|Sep 5, 2019|
|1 year 7 months ||Minor tweak to CsvIO
|Aug 31, 2019|
|1 year 7 months ||Removal of two debugging loops. Why is this always so easy to see
once it's online?
|1 year 7 months ||Singular value decomposition implemented, albeit the most easy one,
and therefore probably still not enough to tackle matrices that
|Aug 30, 2019|
|1 year 7 months ||One step towards singular values decomposition: Eigenvalues and
vectors of symmetrical matrices via Jacobi. My code, so no license
troubles. Further styling and updating a dead link.
|1 year 7 months ||The handling of location information in the data section had been
butchered by the latest changes to location handling, it is now
|1 year 7 months ||Cosmetic changes: making sure we use ii, jj, kk etc indices so
searches work better
|1 year 7 months ||Minor code cleanup in libsmet (factoring a little bit more the
append/not append code). Improved documentation in Date as well as
making internal arrays private. White spaces changes in Coords.
|Aug 29, 2019|
|1 year 7 months ||Offering an alternative, completely independent, solver for linear
systems via Gauss elimination with partial pivoting. Matrix
inversion is about 2.4 times slower than the Doolittle version but
it can handle rogue zeros at the diagonal.
|1 year 7 months ||Interestingly, the reference lat/lon in the tests were slightly off,
leading to errors when comparing the provided lat/lon with the
re-computed lat/lon from easting/northing...
|1 year 7 months ||So, the mysterious compilation messages related to strerror now make
sense: depending on the include, it is either std::strerror or not
part of std! It has been changed to be consistent...
|Aug 28, 2019|
|1 year 7 months ||Trying to address issue 774: now when both lat/lon and east/north
coordinates are provided, their consistency is checked (it used to
be this way a while ago but somehow this behavior ceased, probably
because of some refactoring in Coords).
|1 year 7 months ||Fixed some missing includes and wrong namespace that prevented
|Aug 27, 2019|
|1 year 7 months ||Ciao Eigen, you were just too big. Maybe we can learn from you that
all arrays should have the same base class.
|1 year 7 months ||Switching the particle and Kalman filter to the MeteoIO matrix
class. The results are indistinguishable in the example program, and
it's even 1/3rd faster. I hope I got all index increment
transitions, needs retesting of the more 'obscure' features.
|1 year 7 months ||A few matrix class enhancements to retrieve and set rows and cols by
index, get the diagonal, the max coefficient with index, and resize
with a data vector to init. Naive implementations, needs some love
at some point...
|1 year 7 months ||Adding a filter (RHWATERTOICE) to correct relative humidity over
water for the full temperature range to relative humidity over ice
when T below triple point and over water otherwise. For example
useful for weather stations that report relative humi…
|1 year 7 months ||Removing unnecessary Proj4 dependency of statistical_filters
example, refining example in FilterMaths, mini cleanup.
|Aug 26, 2019|
|1 year 7 months ||Renaming the traditional Makefile in order to make it clearer it is
only an example... See issue 777
|1 year 7 months ||One more memory usage optimization: we can actually know how many
data points will be used when we encounter a new stations, so we can
reserve the memory. Moreover, the debug symbols are now always
generated when compiling the examples.
|Aug 23, 2019|
|1 year 7 months ||Trying to reduce memory consumption (issue 769) as well as memory
re-allocations (is growing a vector because it was not sized at
start). Extracting and converting one year of half-hourly data now
takes 25 to 30% less memory. This has been achieved …