Base

The base module provides fundamental data structures, utilities, and mathematical operations that form the foundation of GTSAM. Here are the key header files:

Core Mathematical Concepts¶

Utilities to define group, manifold, and Lie group classes, as well as concept checks for them.

• Manifold.h - Defines the concept for differentiable manifolds with retract/local operations. See Manifold docs.

• Lie.h - Implements the Lie group concept combining group operations with manifold structure. See LieGroup docs.

• MatrixLieGroup.h - Base class for matrix Lie groups (SO(n), SE(n), SL(n), etc.) with Hat/Vee operators. See MatrixLieGroup docs.

And, less common:

• Group.h - Defines the group concept with identity, inverse, and composition operations. See Group docs.

• VectorSpace.h - Defines the vector space concept. See VectorSpace docs.

• lieProxies.h - Proxy functions for testing

Constructed Lie Groups¶

These templates build larger Lie groups from existing group types while preserving GTSAM’s traits, tangent-vector ordering, and optional-Jacobian conventions:

• ProductLieGroup.h constructs independent direct products G × H. See the ProductLieGroup guide for its componentwise operations, adjoints, and dynamic-dimension behavior.

• SemidirectLieGroup.h constructs action-coupled products G ⋉ H. See the SemidirectLieGroup guide for action requirements, matrix-function kernels, and coupled Jacobians.

• TangentLieGroup.h constructs the tangent group TG = G ⋉ 𝔤, where G acts on its algebra through the adjoint action. See the TangentLieGroup guide for its structured Jacobians, recursive use, and optimized SO(3) specialization.

Linear Algebra¶

Linear algebra operations optimized for robotics applications. Vector and Matrix are essentially typedefs and wrappers for Eigen:

• Vector.h - Vector operations and mathematical functions

• Matrix.h - Essential matrix operations and linear algebra utilities built on Eigen

• cholesky.h - Cholesky decomposition algorithms for solving linear systems

These are used in Jacobian and Hessian factors:

• VerticalBlockMatrix.h - Specialized block matrix structure for efficient sparse operations

• SymmetricBlockMatrix.h - Symmetric block matrix implementation for optimization problems

Jacobians¶

• numericalDerivative.h - Numerical differentiation utilities for automatic derivative computation

• OptionalJacobian.h - Optional Jacobian computation wrapper for optimization efficiency

Container and Data Structures¶

• High-performance container types (FastSet, FastMap, etc.) optimized for GTSAM (unclear whether still relevant for recent c++ compilers)

• DSFMap.h - Disjoint Set Forest implementation using maps for union-find operations

• DSFVector.h - Disjoint Set Forest implementation using vectors for better performance

• TreeTraversal.h - Tree traversal algorithms and utilities for factor graphs

• ConcurrentMap.h - Thread-safe concurrent map implementation

Debugging and Development Tools¶

• debug.h - Debugging utilities, assertion macros, and conditional compilation flags

• timing.h - Performance timing and profiling utilities for benchmarking

• TestableAssertions.h - Assertion macros for unit testing with tolerance checking

• chartTesting.h - Utilities for testing manifold chart operations

Sampling and Statistics¶

• WeightedSampler.h - Weighted random sampling utilities for Monte Carlo methods

• sampler.h - General sampling utilities and random number generation

Graph Algorithms¶

• Kruskal.h - Kruskal’s algorithm for minimum spanning trees in factor graphs

• BTree.h - Balanced tree implementation for efficient storage and retrieval

Type System and Traits¶

• types.h - Common type definitions, aliases, and forward declarations

• concepts.h - C++ concept definitions for type checking and constraints

• Testable.h - Base class and concept for objects that can be tested for equality

Serialization Support¶

• SerializationBase.h - Base serialization functionality for saving/loading objects

• StdOptionalSerialization.h - Serialization support for std::optional types

Memory Management and Performance¶

• utilities.h - General utility functions, memory management helpers, and common operations

• FastDefaultAllocator.h - Custom allocator for improved performance in critical paths

• Value.h - Type-erased value wrapper for heterogeneous containers

Template Metaprogramming¶

• treeTraversal-inst.h - Template instantiations for tree traversal algorithms

• concepts.h - SFINAE utilities and type trait helpers