# RELEASE NOTES

## spatstat 1.56-0

### 15 June 2018

   We thank Sebastian Meyer and Suman Rakshit for contributions.

### OVERVIEW

 * Kernel estimate of intensity as a function(x,y)

 * Extract discrete and continuous components of a measure.

 * Improvements and extensions to leverage and influence code.

 * Plot a line segment pattern using line widths.

 * Find connected components of each tile in a tessellation.

 * Geometrical operations on 'distfun' objects.

 * Join vertices in a linear network.

 * Distance map and contact distribution for rectangular structuring element.

 * An infinite number of infinitesimal bugs has been detected and fixed.

 * Version nickname: "Bondi Tram"

### NEW FUNCTIONS

 * densityfun, densityfun.ppp  
    Compute a kernel estimate of intensity of a point pattern  
    and return it as a function of spatial location.

 * as.im.densityfun  
    Convert function(x,y) to a pixel image.

 * measureDiscrete, measureContinuous  
    Extract the discrete and continuous components of a measure.

 * connected.tess  
    Find connected components of each tile in a tessellation  
    and make a new tessellation composed of these pieces.

 * dffit.ppm  
    Effect change diagnostic DFFIT for spatial point process models.

 * shift.distfun, rotate.distfun, reflect.distfun, flipxy.distfun,  
    affine.distfun, scalardilate.distfun  
    Methods for geometrical operations on 'distfun' objects.

 * rescale.distfun  
    Change the unit of length in a 'distfun' object.

 * plot.indicfun  
    Plot method for indicator functions created by as.function.owin.

 * Smooth.leverage.ppm, Smooth.influence.ppm  
    Smooth a leverage function or an influence measure.

 * integral.leverage.ppm, integral.influence.ppm  
    Compute the integral of a leverage function or an influence measure.

 * mean.leverage.ppm  
    Compute the mean value of a leverage function.

 * rectdistmap  
    Distance map using rectangular metric.

 * rectcontact  
    Contact distribution function using rectangular structuring element.

 * joinVertices  
    Join specified vertices in a linear network.

### SIGNIFICANT USER-VISIBLE CHANGES

 * plot.psp  
    Segments can be plotted with widths proportional to their mark values.  
    New argument 'style'.

 * msr  
    Infinite and NA values are now detected (if check=TRUE)  
    and are reset to zero, with a warning.

 * leverage.ppm, influence.ppm, dfbetas.ppm  
    Faster computation in some cases.

 * as.im.nnfun, as.im.smoothfun  
    New argument 'approx' chooses between a fast, approximate algorithm  
    and a slow, exact algorithm.

 * cdf.test  
    More jittering is applied when jitter=TRUE.  
    Warnings about tied values should not occur any more.

 * plot.im  
    Improved behaviour when all pixel values are NA.

 * plot.tess  
    Now generates a separate plot panel for each column of marks,  
    if do.col=TRUE.  
    New argument 'multiplot'.

 * pcfinhom  
    Now handles correction='good'

 * solist  
    New argument '.NameBase'

 * runifpointOnLines, rpoisppOnLines  
    New argument 'drop'

 * plot.studpermutest  
    This existing function now has a help file.

 * linnet  
    When argument 'edges' is specified, the code now checks whether  
    any edges are duplicated.

### BUG FIXES

 1. kernel.moment  
     Result was incorrect for kernel="cosine" and kernel="optcosine".  
     Fixed.

 2. "[.msr"  
     Format was mangled if the subset contained exactly one quadrature point.  
     Fixed.

 3. tess  
     If a list of tiles was given, and the tiles were pixel images or masks,  
     their pixel resolutions were ignored, and reset to the default 128x128.  
     Fixed.

 4. plot.linim  
     Ignored argument 'legend' when style="colour".  
     [Spotted by Suman Rakshit.]  
     Fixed.

 5. leverage.ppm, influence.ppm, dfbetas.ppm  
     Crashed sometimes if the model was fitted by method="logi".  
     Fixed.

 6. Smooth.ppp  
     Crashed with some combinations of the arguments varcov and weights  
     when X had several columns of marks.  
     Fixed.

 7. plot.tess  
     Crashed sometimes when do.col=TRUE and values=NULL.  
     Fixed.

 8. dilation.ppp  
     Crashed if polygonal=FALSE.

 9. plot.tess  
     Ignored the marks of a tessellation defined by a pixel image.  
     Fixed.

10. predict.ppm  
     Crashed when computing a confidence interval or prediction interval  
     if 'window' was a tessellation.  
     Fixed.

11. tiles  
     names(tiles(x)) was not always identical to tilenames(x).  
     Fixed.

12. model.images.lppm  
     Crashed in many cases.  
     Fixed.

13. model.images.lppm  
     Names or dimnames were missing in some cases.  
     Fixed.

14. nncross.ppp  
     Result had the wrong format if npoints(X) = 0 and length(what) = 1.  
     (Spotted by Sebastian Meyer).  
     Fixed.

15. plot.colourmap  
     Crashed if the colourmap was defined on an interval of dates or times.  
     Fixed.

16. StraussHard, leverage.ppm, influence.ppm, dfbetas.ppm  
     Leverage and influence calculations generated an error   
     when applied to models fitted with the StraussHard interaction.  
     Fixed.

17. "[.ppp"  
     Crashed if there were several columns of marks,  
     some of which were factors, and drop=TRUE.  
     Fixed.