Gstlearn

=============
- MacOS compatibility (Intel and Silicon architectures)
- New dependency: Eigen library for matrices (dense and sparse)
- Fix issue 88, 98, 102, 101, 119, 115, 111
- loadDoc/loadData function convention improvement
- Efficiency improvements (variography, moving neighborhood, kriging systems...)
- Adding new tutorial (with widgets) to describe the calculation of experimental variograms and use of model for Cvv
- Adding description of main classes (Db, Vario and daughter classes, Model and daughter classes).
- Improve the Tutorial on widgets
- Improve the Tutorial on Neighborhood to treat data along profiles
- New extension package (for R only): io.R (interface with 'sf' and 'terra')
- Improving the Tutorial_Model (building a Model by pieces)
- Fix createFromCSV (accept all end-of-line conventions 'CR', 'LF', 'CRLF')
- Fix input/output text file path creation under windows (input data CSV files)
- Improve numerical precision of double values in neutral files (15 digits)
- Function db_model() has been renamed as Model::buildVmapOnDbGrid()
- Function model_stabilize() becomes Model::stabilize()
- Function model_normalize() becomes Model::standardize()
- Function model_get_nonugget_cova() is now included in Model::getCovaNumber(true)
- Function db_vmap_compute() is renamed into db_vmap()
- Function db_variogram_cloud() is renamed as db_vcloud()
- Function variogram_trans_cut() is renamed as Vario::transformCut()
- Function correlation_ident() is renamed as correlationIdentify()
- Function variogram_mlayers() is renamed as Vario::computeGeometryMLayers()
- Function variovect_compute() is renamed as Vario::computeVarioVect()
- New functions for Squeeze-and-Stretch in Grid format:
DbGrid::createSqueezeAndStretchForward() for forward processing and
DbGrid::createSqueezeAndStretchBackward() for backward processing
- New method 'select' for ANeigh (gain of memory and time)
- Pluging the library "eigen" for faster and more accurate algebra.
- Adding the display of known mean (when relevant)
- The class Neigh/ANeigh is introduced (generic for any neighborhood)
- New class Geometry/ABiPointCheck checks if 2 points are connected (for Neighborhood, for Variogram, ...)
- Class PolySet is now called PolyElem
- The Python function display_file.py has been renamed into show_file.py
- reduceComplexity in Polygons
- New name for argument of migrate* functions: ldmax -> distType
and change type for flag_inter and flag_fill (become boolean)
- Correction in kribayes for multivariate aspect.
- In DbGrid::createCoveringDb, argument names have changed: nodes -> nx, dcell -> dx, origin -> x0
- Some Performance tests have been added (or renamed) in TestInter such as:
. Bench_KrigingU.cpp
. Bench_KrigingI.cpp
- db_tool_duplicate() is renamed as DbHelper::findDuplicates()
- db_normalize() is renamed as DbHelper::normalizeVariables()
- db_grid_fill() is renamed as DbHelper::gridFilling()
- Introducing the generic method "fromTL" to convert R object into gstlearn ones

=============
- First version generated from a public github repo
- Remove geopandas python package dependency (issue 68) - No more gstlearn.proj module!!
- Fix Mac OS compilation
- Fix python and R readme
- Create Polygons from WKT files (QGIS export)
- New world.poly data file built from from NaturalEarth.com data
- Fix default R plots options usage
- Update credits in readmes
- Fix global estimation result access
- Factorization of courses content in MD files
- Fix filtering in kriging feature
- New function for creating experimental variogram from R data frames (Vario_fromDF)
- Improve experimental variogram calculation efficiency
- Fix initGeographic when using figsize in Python (locdims was not defined)
- Make db_duplicate and db_tool_duplicate available thru swig
- Add loadData, loadDoc and setNoScroll functions in R and Python
- Fix inline equations rendering in html from jupyter notebooks
- The ways to render MD documentation files has changed in Python and R (as an indented note)

=============
- Prevent NamingConvention from creating variables with empty name

=============
- Update R readme
- Add severity check for model validity (used before kriging)
- Standardize graphics in R and Python
- Rename name1 and name2 into namex and namey for correlation representations
- Parameterize the amount of Nugget (technically) added in SPDE
- Corrections in the API functions of SPDE method
- New test (cpp) for benchmarking SPDE features
- Allow processing target variable with NA in SPDE API functions
- Suppress class MatrixSquareDiagonalCst
- Suppress class MatrixSquareDiagonal
- Important warning: the Drift part of the Model Neutral File has changed.
- Drift* is not a serializable class anymore (only Model class is)
- Adding Legend to plot.varmod for plot.R
- flagCste replaced by flagCst everywhere
- The Enum called EDrift is abandoned
- Function dbRegressionByColIdx() has been abandoned (use dbRegression instead)
- Function regressionByUID() has been abandoned (use regression instead)
- Improvement of Table printout (row name not truncated anymore)
- New class 'ResReg' for storing Regression formula (AStringable)
- Integrate R documentation for functions in plot.R into the R package
- Names in plot.R and plot.py are standardized: using camel-writing systematically
- New method Db_fromPanda for creating Db from Panda Frame (numerical columns only)
- New method Db_fromDF for creating Db from data.frame (numerical columns only)
- Fix functions having "String()" default value in r (replaced by "")
- dbStatististics* functions refactoring (same function for printing and calculating)
- New design for VarioParam / Vario
- Possibility to convert variogram to external language (Python or R) for In and Out

=============
- Fix Issues 76 & 77
- Update official website URL
- Improve SPDE class API

=============
- Add How to cite in READMEs
- Add publish demos and publish doxygen workflows