bug: merge proposals and check cycles (#258)

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

src/deep_neurographs/inference.py

@@ -479,14 +479,9 @@ def run(self, neurograph, proposals):
                 preds = self.predict(dataset)
 
                 # Update graph
-                batch_accepts = get_accepted_proposals(
-                    neurograph, preds, self.threshold
-                )
-                for proposal in batch_accepts:
-                    neurograph.merge_proposal(proposal)
-
-                # Finish
-                accepts.extend(batch_accepts)
+                for p in get_accepts(neurograph, preds, self.threshold):
+                    neurograph.merge_proposal(p)
+                    accepts.append(p)
                 pbar.update(len(batch["proposals"]))
             neurograph.absorb_reducibles()
         return neurograph, accepts
@@ -591,7 +586,7 @@ def predict(self, dataset):
 
 
 # --- Accepting Proposals ---
-def get_accepted_proposals(neurograph, preds, threshold, high_threshold=0.9):
+def get_accepts(neurograph, preds, threshold, high_threshold=0.9):
     """
     Determines which proposals to accept based on prediction scores and the
     specified threshold.
@@ -623,6 +618,7 @@ def get_accepted_proposals(neurograph, preds, threshold, high_threshold=0.9):
     accepts = list()
     accepts.extend(filter_proposals(neurograph, best_proposals))
     accepts.extend(filter_proposals(neurograph, proposals))
+    neurograph.remove_edges_from(map(tuple, accepts))
     return accepts
 
 
@@ -685,17 +681,9 @@ def filter_proposals(graph, proposals):
     """
     accepts = list()
     for i, j in proposals:
-        nodes_i = set(gutil.get_component(graph, i))
-        nodes_j = set(gutil.get_component(graph, j))
-        if nodes_i.isdisjoint(nodes_j):
-            subgraph_i = graph.subgraph(nodes_i)
-            subgraph_j = graph.subgraph(nodes_j)
-            subgraph = nx.union(subgraph_i, subgraph_j)
-            created_cycle, _ = gutil.creates_cycle(subgraph, (i, j))
-            if not created_cycle:
-                graph.add_edge(i, j)
-                accepts.append(frozenset({i, j}))
-    graph.remove_edges_from(map(tuple, accepts))
+        if not nx.has_path(graph, i, j):
+            graph.add_edge(i, j)
+            accepts.append(frozenset({i, j}))
     return accepts
 
 

src/deep_neurographs/neurograph.py

@@ -224,10 +224,13 @@ def absorb_reducibles(self):
         for i in nodes:
             nbs = list(self.neighbors(i))
             if len(nbs) == 2 and len(self.nodes[i]["proposals"]) == 0:
-                # Get attributes
+                # Concatenate attributes
+                len_1 = self.edges[i, nbs[0]]["length"]
+                len_2 = self.edges[i, nbs[1]]["length"]
                 xyz = self.get_branches(i, key="xyz")
                 radius = self.get_branches(i, key="radius")
                 attrs = {
+                    "length": len_1 + len_2,
                     "radius": concatenate([np.flip(radius[0]), radius[1]]),
                     "xyz": concatenate([np.flip(xyz[0], axis=0), xyz[1]]),
                 }
@@ -649,7 +652,7 @@ def proposal_directionals(self, proposal, window):
     def merge_proposal(self, proposal):
         i, j = tuple(proposal)
         somas_check = not (self.is_soma(i) and self.is_soma(j))
-        degrees_check = self.degree[i] == 2 and self.degree[j] == 2
+        degrees_check = not (self.degree[i] == 2 and self.degree[j] == 2)
         if somas_check and degrees_check:
             # Dense attributes
             attrs = dict()
@@ -665,10 +668,10 @@ def merge_proposal(self, proposal):
                 attrs["length"] = len_ij
             elif self.degree[i] == 2:
                 e_j = (j, self.leaf_neighbor(j))
-                attrs["length"] = self.edges[e_i]["length"]
+                attrs["length"] = self.edges[e_j]["length"]
             else:
                 e_i = (i, self.leaf_neighbor(i))
-                attrs["length"] = self.edges[e_j]["length"]
+                attrs["length"] = self.edges[e_i]["length"]
 
             swc_id_i = self.nodes[i]["swc_id"]
             swc_id_j = self.nodes[j]["swc_id"]

src/deep_neurographs/utils/graph_util.py

@@ -798,32 +798,6 @@ def upd_node_attrs(swc_dict, leafs, junctions, i):
 
 
 # -- miscellaneous --
-def creates_cycle(graph, edge):
-    """
-    Checks whether adding "edge" to "graph" creates a cycle.
-
-    Paramaters
-    ----------
-    graph : networkx.Graph
-        Graph to be checked for cycles.
-    edge : tuple
-        Edge to be added to "graph"
-
-    Returns
-    -------
-    bool
-        Indication of whether adding "edge" to graph creates a cycle.
-
-    """
-    graph.add_edges_from([edge])
-    exists = cycle_exists(graph)
-    graph.remove_edges_from([edge])
-    if exists:
-        return True, edge
-    else:
-        return False, edge
-
-
 def cycle_exists(graph):
     """
     Checks whether a cycle exists in "graph".
@@ -900,14 +874,13 @@ def get_component(graph, root):
 
     """
     queue = [root]
-    component = set()
+    visited = set()
     while len(queue):
         i = queue.pop()
-        component.add(i)
-        for j in [j for j in graph.neighbors(i) if j not in component]:
-            if (i, j) in graph.edges:
-                queue.append(j)
-    return component
+        visited.add(i)
+        for j in [j for j in graph.neighbors(i) if j not in visited]:
+            queue.append(j)
+    return visited
 
 
 def count_components(graph):