spatialwarnings v3.1.0 (Release date: unknown) ============================================== This release has received external contributions from K. Siteur Improvements: * The type of neighborhood can now be specified in all patch-related functions, including high-level functions (e.g. `patchdistr_sews`). Neighborhood can also be specified using the strings "von_neumann" or "moore", on top of a 3x3 matrix New indicators: * Clustering of pairs (e.g. Schneider et al. 2016) * Indicators related to the LSW distribution (Siteur et al. 2023) Bug fixes and misc changes: * Many documentation improvements and minor fixes * Old defunct functions have been removed spatialwarnings v3.0.3 (Release date: 2022-03-21) ================================================= This release has received external contributions from J. Guerber. Improvements: * In patch size distribution plots, fits are now rescaled when xmin is above 1 so they overlay nicely on the observed distribution * Better handling of errors in segmented::segmented() when fitting variograms Bug fixes: * Fixed an error occurring when using xmin = "estimate" in patchdistr_sews * Remove calls to assert() that were leftover from testing (CRAN policy) spatialwarnings v3.0.2 (Release date: 2021-12-16) ================================================= Improvements: * Matrix attributes are now preserved when generating null matrices (except when shuffling the original matrix, null_method = 'perm') Bug fixes: * Computation of Moran's I is now accurate for small matrices spatialwarnings v3.0.0 (Release date: 2021-05-19) ================================================= Improvements: * Documentation extended and improved * Fitting of distributions now uses 'plexpo' and 'trunc' to refer to the exponent of a power-law ("slope") and its exponential truncation (from "expo" and "rate", which were ambiguous) * EXPERIMENTAL: Variogram-based indicators have been added. Note that this deserves more testing and application to real-world data to make sure that the computation of variograms are accurate enough. * EXPERIMENTAL: Significance of flow length can now be assessed using the analytical approximation described in Rodriguez et al. (2017) Bug fixes: * Fixed a bug where available methods were not displayed in `summary()` * Fixed a bug where the r-spectrum was wrong when the matrix had an odd number of rows or columns Misc changes: * NAs in provided matrices now produce warnings instead of errors * Dependency to tidyr has been removed (this fixes occasional NOTEs on CRAN automated checking) Removals: * All deprecated functions in spatialwarnings v2.0.0 are now defunct spatialwarnings v2.0.0 (Release date: 2020-05-14) ================================================= Improvements: * New methods are available to produce null matrices, on top of shuffling the original matrix (e.g. based on smoothing the original matrix). * Significance of Power-law range can now be tested using `indictest()`. Using `plot_distr()` on the resulting objects will display the 0.05/0.95 quantiles of the null patch size distributions. * The package gains a generic function `display_matrix`, to eyeball the matrices being used in `spatialwarnings` objects using ggplot2 * Improved the fitting of distributions, which should be more robust now. * Speed improvements in label() Bug fixes: * Fixed a bug where the normalization constant for truncated power-laws was miscalculated Removals: * All the deprecated `*_spews` functions are now defunct (removed). * Most `indicator_*` functions are now deprecated. Misc changes: * Lots of duplicated code has been removed * Minor changes in print/summary methods * The FAQ has been expanded and improved. See the following page: https://alex.lecairn.org/spatialwarnings-faq.html * Dependency to VGAM is now removed spatialwarnings v1.3.1 (Release date: 2018-12) ============================================== New indicators: * Planar flowlength (Mayor et al. 2013, Rodriguez et al. 2017) * Kolmogorov complexity based on Block Decomposition Method (Dakos and Soler-Toscano 2016) Improvements: * Enable parallel computation of patch size distributions * Added a dataset of aerial view of vegetation in Arizona ('arizona') * Added functions to compute the coarse-grained variance/skewness on a single matrix Bug fixes and code improvements: * Added missing methods exports for custom indicators * Fixed the patch labelling for non-square images * General code cleanup and improvement Documentation and description changes: * Updated references to reflect the publication of new paper presenting the package spatialwarnings v1.2 (Release date: 2018-06) =============================================== Bug fixes: * Compilation errors should be fixed on Solaris * Fixed coarse-graining bug when input values are non-integer spatialwarnings v1.1 (Release date: 2018-06) ============================================ This release provides changes as to satisfy referees' comments prior to the publication of the package, as long as minor improvements in documentation. New features: * Support for custom indicators (see ?create_indicator) Minor changes: * Documentation improvements * Safer handling of small matrices in SDR computation Name changes * `*_spews` functions are now deprecated in favor of `*_sews` functions Bug fixes * Counting patches in matrices with 1 line or 1 column does not crash R anymore * The R package should now build on Solaris spatialwarnings v1.0 (Release date: 2017-11-03) =============================================== Initial release