docs/changes/newsfragments/257.change

@@ -0,0 +1 @@
+``ParcellationRegistry.get`` now returns labels as dictionary mapping from values to labels instead of a list of labels by `Synchon Mandal`_

docs/changes/newsfragments/257.enh

@@ -0,0 +1 @@
+Allow parcellation registration with non-continuous ranges as values by `Synchon Mandal`_

junifer/data/_dispatch.py

@@ -120,7 +120,8 @@ def get_data(
     extra_input: Optional[dict[str, Any]] = None,
 ) -> Union[
     tuple[ArrayLike, list[str]],  # coordinates
-    tuple["Nifti1Image", list[str]],  # parcellation / maps
+    tuple["Nifti1Image", dict[int, str]],  # parcellation
+    tuple["Nifti1Image", list[str]],  # maps
     "Nifti1Image",  # mask
 ]:
     """Get tailored ``kind`` for ``target_data``.
@@ -142,7 +143,7 @@ def get_data(
     Returns
     -------
     tuple of numpy.ndarray, list of str; \
-    tuple of nibabel.nifti1.Nifti1Image, list of str; \
+    tuple of nibabel.nifti1.Nifti1Image, dict or list of str; \
     nibabel.nifti1.Nifti1Image
 
     Raises

junifer/data/parcellations/_parcellations.py

@@ -403,7 +403,7 @@ class ParcellationRegistry(BasePipelineDataRegistry):
         if not path_only:
             # Load image via nibabel
             parcellation_img = nib.load(parcellation_fname)
-            # Get unique values
+            # Get unique values (returns sorted result)
             parcel_values = np.unique(parcellation_img.get_fdata())
             # Check for dimension
             if len(parcel_values) - 1 != len(parcellation_labels):
@@ -411,14 +411,6 @@ class ParcellationRegistry(BasePipelineDataRegistry):
                     f"Parcellation {name} has {len(parcel_values) - 1} "
                     f"parcels but {len(parcellation_labels)} labels."
                 )
-            # Sort values
-            parcel_values.sort()
-            # Check if value range is invalid
-            if np.any(np.diff(parcel_values) != 1):
-                raise_error(
-                    f"Parcellation {name} must have all the values in the "
-                    f"range [0, {len(parcel_values)}]"
-                )
 
         return parcellation_img, parcellation_labels, parcellation_fname, space
 
@@ -427,7 +419,7 @@ class ParcellationRegistry(BasePipelineDataRegistry):
         parcellations: Union[str, list[str]],
         target_data: dict[str, Any],
         extra_input: Optional[dict[str, Any]] = None,
-    ) -> tuple["Nifti1Image", list[str]]:
+    ) -> tuple["Nifti1Image", dict[int, str]]:
         """Get parcellation, tailored for the target image.
 
         Parameters
@@ -446,8 +438,8 @@ class ParcellationRegistry(BasePipelineDataRegistry):
         -------
         Nifti1Image
             The parcellation image.
-        list of str
+        dict
-            Parcellation labels.
+            Parcellation value to label mappings.
 
         Raises
         ------
@@ -561,7 +553,16 @@ class ParcellationRegistry(BasePipelineDataRegistry):
         # Avoid merging if there is only one parcellation
         if len(all_parcellations) == 1:
             resampled_parcellation_img = all_parcellations[0]
-            labels = all_labels[0]
+            labels = dict(
+                zip(
+                    np.trim_zeros(
+                        np.unique(
+                            resampled_parcellation_img.get_fdata().astype(int)
+                        )
+                    ),
+                    all_labels[0],
+                )
+            )
         # Parcellations are already transformed to target standard space
         else:
             logger.debug("Merging parcellations.")
@@ -1310,7 +1311,7 @@ def merge_parcellations(
     parcellations_list: list["Nifti1Image"],
     parcellations_names: list[str],
     labels_lists: list[list[str]],
-) -> tuple["Nifti1Image", list[str]]:
+) -> tuple["Nifti1Image", dict[int, str]]:
     """Merge multiple parcellations.
 
     Parameters
@@ -1328,8 +1329,8 @@ def merge_parcellations(
     Niimg-like object
         The parcellation that results from merging the list of input
         parcellations.
-    list of str
+    dict of int and str
-        List of labels for the resultant parcellation.
+        Merged parcellation value to label mappings.
 
     """
     # Check for duplicated labels
@@ -1346,34 +1347,54 @@ def merge_parcellations(
             ]
     overlapping_voxels = False
     ref_parc = parcellations_list[0]
-    parc_data = ref_parc.get_fdata()
+    ref_parc_data = ref_parc.get_fdata()
 
-    labels = labels_lists[0]
+    # Get max value for the parcellation ROIs to get a reference for ROI value
+    # increment
+    max_val = np.max(
+        np.concatenate(
+            [np.unique(p.get_fdata().astype(int)) for p in parcellations_list]
+        )
+    )
+    # Setup parcellation value to label mapping
+    val_label_map = dict(
+        zip(
+            np.trim_zeros(np.unique(ref_parc_data.astype(int))),
+            labels_lists[0],
+        ),
+    )
 
-    for t_parc, t_labels in zip(parcellations_list[1:], labels_lists[1:]):
+    for idx, (parc, labs) in enumerate(
-        if t_parc.shape != ref_parc.shape:
+        zip(parcellations_list[1:], labels_lists[1:])
+    ):
+        # Resample to reference (1st in the list) parcellation
+        if parc.shape != ref_parc.shape:
             warn_with_log(
-                "The parcellations have different resolutions!"
+                "The parcellations have different resolutions. "
                 "Resampling all parcellations to the first one in the list."
             )
-            t_parc = nimg.resample_to_img(
+            parc = nimg.resample_to_img(
-                t_parc, ref_parc, interpolation="nearest", copy=True
+                source_img=parc,
+                target_img=ref_parc,
+                interpolation="nearest",
+                copy=True,
             )
-
         # Get the data from this parcellation
-        t_parc_data = t_parc.get_fdata().copy()  # must be copied
+        parc_data = parc.get_fdata().copy()  # must be copied
         # Increase the values of each ROI to match the labels
-        t_parc_data[t_parc_data != 0] += len(labels)
+        # and update label mapping
-
+        parc_data[parc_data != 0] += (idx + 1) * max_val
+        val_label_map.update(
+            dict(zip(np.trim_zeros(np.unique(parc_data.astype(int))), labs))
+        )
         # Only set new values for the voxels that are 0
         # This makes sure that the voxels that are in multiple
         # parcellations are assigned to the parcellation that was
         # first in the list.
-        if np.any(parc_data[t_parc_data != 0] != 0):
+        if np.any(ref_parc_data[parc_data != 0] != 0):
             overlapping_voxels = True
 
-        parc_data[parc_data == 0] += t_parc_data[parc_data == 0]
+        ref_parc_data[ref_parc_data == 0] += parc_data[ref_parc_data == 0]
-        labels.extend(t_labels)
 
     if overlapping_voxels:
         warn_with_log(
@@ -1382,6 +1403,9 @@ def merge_parcellations(
             "parcellation that was first in the list."
         )
 
-    parcellation_img_res = nimg.new_img_like(parcellations_list[0], parc_data)
+    parcellation_img_res = nimg.new_img_like(
+        ref_niimg=parcellations_list[0],
+        data=ref_parc_data,
+    )
 
-    return parcellation_img_res, labels
+    return parcellation_img_res, val_label_map

junifer/data/parcellations/tests/test_parcellations.py

@@ -7,7 +7,6 @@
 
 from pathlib import Path
 
-import nibabel as nib
 import numpy as np
 import pytest
 from nilearn.image import new_img_like, resample_to_img
@@ -153,46 +152,6 @@ def test_parcellation_wrong_labels_values(tmp_path: Path) -> None:
             target_space="MNI152NLin6Asym",
         )
 
-    schaefer_data = schaefer.get_fdata().copy()
-    schaefer_data[schaefer_data == 50] = 0
-    new_schaefer_path = tmp_path / "new_schaefer.nii.gz"
-    new_schaefer_img = new_img_like(schaefer, schaefer_data)
-    nib.save(new_schaefer_img, new_schaefer_path)
-
-    register_data(
-        kind="parcellation",
-        name="WrongValues",
-        parcellation_path=new_schaefer_path,
-        parcels_labels=labels[:-1],
-        space="MNI152Lin",
-    )
-    with pytest.raises(ValueError, match=r"must have all the values in the"):
-        load_data(
-            kind="parcellation",
-            name="WrongValues",
-            target_space="MNI152NLin6Asym",
-        )
-
-    schaefer_data = schaefer.get_fdata().copy()
-    schaefer_data[schaefer_data == 50] = 200
-    new_schaefer_path = tmp_path / "new_schaefer2.nii.gz"
-    new_schaefer_img = new_img_like(schaefer, schaefer_data)
-    nib.save(new_schaefer_img, new_schaefer_path)
-
-    register_data(
-        kind="parcellation",
-        name="WrongValues2",
-        parcellation_path=new_schaefer_path,
-        parcels_labels=labels,
-        space="MNI152Lin",
-    )
-    with pytest.raises(ValueError, match=r"must have all the values in the"):
-        load_data(
-            kind="parcellation",
-            name="WrongValues2",
-            target_space="MNI152NLin6Asym",
-        )
-
 
 @pytest.mark.parametrize(
     "name, parcellation_path, parcels_labels, space, overwrite",
@@ -1172,7 +1131,7 @@ def test_get_single() -> None:
             tailored_parcellation.get_fdata(),
             resampled_raw_parcellation.get_fdata(),
         )
-        assert tailored_labels == raw_labels
+        assert list(tailored_labels.values()) == raw_labels
 
 
 def test_get_multi_same_space() -> None:

junifer/markers/parcel_aggregation.py

@@ -213,7 +213,7 @@ class ParcelAggregation(BaseMarker):
         logger.debug("Computing ROI means")
         out_values = []
         # Iterate over the parcels (existing)
-        for t_v in range(1, len(labels) + 1):
+        for t_v in labels.keys():
             t_values = agg_func(data[:, parcellation_values == t_v], axis=-1)
             out_values.append(t_values)
             # Update the labels just in case a parcel has no voxels
@@ -238,6 +238,6 @@ class ParcelAggregation(BaseMarker):
         return {
             "aggregation": {
                 "data": out_values,
-                "col_names": labels,
+                "col_names": list(labels.values()),
             },
         }

junifer/markers/tests/test_parcel_aggregation.py

@@ -408,8 +408,8 @@ def test_ParcelAggregation_3D_multiple_non_overlapping(tmp_path: Path) -> None:
         parcellation2_data = parcellation_data.copy()
         parcellation2_data[parcellation2_data <= 8] = 0
         parcellation2_data[parcellation2_data > 0] -= 8
-        labels1 = labels[:8]
+        labels1 = list(labels.values())[:8]
-        labels2 = labels[8:]
+        labels2 = list(labels.values())[8:]
 
         parcellation1_img = new_img_like(
             testing_parcellation, parcellation1_data
@@ -512,8 +512,9 @@ def test_ParcelAggregation_3D_multiple_overlapping(tmp_path: Path) -> None:
         # Make the second parcellation overlap with the first
         parcellation2_data[parcellation2_data <= 6] = 0
         parcellation2_data[parcellation2_data > 0] -= 6
-        labels1 = [f"low_{x}" for x in labels[:8]]  # Change the labels
+        # Change the labels
-        labels2 = [f"high_{x}" for x in labels[6:]]  # Change the labels
+        labels1 = [f"low_{x}" for x in list(labels.values())[:8]]
+        labels2 = [f"high_{x}" for x in list(labels.values())[6:]]
 
         parcellation1_img = new_img_like(
             testing_parcellation, parcellation1_data
@@ -621,8 +622,8 @@ def test_ParcelAggregation_3D_multiple_duplicated_labels(
         parcellation2_data = parcellation_data.copy()
         parcellation2_data[parcellation2_data <= 8] = 0
         parcellation2_data[parcellation2_data > 0] -= 8
-        labels1 = labels[:8]
+        labels1 = list(labels.values())[:8]
-        labels2 = labels[7:-1]  # One label is duplicated
+        labels2 = list(labels.values())[7:-1]  # One label is duplicated
 
         parcellation1_img = new_img_like(
             testing_parcellation, parcellation1_data