9     def __call__(self, tao, x):  argument
14     def __call__(self, tao, x, g):  argument
28     def create(self, tao):  argument
32     def destroy(self, tao):  argument
36     def setFromOptions(self, tao):  argument
39     def setUp(self, tao):  argument
41         self.testvec = tao.getSolution().duplicate()
43     def solve(self, tao):  argument
46     def step(self, tao, x, g, s):  argument
48         tao.computeGradient(x, g)
52     def preStep(self, tao):  argument
55     def postStep(self, tao):  argument
58     def monitor(self, tao):  argument
64         self.tao = PETSc.TAO()
65         self.tao.createPython(MyTao(), comm=PETSc.COMM_SELF)
74         self.tao.destroy()
75         self.tao = None
80         ctx = self.tao.getPythonContext()
82         self.assertTrue(self.tao.getPythonType() == pytype)
85         tao = self.tao
86         ctx = tao.getPythonContext()
87         x = PETSc.Vec().create(tao.getComm())
92         tao.setObjective(Objective())
93         tao.setGradient(Gradient(), None)
94         tao.setMonitor(ctx.monitor)
95         tao.setFromOptions()
96         tao.setMaximumIterations(3)
98         def _update(tao, it, cnt):  argument
102         tao.setUpdate(_update, (cnt_up,))
103         tao.setSolution(x)
107         tao.solve()
108         n = tao.getIterationNumber()
114         tao.solve()
115         n = tao.getIterationNumber()
116         self.assertGreater(tao.getConvergedReason(), 0)
124         tao.solve()
125         n = tao.getIterationNumber()
126         self.assertGreater(tao.getConvergedReason(), 0)
140         tao.cancelMonitor()
143         return self.tao.getPythonContext()
150     def create(self, tao):  argument
152         self._ls.useTAORoutine(tao)
156     def destroy(self, tao):  argument
159     def setUp(self, tao):  argument
162     def solve(self, tao):  argument
163         x = tao.getSolution()
164         gradient = tao.getGradient()[0]
166         for it in range(tao.getMaximumIterations()):
167             tao.setIterationNumber(it)
170             tao.computeGradient(x, gradient)
177             tao.monitor(f=f, res=gradient.norm())
182             if tao.checkConverged() > 0:
185     def step(self, tao, x, g, s):  argument
188     def preStep(self, tao):  argument
191     def postStep(self, tao):  argument
197         self.tao = PETSc.TAO()
198         self.tao.createPython(MyGradientDescent(), comm=PETSc.COMM_SELF)
201         self.tao.destroy()
202         self.tao = None
205         tao = self.tao
211         tao.setOptionsPrefix('test_tao_python_optimiser_')
212         tao.setFromOptions()
214         x = PETSc.Vec().createSeq(2, comm=tao.getComm())
217         tao.setSolution(x)
218         tao.setObjective(Objective())
219         tao.setGradient(Gradient(), x.copy())
221         tao.solve()
223         self.assertEqual(tao.getMaximumIterations(), 100)
224         self.assertAlmostEqual(tao.getTolerances()[0], 1e-6)
225         self.assertGreater(tao.getIterationNumber(), 0)
226         self.assertGreater(tao.getConvergedReason(), 0)
229         self.assertGreater(tao.getObjectiveValue(), 0)
230         self.assertAlmostEqual(tao.getObjectiveValue(), 0, places=5)