Deprecated unwanted for compute_test_value

Author: Dhruvanshu-Joshi

tests/graph/test_destroyhandler.py

@@ -2,7 +2,6 @@
 
 import pytest
 
-from pytensor.configdefaults import config
 from pytensor.graph.basic import Apply, Constant, Variable, clone
 from pytensor.graph.destroyhandler import DestroyHandler
 from pytensor.graph.features import ReplaceValidate
@@ -408,7 +407,6 @@ def test_value_repl():
     assert g.consistent()
 
 
-@config.change_flags(compute_test_value="off")
 def test_value_repl_2():
     x, y, z = inputs()
     sy = sigmoid(y)

tests/graph/test_op.py

@@ -168,7 +168,6 @@ def test_test_value_op():
     assert np.allclose(v, np.zeros((5, 5)))
 
 
-@config.change_flags(compute_test_value="off")
 def test_get_test_values_no_debugger():
     """Tests that `get_test_values` returns `[]` when debugger is off."""
 

tests/scan/test_basic.py

@@ -2244,7 +2244,7 @@ def test_compute_test_value_grad():
     """
     # WEIGHT = np.array([1, 2, 1, 3, 4, 1, 5, 6, 1, 7, 8, 1], dtype="float32")
 
-    with config.change_flags(compute_test_value="raise", exception_verbosity="high"):
+    with config.change_flags(exception_verbosity="high"):
         W_flat = fvector(name="W")
         W = W_flat.reshape((2, 2, 3))
 
@@ -2282,28 +2282,23 @@ def test_compute_test_value_grad_cast():
 
     See https://groups.google.com/d/topic/theano-users/o4jK9xDe5WI/discussion
     """
-    with config.change_flags(compute_test_value="raise"):
-        h = matrix("h")
-        with pytest.warns(FutureWarning):
-            h.tag.test_value = np.array(
-                [[1, 2, 3, 4], [5, 6, 7, 8]], dtype=config.floatX
-            )
+    h = matrix("h")
+    with pytest.warns(FutureWarning):
+        h.tag.test_value = np.array([[1, 2, 3, 4], [5, 6, 7, 8]], dtype=config.floatX)
 
-        w = shared(
-            np.random.default_rng(utt.fetch_seed())
-            .random((4, 3))
-            .astype(config.floatX),
-            name="w",
-        )
+    w = shared(
+        np.random.default_rng(utt.fetch_seed()).random((4, 3)).astype(config.floatX),
+        name="w",
+    )
 
-        outputs, _ = scan(
-            lambda i, h, w: (dot(h[i], w), i),
-            outputs_info=[None, 0],
-            non_sequences=[h, w],
-            n_steps=3,
-        )
+    outputs, _ = scan(
+        lambda i, h, w: (dot(h[i], w), i),
+        outputs_info=[None, 0],
+        non_sequences=[h, w],
+        n_steps=3,
+    )
 
-        grad(outputs[0].sum(), w)
+    grad(outputs[0].sum(), w)
 
 
 def test_constant_folding_n_steps():

tests/tensor/random/rewriting/test_basic.py

@@ -146,7 +146,6 @@ def test_inplace_rewrites(rv_op):
     assert check_stack_trace(f)
 
 
-@config.change_flags(compute_test_value="raise")
 @pytest.mark.parametrize(
     "dist_op, dist_params, size",
     [
@@ -417,7 +416,6 @@ def test_local_rv_size_lift(dist_op, dist_params, size):
         ),
     ],
 )
-@config.change_flags(compute_test_value_opt="raise", compute_test_value="raise")
 def test_DimShuffle_lift(ds_order, lifted, dist_op, dist_params, size, rtol):
     rng = shared(np.random.default_rng(1233532), borrow=False)
 
@@ -803,7 +801,6 @@ def rand_bool_mask(shape, rng=None):
         ),
     ],
 )
-@config.change_flags(compute_test_value_opt="raise", compute_test_value="raise")
 def test_Subtensor_lift(indices, lifted, dist_op, dist_params, size):
     from pytensor.tensor.subtensor import as_index_constant
 

tests/tensor/random/test_basic.py

@@ -752,7 +752,6 @@ def test_dirichlet_infer_shape(M, size):
     assert tuple(rv_shape_val) == tuple(rv_val.shape)
 
 
-@config.change_flags(compute_test_value="raise")
 def test_dirichlet_ShapeFeature():
     """Make sure `RandomVariable.infer_shape` works with `ShapeFeature`."""
     with pytest.warns(FutureWarning):
@@ -1600,7 +1599,6 @@ def test_unnatural_batched_dims(batch_dims_tester):
     batch_dims_tester()
 
 
-@config.change_flags(compute_test_value="off")
 def test_pickle():
     # This is an interesting `Op` case, because it has a conditional dtype
     sample_a = choice(5, replace=False, size=(2, 3))

tests/tensor/random/test_utils.py

@@ -17,7 +17,7 @@
 def set_pytensor_flags():
     rewrites_query = RewriteDatabaseQuery(include=[None], exclude=[])
     py_mode = Mode("py", rewrites_query)
-    with config.change_flags(mode=py_mode, compute_test_value="warn"):
+    with config.change_flags(mode=py_mode):
         yield
 
 

tests/tensor/test_basic.py

@@ -3550,11 +3550,11 @@ def test_assert(self):
         a = Assert()(c, c > 1)
         assert get_underlying_scalar_constant_value(a) == 2
 
-        with config.change_flags(compute_test_value="off"):
-            # condition is always False
-            a = Assert()(c, c > 2)
-            with pytest.raises(NotScalarConstantError):
-                get_underlying_scalar_constant_value(a)
+        # with config.change_flags(compute_test_value="off"):
+        # condition is always False
+        a = Assert()(c, c > 2)
+        with pytest.raises(NotScalarConstantError):
+            get_underlying_scalar_constant_value(a)
 
         # condition is not constant
         a = Assert()(c, c > x)

tests/tensor/test_subtensor.py

@@ -400,7 +400,6 @@ def test_err_invalid(self):
         with pytest.raises(IndexError):
             n.__getitem__(0)
 
-    @config.change_flags(compute_test_value="off")
     def test_err_bounds(self):
         n = self.shared(np.ones(3, dtype=self.dtype))
         t = n[7]
@@ -481,7 +480,6 @@ def test_ok_strided(self):
         assert tval.shape == (2,)
         assert (tval == [0.0, 2.0]).all()
 
-    @config.change_flags(compute_test_value="off")
     def test_err_bounds0(self):
         n = self.shared(np.ones((2, 3), dtype=self.dtype) * 5)
         for idx in [(0, 4), (0, -4)]:
@@ -497,7 +495,6 @@ def test_err_bounds0(self):
             finally:
                 _logger.setLevel(oldlevel)
 
-    @config.change_flags(compute_test_value="off")
     def test_err_bounds1(self):
         n = self.shared(np.ones((2, 3), dtype=self.dtype) * 5)
         t = n[4:5, 3]
@@ -719,71 +716,70 @@ def numpy_inc_subtensor(x, idx, a):
             inc_subtensor(n4[test_array > 2, ..., 0, 1], 1).eval(),
         )
 
-        with config.change_flags(compute_test_value="off"):
-            # the boolean mask should have the correct shape
-            # - too large, padded with True
-            mask = np.array([True, False, True])
-            with pytest.raises(IndexError):
-                test_array[mask].eval()
-            with pytest.raises(IndexError):
-                test_array[mask, ...].eval()
-            with pytest.raises(IndexError):
-                inc_subtensor(test_array[mask], 1).eval()
-            with pytest.raises(IndexError):
-                inc_subtensor(test_array[mask, ...], 1).eval()
-            mask = np.array([[True, False, False, True], [False, True, False, True]])
-            with pytest.raises(IndexError):
-                test_array[mask].eval()
-            with pytest.raises(IndexError):
-                inc_subtensor(test_array[mask], 1).eval()
-            # - too large, padded with False (this works in NumPy < 0.13.0)
-            mask = np.array([True, False, False])
-            with pytest.raises(IndexError):
-                test_array[mask].eval()
-            with pytest.raises(IndexError):
-                test_array[mask, ...].eval()
-            with pytest.raises(IndexError):
-                inc_subtensor(test_array[mask], 1).eval()
-            with pytest.raises(IndexError):
-                inc_subtensor(test_array[mask, ...], 1).eval()
-            mask = np.array([[True, False, False, False], [False, True, False, False]])
-            with pytest.raises(IndexError):
-                test_array[mask].eval()
-            with pytest.raises(IndexError):
-                inc_subtensor(test_array[mask], 1).eval()
-            # - mask too small (this works in NumPy < 0.13.0)
-            mask = np.array([True])
-            with pytest.raises(IndexError):
-                test_array[mask].eval()
-            with pytest.raises(IndexError):
-                test_array[mask, ...].eval()
-            with pytest.raises(IndexError):
-                inc_subtensor(test_array[mask], 1).eval()
-            with pytest.raises(IndexError):
-                inc_subtensor(test_array[mask, ...], 1).eval()
-            mask = np.array([[True], [True]])
-            with pytest.raises(IndexError):
-                test_array[mask].eval()
-            with pytest.raises(IndexError):
-                inc_subtensor(test_array[mask], 1).eval()
-            # - too many dimensions
-            mask = np.array([[[True, False, False], [False, True, False]]])
-            with pytest.raises(IndexError):
-                test_array.__getitem__(mask)
-            with pytest.raises(IndexError):
-                test_array.__getitem__(mask)
-
-            # special cases: Python bools and bools nested in Python arrays are not supported
-            with pytest.raises(TypeError):
-                test_array.__getitem__((True,))
-            with pytest.raises(TypeError):
-                test_array.__getitem__((False,))
-            with pytest.raises(TypeError):
-                test_array.__getitem__((True, False))
-            with pytest.raises(TypeError):
-                test_array.__getitem__(([0, 1], [0, False]))
-            with pytest.raises(TypeError):
-                test_array.__getitem__(([0, 1], [0, pytensor.shared(True)]))
+        # the boolean mask should have the correct shape
+        # - too large, padded with True
+        mask = np.array([True, False, True])
+        with pytest.raises(IndexError):
+            test_array[mask].eval()
+        with pytest.raises(IndexError):
+            test_array[mask, ...].eval()
+        with pytest.raises(IndexError):
+            inc_subtensor(test_array[mask], 1).eval()
+        with pytest.raises(IndexError):
+            inc_subtensor(test_array[mask, ...], 1).eval()
+        mask = np.array([[True, False, False, True], [False, True, False, True]])
+        with pytest.raises(IndexError):
+            test_array[mask].eval()
+        with pytest.raises(IndexError):
+            inc_subtensor(test_array[mask], 1).eval()
+        # - too large, padded with False (this works in NumPy < 0.13.0)
+        mask = np.array([True, False, False])
+        with pytest.raises(IndexError):
+            test_array[mask].eval()
+        with pytest.raises(IndexError):
+            test_array[mask, ...].eval()
+        with pytest.raises(IndexError):
+            inc_subtensor(test_array[mask], 1).eval()
+        with pytest.raises(IndexError):
+            inc_subtensor(test_array[mask, ...], 1).eval()
+        mask = np.array([[True, False, False, False], [False, True, False, False]])
+        with pytest.raises(IndexError):
+            test_array[mask].eval()
+        with pytest.raises(IndexError):
+            inc_subtensor(test_array[mask], 1).eval()
+        # - mask too small (this works in NumPy < 0.13.0)
+        mask = np.array([True])
+        with pytest.raises(IndexError):
+            test_array[mask].eval()
+        with pytest.raises(IndexError):
+            test_array[mask, ...].eval()
+        with pytest.raises(IndexError):
+            inc_subtensor(test_array[mask], 1).eval()
+        with pytest.raises(IndexError):
+            inc_subtensor(test_array[mask, ...], 1).eval()
+        mask = np.array([[True], [True]])
+        with pytest.raises(IndexError):
+            test_array[mask].eval()
+        with pytest.raises(IndexError):
+            inc_subtensor(test_array[mask], 1).eval()
+        # - too many dimensions
+        mask = np.array([[[True, False, False], [False, True, False]]])
+        with pytest.raises(IndexError):
+            test_array.__getitem__(mask)
+        with pytest.raises(IndexError):
+            test_array.__getitem__(mask)
+
+        # special cases: Python bools and bools nested in Python arrays are not supported
+        with pytest.raises(TypeError):
+            test_array.__getitem__((True,))
+        with pytest.raises(TypeError):
+            test_array.__getitem__((False,))
+        with pytest.raises(TypeError):
+            test_array.__getitem__((True, False))
+        with pytest.raises(TypeError):
+            test_array.__getitem__(([0, 1], [0, False]))
+        with pytest.raises(TypeError):
+            test_array.__getitem__(([0, 1], [0, pytensor.shared(True)]))
 
     def test_grad_1d(self):
         subi = 0
@@ -2492,13 +2488,12 @@ def test_AdvancedSubtensor(self):
             bivec.tag.test_value = bivec_val
 
         # Make sure it doesn't complain about test values
-        with config.change_flags(compute_test_value="raise"):
-            self._compile_and_check(
-                [admat, aivec],
-                [admat[1:3, aivec]],
-                [admat_val, aivec_val],
-                AdvancedSubtensor,
-            )
+        self._compile_and_check(
+            [admat, aivec],
+            [admat[1:3, aivec]],
+            [admat_val, aivec_val],
+            AdvancedSubtensor,
+        )
 
     def test_AdvancedSubtensor_bool(self):
         n = dmatrix()

tests/test_gradient.py

@@ -910,7 +910,6 @@ def test_grad_scale():
     assert np.allclose(out, (8, 4))
 
 
-@config.change_flags(compute_test_value="off")
 def test_undefined_grad_opt():
     # Make sure that undefined grad get removed in optimized graph.
     random = RandomStream(np.random.default_rng().integers(1, 2147462579))

tests/test_ifelse.py

@@ -500,12 +500,11 @@ def test_merge_ifs_true_false(self):
 
     def test_grad_test_values(self):
         # Regression test for test values of `ifelse` gradient.
-        with pytensor.config.change_flags(compute_test_value="raise"):
-            x = scalar("x")
-            with pytest.warns(FutureWarning):
-                x.tag.test_value = 1
-            # Used to crash due to undefined test value.
-            pytensor.grad(ifelse(0, x, x), x)
+        x = scalar("x")
+        with pytest.warns(FutureWarning):
+            x.tag.test_value = 1
+        # Used to crash due to undefined test value.
+        pytensor.grad(ifelse(0, x, x), x)
 
     def test_grad_int_value(self):
         w = pytensor.shared(np.random.random(10))