Display color raster image using a matrix of colors

Display color raster image using a matrix of colors

imageByColors(
  x,
  useRaster = FALSE,
  fixRasterRatio = TRUE,
  maxRatioFix = 100,
  xaxt = "s",
  yaxt = "s",
  doPlot = TRUE,
  cellnote = NULL,
  cexCellnote = 1,
  srtCellnote = 0,
  fontCellnote = 1,
  groupCellnotes = TRUE,
  groupBy = c("column", "row"),
  groupByColors = TRUE,
  adjBy = c("column", "row"),
  adjustMargins = FALSE,
  interpolate = getOption("interpolate", TRUE),
  verbose = FALSE,
  xpd = NULL,
  bty = par("bty"),
  flip = c("none", "y", "x", "xy"),
  keepTextAlpha = FALSE,
  doTest = FALSE,
  add = FALSE,
  ...
)

Arguments

x

matrix or data.frame containing colors

useRaster

logical sent to imageDefault to enable raster rendering, as opposed to polygon rendering. This parameter is highly recommended when the matrix is large (>50 columns or rows).

fixRasterRatio

logical sent to imageDefault.

maxRatioFix

numeric sent to imageDefault.

xaxt, yaxt

character values compatible with par to determine whether x- and y-axes are plotted. Set both to "n" to suppress display of axes.

doPlot

logical whether to create a plot, or simply return data which would have been used to create the plot.

cellnote

matrix or data.frame of labels to be displayed on the image. If groupCellnotes==TRUE labels will be placed in the center of consecutive cells with the same label and identical color. Currently, cell text is colored using setTextContrastColor which uses either white or black depending upon the brightness of the background color.

cexCellnote, srtCellnote, fontCellnote

numeric vectors, with values applied to cellnote text to be compatible with par("cex"), par("srt"), and par("font"), respectively. If supplied a matrix or data.frame with it is used as-is or expanded to equivalent dimensions of x. If the vector is named by colnames(x) then it is applied by column in order, otherwise it is applied by row, with values recycled to the number of columns or rows, respectively. Note cexCellnote can also be a list, with the list elements being applied to individual cells in order. If the list is named by colnames(x), each list element is applied to values in each column, in order. In future this parameter may also accept a matrix of cex values as input. Final note: values are applied to each cell, but when cell labels are combined with groupCellnotes==TRUE, the value for the first matching cell is used. Remember that values are placed by coordinate, bottom-to-top on the y-axis, and left-to-right on the x-axis.

groupCellnotes

logical whether to group labels where consecutive cells contain the same label and identical cell colors, thus only displaying one label in the center of these groups.

groupBy

character value indicating the direction to group cellnotes, when groupCellnotes=TRUE: "row" will group cellnote values by row; "column" will group cellnote values by column. By default, it will first group cellnotes by "row" then by "column".

groupByColors

logical indicating whether the cellnote grouping should also include the cell color. When groupByColors=FALSE, cellnote values will be grouped together regardless whether the underlying colors change, which may be preferred when applying text label to topographical data.

adjBy

character value indicating how to apply adjustments for cexCellnote, srtCellnote, and fontCellnote, as described above.

adjustMargins

logical indicating whether to adjust the axis label margins to ensure enough room to draw the text rownames and colnames.

interpolate

logical whether to implement image interpolation, by default TRUE when useRaster=TRUE.

verbose

logical whether to print verbose output.

xpd

NULLL or logical used for par("xpd") to define whether to crop displayed output to the plot area. If xpd=NULL then par("xpd") will not be modified, otherwise par("xpd"=xpd) will be defined while adding any cell notes, then reverted to its previous value afterward. This parameter is mainly useful when cellnote labels may overhang the plot space, and would be cropped and not visible if par("xpd"=TRUE).

keepTextAlpha

logical passed to setTextContrastColor indicating whether the text label color should inherit the alpha transparency from the background color. If TRUE then fully transparent background colors will not have a visible label.

doTest

logical whether to run a test showing basic features of imageByColors.

Details

This function is similar to image except that it takes a matrix which already has colors defined for each cell. This function calls imageDefault which enables updated use of the useRaster functionality.

Additionally, if cellnote is supplied, which contains a matrix of labels for the image cells, those labels will also be displayed. By default, labels are grouped, so that only one label is displayed whenever two or more labels appear in consecutive cells. This behavior can be disabled with groupCellnotes=FALSE.

The groupCellnotes behavior uses breaksByVector() to determine where to place consecutive labels, and it applies this logic starting with rows, then columns. Note that labels are only grouped when both the cell color and the cell label are identical for consecutive cells.

In general, if a large rectangular set of cells contains the same label, and cell colors, the resulting label will be positioned in the center. However, when the square is not symmetric, the label will be grouped only where consecutive columns contain the same groups of consecutive rows for a given label. In theory one could use polygon functions from the sp or rgeos package to detect contiguous polygons, and position the label in the center of those polygons.

It is helpful to rotate labels partially to prevent overlaps, e.g. srtCellnote=10 or srtCellnote=80.

To do:

• Detect the size of the area being labeled and determine whether to rotate the label sideways.

• Detect the size of the label, compared to its bounding box, and resize the label to fit the available space.

• Optionally draw border around contiguous colored and labeled polygons. Whether to draw border based only upon color, or color and label, or just label... it may get confusing.

• Label proper contiguous polygons based upon color and label, especially when color and label are present on multiple rows and columns, but not always the same columns per row.

See also

Other jam plot functions: adjustAxisLabelMargins(), coordPresets(), decideMfrow(), drawLabels(), getPlotAspect(), groupedAxis(), imageDefault(), minorLogTicksAxis(), nullPlot(), plotPolygonDensity(), plotRidges(), plotSmoothScatter(), shadowText_options(), shadowText(), showColors(), smoothScatterJam(), sqrtAxis(), usrBox()

Examples

a1 <- c("red4","blue")[c(1,1,2)];
b1 <- c("yellow","orange")[c(1,2,2)];
c1 <- c("purple","orange")[c(1,2,2)];
d1 <- c("purple","green4")[c(1,2,2)];
df1 <- data.frame(a=a1, b=b1, c=c1, d=d1);

# default using polygons
imageByColors(df1, cellnote=df1);


# useRaster=TRUE, edges are slightly blurred with small tables
imageByColors(df1, cellnote=df1, useRaster=TRUE);


# some text features, rotation, font size, etc
imageByColors(df1, cellnote=df1, useRaster=TRUE, adjBy="column",
   cexCellnote=list(c(1.5,1.5,1), c(1,1.5), c(1.6,1.2), c(1.6,1.5)),
   srtCellnote=list(c(90,0,0), c(0,45), c(0,0,0), c(0,90,0)));