Add hetero_gnn forward_op test

tests/test_hetero_gnn.py

@@ -0,0 +1,28 @@
+import math
+import torch
+import unittest
+
+from torch import nn
+from deepsnap.hetero_gnn import forward_op
+
+class TestHeteroGNN(unittest.TestCase):
+
+    def test_hetero_gnn_forward(self):
+        xs = {}
+        layers = nn.ModuleDict()
+        emb_dim = 5
+        feat_dim = 10
+        num_samples = 8
+        keys = ['a', 'b', 'c']
+
+        for key in keys:
+            layers[key] = nn.Linear(feat_dim, emb_dim)
+            xs[key] = torch.ones(num_samples, feat_dim)
+
+        ys = forward_op(xs, layers)
+        for key in keys:
+            self.assertEqual(ys[key].shape[0], num_samples)
+            self.assertEqual(ys[key].shape[1], emb_dim)
+
+if __name__ == "__main__":
+    unittest.main()