Add support to current numpy __array__ ndarray method

rpyc/core/netref.py

@@ -248,13 +248,14 @@ def method(self, start, stop, *args):
         method.__doc__ = doc
         return method
     elif name == "__array__":
-        def __array__(self):
+        def __array__(self, *args, **kwargs):
             # Note that protocol=-1 will only work between python
             # interpreters of the same version.
             if not object.__getattribute__(self,'____conn__')._config["allow_pickle"]:
                 # Security check that server side allows pickling per #551
                 raise ValueError("pickling is disabled")
-            return pickle.loads(syncreq(self, consts.HANDLE_PICKLE, -1))
+            array = pickle.loads(syncreq(self, consts.HANDLE_PICKLE, -1))
+            return array.__array__(*args, **kwargs)
         __array__.__doc__ = doc
         return __array__
     else:

tests/test_numpy.py

@@ -0,0 +1,35 @@
+from __future__ import with_statement
+import rpyc
+import unittest
+try:
+    import numpy as np
+    _numpy_import_failed = False
+except Exception:
+    _numpy_import_failed = True
+
+
+class MyService(rpyc.Service):
+
+    def exposed_create_array(self, array):
+        return np.array(array, dtype=np.int64, copy=True)
+
+
+@unittest.skipIf(_numpy_import_failed, "Skipping since numpy cannot be imported")
+class TestNumpy(unittest.TestCase):
+    def setUp(self):
+        self.server = rpyc.utils.server.OneShotServer(MyService, port=0, protocol_config={"allow_pickle":True})
+        self.server.logger.quiet = False
+        self.server._start_in_thread()
+        self.conn = rpyc.connect("localhost", port=self.server.port, config={"allow_pickle":True})
+
+    def tearDown(self):
+        self.conn.close()
+
+    def test_numpy(self):
+        remote_array = self.conn.root.create_array(np.array([0.]))
+        self.assertEqual(remote_array[0], 0)
+        self.assertIsInstance(remote_array[0], np.int64)
+
+
+if __name__ == "__main__":
+    unittest.main()