Highdicom

Bug fixes

* Correct code name of Specimen Preparation Step (108)
* Correct attribute name of legacy converted images (105)

Documentation

* Add example on how to create secondary capture image (109)

Bug fixes

* Allow inclusion of content items without Relationship Type into content sequences defined by Acquisition and Protocol templates (107)

New features

* Implement Parametric Map IOD (89)
* Enable construction of *highdicom* objects from existing *pydicom* objects (93)
python
dataset = pydicom.filereader.dcmread(...)
segmentation = hd.seg.Segmentation.from_dataset(dataset)

* Add properties and methods to classes derived from `highdicom.sr.ContentItem` and `highdicom.sr.Template` to facilitate data access
python
document = hd.sr.ComprehensiveSR(...)
measurement_groups = document.content.get_planar_roi_measurements()


Improvements

* Add unit tests
* Improve description of parameters in docstrings
* Add docstrings to enums
* Add image library entries for SR documents based on TID 1500 (77)

Bug fixes

* Fix many mypy errors
* Record `evidence` correctly (Current Requested Procedure Evidence Sequence versus Pertinent Other Evidence Sequence)
* Fix `highdicom.sr.RelationshipType`
* Fix default value for Pixel Origin Interpretation
* Allow Concept Name Code Sequence to be absent
* Write error traceback to stderr instead of stdout

Major changes

* Change implementation of `highdicom.sr.QualitativeEvaluation` (is now derived from `highdicom.sr.Template` rather than `highdicom.sr.ContentItem` for consistency with `highdicom.sr.Measurement`)
* Make `value` and `unit` parameters of `highdicom.sr.NumContentItem` required (the corresponding attributes are required by the standard)
* Change fields of `highdicom.PatientSexValues` to single letters (for consistency with other enums)
* Add references to Current Requested Procedure Evidence Sequence rather than Pertinent Other Evidence Sequence if instances are referenced in the SR document content tree

Bug fixes

* Fix Dimension Organization of Segmentation instances (92).

Major changes

* Deprecate `Segmentation.add_segments()` method (92). All segments now need to be passed to the constructor method. See more detailed release notes about this issue [here](https://highdicom.readthedocs.io/en/latest/release_notes.html#deprecation-of-add-segments-method).

Improvements

* Improve import logic for subpackages and modules to allow:
python
import highdicom as hd
seg_image = hd.seg.Segmentation(...)

* Update API documentation to reflect namespaces of public API
* Add unit tests and increase test coverage

Improvements

* Increase test coverage of `highdicom.sr.templates` and `highdicom.sr.content` modules
* Add hyperlinks for context groups (CIDs) and templates (TIDs) to docstrings
* Add checks to ensure that argument `segment_number` of `highdicom.seg.sop.Segmentation.add_segments` method is a positive number
* Add checks to ensure that segments added to Segmentation instance via `highdicom.seg.sop.Segmentation.add_segments` are ordered correctly
* Add checks to prevent Content Sequence in Finding Sites from being empty
* Make tests stricter by enforcing valid value representation in pydicom
* Add alternative constructor methods for `highdicom.sr.content` classes

Bug fixes

* Make `algorithm_identification` parameter of `highdicom.seg.content.SegmentDescription` constructor optional to support manual segmentation
* Make `laterality` parameter `highdicom.sc.sop.SCImage` constructor optional
* Fix referencing segments in Planar/Volumetric ROI Measurements and Qualitative Evaluations templates
* Fix Code Value in `highdicom.sr.templates.DeviceObserverIdentifyingAttributes`
* Include documentation of `highdicom.uid` module
* Fix constructor of `highdicom.sr.templates.TimePointContext`