GANDLF 0.1.4.dev20250609__py3-none-any.whl → 0.1.4.dev20250611__py3-none-any.whl

GANDLF/cli/generate_metrics.py

@@ -10,7 +10,11 @@ import SimpleITK as sitk
 import numpy as np
 
 from GANDLF.config_manager import ConfigManager
-from GANDLF.utils import find_problem_type_from_parameters, one_hot
+from GANDLF.utils import (
+    find_problem_type_from_parameters,
+    one_hot,
+    sanity_check_on_file_readers,
+)
 from GANDLF.metrics import (
     overall_stats,
     structural_similarity_index,
@@ -176,6 +180,9 @@ def generate_metrics_dict(
         for _, row in tqdm(input_df.iterrows(), total=input_df.shape[0]):
             current_subject_id = row["SubjectID"]
             overall_stats_dict[current_subject_id] = {}
+            sanity_check_on_file_readers(
+                row["Target"], row["Prediction"], current_subject_id
+            )
             label_image = torchio.LabelMap(row["Target"])
             pred_image = torchio.LabelMap(row["Prediction"])
             label_tensor = label_image.data
@@ -264,20 +271,23 @@ def generate_metrics_dict(
         for _, row in tqdm(input_df.iterrows(), total=input_df.shape[0]):
             current_subject_id = row["SubjectID"]
             overall_stats_dict[current_subject_id] = {}
+            sanity_check_on_file_readers(
+                row["Target"], row["Prediction"], current_subject_id
+            )
             label_image = torchio.LabelMap(row["Target"])
             pred_image = torchio.LabelMap(row["Prediction"])
             label_tensor = label_image.data
             pred_tensor = pred_image.data
             spacing = label_image.spacing
-            if label_tensor.data.shape[-1] == 1:
-                spacing = spacing[0:2]
-            # add dimension for batch
-            parameters["subject_spacing"] = torch.Tensor(spacing).unsqueeze(0)
-            label_array = label_tensor.unsqueeze(0).numpy()
-            pred_array = pred_tensor.unsqueeze(0).numpy()
+            # if label_tensor.data.shape[-1] == 1:
+            #     spacing = spacing[0:2]
+            # remove dimension to ensure 3D tensors
+            if label_tensor.data.ndim == 4:
+                label_array = label_tensor.squeeze(0).numpy().astype(int)
+                pred_array = pred_tensor.squeeze(0).numpy().astype(int)
 
             overall_stats_dict[current_subject_id] = generate_instance_segmentation(
-                prediction=pred_array, target=label_array
+                prediction=pred_array, target=label_array, parameters=parameters
             )
 
     elif problem_type == "synthesis":
@@ -340,6 +350,9 @@ def generate_metrics_dict(
         for _, row in tqdm(input_df.iterrows(), total=input_df.shape[0]):
             current_subject_id = row["SubjectID"]
             overall_stats_dict[current_subject_id] = {}
+            sanity_check_on_file_readers(
+                row["Target"], row["Prediction"], current_subject_id
+            )
             target_image = __fix_2d_tensor(torchio.ScalarImage(row["Target"]).data)
             pred_image = __fix_2d_tensor(torchio.ScalarImage(row["Prediction"]).data)
             # if "Mask" is not in the row, we assume that the whole image is the mask

GANDLF/metrics/panoptica_config_brats.yaml

@@ -13,7 +13,10 @@ edge_case_handler: !EdgeCaseHandler
     !Metric IOU: !MetricZeroTPEdgeCaseHandling {empty_prediction_result: !EdgeCaseResult ZERO,
       empty_reference_result: !EdgeCaseResult ZERO, no_instances_result: !EdgeCaseResult NAN,
       normal: !EdgeCaseResult ZERO}
-    !Metric ASSD: !MetricZeroTPEdgeCaseHandling {empty_prediction_result: !EdgeCaseResult INF,
+    !Metric NSD: !MetricZeroTPEdgeCaseHandling {empty_prediction_result: !EdgeCaseResult INF,
+      empty_reference_result: !EdgeCaseResult INF, no_instances_result: !EdgeCaseResult NAN,
+      normal: !EdgeCaseResult INF}
+    !Metric HD95: !MetricZeroTPEdgeCaseHandling {empty_prediction_result: !EdgeCaseResult INF,
       empty_reference_result: !EdgeCaseResult INF, no_instances_result: !EdgeCaseResult NAN,
       normal: !EdgeCaseResult INF}
     !Metric RVD: !MetricZeroTPEdgeCaseHandling {empty_prediction_result: !EdgeCaseResult NAN,
@@ -27,24 +30,27 @@ global_metrics: [!Metric DSC]
 instance_approximator: !ConnectedComponentsInstanceApproximator {cca_backend: null}
 instance_matcher: !NaiveThresholdMatching {allow_many_to_one: false, matching_metric: !Metric IOU,
   matching_threshold: 0.5}
-instance_metrics: [!Metric DSC, !Metric IOU, !Metric ASSD, !Metric RVD]
+instance_metrics: [!Metric DSC, !Metric IOU, !Metric RVD, !Metric NSD, !Metric HD95]
 log_times: false
 save_group_times: false
 segmentation_class_groups: !SegmentationClassGroups
   groups:
-    ed: !LabelGroup
+    snfh: !LabelGroup
       single_instance: false
       value_labels: [2]
     et: !LabelGroup
       single_instance: false
       value_labels: [3]
-    net: !LabelGroup
+    netc: !LabelGroup
       single_instance: false
       value_labels: [1]
+    rc: !LabelGroup
+      single_instance: false
+      value_labels: [4]
     tc: !LabelMergeGroup
       single_instance: false
-      value_labels: [1, 3]
+      value_labels: [1, 3, 4]
     wt: !LabelMergeGroup
       single_instance: false
-      value_labels: [1, 2, 3]
+      value_labels: [1, 2, 3, 4]
 verbose: false

GANDLF/metrics/segmentation_panoptica.py

@@ -1,4 +1,5 @@
 from pathlib import Path
+import tempfile
 
 import numpy as np
 
@@ -6,7 +7,10 @@ from panoptica import Panoptica_Evaluator
 
 
 def generate_instance_segmentation(
-    prediction: np.ndarray, target: np.ndarray, panoptica_config_path: str = None
+    prediction: np.ndarray,
+    target: np.ndarray,
+    parameters: dict = None,
+    panoptica_config_path: str = None,
 ) -> dict:
     """
     Evaluate a single exam using Panoptica.
@@ -14,6 +18,7 @@ def generate_instance_segmentation(
     Args:
         prediction (np.ndarray): The input prediction containing objects.
         label_path (str): The path to the reference label.
+        target (np.ndarray): The input target containing objects.
         panoptica_config_path (str): The path to the Panoptica configuration file.
 
     Returns:
@@ -21,11 +26,21 @@ def generate_instance_segmentation(
     """
 
     cwd = Path(__file__).parent.absolute()
-    panoptica_config_path = (
-        cwd / "panoptica_config_brats.yaml"
-        if panoptica_config_path is None
-        else panoptica_config_path
-    )
+    # the parameters dict takes precedence over the panoptica_config_path
+    panoptica_config = parameters.get("panoptica_config", None)
+    if panoptica_config is None:
+        panoptica_config_path = (
+            cwd / "panoptica_config_brats.yaml"
+            if panoptica_config_path is None
+            else panoptica_config_path
+        )
+    else:
+        # write the panoptica config to a file
+        panoptica_config_path = tempfile.NamedTemporaryFile(
+            mode="w", delete=False, suffix=".yaml"
+        ).name
+        with open(panoptica_config_path, "w") as f:
+            f.write(panoptica_config)
     evaluator = Panoptica_Evaluator.load_from_config(panoptica_config_path)
 
     # call evaluate

GANDLF/utils/__init__.py

@@ -7,6 +7,7 @@ from .imaging import (
     resize_image,
     resample_image,
     perform_sanity_check_on_subject,
+    sanity_check_on_file_readers,
     write_training_patches,
     get_correct_padding_size,
     applyCustomColorMap,

GANDLF/utils/imaging.py

@@ -123,6 +123,56 @@ def softer_sanity_check(
     return result
 
 
+def sanity_check_on_file_readers(
+    file_reader_0: Union[str, sitk.ImageFileReader],
+    file_reader_1: Union[str, sitk.ImageFileReader],
+    subject_id: Optional[str] = "",
+    threshold: Optional[float] = 0.00001,
+) -> bool:
+    """
+    This function performs a sanity check on the file readers to ensure that the properties of the images are consistent.
+
+    Args:
+        file_reader_0 (Union[str, sitk.ImageFileReader]): The first file reader.
+        file_reader_1 (Union[str, sitk.ImageFileReader]): The second file reader.
+        threshold (Optional[float], optional): The threshold for comparison. Defaults to 0.00001.
+
+    Returns:
+        bool: True if the sanity check passes.
+    """
+
+    if isinstance(file_reader_0, str):
+        temp_file = file_reader_0
+        file_reader_0 = sitk.ImageFileReader()
+        file_reader_0.SetFileName(temp_file)
+        file_reader_0.ReadImageInformation()
+
+    if isinstance(file_reader_1, str):
+        temp_file = file_reader_1
+        file_reader_1 = sitk.ImageFileReader()
+        file_reader_1.SetFileName(temp_file)
+        file_reader_1.ReadImageInformation()
+    # this check needs to be absolute
+    assert file_reader_0.GetDimension() == file_reader_1.GetDimension(), (
+        "Dimensions for Subject '" + subject_id + "' are not consistent."
+    )
+
+    # other checks can be softer
+    assert softer_sanity_check(
+        file_reader_0.GetOrigin(), file_reader_1.GetOrigin(), threshold=threshold
+    ), ("Origin for Subject '" + subject_id + "' are not consistent.")
+
+    assert softer_sanity_check(
+        file_reader_0.GetDirection(), file_reader_1.GetDirection(), threshold=threshold
+    ), ("Orientation for Subject '" + subject_id + "' are not consistent.")
+
+    assert softer_sanity_check(
+        file_reader_0.GetSpacing(), file_reader_1.GetSpacing(), threshold=threshold
+    ), ("Spacing for Subject '" + subject_id + "' are not consistent.")
+
+    return True
+
+
 def perform_sanity_check_on_subject(subject: torchio.Subject, parameters: dict) -> bool:
     """
     This function performs a sanity check on the image modalities in input subject to ensure that they are consistent.
@@ -169,39 +219,11 @@ def perform_sanity_check_on_subject(subject: torchio.Subject, parameters: dict)
             else:
                 file_reader_current = _get_itkimage_or_filereader(subject[str(key)])
 
-                # this check needs to be absolute
-                assert (
-                    file_reader_base.GetDimension()
-                    == file_reader_current.GetDimension()
-                ), (
-                    "Dimensions for Subject '"
-                    + subject["subject_id"]
-                    + "' are not consistent."
-                )
-
-                # other checks can be softer
-                assert softer_sanity_check(
-                    file_reader_base.GetOrigin(), file_reader_current.GetOrigin()
-                ), (
-                    "Origin for Subject '"
-                    + subject["subject_id"]
-                    + "' are not consistent."
-                )
-
-                assert softer_sanity_check(
-                    file_reader_base.GetDirection(), file_reader_current.GetDirection()
-                ), (
-                    "Orientation for Subject '"
-                    + subject["subject_id"]
-                    + "' are not consistent."
-                )
-
-                assert softer_sanity_check(
-                    file_reader_base.GetSpacing(), file_reader_current.GetSpacing()
-                ), (
-                    "Spacing for Subject '"
-                    + subject["subject_id"]
-                    + "' are not consistent."
+                sanity_check_on_file_readers(
+                    file_reader_base,
+                    file_reader_current,
+                    subject_id=subject["subject_id"],
+                    threshold=parameters.get("sanity_check_threshold", 0.00001),
                 )
 
     return True

GANDLF/version.py

@@ -1 +1 @@
-__version__ = "0.1.4-dev20250609"
+__version__ = "0.1.4-dev20250611"

{gandlf-0.1.4.dev20250609.dist-info → gandlf-0.1.4.dev20250611.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: GANDLF
-Version: 0.1.4.dev20250609
+Version: 0.1.4.dev20250611
 Summary: PyTorch-based framework that handles segmentation/regression/classification using various DL architectures for medical imaging.
 Author: MLCommons
 Author-email: gandlf@mlcommons.org

{gandlf-0.1.4.dev20250609.dist-info → gandlf-0.1.4.dev20250611.dist-info}/RECORD

@@ -5,14 +5,14 @@ GANDLF/logger.py,sha256=oamQ1SOTTpAnC8vQ67o211Q6_bExGg_hAuqlHGlJfAI,2951
 GANDLF/logging_config.yaml,sha256=9XxRxAKtLn5ehT1khpR8wEiJGW64sx1lylAspM5KaWk,1337
 GANDLF/parseConfig.py,sha256=jO-ybIPxLw23OWDvFdTukbft0ZM8UOofGnoL2C5CEps,754
 GANDLF/training_manager.py,sha256=AZlf-fl7KAwZgyre8-0M5lAyma6NvtiGX6XT51AJdxU,11436
-GANDLF/version.py,sha256=6EFUxHuz3VK2LWZLY5x_Kd_z9--zoPRJAaUKr9ccvB0,34
+GANDLF/version.py,sha256=IfZFHfeewJgD5R9DgFa2wSz2OhAyTh720DH2OZIIm64,34
 GANDLF/anonymize/__init__.py,sha256=Nxig-jM-a-aKlK09PNi1zhNulEpLTyjnsY_oGQKdjhQ,1953
 GANDLF/anonymize/convert_to_nifti.py,sha256=MOfSDncFGJGb-EQP9sFGn0yuKpX010Ioi2KNwttaex8,1339
 GANDLF/cli/__init__.py,sha256=F05eyL2HKyWkHczRZuPE_Z2Yg685P9ARYxTwz6njGeQ,784
 GANDLF/cli/config_generator.py,sha256=_j0aRV3puiIItJ1WOHRNVII_TkGJuH9ep6SuwPblcnc,4502
 GANDLF/cli/data_split_saver.py,sha256=72ygy9s3INt2NFAfSW4j9dxxmuuvfzfBTF0Hwh0nYBU,1720
 GANDLF/cli/deploy.py,sha256=7yPqRox6e5MB-VzHxmA4kt-W5dWwElxDm1mm019jN74,15155
-GANDLF/cli/generate_metrics.py,sha256=TxKJsDHtGBgKc-Fyx4F_tc_CMu5hSsO8xlc5Te86DGk,19973
+GANDLF/cli/generate_metrics.py,sha256=ySQ5s0sGajXbr__HLOzsDtULhuZ3x7plqaC1Lz1oZto,20429
 GANDLF/cli/huggingface_hub_handler.py,sha256=vGQYwDiT3Y8yk_xLNkB8DgW1rYpEcwYp0TCdCO5oWTs,4631
 GANDLF/cli/main_run.py,sha256=gYNJQGvmwqBjj3ARRxj-y3R1PMOP4qTwcwFSRBbkh0k,4352
 GANDLF/cli/patch_extraction.py,sha256=30RpBfjpwE9GLRSSD7JRqwoHngESQL0eOt09n_5r18Q,3555
@@ -108,10 +108,10 @@ GANDLF/metrics/__init__.py,sha256=k8xcpujpUg22R33keOUNetzrLnGSHsEAmClkFeDHZxc,44
 GANDLF/metrics/classification.py,sha256=6rVZjXxHwfJFqw19UiRMYK3uF656QgfMwWDz-LOHUaw,4692
 GANDLF/metrics/generic.py,sha256=N-h9sDwuAsskK7AdZ2N0otcbFTmKxh9kdyi6MQkDO60,4983
 GANDLF/metrics/metric_calculators.py,sha256=c-NAl84yniTWjMKePBlf0TsCIJnYHODplt8HeEk9Uac,3610
-GANDLF/metrics/panoptica_config_brats.yaml,sha256=W2EBe-NRHNYc7vRZdxnAyKy0SgXQ21F3K3Nzb0EvwE0,2368
+GANDLF/metrics/panoptica_config_brats.yaml,sha256=mTGSQuJoENuy84gZd2tHV0Dfl7GX5nIeE_46q5Vczx4,2685
 GANDLF/metrics/regression.py,sha256=Ca58jo9OL1GdjB2ZvQCcrKWz9F67iY5rcbnQxT-LpcU,4584
 GANDLF/metrics/segmentation.py,sha256=zqWEI6LuphFM0Nsm0sG2XeqAxpNJnoK1Z9q4FJt58c8,25645
-GANDLF/metrics/segmentation_panoptica.py,sha256=ExbobmLhlyKYRKE7zEhZmuHtu5fdIUDs-53K-u4AKX8,1026
+GANDLF/metrics/segmentation_panoptica.py,sha256=aFGF3q_Fnf5vIWaJDMHcs2Az_fb23l2cvvifnCkgPO0,1611
 GANDLF/metrics/synthesis.py,sha256=IfYit-R0O5ZUfmsWfLD6BqcLjjoEiotx0lUdsXWlbao,6730
 GANDLF/models/MSDNet.py,sha256=mzBkw_kQigSDTxaR7tojhdI4ECIQ65i_qiCrNaZmBHI,3442
 GANDLF/models/__init__.py,sha256=3SwskWonRW1AduPe6CYQWucU8ZBkkfJUpEF34I0NWP0,4286
@@ -169,22 +169,22 @@ GANDLF/privacy/opacus/training_utils.py,sha256=ejXUIhLNjRzTebF-7uCYUluK7jyINDOpL
 GANDLF/schedulers/__init__.py,sha256=jMTea7kX7zawFrtjVDKHDLIfBgt7GD-G2z9_eKjI1_A,1320
 GANDLF/schedulers/wrap_monai.py,sha256=_kP3ArUIFK3x3u0nQV8Gi2oPruMIGGTSZxa4bSygGPs,400
 GANDLF/schedulers/wrap_torch.py,sha256=NuCAmYMKgYeFzO1PEiHfnFaHpzuQU0DsK5NOi-gVxOo,6746
-GANDLF/utils/__init__.py,sha256=QpNBl-fTrFUDyChAnaidPXdT-pY080pMG9kUTrlJ9OY,1691
+GANDLF/utils/__init__.py,sha256=F74QIQ6dZL8cOIsZkU-VfDdcHewUW959GXRvDfphGCk,1725
 GANDLF/utils/data_splitter.py,sha256=IBVsLYo4y28G6LedbLPYWyiAUiaLxD4wWZD4EZC2D6I,10174
 GANDLF/utils/exceptions.py,sha256=SmEGzCbc5mXxjylmA9XE4PXZdmkkfOjrEh473cWTAvA,100
 GANDLF/utils/gandlf_logging.py,sha256=a1yV4244BGO5gp4OuAIKJUZgR21U1OPdovJ5nH1uUrs,2162
 GANDLF/utils/generic.py,sha256=NyJgaeCdryBlHMziPPLmWtIk450u042NWQgs_qh7knw,11744
 GANDLF/utils/handle_collisions.py,sha256=UFgfrddQqeipOdpKk0Mo-wwJ-TYTESNZyXzEyQeaLdI,3082
-GANDLF/utils/imaging.py,sha256=FZYtRXMHYUkL_7bzVtwsnrYyrYAm6Hvs5R-tFnEbYkY,15083
+GANDLF/utils/imaging.py,sha256=_qCgvIXHZW_k_IUYLU95G5L5yoozxCRCYxo65TI-95U,15983
 GANDLF/utils/modelbase.py,sha256=MJ1ufJGQ-ZgfYWTg7o4f-Iz78d7SYiyxN7MT8bnNDaw,950
 GANDLF/utils/modelio.py,sha256=9koLUn3wCODtNOYA2abtxkqxfLgEG07Bhh-NYIrtWOc,8084
 GANDLF/utils/parameter_processing.py,sha256=DA7ZEsizWWLJZnCxBnDNh1NyA-bw5oirOvodiZZWbIk,5675
 GANDLF/utils/pred_target_processors.py,sha256=aatXJ6jdJaNAHa_tPzHfC1gOQrPYJLtg-cYeUFvkM_s,2701
 GANDLF/utils/tensor.py,sha256=AOwNTQfw9mnsomGwOF2Q_rdDS1WANlIatB0hhZN0gSg,21504
 GANDLF/utils/write_parse.py,sha256=HROxskhet-uIdUJmz16z_7p9r0mf8hWsqQFygwZ8ap0,9202
-gandlf-0.1.4.dev20250609.dist-info/licenses/LICENSE,sha256=GlZPAfA4eckod8IVayhBXkqCpESXf6cc9BGli_Jwims,11357
-gandlf-0.1.4.dev20250609.dist-info/METADATA,sha256=CmAfClLNUVdbXqogPnBI4TSWilLvYzPzS_fzMUhHOgY,9932
-gandlf-0.1.4.dev20250609.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-gandlf-0.1.4.dev20250609.dist-info/entry_points.txt,sha256=agwocNI7Upi-sFDe1rMl71dGN8YhCBB7WJmtBHRF4jg,902
-gandlf-0.1.4.dev20250609.dist-info/top_level.txt,sha256=i5D9wEbQhl085_9Lx2m7x-9Zu6nlx1tjYYbuSihG09E,7
-gandlf-0.1.4.dev20250609.dist-info/RECORD,,
+gandlf-0.1.4.dev20250611.dist-info/licenses/LICENSE,sha256=GlZPAfA4eckod8IVayhBXkqCpESXf6cc9BGli_Jwims,11357
+gandlf-0.1.4.dev20250611.dist-info/METADATA,sha256=EKvbL6HRwo773sCkUvDapy-iUNoGoTfER_893QLLZgs,9932
+gandlf-0.1.4.dev20250611.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+gandlf-0.1.4.dev20250611.dist-info/entry_points.txt,sha256=agwocNI7Upi-sFDe1rMl71dGN8YhCBB7WJmtBHRF4jg,902
+gandlf-0.1.4.dev20250611.dist-info/top_level.txt,sha256=i5D9wEbQhl085_9Lx2m7x-9Zu6nlx1tjYYbuSihG09E,7
+gandlf-0.1.4.dev20250611.dist-info/RECORD,,