Add draft loaders for suite2p processed data using pynapple 🍍 and NWB

derotation/postprocessing/load_processed_data.py

@@ -0,0 +1,48 @@
+from pathlib import Path
+
+import numpy as np
+import pynapple as nap
+from skimage.measure import find_contours
+
+from derotation.postprocessing.neuropil_subtraction import neuropil_subtraction
+
+
+def load_suite2p_data(path: Path) -> nap.NWBFile:
+    suite2p_data = nap.load_file(path)
+    return suite2p_data
+
+
+def get_dff(suite2p_data: nap.NWBFile) -> tuple:
+    raw_fluorescence = suite2p_data["RoiResponseSeries"]
+
+    neuropil = suite2p_data["Neuropil"]
+
+    dff, r = neuropil_subtraction(
+        raw_fluorescence[:].values.T, neuropil[:].values.T
+    )
+
+    dff = nap.Tsd(t=raw_fluorescence.t, d=dff.T)
+    timebase = raw_fluorescence.t
+
+    return dff, timebase
+
+
+def get_plane_segmentation(suite2p_data: nap.NWBFile) -> tuple:
+    plane_seg = (
+        suite2p_data.nwb.processing["ophys"]
+        .data_interfaces["ImageSegmentation"]
+        .plane_segmentations["PlaneSegmentation"]
+    )
+
+    ROI_centroids = plane_seg.ROICentroids[:]
+    is_cell = plane_seg.Accepted[:]
+    labels = plane_seg.image_mask[:]
+
+    contours = [find_contours(c)[0] for c in labels[is_cell.astype(bool)]]
+
+    return ROI_centroids, is_cell, labels, contours
+
+
+def load_registered_binary(path: Path, shape: tuple) -> np.memmap:
+    registered = np.memmap(path, shape=shape, dtype="int16")
+    return registered