|Sep 5, 2019|
|1 year 10 months ||Minor tweak to CsvIO
|Aug 31, 2019|
|1 year 10 months ||Removal of two debugging loops. Why is this always so easy to see
once it's online?
|1 year 10 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 10 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 10 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 10 months ||Cosmetic changes: making sure we use ii, jj, kk etc indices so
searches work better
|1 year 10 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 10 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 10 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 10 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 10 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 10 months ||Fixed some missing includes and wrong namespace that prevented
|Aug 27, 2019|
|1 year 10 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 10 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 10 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 10 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 10 months ||Removing unnecessary Proj4 dependency of statistical_filters
example, refining example in FilterMaths, mini cleanup.
|Aug 26, 2019|
|1 year 11 months ||Renaming the traditional Makefile in order to make it clearer it is
only an example... See issue 777
|1 year 11 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 11 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 …