C++ Layer

While GalSim is primarily a Python package, much of the implementation is done in C++ for improved speed for the hard-core numerical calculations.

If you would like to use the C++-layer functions in your own C++ project, you can do so with the caveat that these are officially implementation details, so we don’t strictly enforce semantic versioning for the C++-layer API. That is, function signatures may change on minor version updates (e.g. 2.3.x to 2.4.0). We don’t often make huge changes to the C++ API, so most of the time you will be fine upgrading, but you should be prepared that you may need to update your code when upgrading GalSim. (We do guarantee that we won’t change the C++ API for bugfix updates, e.g. 2.4.1 to 2.4.2.)

Note

While we don’t intend to break ABI compatibility on bugfix releases, we don’t currently have any mechanisms in place to ensure ABI compatibility. So it is possible that you will need to recompile your code when updating to a new GalSim version.

The other caveat is that we haven’t put much energy into documenting the C++ layer functions. The following comes from a combination of Doxygen and Breathe to shoehorn it into the Sphinx structure. But the docs are pretty bare bones in places. Sorry about that. If you use these and want to pretty up these docs, a PR doing so would be much appreciated. :)

When compiling your code, all of the public functionality should be included simply by using #include "GalSim.h" with the appropriate -I directive when compiling to find the GalSim include directory. The appropriate directory name is accessible from python by running the command:

python -c "import galsim; print(galsim.include_dir)"

• C++ Images
• C++ Surface Brightness Profiles
• Positions and Bounds
• Noise-related Functionality
• Photons and Sensor Effects
• Interpolation Tools
• HSM Implementation
• Math

Linking Your Code

If you install GalSim using conda (see Installing With Conda), then the appropriate C++ library file is included in the conda packaging. It should be installed into either the main conda lib directory (e.g. /anaconda/lib) or the one for your conda environment (e.g. /anaconda/envs/myenv/lib).

If you don’t use conda, then you will need to build the lib file yourself, since we don’t include it in the pip package. See Installing the C++ Shared Library for instructions on installing it.

There are both versioned and unversioned copies of the library. On OSX, these are libgalsim.M.m.dylib and libgalsim.dylib. On Linux, they are libgalsim.M.m.so and libgalsim.so.

You should link by specifying the appropriate directory with -L and link with -lgalsim.

Version control

We provide a number of functions to help you ensure that your code remains compatible with updates to GalSim.

First, there are 3 MACROS that you can use to make a compile-time assert that you are coding to the right version:

GALSIM_MAJOR

GALSIM_MINOR

GALSIM_REVISION

Then there are three functions that return the compiled versions of those same numbers:

int galsim::major_version()

int galsim::minor_version()

int galsim::revision()

One can also get the full three-number version as a string (e.g. “1.4.2”)

std::string galsim::version()

And finally, we provide a fucntion that checks that the header file being included matches the compiled values in the library being linked to:

inline bool galsim::check_version()