openModeller

This release fixes an error which occurred during normalization offset
calculation and made the DG_GARP and DG_GARP_BS algorithms produce null
models.

The serialization/deserialization mechanism was
reimplemented using generic configuration objects
(all available algorithms are now serializable).
om_console accepts two new keywords: "Output
model" and "Input model". A new framework for test
cases is available. Various problems with memory
leaks and uninitialized values were fixed. CSM
Kaiser-Gutman was removed from the build. The
SWIG/Java binding was removed from the build.
Projected maps now have the mask extent and cell
size of a specified "map format". Major
restructuring of directories, file locations, and
file names was done.

A DesktopGarp version of GARP (as GARP 2.1) has been added. GARP 3.x has been removed from the build due to instability under high dimensional datasets. The Best Subset Procedure was replaced with a new generic version that can be extended to encapsulate any GARP version. A bug in the Bioclim distance algorithm has been fixed (probabilities were greater than zero outside the envelope depending on the cutoff parameter). A bug in the sampler class has been fixed (when absence data was available, it was splitting train samplers once again and test samplers were coming out empty).

A new implementation of the GARP algorithm
(Genetic Algorithm for Rule-set Production) was
added, including the "best subsets procedure". The
CSM (Climate Space Model) algorithm now comes in
two flavors: using Kaiser-Gutman or Broken Stick
cutoff methods. Bioclim was refactored and split
into two different implementations. A SWIG
interface is available for Python. A prototype
SOAP interface is available. This is compatible
with GDAL 1.2.5. Reprojection capabilities were
added. A tool to visualize models in a
two-dimensional environment space was added.