From 43038ecbcfe3c6b7e56db12dbf5e0e7a0fe1629b Mon Sep 17 00:00:00 2001
From: Tom Close <tom.g.close@gmail.com>
Date: Sat, 28 Sep 2024 10:02:20 +1000
Subject: [PATCH] added MedicalImagingData base class

---
 .../extras/medimage/tests/test_deidentify.py  |  4 +++
 fileformats/medimage/__init__.py              |  3 +-
 fileformats/medimage/base.py                  | 32 +++++++++++--------
 fileformats/medimage/raw/pet/base.py          |  3 +-
 4 files changed, 27 insertions(+), 15 deletions(-)

diff --git a/extras/fileformats/extras/medimage/tests/test_deidentify.py b/extras/fileformats/extras/medimage/tests/test_deidentify.py
index 16bc0b3..241e3bf 100644
--- a/extras/fileformats/extras/medimage/tests/test_deidentify.py
+++ b/extras/fileformats/extras/medimage/tests/test_deidentify.py
@@ -21,8 +21,12 @@ def dicom(request):
 
 def test_deidentify_dicom(dicom):
     assert str(dicom.metadata["PatientName"]) == "Doe^John"
+    assert dicom.metadata["InstitutionAddress"]
+    assert not dicom.metadata["PatientBirthDate"].endswith("0101")
     deidentified = dicom.deidentify()
     assert str(deidentified.metadata["PatientName"]) == "Anonymous^Anonymous"
+    assert deidentified.metadata["InstitutionAddress"] == ""
+    assert deidentified.metadata["PatientBirthDate"] == "19800101"
 
 
 def test_nifti_deidentify():
diff --git a/fileformats/medimage/__init__.py b/fileformats/medimage/__init__.py
index d4eeab1..5df25f4 100644
--- a/fileformats/medimage/__init__.py
+++ b/fileformats/medimage/__init__.py
@@ -1,5 +1,5 @@
 from ._version import __version__
-from .base import MedicalImage
+from .base import MedicalImagingData, MedicalImage
 
 
 from .misc import (
@@ -128,6 +128,7 @@
 
 __all__ = [
     "__version__",
+    "MedicalImagingData",
     "MedicalImage",
     "DicomImage",
     "Analyze",
diff --git a/fileformats/medimage/base.py b/fileformats/medimage/base.py
index 0e8d89f..43a43b0 100644
--- a/fileformats/medimage/base.py
+++ b/fileformats/medimage/base.py
@@ -33,7 +33,25 @@
 )  # In Py<3.9 this is problematic "numpy.typing.NDArray[typing.Union[numpy.floating[typing.Any], numpy.integer[typing.Any]]]"
 
 
-class MedicalImage(WithClassifiers, FileSet):
+class MedicalImagingData(FileSet):
+    """Base class for all medical imaging data including pre-image raw data and
+    associated data"""
+
+    contains_phi: bool = True
+
+    @extra
+    def deidentify(
+        self,
+        out_dir: ty.Optional[Path] = None,
+        new_stem: ty.Optional[str] = None,
+        copy_mode: FileSet.CopyMode = FileSet.CopyMode.copy,
+    ) -> Self:
+        """Returns a new copy of the image with any subject-identifying information
+        stripped from the from the image header"""
+        raise NotImplementedError
+
+
+class MedicalImage(WithClassifiers, MedicalImagingData):
 
     INCLUDE_HDR_KEYS: ty.Optional[ty.Tuple[str, ...]] = None
     IGNORE_HDR_KEYS: ty.Optional[ty.Tuple[str, ...]] = None
@@ -42,7 +60,6 @@ class MedicalImage(WithClassifiers, FileSet):
     image_contents = ()
     allowed_classifiers = (ContentsClassifier,)
     exclusive_classifiers = (ImagingModality, AnatomicalEntity, Derivative)
-    contains_phi: bool = True
 
     @extra
     def read_array(self) -> DataArrayType:
@@ -63,14 +80,3 @@ def vox_sizes(self) -> ty.Tuple[float, float, float]:
     def dims(self) -> ty.Tuple[int, int, int]:
         """The dimensions of the image"""
         raise NotImplementedError
-
-    @extra
-    def deidentify(
-        self,
-        out_dir: ty.Optional[Path] = None,
-        new_stem: ty.Optional[str] = None,
-        copy_mode: FileSet.CopyMode = FileSet.CopyMode.copy,
-    ) -> Self:
-        """Returns a new copy of the image with any subject-identifying information
-        stripped from the from the image header"""
-        raise NotImplementedError
diff --git a/fileformats/medimage/raw/pet/base.py b/fileformats/medimage/raw/pet/base.py
index 407fe8c..1ad0ef8 100644
--- a/fileformats/medimage/raw/pet/base.py
+++ b/fileformats/medimage/raw/pet/base.py
@@ -2,6 +2,7 @@
 import typing as ty
 from pathlib import Path
 from fileformats.core import FileSet, extra
+from fileformats.medimage import MedicalImagingData
 from fileformats.generic import BinaryFile
 
 if sys.version_info >= (3, 12):
@@ -10,7 +11,7 @@
     from typing_extensions import Self
 
 
-class PetRawData(BinaryFile):
+class PetRawData(BinaryFile, MedicalImagingData):
     """Base class for raw PET data files"""
 
     @extra