major upd : feature_extraction

doc_template/source/conf.py

@@ -41,10 +41,7 @@
 html_theme = "furo"
 html_static_path = ["_static"]
 html_favicon = "_static/favicon.ico"
-html_theme_options = {
-    "light_logo": "light-logo.svg",
-    "dark_logo": "dark-logo.svg",
-}
+html_theme_options = {"light_logo": "light-logo.svg", "dark_logo": "dark-logo.svg"}
 
 # If true, "Created using Sphinx" is shown in the HTML footer. Default is True.
 html_show_sphinx = False

src/deep_neurographs/deep_learning/models.py

@@ -127,11 +127,7 @@ def _init_conv_layer(self, in_channels, out_channels):
         """
         conv_layer = nn.Sequential(
             nn.Conv3d(
-                in_channels,
-                out_channels,
-                kernel_size=(3, 3, 3),
-                stride=1,
-                padding=0,
+                in_channels, out_channels, kernel_size=(3, 3, 3), stride=1, padding=0
             ),
             nn.BatchNorm3d(out_channels),
             nn.LeakyReLU(),
@@ -231,11 +227,7 @@ def _init_conv_layer(self, in_channels, out_channels):
         """
         conv_layer = nn.Sequential(
             nn.Conv3d(
-                in_channels,
-                out_channels,
-                kernel_size=(3, 3, 3),
-                stride=1,
-                padding=0,
+                in_channels, out_channels, kernel_size=(3, 3, 3), stride=1, padding=0
             ),
             nn.BatchNorm3d(out_channels),
             nn.LeakyReLU(),

src/deep_neurographs/deep_learning/train.py

@@ -36,7 +36,7 @@
 BATCH_SIZE = 32
 NUM_WORKERS = 0
 SHUFFLE = True
-SUPPORTED_CLFS = [
+SUPPORTED_MODELS = [
     "AdaBoost",
     "RandomForest",
     "FeedForwardNet",
@@ -45,7 +45,7 @@
 ]
 
 
-# Training
+# -- Cross Validation --
 def get_kfolds(train_data, k):
     folds = []
     samples = set(train_data)
@@ -60,8 +60,9 @@ def get_kfolds(train_data, k):
     return folds
 
 
+# -- Training --
 def get_clf(key, data=None, num_features=None):
-    assert key in SUPPORTED_CLFS
+    assert key in SUPPORTED_MODELS
     if key == "AdaBoost":
         return AdaBoostClassifier()
     elif key == "RandomForest":
@@ -104,10 +105,7 @@ def train_network(
         shuffle=SHUFFLE,
     )
     valid_loader = DataLoader(
-        valid_set,
-        batch_size=BATCH_SIZE,
-        num_workers=NUM_WORKERS,
-        pin_memory=True,
+        valid_set, batch_size=BATCH_SIZE, num_workers=NUM_WORKERS, pin_memory=True
     )
 
     # Configure trainer
@@ -149,7 +147,7 @@ def eval_network(X, model):
     return np.array(y_pred)
 
 
-# Lightning Module
+# -- Lightning Module --
 class LitNeuralNet(pl.LightningModule):
     def __init__(self, net=None, lr=10e-3):
         super().__init__()

src/deep_neurographs/evaluation.py

@@ -9,13 +9,25 @@
 """
 import numpy as np
 
-METRICS_LIST = [
-    "precision",
-    "recall",
-    "f1",
-    "# splits fixed",
-    "# merges created",
-]
+METRICS_LIST = ["precision", "recall", "f1", "# splits fixed", "# merges created"]
+
+
+def init_stats():
+    """
+    Initializes a dictionary that stores stats computes by routines in this
+    module.
+
+    Parameters
+    ----------
+    None
+
+    Returns
+    -------
+    dict
+        Dictionary that stores stats computes by routines in this module.
+
+    """
+    return dict([(metric, []) for metric in METRICS_LIST])
 
 
 def run_evaluation(neurographs, blocks, pred_edges):
@@ -40,11 +52,7 @@ def run_evaluation(neurographs, blocks, pred_edges):
         metrics contained in this dictionary are identical to "METRICS_LIST"].
 
     """
-    stats = {
-        "Overall": dict([(metric, []) for metric in METRICS_LIST]),
-        "Simple": dict([(metric, []) for metric in METRICS_LIST]),
-        "Complex": dict([(metric, []) for metric in METRICS_LIST]),
-    }
+    stats = {"Overall": init_stats(), "Simple": init_stats(), "Complex": init_stats()}
     for block_id in blocks:
         # Compute accuracy
         overall_stats_i = get_stats(