Merge pull request #83 from smasuda/SUPPORT_MULTIPLE_DICOM_VERSIONS

Support multiple dicom versions

Author: pchoisel

README.md

@@ -1,7 +1,7 @@
 # DicomAnonymizer
 
 Python package to anonymize DICOM files.
-The anonymization answer to the standard . More information about dicom fields for anonymization can be found [here](http://dicom.nema.org/dicom/2013/output/chtml/part15/chapter_E.html#table_E.1-1).
+The anonymization answer to the standard . More information about dicom fields for anonymization can be found [here](http://dicom.nema.org/dicom/2023/output/chtml/part15/chapter_E.html#table_E.1-1).
 
 The default behaviour of this package is to anonymize DICOM fields referenced in [dicomfields](dicomanonymizer/dicomfields.py).
 

dicomanonymizer/dicomfields.py renamed to dicomanonymizer/dicom_anonymization_databases/dicomfields_2023.py

@@ -1,6 +1,6 @@
 # Tags anonymized in DICOM standard
 # Documentation for groups meaning can be found in default associated actions.
-# http://dicom.nema.org/dicom/2013/output/chtml/part15/chapter_E.html#table_E.1-1
+# http://dicom.nema.org/dicom/2023/output/chtml/part15/chapter_E.html#table_E.1-1
 
 # Replaced tags
 D_TAGS = [

dicomanonymizer/dicom_anonymization_databases/dicomfields_2024b.py

@@ -0,0 +1,39 @@
+import importlib
+
+ANONYMIZATION_CATEGORIES = [
+    "D_TAGS",
+    "Z_TAGS",
+    "X_TAGS",
+    "U_TAGS",
+    "Z_D_TAGS",
+    "X_Z_TAGS",
+    "X_D_TAGS",
+    "X_Z_D_TAGS",
+    "X_Z_U_STAR_TAGS",
+    "ALL_TAGS",
+]
+
+
+def dicom_anonymization_database_selector(
+    dicom_version: str = "dicomfields_2023",
+) -> dict:
+    try:
+        dicom_anonymization_database = importlib.import_module(
+            f"dicomanonymizer.dicom_anonymization_databases.{dicom_version}"
+        )
+    except ModuleNotFoundError:
+        raise ValueError(f"Unknown DICOM anonymization database: {dicom_version}")
+
+    try:
+        dicom_anonymization_dict = {
+            anonymization_category: getattr(
+                dicom_anonymization_database, anonymization_category
+            )
+            for anonymization_category in ANONYMIZATION_CATEGORIES
+        }
+    except AttributeError:
+        print(
+            f"Anonymization database {dicom_version} is missing a category, please check it has them all."
+        )
+        raise
+    return dicom_anonymization_dict

dicomanonymizer/dicomfields_selector.py

@@ -2,23 +2,13 @@
 import re
 
 from enum import Enum
-from typing import List, Union
+from typing import Callable, List, Union
 from dataclasses import dataclass
 
-from dicomanonymizer.dicomfields import (
-    D_TAGS,
-    Z_TAGS,
-    X_TAGS,
-    U_TAGS,
-    Z_D_TAGS,
-    X_Z_TAGS,
-    X_D_TAGS,
-    X_Z_D_TAGS,
-    X_Z_U_STAR_TAGS,
-)
+from dicomanonymizer.dicomfields_selector import dicom_anonymization_database_selector
 from dicomanonymizer.format_tag import tag_to_hex_strings
 
-
+# keeps the mapping from old UID to new UID
 dictionary = {}
 
 
@@ -344,33 +334,47 @@ class ActionsMapNameFunctions(Enum):
     regexp = Action(regexp, 2)
 
 
-def initialize_actions() -> dict:
+def initialize_actions(dicom_version: str = "dicomfields_2023") -> dict:
     """
     Initialize anonymization actions with DICOM standard values
 
+    :param dicom_version: DICOM version to use
     :return Dict object which map actions to tags
     """
-    anonymization_actions = {tag: replace for tag in D_TAGS}
-    anonymization_actions.update({tag: empty for tag in Z_TAGS})
-    anonymization_actions.update({tag: delete for tag in X_TAGS})
-    anonymization_actions.update({tag: replace_UID for tag in U_TAGS})
-    anonymization_actions.update({tag: empty_or_replace for tag in Z_D_TAGS})
-    anonymization_actions.update({tag: delete_or_empty for tag in X_Z_TAGS})
-    anonymization_actions.update({tag: delete_or_replace for tag in X_D_TAGS})
+    tags = dicom_anonymization_database_selector(dicom_version)
+
+    anonymization_actions = {tag: replace for tag in tags["D_TAGS"]}
+    anonymization_actions.update({tag: empty for tag in tags["Z_TAGS"]})
+    anonymization_actions.update({tag: delete for tag in tags["X_TAGS"]})
+    anonymization_actions.update({tag: replace_UID for tag in tags["U_TAGS"]})
+    anonymization_actions.update({tag: empty_or_replace for tag in tags["Z_D_TAGS"]})
+    anonymization_actions.update({tag: delete_or_empty for tag in tags["X_Z_TAGS"]})
+    anonymization_actions.update({tag: delete_or_replace for tag in tags["X_D_TAGS"]})
     anonymization_actions.update(
-        {tag: delete_or_empty_or_replace for tag in X_Z_D_TAGS}
+        {tag: delete_or_empty_or_replace for tag in tags["X_Z_D_TAGS"]}
     )
     anonymization_actions.update(
-        {tag: delete_or_empty_or_replace_UID for tag in X_Z_U_STAR_TAGS}
+        {tag: delete_or_empty_or_replace_UID for tag in tags["X_Z_U_STAR_TAGS"]}
     )
     return anonymization_actions
 
 
+def initialize_actions_2024b() -> dict:
+    """
+    Initialize anonymization actions with DICOM standard values of 2024b.
+    If you want to use 2024b version of anonymization, call anonymize_dataset with base_rules_gen=initialize_actions_2024b.
+
+    :return Dict object which map actions to tags
+    """
+    return initialize_actions("dicomfields_2024b")
+
+
 def anonymize_dicom_file(
     in_file: str,
     out_file: str,
     extra_anonymization_rules: dict = None,
     delete_private_tags: bool = True,
+    base_rules_gen: Callable = initialize_actions,
 ) -> None:
     """
     Anonymize a DICOM file by modifying personal tags
@@ -384,7 +388,9 @@ def anonymize_dicom_file(
     """
     dataset = pydicom.dcmread(in_file)
 
-    anonymize_dataset(dataset, extra_anonymization_rules, delete_private_tags)
+    anonymize_dataset(
+        dataset, extra_anonymization_rules, delete_private_tags, base_rules_gen
+    )
 
     # Store modified image
     dataset.save_as(out_file)
@@ -450,15 +456,17 @@ def anonymize_dataset(
     dataset: pydicom.Dataset,
     extra_anonymization_rules: dict = None,
     delete_private_tags: bool = True,
+    base_rules_gen: Callable = initialize_actions,
 ) -> None:
     """
     Anonymize a pydicom Dataset by using anonymization rules which links an action to a tag
 
     :param dataset: Dataset to be anonymize
+    :param base_rules_gen: Function to generate the base rules
     :param extra_anonymization_rules: Rules to be applied on the dataset
     :param delete_private_tags: Define if private tags should be delete or not
     """
-    current_anonymization_actions = initialize_actions()
+    current_anonymization_actions = base_rules_gen()
 
     if extra_anonymization_rules is not None:
         current_anonymization_actions.update(extra_anonymization_rules)

dicomanonymizer/simpledicomanonymizer.py

@@ -1,6 +1,6 @@
 import argparse
 
-from dicomanonymizer.dicomfields import ALL_TAGS
+from dicomanonymizer.dicom_anonymization_databases.dicomfields_2023 import ALL_TAGS
 from dicomanonymizer import anonymize, keep
 
 

examples/anonymize_extra_rules.py

@@ -115,8 +115,9 @@ def create_DICOM_fields(profiles):
 
 def main(
     url="https://dicom.nema.org/medical/dicom/current/output/chtml/part15/chapter_e.html",
-    output_path="dicomanonymizer/dicomfields.py",
+    output_path="dicomanonymizer/dicomfields_2024b.py",
 ):
+    # As of 2024.05.14, the current version of DICOM spec is 2024b.
     profiles = scrap_profiles(url)
     file_content = create_DICOM_fields(profiles=profiles)
     with open(output_path, "w") as file:

scripts/scrap_DICOM_fields.py

@@ -8,7 +8,7 @@
 from pydicom.data import get_testdata_files
 
 from dicomanonymizer.simpledicomanonymizer import anonymize_dataset
-from dicomanonymizer import dicomfields
+from dicomanonymizer.dicom_anonymization_databases import dicomfields_2023
 
 # Ignore warnings from pydicom validation
 settings.writing_validation_mode = IGNORE
@@ -51,7 +51,7 @@ def orig_anon_dataset(request):
 
 def test_deleted_tags_are_removed(orig_anon_dataset):
     orig_ds, anon_ds = orig_anon_dataset
-    deleted_tags = dicomfields.X_TAGS
+    deleted_tags = dicomfields_2023.X_TAGS
 
     for tt in deleted_tags:  # sourcery skip: no-loop-in-tests
         if (
@@ -66,12 +66,12 @@ def test_deleted_tags_are_removed(orig_anon_dataset):
 
 
 changed_tags = (
-    dicomfields.U_TAGS
-    + dicomfields.D_TAGS
-    + dicomfields.Z_D_TAGS
-    + dicomfields.X_D_TAGS
-    + dicomfields.X_Z_D_TAGS
-    + dicomfields.X_Z_U_STAR_TAGS
+    dicomfields_2023.U_TAGS
+    + dicomfields_2023.D_TAGS
+    + dicomfields_2023.Z_D_TAGS
+    + dicomfields_2023.X_D_TAGS
+    + dicomfields_2023.X_Z_D_TAGS
+    + dicomfields_2023.X_Z_U_STAR_TAGS
 )
 
 empty_values = (0, "", "00010101", "000000.00", "ANONYMIZED")
@@ -104,7 +104,7 @@ def test_changed_tags_are_replaced(orig_anon_dataset):
             ), f"({tt[0]:04X},{tt[1]:04x}):{orig_ds[tt].value} not replaced"
 
 
-empty_tags = dicomfields.Z_TAGS + dicomfields.X_Z_TAGS
+empty_tags = dicomfields_2023.Z_TAGS + dicomfields_2023.X_Z_TAGS
 
 
 def is_elem_empty(elem) -> bool:

tests/test_anon.py

@@ -1,7 +1,11 @@
 import pydicom
 
 from dicomanonymizer import anonymize_dataset
-from dicomanonymizer.simpledicomanonymizer import empty
+from dicomanonymizer.simpledicomanonymizer import (
+    empty,
+    initialize_actions,
+    initialize_actions_2024b,
+)
 
 
 def test_anonymization_without_dicom_file():
@@ -74,3 +78,36 @@ def test_anonymization_of_ranged_tags_without_dicom_file():
     # Check that the dataset has been anonymized
     assert (0x5011, 0x0110) not in anon_ds
     assert (0x5012, 0x0112) not in anon_ds
+
+
+def test_switching_dicom_versions():
+    """To confirm the different behavior of annonymization beteen dicom versions of 2023 and 2024b."""
+    fields = [
+        {  # Replaced by Anonymized
+            "id": (0x0010, 0x0020),
+            "type": "LO",
+            "value": "Test Patient ID",
+        },
+    ]
+
+    # Create a readable dataset for pydicom
+    data = pydicom.Dataset()
+    data_2023 = pydicom.Dataset()
+    data_2024b = pydicom.Dataset()
+
+    for field in fields:  # sourcery skip: no-loop-in-tests
+        data.add_new(field["id"], field["type"], field["value"])
+        data_2023.add_new(field["id"], field["type"], field["value"])
+        data_2024b.add_new(field["id"], field["type"], field["value"])
+
+    anonymize_dataset(data, base_rules_gen=initialize_actions)
+    anonymize_dataset(
+        data_2023, base_rules_gen=lambda: initialize_actions("dicomfields_2023")
+    )
+    anonymize_dataset(data_2024b, base_rules_gen=initialize_actions_2024b)
+
+    assert data[(0x0010, 0x0020)].value == ""  # default behavior which is DICOM 2023.
+    assert data_2023[(0x0010, 0x0020)].value == ""  # same as the default.
+    assert (
+        data_2024b[(0x0010, 0x0020)].value == "ANONYMIZED"
+    )  # 2024b differs from the default

tests/test_anonymization_without_dicom.py

@@ -0,0 +1,44 @@
+from dicomanonymizer.dicom_anonymization_databases import dicomfields_2024b
+from dicomanonymizer.dicomfields_selector import dicom_anonymization_database_selector
+
+from dicomanonymizer.dicom_anonymization_databases import dicomfields_2023
+
+
+def test_selector():
+    assert dicom_anonymization_database_selector("dicomfields_2023") == {
+        "D_TAGS": dicomfields_2023.D_TAGS,
+        "Z_TAGS": dicomfields_2023.Z_TAGS,
+        "X_TAGS": dicomfields_2023.X_TAGS,
+        "U_TAGS": dicomfields_2023.U_TAGS,
+        "Z_D_TAGS": dicomfields_2023.Z_D_TAGS,
+        "X_Z_TAGS": dicomfields_2023.X_Z_TAGS,
+        "X_D_TAGS": dicomfields_2023.X_D_TAGS,
+        "X_Z_D_TAGS": dicomfields_2023.X_Z_D_TAGS,
+        "X_Z_U_STAR_TAGS": dicomfields_2023.X_Z_U_STAR_TAGS,
+        "ALL_TAGS": dicomfields_2023.ALL_TAGS,
+    }
+    assert dicom_anonymization_database_selector("dicomfields_2024b") == {
+        "D_TAGS": dicomfields_2024b.D_TAGS,
+        "Z_TAGS": dicomfields_2024b.Z_TAGS,
+        "X_TAGS": dicomfields_2024b.X_TAGS,
+        "U_TAGS": dicomfields_2024b.U_TAGS,
+        "Z_D_TAGS": dicomfields_2024b.Z_D_TAGS,
+        "X_Z_TAGS": dicomfields_2024b.X_Z_TAGS,
+        "X_D_TAGS": dicomfields_2024b.X_D_TAGS,
+        "X_Z_D_TAGS": dicomfields_2024b.X_Z_D_TAGS,
+        "X_Z_U_STAR_TAGS": dicomfields_2024b.X_Z_U_STAR_TAGS,
+        "ALL_TAGS": dicomfields_2024b.ALL_TAGS,
+    }
+
+    # check default selector
+    assert (
+        dicom_anonymization_database_selector()
+        == dicom_anonymization_database_selector("dicomfields_2023")
+    )
+
+    try:
+        dicom_anonymization_database_selector("2019")
+    except ValueError as e:
+        assert str(e) == "Unknown DICOM anonymization database: 2019"
+    else:
+        assert False