# RELEASE NOTES

## spatstat 1.55-1

### 05 April 2018

   We thank Jens Astrom, Ines Moncada, Mehdi Moradi and Nicholas Read for contributions.

### OVERVIEW

 * More support for tessellations.

 * Improved support for linear networks.

 * Fixed longstanding bug in leverage and influence diagnostics.

 * Minor improvements and bug fixes.

 * Version nickname: "Gamble Responsibly"

### NEW FUNCTIONS

 * summary.ssf  
    Summary method for a spatially sampled function (class 'ssf').

 * unstack.tess  
    Given a tessellation with multiple columns of marks,  
    take the columns one at a time, and return a list of tessellations,  
    each carrying only one of the original columns of marks.

### SIGNIFICANT USER-VISIBLE CHANGES

 * plot.tess  
    This plot method can now fill each tile with a different colour.  
    New arguments 'do.col', 'values', 'col' and 'ribargs'.  
    Old argument 'col' has been renamed 'border' for consistency.

 * integral.linim, integral.linfun  
    Argument 'domain' can now be a tessellation.

 * integral.ssf  
    Argument 'domain' can now be a tessellation.

 * as.owin.default  
    Now accepts a structure with entries named 'xmin,ymin,xmax,ymax'  
    in any order. This handles objects of class 'bbox' in the 'sf' package.

 * as.owin.default  
    Now detects objects of class "SpatialPolygons"  
    and issues a more helpful error message.

 * pseudoR2.ppm, pseudoR2.lppm  
    The null model now includes any offset terms, by default.  
    [Suggested by Jens Astrom.]  
    New argument 'keepoffset'.

 * closepairs.ppp  
    New argument 'periodic'

 * fitted.ppm  
    New argument 'ignore.hardcore'.

 * predict.ppm  
    New argument 'ignore.hardcore'.

 * leverage.ppm, influence.ppm, dfbetas.ppm  
    Computation has been vastly accelerated for models with Geyer interaction  
    fitted using isotropic or translation edge corrections.

 * leverage.ppm, influence.ppm, dfbetas.ppm  
    Virtually all models and edge corrections are now supported,  
    using a "brute force" algorithm. This can be slow in some cases.

 * cdf.test  
    Monte Carlo test runs faster.

 * summary.distfun, summary.funxy  
    Pixel resolution can now be controlled.

 * persp.funxy  
    Improved z-axis label.

 * plot.ppp  
    Improved placement of symbol legend when argument 'symap' is given.

 * plot.msr  
    Changed the default rule for bandwidth for smoothing the density.

### BUG FIXES

 1. nnmark, as.im.ssf  
     if marks(X) was a matrix rather than a data frame,   
     the results were completely incorrect (and had completely wrong format).  
     Fixed.

 2. predict.mppm  
     If the model included random effects, and if the library 'MASS' was not loaded,  
     the predictions were on the log scale (i.e. they were logarithms of the correct values).  
     [Spotted by Nicholas Read.]  
     Fixed.

 3. leverage.ppm, influence.ppm, dfbetas.ppm  
     Calculations were slightly incorrect for models with a hard core.  
     Fixed.

 4. leverage.ppm  
     The mean leverage value (shown as a contour line in plot.leverage.ppm)  
     was slightly incorrect for Gibbs models.  
     Fixed.

 5. Ops.msr  
     If the input data contained an auxiliary pixel image of  
     the density component of the measure (attribute "smoothdensity")  
     this image was not updated; it was copied to the output unchanged.  
     Plots of the resulting measure were incorrect, but calculations  
     with the measure were correct.  
     Fixed.

 6. integral.msr  
     If the result was a matrix, it was the transpose of the correct answer.  
     Fixed.

 7. "[.linim"  
     The result sometimes had the wrong class.  
     Fixed.

 8. "[.linnet"  
     In calculating L[W] where W is a window, the code ignored  
     segments of L that crossed W without having a vertex in W.  
     Fixed.

 9. nnmap  
     Crashed if W = NULL.  
     Fixed.

10. density.lpp, nncross.lpp  
     Crashed sometimes with an obscure message about "z$which".  
     [Spotted by Ines Moncada.]  
     Fixed.

11. as.im.distfun  
     Crashed, for the distfun of a point pattern, if approx=FALSE.  
     Fixed.

12. as.solist  
     Crashed when x was a 'layered' object.  
     Fixed.

13. linnet  
     Crashed in some trivial cases where there were no points or lines.  
     Fixed.