Summary

This PR exposes the PriorFactor<EssentialMatrix> template instantiation to the Python wrapper in gtsam/nonlinear/nonlinear.i.

While gtsam::EssentialMatrix is a fully defined 5-DoF manifold in the C++ core (gtsam/geometry/EssentialMatrix.h), its corresponding prior factor (PriorFactor) was previously omitted from the Python interface. Exposing this class allows researchers and engineers working in Python to place a direct prior anchor constraint on an Essential Matrix (representing relative camera rotation and unit translation direction) during non-linear optimization.

Changes

• Added gtsam::EssentialMatrix to the template type parameter list T of PriorFactor in gtsam/nonlinear/nonlinear.i.
• This triggers the auto-generation of the gtsam.PriorFactorEssentialMatrix binding class during the build phase.

How This Was Tested

1. Compiled GTSAM from source locally with Python bindings enabled (-DGTSAM_BUILD_PYTHON=ON).
2. Verified that the PriorFactorEssentialMatrix class is correctly generated, exposed, and functional inside a Python 3.10 interpreter:

import gtsam

# Create a key
key = 1

# Create an Essential Matrix (Identity rotation, unit translation along X)
rot = gtsam.Rot3()
direction = gtsam.Unit3(1.0, 0.0, 0.0)
essential = gtsam.EssentialMatrix(rot, direction)

# Create a 5-DoF noise model for the Essential Matrix tangent space
noise = gtsam.noiseModel.Isotropic.Precision(5, 100.0)

# Instantiate the newly exposed PriorFactor!
prior = gtsam.PriorFactorEssentialMatrix(key, essential, noise)

print(prior)
# Output:
# PriorFactor on 1
#   prior mean: R:
#  [
# 	1, 0, 0;
# 	0, 1, 0;
# 	0, 0, 1
# ]
# d: :1
# 0
# 0
# isotropic dim=5 sigma=0.1