doc: skeleton metrics and swc_loading (#157)

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

Author: anna-grim

src/segmentation_skeleton_metrics/data_handling/graph_loading.py

@@ -65,6 +65,7 @@ def load_groundtruth(self, swc_pointer, label_mask):
             label_handler=self.label_handler,
             label_mask=label_mask,
             use_anisotropy=False,
+            verbose=self.verbose
         )
         return graph_loader.run(swc_pointer)
 
@@ -99,6 +100,7 @@ def load_fragments(self, swc_pointer, gt_graphs):
             label_handler=self.label_handler,
             selected_ids=selected_ids,
             use_anisotropy=self.use_anisotropy,
+            verbose=self.verbose
         )
         return graph_loader.run(swc_pointer)
 
@@ -110,6 +112,7 @@ def get_all_node_labels(self, graphs):
         Parameters
         ----------
         graphs : Dict[str, SkeletonGraph]
+            Graph to be searched.
 
         Returns
         -------
@@ -135,6 +138,7 @@ def __init__(
         label_mask=None,
         selected_ids=None,
         use_anisotropy=True,
+        verbose=True
     ):
         """
         Instantiates a GraphLoader object.
@@ -156,12 +160,15 @@ def __init__(
             Indication of whether coordinates in SWC files should be converted
             from physical to image coordinates using the given anisotropy.
             Default is True.
+        verbose : bool, optional
+            Indication of whether to display a progress bar. Default is True.
         """
         # Instance attributes
         self.anisotropy = anisotropy
         self.is_groundtruth = is_groundtruth
         self.label_handler = label_handler
         self.label_mask = label_mask
+        self.verbose = verbose
 
         # Reader
         anisotropy = anisotropy if use_anisotropy else (1.0, 1.0, 1.0)
@@ -210,17 +217,19 @@ def _build_graphs_from_swcs(self, swc_pointer):
         """
         # Initializations
         swc_dicts = self.swc_reader.read(swc_pointer)
-        pbar = tqdm(total=len(swc_dicts), desc="Build Graphs")
+        if self.verbose:
+            pbar = tqdm(total=len(swc_dicts), desc="Build Graphs")
 
         # Main
         graph_dict = dict()
         if len(swc_dicts) > 10 ** 4:
             while len(swc_dicts) > 0:
                 swc_dict = swc_dicts.pop()
                 graph_dict.update(self.to_graph(swc_dict))
-                pbar.update(1)
+                if self.verbose:
+                    pbar.update(1)
         else:
-            with ProcessPoolExecutor() as executor:
+            with ProcessPoolExecutor(max_workers=1) as executor:
                 # Assign processes
                 processes = list()
                 while len(swc_dicts) > 0:
@@ -230,7 +239,8 @@ def _build_graphs_from_swcs(self, swc_pointer):
                 # Store results
                 for process in as_completed(processes):
                     graph_dict.update(process.result())
-                    pbar.update(1)
+                    if self.verbose:
+                        pbar.update(1)
         return graph_dict
 
     def to_graph(self, swc_dict):

src/segmentation_skeleton_metrics/data_handling/swc_loading.py

@@ -429,6 +429,20 @@ def read_from_gcs_zip(self, bucket_name, path):
         return swc_dicts
 
     def read_from_s3(self, s3_path):
+        """
+        Reads and parses SWC files from an S3 directory.
+
+        Parameters
+        ----------
+        s3_path : str
+            Path to a directory in an S3 bucket containing SWC files.
+
+        Returns
+        -------
+        Dequeue[dict]
+            Dictionaries whose keys and values are the attribute
+            names and values from an SWC file.
+        """
         # List filenames
         bucket_name, prefix = util.parse_cloud_path(s3_path)
         swc_paths = util.list_s3_paths(bucket_name, prefix, extension=".swc")

src/segmentation_skeleton_metrics/evaluate.py

@@ -161,7 +161,22 @@ def init_results(self, gt_graphs):
         return results
 
     def report_summary(self, results):
-        pass
+        # Averaged results
+        filename = f"{self.results_filename}-overview.txt"
+        path = os.path.join(self.output_dir, filename)
+        util.update_txt(path, "Average Results...")
+        for column in results.columns:
+            if column != "SWC Run Length" and column != "SWC Name":
+                avg = util.compute_weighted_avg(results, column)
+                util.update_txt(path, f"  {column}: {avg:.4f}")
+
+        # Total results
+        n_splits = results["# Splits"].sum()
+        util.update_txt(path, "\nTotal Results...")
+        util.update_txt(path, f"  # Splits: {n_splits}")
+        if "# Merges" in results.columns:
+            n_merges = results["# Merges"].sum()
+            util.update_txt(path, f"  # Merges: {n_merges}")
 
     # --- Writers ---
     def save_fragments(self):