Merge pull request #60 from radexperts/dicognito-0.17.0

Upgrade dicognito to v0.17.0
openradx · Nov 20, 2023 · b09682e · b09682e
2 parents adb867a + 7d881d6
commit b09682e
Show file tree

Hide file tree

Showing 5 changed files with 39 additions and 46 deletions.
diff --git a/KNOWLEDGE.md b/KNOWLEDGE.md
@@ -34,6 +34,8 @@
 
 ## DICOM
 
+- All available DICOM tags: <https://dicom.nema.org/medical/dicom/current/output/chtml/part06/chapter_6.html>
+
 ### pydicom and datetime
 
 - To automatically convert dates to the datetime.date class this config must be set explicitly (default is False): config.datetime_conversion = True

diff --git a/adit/core/utils/transfer_utils.py b/adit/core/utils/transfer_utils.py
@@ -5,9 +5,10 @@
 from datetime import datetime
 from functools import partial
 from pathlib import Path
-from typing import Any, Callable
+from typing import Callable
 
 from dicognito.anonymizer import Anonymizer
+from dicognito.value_keeper import ValueKeeper
 from django.conf import settings
 from pydicom import Dataset
 
@@ -94,6 +95,15 @@ def _transfer_to_folder(self) -> None:
         download_folder = dicom_folder / self._create_destination_name()
         self._download_dicoms(download_folder)
 
+    def _create_destination_name(self) -> str:
+        transfer_job = self.transfer_task.job
+        name = "adit_"
+        name += transfer_job._meta.app_label + "_"
+        name += str(transfer_job.id) + "_"
+        name += transfer_job.created.strftime("%Y%m%d") + "_"
+        name += transfer_job.owner.username
+        return sanitize_dirname(name)
+
     def _download_dicoms(
         self,
         download_folder: Path,
@@ -134,7 +144,8 @@ def _download_dicoms(
         study_folder = patient_folder / f"{prefix}-{modalities}"
         os.makedirs(study_folder, exist_ok=True)
 
-        anonymizer = Anonymizer()
+        anonymizer = self._setup_anonymizer()
+
         modifier = partial(
             self._modify_dataset,
             anonymizer,
@@ -159,15 +170,6 @@ def _download_dicoms(
 
         return patient_folder
 
-    def _create_destination_name(self) -> str:
-        transfer_job = self.transfer_task.job
-        name = "adit_"
-        name += transfer_job._meta.app_label + "_"
-        name += str(transfer_job.id) + "_"
-        name += transfer_job.created.strftime("%Y%m%d") + "_"
-        name += transfer_job.owner.username
-        return sanitize_dirname(name)
-
     def _fetch_patient(self) -> ResultDataset:
         patients = list(
             self.source_operator.find_patients(
@@ -228,6 +230,12 @@ def _fetch_series_list(self, series_uid: str) -> list[ResultDataset]:
 
         return results
 
+    def _setup_anonymizer(self) -> Anonymizer:
+        anonymizer = Anonymizer()
+        for element in settings.SKIP_ELEMENTS_ANONYMIZATION:
+            anonymizer.add_element_handler(ValueKeeper(element))
+        return anonymizer
+
     def _download_study(
         self,
         patient_id: str,
@@ -262,38 +270,8 @@ def _modify_dataset(
         """Optionally pseudonymize an incoming dataset with the given pseudonym
         and add the trial ID and name to the DICOM header if specified."""
         if pseudonym:
-            # All dates get anonymized by dicognito, but we want to retain some dates.
-            # For that we save them here to restore them after anonymization as dicognito
-            # currently does not provide that functionality.
-            # TODO: Use dicognito for this when issue is fixed:
-            # https://github.com/blairconrad/dicognito/issues/155
-            # TODO: Make this configurable.
-            attributes_to_retain = [
-                "StudyDate",
-                "StudyTime",
-                "SeriesDate",
-                "SeriesTime",
-                "AcquisitionDate",
-                "AcquisitionTime",
-                "ImageDate",
-                "ImageTime",
-                "ContentDate",
-                "ContentTime",
-                "DateOfLastCalibration",
-                "TimeOfLastCalibration",
-                "AcquisitionDateTime",
-            ]
-            saved_attributes: dict[str, Any] = {}
-            for attr in attributes_to_retain:
-                value = ds.get(attr)
-                if value is not None:
-                    saved_attributes[attr] = value
-
             anonymizer.anonymize(ds)
 
-            for attr in saved_attributes:
-                setattr(ds, attr, saved_attributes[attr])
-
             ds.PatientID = pseudonym
             ds.PatientName = pseudonym
 

diff --git a/adit/settings/base.py b/adit/settings/base.py
@@ -422,3 +422,16 @@
 # DicomWeb Settings
 DEFAULT_BOUNDARY = "adit-boundary"
 ERROR_MESSAGE = "Processing your DicomWeb request failed."
+
+# Elements to keep during pseudonymization
+SKIP_ELEMENTS_ANONYMIZATION = [
+    "AcquisitionDate",
+    "AcquisitionDateTime",
+    "AcquisitionTime",
+    "ContentDate",
+    "ContentTime",
+    "SeriesDate",
+    "SeriesTime",
+    "StudyDate",
+    "StudyTime",
+]
diff --git a/poetry.lock b/poetry.lock
diff --git a/pyproject.toml b/pyproject.toml
@@ -14,7 +14,7 @@ channels = "^4.0.0"
 channels-redis = "^4.0.0"
 cryptography = "^41.0.1"
 daphne = "^4.0.0"
-dicognito = "^0.16.0"
+dicognito = "^0.17.0"
 dicomweb-client = "^0.59.1"
 Django = "^4.0.6"
 django-crispy-forms = "^2.0"