v0.5.2

• Fix color option in solution plots
• Fix color option in empirical variogram plots
• Check worker processes with nworkers() instead of nprocs()
• Fix type inconsistencies in theoretical variograms
• Fix type inconsistencies in Kriging estimators
• Major rewrite of Kriging estimators internals
• ImageQuilting.ipynb example notebook
• h-scatter plots
• Update documentation

v0.5.1

• Fix #17: Add Ellipsoidal distance back meanwhile JuliaStats/Distances.jl#81 is not merged
• Add EmpiricalDistribution meanwhile JuliaStats/Distributions.jl#661 is not merged
• Parallelize visual comparison and k-fold cross-validation

v0.5.0

This is a breaking release. We kindly ask users to update their code accordingly.

New features

• Support to any metric defined in Distances.jl
• Empirical cross-variogram implementation
• Visual comparison and k-fold cross-validation (see SolverComparison.ipynb)
• LocallyWeightedRegression.jl estimation solver

Breaking changes

• EuclideanDistance, HaversineDistance and EllipsoidDistance were revised for performance and are now part of Distances.jl under the names Euclidean, Haversine and Ellipsoidal

• DataFrames v0.11.0 was released with breaking changes and GeoStats.jl was updated accordingly

Other improvements

• Updated examples and documentation
• Review of README with links to solver build status

v0.4.4

• Add coefficient of 3 to Gaussian and exponential variograms. We ask users to please update their code for consistency with other Geostatistical software.
• Minor fixes to plot recipes.

v0.4.3

• Import solve_single to make it visible to external solvers
• Remove assertions in Kriging estimators for performance

v0.4.2

• Make use of GeoStatsDevTools.jl and new GeoStatsBase.jl specifications
• All index locations are now of type Int instead of arbitrary Integer

v0.4.1

• Migrate problem and solution interface to GeoStatsBase.jl
• Sequential Gaussian simulation on point collections

v0.4.0

This is an important release of the project with a whole new infrastructure for general estimation and stochastic simulation problems over general domains. It is also an informal call to researchers interested in contributing to the GeoStats.jl ecosystem.

New features

• Abstraction of problems and solvers for geostatistical estimation/simulation
• New base package GeoStatsBase.jl for solver writers
• Kriging and Sequential Gaussian simulation solvers
• Parallel stochastic simulation

v0.3.7

• Solve covariance instead of variogram if the latter possesses the 2nd-order stationarity property
• Add power, cubic, pentaspherical and sine hole theoretical variogram models
• Add method isstationary for theoretical variogram models
• Add Kriging with external drift
• Add documentation for plotting

v0.3.6

• Add function GeoStats.examples() for launching notebooks locally
• Add visual regression tests for plot recipes