Merge pull request #345 from borglab/feature/logging_optimizer

Add logging (hooked) optimizer

cython/gtsam/tests/test_logging_optimizer.py

@@ -0,0 +1,79 @@
+"""
+Unit tests for optimization that logs to comet.ml.
+Author: Jing Wu and Frank Dellaert
+"""
+# pylint: disable=invalid-name
+
+import unittest
+from datetime import datetime
+
+import gtsam
+import numpy as np
+from gtsam import Rot3
+from gtsam.utils.test_case import GtsamTestCase
+
+from gtsam.utils.logging_optimizer import gtsam_optimize
+
+KEY = 0
+MODEL = gtsam.noiseModel_Unit.Create(3)
+
+
+class TestOptimizeComet(GtsamTestCase):
+    """Check correct logging to comet.ml."""
+
+    def setUp(self):
+        """Set up a small Karcher mean optimization example."""
+        # Grabbed from KarcherMeanFactor unit tests.
+        R = Rot3.Expmap(np.array([0.1, 0, 0]))
+        rotations = {R, R.inverse()}  # mean is the identity
+        self.expected = Rot3()
+
+        graph = gtsam.NonlinearFactorGraph()
+        for R in rotations:
+            graph.add(gtsam.PriorFactorRot3(KEY, R, MODEL))
+        initial = gtsam.Values()
+        initial.insert(KEY, R)
+        self.params = gtsam.GaussNewtonParams()
+        self.optimizer = gtsam.GaussNewtonOptimizer(
+            graph, initial, self.params)
+
+    def test_simple_printing(self):
+        """Test with a simple hook."""
+
+        # Provide a hook that just prints
+        def hook(_, error: float):
+            print(error)
+
+        # Only thing we require from optimizer is an iterate method
+        gtsam_optimize(self.optimizer, self.params, hook)
+
+        # Check that optimizing yields the identity.
+        actual = self.optimizer.values()
+        self.gtsamAssertEquals(actual.atRot3(KEY), self.expected, tol=1e-6)
+
+    @unittest.skip("Not a test we want run every time, as needs comet.ml account")
+    def test_comet(self):
+        """Test with a comet hook."""
+        from comet_ml import Experiment
+        comet = Experiment(project_name="Testing",
+                           auto_output_logging="native")
+        comet.log_dataset_info(name="Karcher", path="shonan")
+        comet.add_tag("GaussNewton")
+        comet.log_parameter("method", "GaussNewton")
+        time = datetime.now()
+        comet.set_name("GaussNewton-" + str(time.month) + "/" + str(time.day) + " "
+                       + str(time.hour)+":"+str(time.minute)+":"+str(time.second))
+
+        # I want to do some comet thing here
+        def hook(optimizer, error: float):
+            comet.log_metric("Karcher error",
+                             error, optimizer.iterations())
+
+        gtsam_optimize(self.optimizer, self.params, hook)
+        comet.end()
+
+        actual = self.optimizer.values()
+        self.gtsamAssertEquals(actual.atRot3(KEY), self.expected)
+
+if __name__ == "__main__":
+    unittest.main()

cython/gtsam/utils/logging_optimizer.py

@@ -0,0 +1,54 @@
+"""
+Optimization with logging via a hook.
+Author: Jing Wu and Frank Dellaert
+"""
+# pylint: disable=invalid-name
+
+from typing import TypeVar
+
+from gtsam import NonlinearOptimizer, NonlinearOptimizerParams
+import gtsam
+
+T = TypeVar('T')
+
+
+def optimize(optimizer: T, check_convergence, hook):
+    """ Given an optimizer and a convergence check, iterate until convergence.
+        After each iteration, hook(optimizer, error) is called.
+        After the function, use values and errors to get the result.
+        Arguments:
+            optimizer (T): needs an iterate and an error function.
+            check_convergence: T * float * float -> bool
+            hook -- hook function to record the error
+    """
+    # the optimizer is created with default values which incur the error below
+    current_error = optimizer.error()
+    hook(optimizer, current_error)
+
+    # Iterative loop
+    while True:
+        # Do next iteration
+        optimizer.iterate()
+        new_error = optimizer.error()
+        hook(optimizer, new_error)
+        if check_convergence(optimizer, current_error, new_error):
+            return
+        current_error = new_error
+
+
+def gtsam_optimize(optimizer: NonlinearOptimizer,
+                   params: NonlinearOptimizerParams,
+                   hook):
+    """ Given an optimizer and params, iterate until convergence.
+        After each iteration, hook(optimizer) is called.
+        After the function, use values and errors to get the result.
+        Arguments:
+                optimizer {NonlinearOptimizer} -- Nonlinear optimizer
+                params {NonlinearOptimizarParams} -- Nonlinear optimizer parameters
+                hook -- hook function to record the error
+    """
+    def check_convergence(optimizer, current_error, new_error):
+        return (optimizer.iterations() >= params.getMaxIterations()) or (
+            gtsam.checkConvergence(params.getRelativeErrorTol(), params.getAbsoluteErrorTol(), params.getErrorTol(),
+                                   current_error, new_error))
+    optimize(optimizer, check_convergence, hook)