 |
IMP
2.2.1
The Integrative Modeling Platform
|
|
IMP
2.2.1
The Integrative Modeling Platform
|
This module contains general purpose algebraic and geometric methods that are expected to be used by a wide variety of IMP modules.
IMP has a number of geometry primitives. They support the following namespace functions as appropriate
In addition, they cannot be compared against one another due to floating point implementation issues (eg Vector3D v=v2 does not imply v==v2).
Geometric primitives are not put into a defined state by their constructor. Such classes mimic POD types (int, float etc) in C++ and are optimized for efficiency. All operations on a default initialized instance other than assigning to it from a non-default initialized instance should be assumed to be invalid.
Many of the geometric primitives and operations in IMP are written to work in any dimension. In C++, this is implemented via templates (such as IMP::algebra::VectorD). In the python side, the different dimensions are named explicitly instead. That means, a 2-D point is IMP::algebra::VectorD<2> in C++, and IMP::algbra::Vector2D in python and the function IMP::algebra::get_basis_vector_d<3>() in C++ becomes IMP.algebra.get_basis_vector_3d() in Python. Similarly, a collection of 2D points is IMP::base::Vector<IMP::algebra::VectorD<2> > in C++ and IMP.algebra.Vector2Ds in python, which as with all collections, look like python lists. For convenience, we provide typedefs in C++ to the IMP::algbra::Vector2D and IMP::algebra::Vector2Ds style names.
Geometry in IMP can be stored in a variety of ways. For example, a point in 3D can be stored using an IMP::algebra::VectorD<3> or using an IMP::core::XYZ particle. It is often useful to be able to write algorithms that work on sets of points without worrying how they are stored, the Generic Geometry layer provides that. It works using a set of functions get_vector_3d() and set_vector_3d() which manipulate the geometry in terms of the IMP::algebra representation of the geometry in question. That is, get_vector_3d() returns a IMP::algebra::VectorD<3> for both an IMP::algebra::Vector3D and a IMP::core::XYZ. Algorithms take their arguments as C++ templates and use the generic geometry methods to manipulate the geometry. And versions of the function for both types of storage are exported to python, so one could also write generic functions in python.
For example, IMP::atom::get_rmsd() takes any combination of IMP::algebra::Vector3Ds or IMP::core::XYZs or IMP::core::XYZsTemp as arguments. Versions for all combinations of those are exported to python.
ANN is a library implementing fast nearest neighbor searches. Certain data structures will be faster if it is installed. While compilation of the library from source is quite straight forward, it is not avaible as a package for common platforms. In addition, ANN must be built as a shared library rather than a static library.
We ship a copy of Eigen as part of IMP as Eigen support varies quite considerably across platforms and it is header-only and so easy to ship. Our version of Eigen has been patched to move everything to the IMP_Eigen namespace and preprocessor symbols to start with IMP_EIGEN_ so that things do not conflict with system installs of eigen.
Author(s): Daniel Russel, Keren Lasker, Ben Webb, Javier Angel Velazquez-Muriel
Maintainer: benmwebb
License: LGPL This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version.
Publications: