Merge pull request #347 from borglab/feature/wrap-symbol-shorthand

Wrapper Updates for Global Functions
release/4.3a0
Varun Agrawal 2020-06-11 11:27:37 -05:00 committed by GitHub
commit 6400c505c6
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
15 changed files with 98 additions and 86 deletions

View File

@ -14,25 +14,18 @@ from __future__ import print_function
import math
import gtsam
import matplotlib.pyplot as plt
import numpy as np
from gtsam import symbol_shorthand_B as B
from gtsam import symbol_shorthand_V as V
from gtsam import symbol_shorthand_X as X
from gtsam.utils.plot import plot_pose3
from mpl_toolkits.mplot3d import Axes3D
import gtsam
from gtsam.utils.plot import plot_pose3
from PreintegrationExample import POSES_FIG, PreintegrationExample
BIAS_KEY = int(gtsam.symbol(ord('b'), 0))
def X(key):
"""Create symbol for pose key."""
return gtsam.symbol(ord('x'), key)
def V(key):
"""Create symbol for velocity key."""
return gtsam.symbol(ord('v'), key)
BIAS_KEY = B(0)
np.set_printoptions(precision=3, suppress=True)

View File

@ -8,27 +8,19 @@ from __future__ import print_function
import math
import matplotlib.pyplot as plt
import numpy as np
from mpl_toolkits.mplot3d import Axes3D # pylint: disable=W0611
import gtsam
import gtsam.utils.plot as gtsam_plot
import matplotlib.pyplot as plt
import numpy as np
from gtsam import (ISAM2, BetweenFactorConstantBias, Cal3_S2,
ConstantTwistScenario, ImuFactor, NonlinearFactorGraph,
PinholeCameraCal3_S2, Point3, Pose3,
PriorFactorConstantBias, PriorFactorPose3,
PriorFactorVector, Rot3, Values)
def X(key):
"""Create symbol for pose key."""
return gtsam.symbol(ord('x'), key)
def V(key):
"""Create symbol for velocity key."""
return gtsam.symbol(ord('v'), key)
from gtsam import symbol_shorthand_B as B
from gtsam import symbol_shorthand_V as V
from gtsam import symbol_shorthand_X as X
from mpl_toolkits.mplot3d import Axes3D # pylint: disable=W0611
def vector3(x, y, z):
@ -115,7 +107,7 @@ def IMU_example():
newgraph.push_back(PriorFactorPose3(X(0), pose_0, noise))
# Add imu priors
biasKey = gtsam.symbol(ord('b'), 0)
biasKey = B(0)
biasnoise = gtsam.noiseModel_Isotropic.Sigma(6, 0.1)
biasprior = PriorFactorConstantBias(biasKey, gtsam.imuBias_ConstantBias(),
biasnoise)

View File

@ -13,9 +13,10 @@ Author: Alex Cunningham (C++), Kevin Deng & Frank Dellaert (Python)
from __future__ import print_function
import numpy as np
import gtsam
import numpy as np
from gtsam import symbol_shorthand_L as L
from gtsam import symbol_shorthand_X as X
# Create noise models
PRIOR_NOISE = gtsam.noiseModel_Diagonal.Sigmas(np.array([0.3, 0.3, 0.1]))
@ -26,11 +27,11 @@ MEASUREMENT_NOISE = gtsam.noiseModel_Diagonal.Sigmas(np.array([0.1, 0.2]))
graph = gtsam.NonlinearFactorGraph()
# Create the keys corresponding to unknown variables in the factor graph
X1 = gtsam.symbol(ord('x'), 1)
X2 = gtsam.symbol(ord('x'), 2)
X3 = gtsam.symbol(ord('x'), 3)
L1 = gtsam.symbol(ord('l'), 4)
L2 = gtsam.symbol(ord('l'), 5)
X1 = X(1)
X2 = X(2)
X3 = X(3)
L1 = L(4)
L2 = L(5)
# Add a prior on pose X1 at the origin. A prior factor consists of a mean and a noise model
graph.add(gtsam.PriorFactorPose2(X1, gtsam.Pose2(0.0, 0.0, 0.0), PRIOR_NOISE))

View File

@ -82,7 +82,7 @@ else:
print ("Done!")
if args.plot:
resultPoses = gtsam.extractPose2(result)
resultPoses = gtsam.utilities_extractPose2(result)
for i in range(resultPoses.shape[0]):
plot.plot_pose2(1, gtsam.Pose2(resultPoses[i, :]))
plt.show()

View File

@ -65,7 +65,7 @@ else:
print ("Done!")
if args.plot:
resultPoses = gtsam.allPose3s(result)
resultPoses = gtsam.utilities_allPose3s(result)
for i in range(resultPoses.size()):
plot.plot_pose3(1, resultPoses.atPose3(i))
plt.show()

View File

@ -1,7 +1,7 @@
These examples are almost identical to the old handwritten python wrapper
examples. However, there are just some slight name changes, for example
`noiseModel.Diagonal` becomes `noiseModel_Diagonal` etc...
Also, annoyingly, instead of `gtsam.Symbol('b', 0)` we now need to say `gtsam.symbol(ord('b'), 0))`
Also, instead of `gtsam.Symbol('b', 0)` we can simply say `gtsam.symbol_shorthand_B(0)` or `B(0)` if we use python aliasing.
# Porting Progress

View File

@ -10,24 +10,20 @@ A structure-from-motion problem on a simulated dataset
"""
from __future__ import print_function
import gtsam
import matplotlib.pyplot as plt
import numpy as np
import gtsam
from gtsam import symbol_shorthand_L as L
from gtsam import symbol_shorthand_X as X
from gtsam.examples import SFMdata
from gtsam.gtsam import (Cal3_S2, DoglegOptimizer,
GenericProjectionFactorCal3_S2, Marginals,
NonlinearFactorGraph, Point3, Pose3,
PriorFactorPoint3, PriorFactorPose3, Rot3,
NonlinearFactorGraph, PinholeCameraCal3_S2, Point3,
Pose3, PriorFactorPoint3, PriorFactorPose3, Rot3,
SimpleCamera, Values)
from gtsam.utils import plot
def symbol(name: str, index: int) -> int:
""" helper for creating a symbol without explicitly casting 'name' from str to int """
return gtsam.symbol(ord(name), index)
def main():
"""
Camera observations of landmarks (i.e. pixel coordinates) will be stored as Point2 (x, y).
@ -74,7 +70,7 @@ def main():
# Add a prior on pose x1. This indirectly specifies where the origin is.
# 0.3 rad std on roll,pitch,yaw and 0.1m on x,y,z
pose_noise = gtsam.noiseModel_Diagonal.Sigmas(np.array([0.3, 0.3, 0.3, 0.1, 0.1, 0.1]))
factor = PriorFactorPose3(symbol('x', 0), poses[0], pose_noise)
factor = PriorFactorPose3(X(0), poses[0], pose_noise)
graph.push_back(factor)
# Simulated measurements from each camera pose, adding them to the factor graph
@ -83,14 +79,14 @@ def main():
for j, point in enumerate(points):
measurement = camera.project(point)
factor = GenericProjectionFactorCal3_S2(
measurement, measurement_noise, symbol('x', i), symbol('l', j), K)
measurement, measurement_noise, X(i), L(j), K)
graph.push_back(factor)
# Because the structure-from-motion problem has a scale ambiguity, the problem is still under-constrained
# Here we add a prior on the position of the first landmark. This fixes the scale by indicating the distance
# between the first camera and the first landmark. All other landmark positions are interpreted using this scale.
point_noise = gtsam.noiseModel_Isotropic.Sigma(3, 0.1)
factor = PriorFactorPoint3(symbol('l', 0), points[0], point_noise)
factor = PriorFactorPoint3(L(0), points[0], point_noise)
graph.push_back(factor)
graph.print_('Factor Graph:\n')
@ -99,10 +95,10 @@ def main():
initial_estimate = Values()
for i, pose in enumerate(poses):
transformed_pose = pose.retract(0.1*np.random.randn(6,1))
initial_estimate.insert(symbol('x', i), transformed_pose)
initial_estimate.insert(X(i), transformed_pose)
for j, point in enumerate(points):
transformed_point = Point3(point.vector() + 0.1*np.random.randn(3))
initial_estimate.insert(symbol('l', j), transformed_point)
initial_estimate.insert(L(j), transformed_point)
initial_estimate.print_('Initial Estimates:\n')
# Optimize the graph and print results

View File

@ -10,8 +10,9 @@ This example will perform a relatively trivial optimization on
a single variable with a single factor.
"""
import numpy as np
import gtsam
import numpy as np
from gtsam import symbol_shorthand_X as X
def main():
@ -33,7 +34,7 @@ def main():
prior = gtsam.Rot2.fromAngle(np.deg2rad(30))
prior.print_('goal angle')
model = gtsam.noiseModel_Isotropic.Sigma(dim=1, sigma=np.deg2rad(1))
key = gtsam.symbol(ord('x'), 1)
key = X(1)
factor = gtsam.PriorFactorRot2(key, prior, model)
"""

View File

@ -13,23 +13,14 @@ Author: Duy-Nguyen Ta (C++), Frank Dellaert (Python)
from __future__ import print_function
import matplotlib.pyplot as plt
import numpy as np
from mpl_toolkits.mplot3d import Axes3D # pylint: disable=W0611
import gtsam
import gtsam.utils.plot as gtsam_plot
import matplotlib.pyplot as plt
import numpy as np
from gtsam import symbol_shorthand_L as L
from gtsam import symbol_shorthand_X as X
from gtsam.examples import SFMdata
def X(i):
"""Create key for pose i."""
return int(gtsam.symbol(ord('x'), i))
def L(j):
"""Create key for landmark j."""
return int(gtsam.symbol(ord('l'), j))
from mpl_toolkits.mplot3d import Axes3D # pylint: disable=W0611
def visual_ISAM2_plot(result):

View File

@ -19,11 +19,8 @@ from gtsam.gtsam import (Cal3_S2, GenericProjectionFactorCal3_S2,
NonlinearFactorGraph, NonlinearISAM, Point3, Pose3,
PriorFactorPoint3, PriorFactorPose3, Rot3,
PinholeCameraCal3_S2, Values)
def symbol(name: str, index: int) -> int:
""" helper for creating a symbol without explicitly casting 'name' from str to int """
return gtsam.symbol(ord(name), index)
from gtsam import symbol_shorthand_L as L
from gtsam import symbol_shorthand_X as X
def main():
@ -58,7 +55,7 @@ def main():
# Add factors for each landmark observation
for j, point in enumerate(points):
measurement = camera.project(point)
factor = GenericProjectionFactorCal3_S2(measurement, camera_noise, symbol('x', i), symbol('l', j), K)
factor = GenericProjectionFactorCal3_S2(measurement, camera_noise, X(i), L(j), K)
graph.push_back(factor)
# Intentionally initialize the variables off from the ground truth
@ -66,7 +63,7 @@ def main():
initial_xi = pose.compose(noise)
# Add an initial guess for the current pose
initial_estimate.insert(symbol('x', i), initial_xi)
initial_estimate.insert(X(i), initial_xi)
# If this is the first iteration, add a prior on the first pose to set the coordinate frame
# and a prior on the first landmark to set the scale
@ -75,12 +72,12 @@ def main():
if i == 0:
# Add a prior on pose x0, with 0.3 rad std on roll,pitch,yaw and 0.1m x,y,z
pose_noise = gtsam.noiseModel_Diagonal.Sigmas(np.array([0.3, 0.3, 0.3, 0.1, 0.1, 0.1]))
factor = PriorFactorPose3(symbol('x', 0), poses[0], pose_noise)
factor = PriorFactorPose3(X(0), poses[0], pose_noise)
graph.push_back(factor)
# Add a prior on landmark l0
point_noise = gtsam.noiseModel_Isotropic.Sigma(3, 0.1)
factor = PriorFactorPoint3(symbol('l', 0), points[0], point_noise)
factor = PriorFactorPoint3(L(0), points[0], point_noise)
graph.push_back(factor)
# Add initial guesses to all observed landmarks
@ -88,7 +85,7 @@ def main():
for j, point in enumerate(points):
# Intentionally initialize the variables off from the ground truth
initial_lj = points[j].vector() + noise
initial_estimate.insert(symbol('l', j), Point3(initial_lj))
initial_estimate.insert(L(j), Point3(initial_lj))
else:
# Update iSAM with the new factors
isam.update(graph, initial_estimate)

View File

@ -15,17 +15,18 @@ from __future__ import print_function
import unittest
import gtsam
import numpy as np
from gtsam import symbol_shorthand_X as X
from gtsam.utils.test_case import GtsamTestCase
import numpy as np
def create_graph():
"""Create a basic linear factor graph for testing"""
graph = gtsam.GaussianFactorGraph()
x0 = gtsam.symbol(ord('x'), 0)
x1 = gtsam.symbol(ord('x'), 1)
x2 = gtsam.symbol(ord('x'), 2)
x0 = X(0)
x1 = X(1)
x2 = X(2)
BETWEEN_NOISE = gtsam.noiseModel_Diagonal.Sigmas(np.ones(1))
PRIOR_NOISE = gtsam.noiseModel_Diagonal.Sigmas(np.ones(1))

View File

@ -33,7 +33,7 @@ class TestDSFMap(GtsamTestCase):
# testing the merge feature of dsf
dsf.merge(pair1, pair2)
self.assertEquals(key(dsf.find(pair1)), key(dsf.find(pair2)))
self.assertEqual(key(dsf.find(pair1)), key(dsf.find(pair2)))
if __name__ == '__main__':

View File

@ -279,7 +279,7 @@ def plot_trajectory(fignum, values, scale=1, marginals=None):
marginals (gtsam.Marginals): Marginalized probability values of the estimation.
Used to plot uncertainty bounds.
"""
pose3Values = gtsam.allPose3s(values)
pose3Values = gtsam.utilities_allPose3s(values)
keys = gtsam.KeyVector(pose3Values.keys())
lastIndex = None

33
gtsam.h
View File

@ -1977,6 +1977,35 @@ size_t symbol(char chr, size_t index);
char symbolChr(size_t key);
size_t symbolIndex(size_t key);
namespace symbol_shorthand {
size_t A(size_t j);
size_t B(size_t j);
size_t C(size_t j);
size_t D(size_t j);
size_t E(size_t j);
size_t F(size_t j);
size_t G(size_t j);
size_t H(size_t j);
size_t I(size_t j);
size_t J(size_t j);
size_t K(size_t j);
size_t L(size_t j);
size_t M(size_t j);
size_t N(size_t j);
size_t O(size_t j);
size_t P(size_t j);
size_t Q(size_t j);
size_t R(size_t j);
size_t S(size_t j);
size_t T(size_t j);
size_t U(size_t j);
size_t V(size_t j);
size_t W(size_t j);
size_t X(size_t j);
size_t Y(size_t j);
size_t Z(size_t j);
}///\namespace symbol
// Default keyformatter
void PrintKeyList (const gtsam::KeyList& keys);
void PrintKeyList (const gtsam::KeyList& keys, string s);
@ -2141,6 +2170,7 @@ class Values {
void insert(size_t j, const gtsam::EssentialMatrix& essential_matrix);
void insert(size_t j, const gtsam::PinholeCameraCal3_S2& simple_camera);
void insert(size_t j, const gtsam::imuBias::ConstantBias& constant_bias);
void insert(size_t j, const gtsam::NavState& nav_state);
void insert(size_t j, Vector vector);
void insert(size_t j, Matrix matrix);
@ -2158,10 +2188,11 @@ class Values {
void update(size_t j, const gtsam::Cal3Bundler& cal3bundler);
void update(size_t j, const gtsam::EssentialMatrix& essential_matrix);
void update(size_t j, const gtsam::imuBias::ConstantBias& constant_bias);
void update(size_t j, const gtsam::NavState& nav_state);
void update(size_t j, Vector vector);
void update(size_t j, Matrix matrix);
template<T = {gtsam::Point2, gtsam::Point3, gtsam::Rot2, gtsam::Pose2, gtsam::SO3, gtsam::SO4, gtsam::SOn, gtsam::Rot3, gtsam::Pose3, gtsam::Cal3_S2, gtsam::Cal3DS2, gtsam::Cal3Bundler, gtsam::EssentialMatrix, gtsam::imuBias::ConstantBias, Vector, Matrix}>
template<T = {gtsam::Point2, gtsam::Point3, gtsam::Rot2, gtsam::Pose2, gtsam::SO3, gtsam::SO4, gtsam::SOn, gtsam::Rot3, gtsam::Pose3, gtsam::Cal3_S2, gtsam::Cal3DS2, gtsam::Cal3Bundler, gtsam::EssentialMatrix, gtsam::imuBias::ConstantBias, gtsam::NavState, Vector, Matrix}>
T at(size_t j);
/// version for double

View File

@ -54,7 +54,16 @@ struct GlobalFunction: public FullyOverloadedFunction {
// function name in Cython pxd
std::string pxdName() const { return "pxd_" + pyRename(name_); }
// function name in Python pyx
std::string pyxName() const { return pyRename(name_); }
std::string pyxName() const {
std::string result = "";
for(size_t i=0; i<overloads[0].namespaces_.size(); i++){
if (i >= 1) {
result += (overloads[0].namespaces_[i] + "_");
}
}
result += pyRename(name_);
return result;
}
// emit cython wrapper
void emit_cython_pxd(FileWriter& pxdFile) const;