1st implementation of protein detection

utils/gridsearch.py

@@ -0,0 +1,2 @@
+def gridsearch():
+

utils/protein_detection.py

@@ -1,9 +1,7 @@
 import numpy as np
-from scipy.ndimage import gaussian_laplace, label, find_objects, center_of_mass
-from scipy.ndimage.measurements import variance
+from skimage.feature import blob_log
 
-
-def protein_detection(heatmap):  # TODO do this properly
+def protein_detection(heatmap): #TODO do this properly
     """
     Detects local maxima and estimates sizes of Gaussians in a 3D heatmap.
 
@@ -15,40 +13,20 @@ def protein_detection(heatmap):  # TODO do this properly
             - 'coordinates': Tuple of (z, y, x) for the local maxima
             - 'size': Estimated size of the Gaussian (sigma equivalent)
     """
-    # Apply Laplacian of Gaussian (LoG) filter to enhance Gaussian-like structures
-    log_filtered = -gaussian_laplace(heatmap, sigma=1)
-
-    # Find local maxima
-    labeled, num_features = label(log_filtered > np.mean(log_filtered))  # Binary threshold
-
-    # Get bounding boxes and compute centers of mass
-    regions = find_objects(labeled)
     detections = []
 
-    for i, region in enumerate(regions):
-        if region is None:
-            continue
-
-        # Extract subregion
-        subregion = labeled[region]
-        sub_heatmap = heatmap[region]
-
-        # Mask specific to the current label
-        mask = (subregion == (i + 1))
-
-        # Compute center of mass as the coordinates of the local maximum
-        com = center_of_mass(sub_heatmap, labels=mask, index=1)
-
-        # Compute size: estimate the variance of the Gaussian
-        size = np.sqrt(variance(sub_heatmap, labels=mask, index=1))
+    data_path=#TODO pass the val data paths
+    threshold = gridsearch(data_path, model) 
+    #smalles protein structure: "beta-amylase": 33.27
+    #bigges protein structure: "ribosome": 109.02,
+    #0.3 is the factor to match the PDB size to the experimental data size
+    adj_factor=0.3 #TODO implement this as an argument, also when creating heatmap
+    pred_coords = blob_log(preds, min_sigma=33.27*adj_factor, max_sigma=109.02*adj_factor, threshold_abs=threshold) 
 
-        # Adjust coordinates to global
-        starts = [r.start for r in region]  # Dynamically handle dimensions
-        com_global = tuple(com[i] + starts[i] for i in range(len(starts)))
 
-        detections.append({
-            'coordinates': com_global,
-            'size': size
-        })
+    detections.append({
+        'coordinates': coordinates,
+        'size': size
+    })
 
-    return detections
+    return detections