Refactor general updates (#147)

* refactor: improved metric defaults

* refactor: improved NaN metric handling

---------

Co-authored-by: anna-grim <[email protected]>

Author: anna-grim

src/segmentation_skeleton_metrics/skeleton_metric.py

@@ -301,6 +301,7 @@ def run(self):
         path = f"{self.output_dir}/{prefix}results.csv"
         if self.fragment_graphs is None:
             self.metrics = self.metrics.drop("# Merges", axis=1)
+            self.metrics = self.metrics.drop("Merge Rate", axis=1)
         self.metrics.to_csv(path, index=False)
 
         # Report results
@@ -337,28 +338,39 @@ def detect_splits(self):
 
             # Store results
             for process in as_completed(pending.keys()):
+                # Update graph
                 key = pending.pop(process)
-                self.graphs[key], n_split_edges = process.result()
-                n_before = self.graphs[key].graph["n_initial_edges"]
-                n_after = self.graphs[key].number_of_edges()
-
-                n_missing_edges = n_before - n_after
-                n_splits = gutil.count_splits(self.graphs[key])
-                p_omit = 100 * (n_missing_edges + n_split_edges) / n_before
-                p_split = 100 * n_split_edges / n_before
-                rl = np.sum(self.graphs[key].run_lengths())
-                gt_rl = self.graphs[key].run_length
-                split_rate = rl / n_splits if n_splits > 0 else np.nan
-
-                self.metrics.at[key, "# Splits"] = n_splits
-                self.metrics.at[key, "Split Rate"] = split_rate
-                self.metrics.loc[key, "% Split Edges"] = round(p_split, 2)
-                self.metrics.at[key, "% Omit Edges"] = round(p_omit, 2)
-                self.metrics.loc[key, "SWC Run Length"] = round(gt_rl, 2)
+                graph, n_split_edges = process.result()
+                self.graphs[key] = graph
 
+                # Compute metrics
+                self.compute_split_metrics(key, n_split_edges)
                 if self.verbose:
                     pbar.update(1)
 
+    def compute_split_metrics(self, key, n_split_edges):
+        # Edge counts
+        n_edges_before = self.graphs[key].graph["n_initial_edges"]
+        n_edges_after = self.graphs[key].number_of_edges()
+        n_edge_missing = n_edges_before - n_edges_after
+
+        # Split statistics
+        n_splits = gutil.count_splits(self.graphs[key])
+        p_omit = 100 * (n_edge_missing + n_split_edges) / n_edges_before
+        p_split = 100 * n_split_edges / n_edges_before
+
+        # Run lengths
+        rl = np.sum(self.graphs[key].run_lengths())
+        gt_rl = self.graphs[key].run_length
+        split_rate = rl / n_splits if n_splits > 0 else np.nan
+
+        # Update metrics
+        self.metrics.at[key, "# Splits"] = n_splits
+        self.metrics.at[key, "Split Rate"] = split_rate
+        self.metrics.at[key, "% Split Edges"] = round(p_split, 2)
+        self.metrics.at[key, "% Omit Edges"] = round(p_omit, 2)
+        self.metrics.at[key, "SWC Run Length"] = round(gt_rl, 2)
+
     # -- Merge Detection --
     def detect_merges(self):
         """
@@ -724,4 +736,3 @@ def to_local_voxels(self, key, i, offset):
         voxel = np.array(self.graphs[key].voxels[i])
         offset = np.array(offset)
         return tuple(voxel - offset)
-

src/segmentation_skeleton_metrics/utils/swc_util.py

@@ -67,21 +67,21 @@ def read(self, swc_pointer):
         ----------
         swc_pointer : str or List[str]
             Object that points to SWC files to be read, must be one of:
-                - swc_path (str): Path to single SWC file.
-                - swc_dir_path (str): Path to local directory containing SWC files.
-                - swc_zip_path (str): Path to a local ZIP archive containing SWC files.
-                - swc_zip_dir_path (str): Path to a local directory of ZIPs with SWC files.
-                - swc_s3_dir_path (str): Path to S3 directory containing SWC files.
-                - swc_gcs_dir_path (str): Path to GCS directory containing SWC files.
-                - swc_gcs_zip_dir_path (str): Path to GCS directory containing ZIP archives of SWC files.
-                - swc_path_list (List[str]): List of paths to local SWC files.
+                - file_path (str): Path to single SWC file
+                - dir_path (str): Path to local directory with SWC files
+                - zip_path (str): Path to local ZIP with SWC files
+                - zip_dir_path (str): Path to local directory of ZIPs with SWC files
+                - s3_dir_path (str): Path to S3 directory with SWC files
+                - gcs_dir_path (str): Path to GCS directory with SWC files
+                - gcs_zip_dir_path (str): Path to GCS directory with ZIPs of SWC files
+                - path_list (List[str]): List of paths to local SWC files
 
         Returns
         -------
         Deque[dict]
-            List of dictionaries whose keys and values are the attribute names
-            and values from the SWC files. Each dictionary contains the
-            following items:
+            Dictionaries whose keys and values are the attribute names and
+            values from the SWC files. Each dictionary contains the following:
+            items:
                 - "id": unique identifier of each node in an SWC file.
                 - "pid": parent ID of each node.
                 - "radius": radius value corresponding to each node.