Release Notes: The precipitation reaccumulation has been redone, and a new filter has been implemented to distribute precipitation over a past period. More undercatch correction algorithms have been implemented, and more rain gauges are supported, including the Chinese Standard Precipitation Gauge and the shielded Geonor gauge. ASCII parsing has been optimized for speed along with other speed optimizations. Several bugs have been fixed related to time zone handling (for solar radiation and dates).
Release Notes: A large fraction of the code has been refactored, enabling new data handling algorithms as well as significant performance improvements. A new stage has been added: data generators based on parametrizations. Lots of micro-optimizations, for SIMD as well as suggested from static analysis have been implemented alongside new numerical approximations. Lots of rarely triggered bugs have been fixed both in the main code and in the data plugins. More precise Dates and Timers have been implemented. The error reporting has been improved and exceptions can now display a graphical message box.
Release Notes: The handling of plugins has been redone in order to be less confusing for the users (there is no PLUGIN_PATH to set anymore). The sun model has been improved and several rare bugs fixed. Several plugins are now handling large files in a much more efficient manner. Several speed optimizations have been implemented through the code, as well as a lot of cleanup in order to conform to C++ best practices. This version compiles for Android.
Release Notes: Temporal resampling and reaccumulation have been redesigned, fixing major behavior inconsistencies and leading to major speed improvements. Inconsistencies in the filters have also been fixed, and new filters added. Two new plugins have been developed to read GRIB files and to write PNGs. The grid handling API has been significantly improved to become more generic. Lots of small bugs have been fixed, and lots of code has been cleaned up and reorganized.
Release Notes: A major performance improvement has been found by slightly changing the API. This leads to an 8 times speed up. Several bugs have been fixed and several plugins improved in order to get the most out of the available data. Some new filters have been implemented. More error messages have been improved for clarity.
Release Notes: This release brings much improved packaging to Windows and Mac. A lot of cleanup has been done in the code as well as the build system. Many legacy hacks have been fixed, leading to bugfixes and some speed improvements. Significant efforts have been put into the statistical methods, implementing a least square fitting class, statistical analysis methods, and a more consistent interface to such methods.
Release Notes: This version now supports Windows with Visual C++ or Code::Blocks. Several plugins have been significantly improved in term of robustness, speed, and error messages. The various array structures have been improved by adding new method and improving existing ones. Speed has been improved and multiple bugs fixed. In order to help applications that provide their own data and need to bypass the data reading by the plugins, data injection capability has been added.
Release Notes: The structure has been deeply modified in order to address several shortcomings and bugs. A new user interface class has been designed. The filtering infrastructure was rewritten. The spatial interpolations were integrated. The buffering was refocused on buffering issues. The meteorological data structures have been merged with the metadata structures in order to make the user code simpler and more compact. A first set of meteorological physical laws has been introduced around sun and standard atmosphere models.
Release Notes: Several new spatial interpolations algorithms have been implemented, and some bugs fixed. The accumulation resampling algorithm has been rewritten from scratch. The JNative spatial interpolations interface has been improved, and now compiles on Linux and Windows. The build system has been improved and the auto-detection of third party libraries is much improved.
Release Notes: Several bugs have been fixed: among others, the highly specific (and quite dangerous) accumulation filter has been replaced by the accumulation resampling, and the DEM azimuths and slopes are now given in degrees. This release also introduces speed optimizations in 2D arrays and spatial interpolations, a matrix class, and several statistical methods for time series. Several new regression fitting models have been implemented, and the linear regression can now automatically remove as many outliers as necessary, depending on the initial size of the data set.