Merge pull request #13 from cgoliver/master

use node index instead of ID in alignment matrix

subgraph_matching/alignment.py

@@ -48,8 +48,8 @@ def gen_alignment_matrix(model, query, target, method_type="order"):
     """
 
     mat = np.zeros((len(query), len(target)))
-    for u in query.nodes:
-        for v in target.nodes:
+    for i, u in enumerate(query.nodes):
+        for j, v in enumerate(target.nodes):
             batch = utils.batch_nx_graphs([query, target], anchors=[u, v])
             embs = model.emb_model(batch)
             pred = model(embs[1].unsqueeze(0), embs[0].unsqueeze(0))
@@ -58,7 +58,7 @@ def gen_alignment_matrix(model, query, target, method_type="order"):
                 raw_pred = torch.log(raw_pred)
             elif method_type == "mlp":
                 raw_pred = raw_pred[0][1]
-            mat[u][v] = raw_pred.item()
+            mat[i][j] = raw_pred.item()
     return mat
 
 def main():