with a value multiplicity greater than 1 from list to tuple.

0.1.3

• Changed Code String (CS) parsing to return verbose value or set of values using the code_strings module. This feature can be disabled by setting the Parser’s verbose_code_strings attribute to False.

• Added warnings for invalid pixel arrays and invalid Code String (CS) values.

• Updated the sequences module to work with verbose Code String (CS) data element values and renamed known MR sequences.

• Refactored the Parser class to a bit to improve readability.

• Fixed CSA header bug for headers with a duplicate “### ASCCONV END ###” pattern.

• Created the DataElement class as a wrapper around pydicom's DataElement to support some custom functionality and provide better integration with django_dicom.

0.1.2

• Improved private tags definition so that the Header model’s get() method will work using the private tag’s keyword.

• Added support for calling len() over a Series instance to return the number of images in it.

• Added support for querying header information from a Series instance using the get() method or the indexing operator ([]) using a str or a tuple.

• Added support for indexing the Image instances from a Series using an int or slice.

0.1.1

• Improved support for accessing CSA headers.

• Added auto-decoding for Siemens mosaic encoded data (applies to data extraction in both Image and Series instances).

  • The decoding method also changes the orientation of the stacked arrays to match the product of conversion to NIfTI using dcm2niix.

0.1.0

First release!

• Type correction based on value-representation (VR).

• Simple Series class for reading DICOM series directories.

• Basic support for reading CSA headers using the CsaHeader class.