Merge pull request #323 from Kenneth-T-Moore/ken2

API change, remove outputs from compute_partials

openmdao/components/exec_comp.py

@@ -232,7 +232,7 @@ def compute(self, inputs, outputs):
         for expr in self._codes:
             exec(expr, _expr_dict, _IODict(outputs, inputs))
 
-    def compute_partials(self, inputs, outputs, partials):
+    def compute_partials(self, inputs, partials):
         """
         Use complex step method to update the given Jacobian.
 
@@ -241,9 +241,6 @@ def compute_partials(self, inputs, outputs, partials):
         inputs : `VecWrapper`
             `VecWrapper` containing parameters. (p)
 
-        outputs : `VecWrapper`
-            `VecWrapper` containing outputs and states. (u)
-
         partials : `Jacobian`
             Contains sub-jacobians.
         """
@@ -273,7 +270,7 @@ def compute_partials(self, inputs, outputs, partials):
                 else:
                     pwrap[param][idx] += step
 
-                uwrap = _TmpDict(outputs, return_complex=True)
+                uwrap = _TmpDict(self._outputs, return_complex=True)
 
                 # solve with complex param value
                 self._residuals.set_const(0.0)

openmdao/components/meta_model.py

@@ -312,16 +312,14 @@ def _vec_to_array2d(self, vec):
 
         return arr
 
-    def compute_partials(self, inputs, outputs, partials):
+    def compute_partials(self, inputs, partials):
         """
         Compute sub-jacobian parts. The model is assumed to be in an unscaled state.
 
         Parameters
         ----------
         inputs : Vector
             unscaled, dimensional input variables read via inputs[key]
-        outputs : Vector
-            unscaled, dimensional output variables read via outputs[key]
         partials : Jacobian
             sub-jac components written to partials[output_name, input_name]
         """

openmdao/core/explicitcomponent.py

@@ -296,7 +296,7 @@ def _linearize(self, do_nl=False, do_ln=False):
                 # negate constant subjacs (and others that will get overwritten)
                 # back to normal
                 self._negate_jac()
-                self.compute_partials(self._inputs, self._outputs, J)
+                self.compute_partials(self._inputs, J)
 
                 # re-negate the jacobian
                 self._negate_jac()
@@ -322,16 +322,14 @@ def compute(self, inputs, outputs):
         """
         pass
 
-    def compute_partials(self, inputs, outputs, partials):
+    def compute_partials(self, inputs, partials):
         """
         Compute sub-jacobian parts. The model is assumed to be in an unscaled state.
 
         Parameters
         ----------
         inputs : Vector
             unscaled, dimensional input variables read via inputs[key]
-        outputs : Vector
-            unscaled, dimensional output variables read via outputs[key]
         partials : Jacobian
             sub-jac components written to partials[output_name, input_name]
         """

openmdao/core/tests/test_approx_derivs.py

@@ -153,7 +153,7 @@ def test_arrray_comp(self):
 
         class DoubleArrayFD(DoubleArrayComp):
 
-            def compute_partials(self, inputs, outputs, partials):
+            def compute_partials(self, inputs, partials):
                 """
                 Override deriv calculation.
                 """
@@ -493,7 +493,7 @@ def test_arrray_comp(self, vec_class):
 
         class DoubleArrayFD(DoubleArrayComp):
 
-            def compute_partials(self, inputs, outputs, partials):
+            def compute_partials(self, inputs, partials):
                 """
                 Override deriv calculation.
                 """

openmdao/core/tests/test_check_derivs.py

@@ -27,7 +27,7 @@ def compute(self, inputs, outputs):
                 """ Doesn't do much. """
                 outputs['y'] = 3.0*inputs['x1'] + 4.0*inputs['x2']
 
-            def compute_partials(self, inputs, outputs, partials):
+            def compute_partials(self, inputs, partials):
                 """Intentionally incorrect derivative."""
                 J = partials
                 J['y', 'x1'] = np.array([4.0])
@@ -74,7 +74,7 @@ def compute(self, inputs, outputs):
                 """ Doesn't do much. """
                 outputs['y'] = 3.0*inputs['x1'] + 4.0*inputs['x2']
 
-            def compute_partials(self, inputs, outputs, partials):
+            def compute_partials(self, inputs, partials):
                 """Intentionally incorrect derivative."""
                 J = partials
                 J['y', 'x1'] = np.array([4.0])
@@ -117,7 +117,7 @@ def compute(self, inputs, outputs):
                 """ Doesn't do much. """
                 outputs['y'] = 3.0*inputs['x1'] + 4.0*inputs['x2']
 
-            def compute_partials(self, inputs, outputs, partials):
+            def compute_partials(self, inputs, partials):
                 """Intentionally incorrect derivative."""
                 J = partials
                 J['y', 'x1'] = np.array([4.0])
@@ -153,7 +153,7 @@ def compute(self, inputs, outputs):
                 """ Doesn't do much. """
                 outputs['y'] = 3.0*inputs['x1'] + 4.0*inputs['x2']
 
-            def compute_partials(self, inputs, outputs, partials):
+            def compute_partials(self, inputs, partials):
                 """Intentionally incorrect derivative."""
                 J = partials
                 J['y', 'x1'] = np.array([4.0])
@@ -193,7 +193,7 @@ def compute(self, inputs, outputs):
                 """ Doesn't do much. """
                 outputs['y'] = 3.0*inputs['x1'] + 4.0*inputs['x2']
 
-            def compute_partials(self, inputs, outputs, partials):
+            def compute_partials(self, inputs, partials):
                 """Intentionally incorrect derivative."""
                 J = partials
                 J['y', 'x1'] = np.array([4.0])
@@ -231,7 +231,7 @@ def compute(self, inputs, outputs):
                 """ Doesn't do much. """
                 outputs['y'] = 3.0*inputs['x1'] + 4.0*inputs['x2']
 
-            def compute_partials(self, inputs, outputs, partials):
+            def compute_partials(self, inputs, partials):
                 """Intentionally left out derivative."""
                 J = partials
                 J['y', 'x1'] = np.array([3.0])
@@ -317,7 +317,7 @@ def setup(self):
 
                 self.run_count = 0
 
-            def compute_partials(self, inputs, outputs, partials):
+            def compute_partials(self, inputs, partials):
                 partials['flow:T', 'T'] = 1.
                 partials['flow:P', 'P'] = 1.
 

openmdao/core/tests/test_expl_comp.py

@@ -27,7 +27,7 @@ def compute(self, inputs, outputs):
 
 class RectanglePartial(RectangleComp):
 
-    def compute_partials(self, inputs, outputs, partials):
+    def compute_partials(self, inputs, partials):
         partials['area', 'length'] = inputs['width']
         partials['area', 'width'] = inputs['length']
 

openmdao/core/tests/test_matmat.py

@@ -224,7 +224,7 @@ def compute(self, inputs, outputs):
         outputs['f_arclength'] = np.sqrt(1 + inputs['yp_lgl']**2)
 
 
-    def compute_partials(self, inputs, outputs, partials):
+    def compute_partials(self, inputs, partials):
         partials['f_arclength', 'yp_lgl'] = inputs['yp_lgl'] / np.sqrt(1 + inputs['yp_lgl']**2)
 
 

openmdao/core/tests/test_parallel_groups.py

@@ -211,7 +211,7 @@ def setup(self):
             def compute(self, inputs, outputs):
                 outputs['y'] = inputs['x'] * self.mult
 
-            def compute_partials(self, inputs, outputs, partials):
+            def compute_partials(self, inputs, partials):
                 partials['y', 'x'] = numpy.array([self.mult])
 
         prob = Problem()

openmdao/core/tests/test_reconf.py

@@ -26,7 +26,7 @@ def setup(self):
     def compute(self, inputs, outputs):
         outputs['y'] = 2 * inputs['x']
 
-    def compute_partials(self, inputs, outputs, jacobian):
+    def compute_partials(self, inputs, jacobian):
         jacobian['y', 'x'] = 2 * np.ones((self.size, 1))
 
 
@@ -39,7 +39,7 @@ def setup(self):
     def compute(self, inputs, outputs):
         outputs['z'] = 3 * inputs['x']
 
-    def compute_partials(self, inputs, outputs, jacobian):
+    def compute_partials(self, inputs, jacobian):
         jacobian['z', 'x'] = 3.0
 
 

openmdao/core/tests/test_reconf_dynamic.py

@@ -28,7 +28,7 @@ def setup(self):
     def compute(self, inputs, outputs):
         outputs['y'] = 2 * inputs['x']
 
-    def compute_partials(self, inputs, outputs, jacobian):
+    def compute_partials(self, inputs, jacobian):
         jacobian['y', 'x'] = 2 * np.ones((self.size, 1))
 
 
@@ -47,7 +47,7 @@ def setup(self):
     def compute(self, inputs, outputs):
         outputs['z'] = 3 * inputs['x']
 
-    def compute_partials(self, inputs, outputs, jacobian):
+    def compute_partials(self, inputs, jacobian):
         jacobian['z', 'x'] = 3.0
 
 

openmdao/core/tests/test_scaling.py

@@ -343,7 +343,7 @@ def setup(self):
             def compute(self, inputs, outputs):
                 outputs['y'] = 2.0*(inputs['x'] + 1.0)
 
-            def compute_partials(self, inputs, outputs, partials):
+            def compute_partials(self, inputs, partials):
                 """
                 Jacobian for Sellar discipline 1.
                 """

openmdao/docs/features/building_components/explicitcomp.rst

@@ -31,10 +31,9 @@ The implementation of each method will be illustrated using a simple explicit co
   .. embed-code::
       openmdao.core.tests.test_expl_comp.RectangleComp.compute
 
-- :code:`compute_partials(inputs, outputs, partials)` :
+- :code:`compute_partials(inputs, partials)` :
 
   [Optional] Compute the :code:`partials` (partial derivatives) given the :code:`inputs`.
-  The :code:`outputs` are also provided for convenience.
 
   .. embed-code::
       openmdao.core.tests.test_expl_comp.RectanglePartial.compute_partials

openmdao/docs/style_guide/sphinx_decorators.rst

@@ -29,16 +29,14 @@ and a method that looks like this and has the decorator applied to it:
    ::
 
     @mydecorator
-    def compute_partials(self, inputs, outputs, partials):
+    def compute_partials(self, inputs, partials):
         """
         Compute sub-jacobian parts / factorization.
 
         Parameters
         ----------
         inputs : Vector
             unscaled, dimensional input variables read via inputs[key]
-        outputs : Vector
-            unscaled, dimensional output variables read via outputs[key]
         partials : Jacobian
             sub-jac components written to partials[output_name, input_name]
         """
@@ -57,16 +55,14 @@ But if we add the call signature of the method to the docstring as the first lin
    ::
 
     @mydecorator
-    def compute_partials(self, inputs, outputs, partials):
+    def compute_partials(self, inputs, partials):
         """
         Compute sub-jacobian parts / factorization.
 
         Parameters
         ----------
         inputs : Vector
             unscaled, dimensional input variables read via inputs[key]
-        outputs : Vector
-            unscaled, dimensional output variables read via outputs[key]
         partials : Jacobian
             sub-jac components written to partials[output_name, input_name]
         """

openmdao/drivers/tests/test_pyoptsparse_driver.py

@@ -59,7 +59,7 @@ def compute(self, inputs, outputs):
         outputs['f_xy'] = (x-3.0)**2 + x*y + (y+4.0)**2 - 3.0
         self.eval_iter_count += 1
 
-    def compute_partials(self, inputs, outputs, partials):
+    def compute_partials(self, inputs, partials):
         """ Jacobian for our paraboloid."""
 
         if self.grad_iter_count == self.grad_fail_at:

openmdao/jacobians/tests/test_jacobian.py

@@ -43,7 +43,7 @@ def compute(self, inputs, outputs):
                            y[0]*17. - y[0]*y[1] + 2.*y[1]
         outputs['f'][1] = outputs['f'][0]*3.0
 
-    def compute_partials(self, inputs, outputs, partials):
+    def compute_partials(self, inputs, partials):
         x = inputs['x']
         y = inputs['y']
         jac1 = self._jac_type(np.array([
@@ -94,7 +94,7 @@ def compute(self, inputs, outputs):
         z = inputs['z']
         outputs['f'] = (w[0]-5.0)**2 + (w[1]+1.0)**2 + w[2]*6. + z*7.
 
-    def compute_partials(self, inputs, outputs, partials):
+    def compute_partials(self, inputs, partials):
         w = inputs['w']
         z = inputs['z']
         jac = self._jac_type(np.array([[
@@ -137,7 +137,7 @@ def setup(self):
         self.add_input('in', val=in_val*scale)
         self.add_output('out', val=out_val*scale)
 
-    def compute_partials(self, inputs, outputs, partials):
+    def compute_partials(self, inputs, partials):
         partials['out', 'in'] = self._constructor(self._value)
 
 

openmdao/jacobians/tests/test_jacobian_features.py

@@ -31,7 +31,7 @@ def compute(self, inputs, outputs):
                         + np.outer(np.ones(2), inputs['y2'])
                         + inputs['x']*np.eye(2))
 
-    def compute_partials(self, inputs, outputs, partials):
+    def compute_partials(self, inputs, partials):
         partials['f', 'x'] = 1.
         partials['f', 'z'] = np.ones((1, 4))
 
@@ -102,7 +102,7 @@ def compute(self, inputs, outputs):
         outputs['f'] = np.sum(inputs['z']) + inputs['x']
         outputs['g'] = np.outer(inputs['y1'] + inputs['y3'], inputs['y2']) + inputs['x'] * np.eye(2)
 
-    def compute_partials(self, inputs, outputs, partials):
+    def compute_partials(self, inputs, partials):
         pass
 
 
@@ -119,7 +119,7 @@ def setup(self):
 
         self.approx_partials('*', '*', **self.kwargs)
 
-    def compute_partials(self, inputs, outputs, partials):
+    def compute_partials(self, inputs, partials):
         pass
 
 
@@ -137,7 +137,7 @@ def setup(self):
         self.approx_partials('g', 'x', **self.kwargs)
         self.approx_partials('g', 'y2', **self.kwargs)
 
-    def compute_partials(self, inputs, outputs, partials):
+    def compute_partials(self, inputs, partials):
         partials['f', 'x'] = 1.
         partials['f', 'z'] = np.ones((1, 4))
 
@@ -157,7 +157,7 @@ def setup(self):
 
         self.declare_partials(**self.partial_kwargs)
 
-    def compute_partials(self, inputs, outputs, partials):
+    def compute_partials(self, inputs, partials):
         pass
 
 
@@ -403,7 +403,7 @@ def setup(self):
                 self.add_output('z', shape=(3,))
                 self.add_input('x', shape=(3,), units='degF')
 
-            def compute_partials(self, inputs, outputs, partials):
+            def compute_partials(self, inputs, partials):
                 partials['y', 'x'] = self.A
                 partials['z', 'x'] = self.A
 
@@ -464,7 +464,7 @@ def setup(self):
 
                 self.declare_partials(of='f', wrt='x', rows=[0,1,1,1], cols=[0,1,2,3])
 
-            def compute_partials(self, inputs, outputs, partials):
+            def compute_partials(self, inputs, partials):
                 pd = partials['f', 'x']
 
                 # Corresponds to the (0, 0) entry
@@ -503,7 +503,7 @@ def setup(self):
 
                 self.declare_partials(of='f', wrt='x', rows=[0, 1, 1, 1], cols=[0, 1, 2, 3])
 
-            def compute_partials(self, inputs, outputs, partials):
+            def compute_partials(self, inputs, partials):
                 # Corresponds to the [(0,0), (1,1), (1,2), (1,3)] entries.
                 partials['f', 'x'] = [1., 2., 3., 4.]
 
@@ -534,7 +534,7 @@ def setup(self):
                                       val=[1. , 2., 3., 4.])
                 self.declare_partials(of='f', wrt='y', val=sp.sparse.eye(2, format='csc'))
 
-            def compute_partials(self, inputs, outputs, partials):
+            def compute_partials(self, inputs, partials):
                 pass
 
         model = Group()

openmdao/solvers/linesearch/tests/test_backtracking.py

@@ -152,7 +152,7 @@ def compute(self, inputs, outputs):
 
                 outputs['f_xy'] = (x-3.0)**2 + x*y + (y+4.0)**2 - 3.0
 
-            def compute_partials(self, inputs, outputs, partials):
+            def compute_partials(self, inputs, partials):
                 """ Jacobian for our paraboloid."""
                 x = inputs['x']
                 y = inputs['y']