change of folder structure

inference/run_protein_detection.py

@@ -9,9 +9,9 @@
 import sys
 sys.path.append("/user/muth9/u12095/czii-protein-challenge")
 
-from utils.prediction import get_prediction_torch_em
-from utils.protein_detection import protein_detection
-from utils.tiling_helper import parse_tiling
+from utils import get_prediction_torch_em
+from utils import protein_detection
+from utils import parse_tiling
 
 def get_volume(input_path):
     zarr_file = zarr.open(os.path.join(input_path, "VoxelSpacing10.000", "denoised.zarr", "0"), mode='r')

training/train.py

@@ -4,8 +4,8 @@
 import sys
 sys.path.append("/user/muth9/u12095/czii-protein-challenge")
 
-from utils.dataset_splits import get_paths  # noqa
-from utils.training import supervised_training  # noqa
+from utils import get_paths  # noqa
+from utils import supervised_training  # noqa
 
 TRAIN_ROOT = "/scratch-grete/projects/nim00007/cryo-et/challenge-data/train/static/"
 LABEL_ROOT = "/scratch-grete/projects/nim00007/cryo-et/challenge-data/train/overlay/"

utils/__init__.py

@@ -1,7 +1,8 @@
-from .data_loader import create_data_loader
-from .dataset_splits import get_paths
-from .heatmap_dataset import HeatmapDataset
-from .training import supervised_training
-from .prediction import get_prediction_torch_em
-from .protein_detection import protein_detection
-from .tiling_helper import parse_tiling
+from .training.data_loader import create_data_loader
+from .training.dataset_splits import get_paths
+from .training.heatmap_dataset import HeatmapDataset
+from .training.training import supervised_training
+from .prediction.prediction import get_prediction_torch_em
+from .inference.protein_detection import protein_detection
+from .training.tiling_helper import parse_tiling
+from .evaluation.evaluation_metrics import metric_coords, get_distance_threshold_from_gridsearch

utils/evaluation/evaluation_metrics.py

@@ -0,0 +1,5 @@
+def metric_coords():
+    return
+
+def get_distance_threshold_from_gridsearch():
+    return

utils/inference/gridsearch.py

@@ -0,0 +1,31 @@
+import pandas
+from tqdm import tqdm
+from skimage.feature import blob_log
+from ..evaluation.evaluation_metrics import metric_coords, get_distance_threshold_from_gridsearch 
+
+
+def gridsearch():
+    data = []
+    for i in tqdm(range(len(idx))):
+        image_path = grid_search_image_paths[i]
+        label_path = grid_search_label_paths[i]
+
+        model, device = load_model(config, in_channels)
+        preds = predict_image(image_path, model, tuple(config["patch_shape"]), device=device)
+        label_coords = get_center_coordinates(label_path)
+
+
+        for thresh in threshes:
+            #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=thresh) 
+
+
+            _, _, f1, _, _, _ = metric_coords(label_coords, pred_coords)
+            data.append([f1, dist, thresh])
+
+    df = pandas.DataFrame(data=data, columns=["f1", "Distance", "Threshold"])
+    dist, thresh = get_distance_threshold_from_gridsearch(df, distances, threshes)
+    return dist, thresh, grid_search_numbers

utils/protein_detection.py → utils/inference/protein_detection.py

@@ -1,5 +1,6 @@
 import numpy as np
 from skimage.feature import blob_log
+from .gridsearch import gridsearch
 
 def protein_detection(heatmap): #TODO do this properly
     """
@@ -15,7 +16,7 @@ def protein_detection(heatmap): #TODO do this properly
     """
     detections = []
 
-    data_path=#TODO pass the val data paths
+    data_path="test"#TODO pass the val data paths
     threshold = gridsearch(data_path, model) 
     #smalles protein structure: "beta-amylase": 33.27
     #bigges protein structure: "ribosome": 109.02,