fix test as now we do not allow to change the optimizer after a mode is created

theano/tensor/tests/test_opt.py

@@ -304,12 +304,11 @@ class test_canonize(unittest.TestCase):
         # We must be sure that the Canonizer is working, but that we don't have other
         # optimisation that could hide bug in the Canonizer as local_elemwise_fusion
         mode = compile.mode.get_default_mode()
-        old_optimizer = mode._optimizer
-        try:
-            mode._optimizer = gof.Query(["canonicalize"])
-            mode._optimizer = mode._optimizer.excluding(
-                'local_elemwise_fusion')
-            for id, [g, sym_inputs, val_inputs, nb_elemwise, out_dtype] in enumerate(cases):
+        opt = gof.Query(["canonicalize"])
+        opt = opt.excluding('local_elemwise_fusion')
+        mode = mode.__class__(linker=mode.linker, optimizer=opt)
+        for id, [g, sym_inputs, val_inputs,
+                 nb_elemwise, out_dtype] in enumerate(cases):
             if isinstance(out_dtype, dict):
                 out_dtype = out_dtype[config.cast_policy]
             f = compile.function(list(sym_inputs), g,
@@ -319,8 +318,6 @@ class test_canonize(unittest.TestCase):
             out = f(*val_inputs)
             assert(len(f.maker.fgraph.toposort()) == nb_elemwise)
             assert(out_dtype == out.dtype)
-        finally:
-            mode._optimizer = old_optimizer
 
     def test_elemwise_multiple_inputs_optimisation2(self):
         """
@@ -455,13 +452,12 @@ class test_canonize(unittest.TestCase):
         # We must be sure that the Canonizer is working, but that we don't have other
         # optimisation that could hide bug in the Canonizer as local_elemwise_fusion
         mode = compile.mode.get_default_mode()
-        old_optimizer = mode._optimizer
         try:
-            mode._optimizer = gof.Query(["canonicalize"])
-            mode._optimizer = mode._optimizer.including('ShapeOpt')
-            mode._optimizer = mode._optimizer.excluding(
+            opt = gof.Query(["canonicalize"])
+            opt = opt.including('ShapeOpt')
+            opt = opt.excluding(
                 'local_elemwise_fusion')
-
+            mode = mode.__class__(linker=mode.linker, optimizer=opt)
             # test x / x -> 1
             for id, (g, sym_inputs, val_inputs, out_dtype) in enumerate([(fx/fx, [fx], [fxv], 'float32'),
                                                            (dx/dx, [dx], [dxv], 'float64'),
@@ -644,7 +640,7 @@ class test_canonize(unittest.TestCase):
                 assert numpy.allclose(out, numpy.sign(val_inputs[0]) * 2 / 3)
                 assert(out_dtype == out.dtype)
         finally:
-            mode._optimizer = old_optimizer
+            pass
 
     def test_abs_mul_div(self):
         """
@@ -705,12 +701,11 @@ class test_canonize(unittest.TestCase):
         # We must be sure that the Canonizer is working, but that we don't have other
         # optimisation that could hide bug in the Canonizer as local_elemwise_fusion
         mode = compile.mode.get_default_mode()
-        old_optimizer = mode._optimizer
         try:
-            mode._optimizer = gof.Query(["canonicalize"])
-            mode._optimizer = mode._optimizer.excluding(
+            opt = gof.Query(["canonicalize"])
+            opt = opt.excluding(
                 'local_elemwise_fusion')
-
+            mode = mode.__class__(linker=mode.linker, optimizer=opt)
     # test fail!
             # test x / y / z -> x / (y * z)
             for (g, sym_inputs, val_inputs, out_dtype) in [
@@ -749,7 +744,7 @@ class test_canonize(unittest.TestCase):
                 assert(out_dtype == out.dtype)
 
         finally:
-            mode._optimizer = old_optimizer
+            pass
 
     def test_dont_merge_if_multiple_client(self):
         """ test those case take from the comment in Canonizer
@@ -3316,6 +3311,8 @@ class test_shapeoptimizer(unittest.TestCase):
         # Register the optimization
         opt.register_specialize(local_identity_noshape_to_identity_shape)
 
+        mode = theano.compile.get_default_mode().including(
+            'ShapeOpt', 'specialize')
         # With the optimization
         # The identity_shape op should not be needed anymore to compute
         # the shape