Plotchart - Simple plotting and charting package
Plotchart is a Tcl-only package that focuses on the easy creation of xy-plots, barcharts and other common types of graphical presentations. The emphasis is on ease of use, rather than flexibility. The procedures that create a plot use the entire canvas window, making the layout of the plot completely automatic.
This results in the creation of an xy-plot in, say, ten lines of code:
package require Plotchart canvas .c -background white -width 400 -height 200 pack .c -fill both # # Create the plot with its x- and y-axes # set s [::Plotchart::createXYPlot .c {0.0 100.0 10.0} {0.0 100.0 20.0}] foreach {x y} {0.0 32.0 10.0 50.0 25.0 60.0 78.0 11.0 } { $s plot series1 $x $y } $s title "Data series"
A drawback of the package might be that it does not do any data management. So if the canvas that holds the plot is to be resized, the whole plot must be redrawn. The advantage, though, is that it offers a number of plot and chart types:
XY-plots like the one shown above with any number of data series.
Stripcharts, a kind of XY-plots where the horizontal axis is adjusted automatically. The result is a kind of sliding window on the data series.
Polar plots, where the coordinates are polar instead of cartesian.
Histograms, for plotting statistical information.
Isometric plots, where the scale of the coordinates in the two directions is always the same, i.e. a circle in world coordinates appears as a circle on the screen.
You can zoom in and out, as well as pan with these plots (Note: this works best if no axes are drawn, the zooming and panning routines do not distinguish the axes), using the mouse buttons with the control key and the arrow keys with the control key.
Piecharts, with automatic scaling to indicate the proportions.
Barcharts, with either vertical or horizontal bars, stacked bars or bars side by side.
Timecharts, where bars indicate a time period and milestones or other important moments in time are represented by triangles.
3D plots (both for displaying surfaces and 3D bars)
With version 1.5 a new command has been introduced: plotconfig, which can be used to configure the plot options for particular types of plots and charts (cf. CONFIGURATION OPTIONS)
You create the plot or chart with one single command and then fill the plot with data:
Create a new xy-plot (configuration type: xyplot).
Name of the existing canvas widget to hold the plot.
A 3-element list containing minimum, maximum and stepsize for the x-axis, in this order. For an inverted axis, where the maximum appears on the left-hand side, use: maximum, minimum and a negative stepsize.
A 3-element list containing minimum, maximum and stepsize for the y-axis, in this order. For an inverted axis, where the maximum appears at the bottom, use: maximum, minimum and a negative stepsize.
Create a new strip chart (configuration type: stripchart). The only difference to a regular XY plot is that the x-axis will be automatically adjusted when the x-coordinate of a new point exceeds the maximum.
Name of the existing canvas widget to hold the plot.
A 3-element list containing minimum, maximum and stepsize for the x-axis, in this order. Note that an inverted x-axis is not supported for this type of plot.
A 3-element list containing minimum, maximum and stepsize for the y-axis, in this order. For an inverted axis, where the maximum appears at the bottom, use: maximum, minimum and a negative stepsize.
Create a new time-x-plot (configuration type: txplot). The horizontal axis represents the date/time of the data and the vertical axis the values themselves.
Name of the existing canvas widget to hold the plot.
A 3-element list containing the minimum and maximum date/time to be shown and the stepsize (in days) for the time-axis, in this order. Note that an inverted time-axis is not supported.
A 3-element list containing minimum, maximum and stepsize for the vertical axis, in this order. For an inverted axis, where the maximum appears at the bottom, use: maximum, minimum and a negative stepsize.
Create a new xy-plot where the y-axis has a logarithmic scale (configuration type: xlogyplot).
The data should be given as for a linear scale, as the logarithmic transformation is taken of internally.
Name of the existing canvas widget to hold the plot.
A 3-element list containing minimum, maximum and stepsize for the x-axis, in this order. For an inverted axis, where the maximum appears on the left-hand side, use: maximum, minimum and a negative stepsize.
A 2-element list containing minimum and maximum for the y-axis, in this order. Note that an inverted logarithmic axis is not supported.
Create a new polar plot (configuration type: polarplot).
Name of the existing canvas widget to hold the plot.
A 2-element list containing maximum radius and stepsize for the radial axis, in this order.
Create a new isometric plot, where the vertical and the horizontal coordinates are scaled so that a circle will truly appear as a circle (configuration type: isometric).
Name of the existing canvas widget to hold the plot.
A 2-element list containing minimum, and maximum for the x-axis, in this order.
A 2-element list containing minimum, and maximum for the y-axis, in this order.
Either the stepsize used by both axes or the keyword noaxes to signal the plot that it should use the full area of the widget, to not draw any of the axes.
Create a new histogram (configuration type: histogram).
Name of the existing canvas widget to hold the plot.
A 3-element list containing minimum, maximum and stepsize for the x-axis, in this order.
A 3-element list containing minimum, maximum and stepsize for the y-axis, in this order.
Create a new 3D plot.
Name of the existing canvas widget to hold the plot.
A 3-element list containing minimum, maximum and stepsize for the x-axis, in this order.
A 3-element list containing minimum, maximum and stepsize for the y-axis, in this order.
A 3-element list containing minimum, maximum and stepsize for the z-axis, in this order.
Create a new piechart (configuration type: piechart).
Name of the existing canvas widget to hold the plot.
Create a new radial chart (the data are drawn as a line connecting the spokes of the diagram) (configuration type: radialchart).
Name of the existing canvas widget to hold the plot.
Names for the spokes.
Scale value to determine the position of the data along the spokes.
Style of the chart (optional). One of:
lines - the default: draw the data as independent polylines.
cumulative - draw the data as polylines where the data are accumulated.
filled - draw the data as filled polygons where the data are accumulated
Create a new barchart with vertical bars (configuration type: vertbars). The horizontal axis will display the labels contained in the argument xlabels. The number of series given by noseries determines both the width of the bars, and the way the series will be drawn.
If the keyword stacked was specified the series will be drawn stacked on top of each other. Otherwise each series that is drawn will be drawn shifted to the right.
The number of series determines the width of the bars, so that there is space of that number of bars. If you use a floating-point number, like 2.2, instead of an integer, like 2, a small gap between the sets of bars will be drawn - the width depends on the fractional part.
Name of the existing canvas widget to hold the plot.
List of labels for the x-axis. Its length also determines the number of bars that will be plotted per series.
A 3-element list containing minimum, maximum and stepsize for the y-axis, in this order.
The number of data series that will be plotted. This has to be an integer number greater than zero (if stacked is not used).
Create a new barchart with horizontal bars (configuration type: horizbars). The vertical axis will display the labels contained in the argument ylabels. The number of series given by noseries determines both the width of the bars, and the way the series will be drawn.
If the keyword stacked was specified the series will be drawn stacked from left to right. Otherwise each series that is drawn will be drawn shifted upward.
Name of the existing canvas widget to hold the plot.
List of labels for the y-axis. Its length also determines the number of bars that will be plotted per series.
A 3-element list containing minimum, maximum and stepsize for the x-axis, in this order.
The number of data series that will be plotted. This has to be an integer number greater than zero (if stacked is not used).
Create a new barchart with 3D vertical bars (configuration type: 3dbars). The horizontal axis will display the labels per bar. The number of bars given by nobars determines the position and the width of the bars. The colours can be varied per bar. (This type of chart was inspired by the Wiki page on 3D bars by Richard Suchenwirth.)
Name of the existing canvas widget to hold the plot.
A 3-element list containing minimum, maximum and stepsize for the y-axis, in this order.
The number of bars that will be plotted.
Create a new "ribbon chart" (configuration type: 3dribbon). This is a chart where the data series are represented as ribbons in a three-dimensional axis system. Along the x-axis (which is "into" the screen) the names are plotted, each representing a single series. The first plot command draws the furthest series, the second draws the series in front of that and so on.
Name of the existing canvas widget to hold the plot.
Names of the series, plotted as labels along the x-axis
A 3-element list containing minimum, maximum and stepsize for the y-axis (drawn horizontally!), in this order.
A 3-element list containing minimum, maximum and stepsize for the z-axis (drawn vertically), in this order.
The number of bars that will be plotted.
Create a new boxplot with horizontal boxes (box-and-whiskers). The y-axis is drawn with labels. The boxes are drawn based on the raw data (see the plot subcommand for this type of plot).
Name of the existing canvas widget to hold the plot.
A 3-element list containing minimum, maximum and stepsize for the y-axis, in this order.
List of labels for the y-axis. Its length also determines the number of boxes that can be plotted. The labels are also used in the plot subcommand.
Create a new timechart (configuration type: timechart). The time axis (= x-axis) goes from time_begin to time_end, and the vertical spacing is determined by the number of items to plot.
Name of the existing canvas widget to hold the plot.
The start time given in a form that is recognised by the clock scan command (e.g. "1 january 2004").
The end time given in a form that is recognised by the clock scan command (e.g. "1 january 2004").
The remaining arguments can be:
The expected/maximum number of items. This determines the vertical spacing. (If given, it must be the first argument after "time_end"
The keyword -barheight and the number of pixels per bar. This is an alternative method to determine the vertical spacing.
The keyword -ylabelwidth and the number of pixels to reserve for the labels at the y-axis.
Create a new Gantt chart (configuration type: ganttchart). The time axis (= x-axis) goes from time_begin to time_end, and the vertical spacing is determined by the number of items to plot. Via the specific commands you can then add tasks and connections between the tasks.
Name of the existing canvas widget to hold the plot.
The start time given in a form that is recognised by the clock scan command (e.g. "1 january 2004").
The end time given in a form that is recognised by the clock scan command (e.g. "1 january 2004").
The remaining arguments can be:
The expected/maximum number of items. This determines the vertical spacing. (If given this way, it must be the first argument after "time_end")
The expected/maximum width of the descriptive text (roughly in characters, for the actual space reserved for the text, it is assumed that a character is about ten pixels wide). Defaults to 20. (If given this way, it must be the second argument after "time_end").
The keyword -barheight and the number of pixels per bar. This is an alternative method to determine the vertical spacing.
The keyword -ylabelwidth and the number of pixels to reserve for the labels at the y-axis.
Create a plot command that will use a right axis instead of the left axis (configuration type: inherited from the existing plot). The widget (w) must already contain an ordinary plot, as the horizontal axis and other properties are reused. To plot data using the right axis, use this new command, to plot data using the left axis, use the original plot command.
Name of the existing canvas widget to hold the plot.
A 3-element list containing minimum, maximum and stepsize for the y-axis, in this order.
Each of the creation commands explained in the last section returns the name of a new object command that can be used to manipulate the plot or chart. The subcommands available to a chart command depend on the type of the chart.
General subcommands for all types of charts. $anyplot is the command returned by the creation command:
Specify the title of the whole chart.
The text of the title to be drawn.
Draws the plot into a file, using PostScript.
Contain the path name of the file to write the plot to.
Optionally you can specify the option -format "some picture format" to store the plot in a different file than a PostScript file. This, however, relies on the Img package to do the actual job.
Note: Because the window holding the plot must be fully visible before Img can successfully grab it, it is raised first. On some systems, for instance Linux with KDE, raising a window is not done automatically, but instead you need to click on the window in the task bar. Similar things happen on Windows XP.
There seems to be something wrong under some circumstances, so instead of waiting for the visibility of the window, the procedure simply waits two seconds. It is not ideal, but it seems to work better.
Specify the title of the (horizontal) x-axis, for those plots that have a straight x-axis.
The text of the x-axis label to be drawn.
Specify the title of the (horizontal) y-axis, for those plots that have a straight y-axis.
The text of the y-axis label to be drawn.
Set one or more configuration parameters for the x-axis. The following options are supported:
Set one or more configuration parameters for the y-axis. This method accepts the same options and values as the method xconfig.
Set the background of a part of the plot
Which part of the plot: "axes" for the axes area and "plot" for the inner part. The interpretation depends on the type of plot. Two further possibilities are:
image, in which case a predefined image is loaded into the background of the plot.
gradient, in which case the background is coloured in different shades of the given colour. The "dir" argument specifies the direction in which the colour gets whiter.
Colour for that part or the name of the image if "part" is "image"
The direction of the gradient. One of: top-down, bottom-up, left-right or right-left.
Draw vertical ticklines at each tick location
Colour of the lines. Specifying an empty colour ("") removes them again. Defaults to "black"
Draw horizontal ticklines at each tick location
Colour of the lines. Specifying an empty colour ("") removes them again Defaults to "black"
Set one or more options for the legend. The legend is drawn as a rectangle with text and graphics inside.
Add an entry to the legend. The series determines which graphical symbol is to be used. (As a side effect the legend is actually drawn.)
Name of the data series. This determines the colour of the line and the symbol (if any) that will be drawn.
Text to be drawn next to the line/symbol.
Add balloon text to the plot (except for 3D plots). The arrow will point to the given x- and y-coordinates. For xy-graphs and such, the coordinates are directly related to the axes; for vertical barcharts the x-coordinate is measured as the number of bars minus 1 and similar for horizontal barcharts.
X-coordinate of the point that the arrow of the balloon will point to.
Y-coordinate of the point that the arrow of the balloon will point to.
Text to be drawn in the balloon.
Direction of the arrow, one of: north, north-east, east, south-east, south, south-west, west or north-west.
Configure the balloon text for the plot. The new settings will be used for the next balloon text.
Note: The commands xconfig and yconfig are currently implemented only for XY-plots and only the option -format has any effect.
For xy plots, stripcharts, histograms and time-x-plots:
Add a data point to the plot.
Name of the data series the new point belongs to.
X-coordinate of the new point. (For time-x plots this must be valid date/time that can be read with the clock scan command).
Y-coordinate of the new point.
Note on histograms:
For histograms the x-coordinate that is given is interpreted to be the x-coordinate of the right side of the bar. The first bar starts at the y-axis on the left. To completely fill the range of the x-axis, you should draw a bar at the maximum x-coordinate.
For xy plots:
Draw or update a trend line using the data given sofar.
Name of the data series the trend line belongs to.
X-coordinate of the new data point
Y-coordinate of the new data point
Draw data in the same way as the plot method, but with two lines added that indicate the expected range (+/- 3*standard deviation) of the data.
Name of the data series the data point belongs to.
X-coordinate of the new data point
Y-coordinate of the new data point
Add a vertical error interval to the plot. The interval is drawn from ymin to ymax. If the ycentr argument is given, a symbol is drawn at that position.
Name of the data series the interval belongs to.
X-coordinate of the interval
Minimum y-coordinate of the interval.
Maximum y-coordinate of the interval.
Y-coordinate to draw the symbol at (optional)
Draw a box and whiskers in the plot. If the argument xcrd is a list of several values and the argument ycrd is a single value, a horizontal box is drawn with the quartiles determined from the list of values contained in xcrd.
If, instead, the argument ycrd contains a list of several values and the argument xcrd a single value, then a vertical box is drawn and the quartiles are determined from ycrd. (There must be exactly one list of several values. Otherwise an error is reported.)
The option -boxwidth (default: 10 pixels) determines the width (or height) of the box.
Name of the data series the box-and-whiskers belongs to.
X-coordinate of the box or a list of values.
Y-coordinate of the box or a list of values.
Draw a vector in the plot. The vector can be given as either cartesian coordinates or as length/angle, where the angle is in degrees and is interpreted according to the mathematical convention or the nautical. (See the vectorconfig subcommand)
Name of the series the vector belongs to. Determines the appearance and interpretation.
X-coordinate of the point where the arrow appears
Y-coordinate of the point where the arrow appears
X-component or the length of the vector
Y-component or the angle of the vector
] Set the vector drawing options for a particular series
Name of the series the vector belongs to.
The options can be one of the following:
Draw a dot in the plot. The size and colour is determined by the value and by the options set for the series it belongs to. (See the dotconfig subcommand)
Name of the series the dot belongs to. Determines size and colour
X-coordinate of the point where the arrow appears
Y-coordinate of the point where the arrow appears
Value determining size and colour
] Set the dot drawing options for a particular series
Name of the series the dot belongs to.
The options can be one of the following:
Draw contour lines for the values given on the grid. The grid is defined by the xcrd and ycrd arguments (they give the x- and y-coordinates of the grid cell corners). The values are given at these corners. The classes determine which contour lines are drawn. If a value on one of the corners is missing, the contour lines in that cell will not be drawn.
List of lists, each value is an x-coordinate for a grid cell corner
List of lists, each value is an y-coordinate for a grid cell corner
List of lists, each value is the value at a grid cell corner
List of class values or a list of lists of two elements (each inner list the class value and the colour to be used). If empty or missing, the classes are determined automatically.
Note: The class values must enclose the whole range of values.
Draw filled contours for the values given on the grid. (The use of this method is identical to the "contourlines" method).
Draw the cells as filled quadrangles. The colour is determined from the average of the values on all four corners.
Set the colours to be used with the contour methods. The argument is either a predefined colourmap (grey/gray, jet, hot or cool) or a list of colours. When selecting the colours for actually drawing the contours, the given colours will be interpolated (based on the HLS scheme).
List of colour names or colour values or one of the predefined maps:
grey or gray: gray colours from dark to light
jet: rainbow colours
hot: colours from yellow via red to darkred
cool: colours from cyan via blue to magenta
Draw the grid cells as lines connecting the (valid) grid points.
List of lists, each value is an x-coordinate for a grid cell corner
List of lists, each value is an y-coordinate for a grid cell corner
For polar plots:
Add a data point to the polar plot.
Name of the data series the new point belongs to.
Radial coordinate of the new point.
Angular coordinate of the new point (in degrees).
For 3D plots:
Plot a function defined over two variables x and y. The resolution is determined by the set grid sizes (see the method gridsize for more information).
Name of the procedure that calculates the z-value for the given x and y coordinates. The procedure has to accept two float arguments (x is first argument, y is second) and return a floating-point value.
Plot a function defined over two variables x and y using the contour levels in contours to colour the surface. The resolution is determined by the set grid sizes (see the method gridsize for more information).
Name of the procedure that calculates the z-value for the given x and y coordinates. The procedure has to accept two float arguments (x is first argument, y is second) and return a floating-point value.
List of values in ascending order that represent the contour levels (the boundaries between the colours in the contour map).
Set the grid size in the two directions. Together they determine how many polygons will be drawn for a function plot.
Number of grid cells in x direction. Has to be an integer number greater than zero.
Number of grid cells in y direction. Has to be an integer number greater than zero.
Plot a matrix of data.
The data to be plotted. The data has to be provided as a nested list with 2 levels. The outer list contains rows, drawn in y-direction, and each row is a list whose elements are drawn in x-direction, for the columns. Example:
set data { {1.0 2.0 3.0} {4.0 5.0 6.0} }
Configure the colours to use for polygon borders and inner area.
The colour to use for filling the polygons.
The colour to use for the border of the polygons.
For xy plots, stripcharts and polar plots:
Set the value for one or more options regarding the drawing of data of a specific series.
Name of the data series whose configuration we are changing.
The following options are allowed:
For piecharts:
Fill a piechart.
A list of pairs (labels and values). The values determine the relative size of the circle segments. The labels are drawn beside the circle.
Set the colours to be used.
The first colour.
The second colour, and so on.
For radial charts:
Draw a new line in the radial chart
A list of data (one for each spoke). The values determine the distance from the centre of the line connecting the spokes.
The colour for the line.
An optional argument for the thickness of the line.
Set the colours to be used.
The first colour.
The second colour, and so on.
For vertical barcharts:
Add a data series to a barchart.
Name of the series the values belong to.
A list of values, one for each x-axis label.
The colour of the bars.
For horizontal barcharts:
Add a data series to a barchart.
Name of the series the values belong to.
A list of values, one for each y-axis label.
The colour of the bars.
For 3D barcharts:
Add the next bar to the barchart.
The label to be shown below the column.
The value that determines the height of the column
The colour of the column.
Set one or more configuration parameters. The following options are supported:
For 3D ribbon charts:
Plot the given xy-pairs as a ribbon in the chart
The pairs of x/y values to be drawn (the series is drawn as a whole)
The colour of the ribbon.
Plot the given xy-pairs as a ribbon with a filled area in front. The effect is that of a box with the data as its upper surface.
The pairs of x/y values to be drawn (the series is drawn as a whole)
The colour of the ribbon/area.
For boxplots:
Add a box-and-whisker to the plot.
The label along the y-axis to which the data belong
List of raw values, the extent of the box and the whiskers will be determined from this list.
For timecharts:
Add a time period to the chart.
The text describing the period.
Start time of the period.
Stop time of the period.
The colour of the bar (defaults to black).
Add a milestone (represented as an point-down triangle) to the chart.
The text describing the milestone.
Time at which the milestone must be positioned.
The colour of the triangle (defaults to black).
Add a vertical line (to indicate the start of the month for instance) to the chart.
The text appearing at the top (an abbreviation of the date/time for instance).
Time at which the line must be positioned.
Connect a horizontal scrollbar to the chart. See also the section on scrolling.
The horizontal scrollbar that is to be connected to the chart
Connect a vertical scrollbar to the chart. See also the section on scrolling.
The vertical scrollbar that is to be connected to the chart
For Gantt charts:
Add a task with its period and level of completion to the chart. Returns a list of canvas items that can be used for further manipulations, like connecting two tasks.
The text describing the task.
Start time of the task.
Stop time of the task.
The percentage of the task that is completed.
Add a milestone (represented as an point-down triangle) to the chart.
The text describing the milestone.
Time at which the milestone must be positioned.
The colour of the triangle (defaults to black).
Add a vertical line (to indicate the start of the month for instance) to the chart.
The text appearing at the top (an abbreviation of the date/time for instance).
Time at which the line must be positioned.
Add an arrow that connects the from task with the to task.
The list of items returned by the "task" command that represents the task from which the arrow starts.
The text summarising the tasks
One or more tasks (the lists returned by the "task" command). They are shifted down to make room for the summary.
The list of items returned by the "task" command that represents the task at which the arrow ends.
Add a summary item that spans all the tasks listed. The graphical representation is a thick bar running from the leftmost task to the rightmost.
Use this command before connecting the tasks, as the arrow would not be shifted down!
The text summarising the tasks
One or more tasks (the lists returned by the "task" command). They are shifted down to make room for the summary.
Set the colour of a part of the Gantt chart. These colours hold for all items of that type.
The keyword indicates which part of the Gantt chart to change:
description - the colour of the descriptive text
completed - the colour of the filled bar representing the completed part of a task
left - the colour for the part that is not yet completed
odd - the background colour for the odd entries
even - the background colour for the even entries
summary - the colour for the summary text
summarybar - the colour for the bar for a summary
The new colour for the chosen items.
Set the font of a part of the Gantt chart. These fonts hold for all items of that type.
The keyword indicates which part of the Gantt chart to change:
description - the font used for descriptive text
summary - the font used for summaries
scale - the font used for the time scale
The new font for the chosen items.
Connect a horizontal scrollbar to the chart. See also the section on scrolling.
The horizontal scrollbar that is to be connected to the chart
Connect a vertical scrollbar to the chart. See also the section on scrolling.
The vertical scrollbar that is to be connected to the chart
For isometric plots (to be extended):
Plot the outlines of a rectangle.
Minimum x coordinate of the rectangle to be drawn.
Minimum y coordinate of the rectangle.
Maximum x coordinate of the rectangle to be drawn.
Maximum y coordinate of the rectangle.
The colour of the rectangle.
Plot a rectangle filled with the given colour.
Minimum x coordinate of the rectangle to be drawn.
Minimum y coordinate of the rectangle.
Maximum x coordinate of the rectangle to be drawn.
Maximum y coordinate of the rectangle.
The colour of the rectangle.
Plot the outline of a circle.
X coordinate of the circle's centre.
Y coordinate of the circle's centre.
The colour of the circle.
Plot a circle filled with the given colour.
X coordinate of the circle's centre.
Y coordinate of the circle's centre.
The colour of the circle.
There are a number of public procedures that may be useful in specific situations: Pro memorie.
Besides the commands that deal with the plots and charts directly, there are a number of commands that can be used to convert world coordinates to pixels and vice versa. These include:
Set the viewport for window w. Should be used in cooperation with ::Plotchart::worldCoordinates.
Name of the window (canvas widget) in question.
Left-most pixel coordinate.
Top-most pixel coordinate (remember: the vertical pixel coordinate starts with 0 at the top!).
Right-most pixel coordinate.
Bottom-most pixel coordinate.
Set the extreme world coordinates for window w. The world coordinates need not be in ascending order (i.e. xmin can be larger than xmax, so that a reversal of the x-axis is achieved).
Name of the window (canvas widget) in question.
X-coordinate to be mapped to left side of viewport.
Y-coordinate to be mapped to bottom of viewport.
X-coordinate to be mapped to right side of viewport.
Y-coordinate to be mapped to top side of viewport.
Set the extreme three-dimensional world coordinates for window w. The world coordinates need not be in ascending order (i.e. xmin can be larger than xmax, so that a reversal of the x-axis is achieved).
Name of the window (canvas widget) in question.
X-coordinate to be mapped to front side of the 3D viewport.
Y-coordinate to be mapped to left side of the viewport.
Z-coordinate to be mapped to bottom of viewport.
X-coordinate to be mapped to back side of viewport.
Y-coordinate to be mapped to right side of viewport.
Z-coordinate to be mapped to top side of viewport.
Return a list of pixel coordinates valid for the given window.
Name of the window (canvas widget) in question.
X-coordinate to be mapped.
Y-coordinate to be mapped.
Return a list of pixel coordinates valid for the given window.
Name of the window (canvas widget) in question.
X-coordinate to be mapped.
Y-coordinate to be mapped.
Z-coordinate to be mapped.
Set the extreme polar coordinates for window w. The angle always runs from 0 to 360 degrees and the radius starts at 0. Hence you only need to give the maximum radius. Note: If the viewport is not square, this procedure will not adjust the extremes, so that would result in an elliptical plot. The creation routine for a polar plot always determines a square viewport.
Name of the window (canvas widget) in question.
Maximum radius.
Wrapper for a call to ::Plotchart::coordsToPixel, which assumes the world coordinates and viewport are set appropriately. Converts polar coordinates to pixel coordinates. Note: To be useful it should be accompanied by a matching ::Plotchart::worldCoordinates procedure. This is automatically taken care of in the creation routine for polar plots.
Name of the window (canvas widget) in question.
Radius of the point.
Angle to the positive x-axis.
Return a list of world coordinates valid for the given window.
Name of the window (canvas widget) in question.
X-pixel to be mapped.
Y-pixel to be mapped.
Return the index of the pie segment containing the pixel coordinates (x,y)
Name of the window (canvas widget) in question, holding a piechart.
X-pixel to be mapped.
Y-pixel to be mapped.
Furthermore there is a routine to determine "pretty" numbers for use with an axis:
Determine "pretty" numbers from the given range and return a list containing the minimum, maximum and stepsize that can be used for a (linear) axis.
Rough minimum value for the scaling
Rough maximum value for the scaling.
Optional argument: if 1, then the returned list produces an inverted axis. Defaults to 0 (the axis will be from minimum to maximum)
Often data that need to be plotted contain gaps - in a series of measurement data, they can occur because the equipment failed, a sample was not collected correctly or for many other reasons. The Plotchart handles these gaps by assuming that one or both coordinates of such data points are an empty string:
# # Create the plot with its x- and y-axes # set s [::Plotchart::createXYPlot .c {0.0 100.0 10.0} {0.0 100.0 20.0}] foreach {x y} {0.0 32.0 10.0 {} 25.0 60.0 78.0 11.0 } { $s plot series1 $x $y }
The effect varies according to the type of plot:
For xy-plots, radial plots and strip charts the missing data point causes a gap in the line through the points.
For barchats, missing values are treated as if a value of zero was given.
For time charts and Gantt charts missing values cause errors - there is no use for them there.
Besides output to the canvas on screen, the module is capable, via canvas postscript, of producing PostScript files. One may wonder whether it is possible to extend this set of output formats and the answer is "yes". This section tries to sum up the aspects of using this module for another sort of output.
One way you can create output files in a different format, is by examining the contents of the canvas after everything has been drawn and render that contents in the right form. This is probably the easiest way, as it involves nothing more than the re-creation of all the elements in the plot that are already there.
The drawback of that method is that you need to have a display, which is not always the case if you run a CGI server or something like that.
An alternative is to emulate the canvas command. For this to work, you need to know which canvas subcommands are used and what for. Obviously, the create subcommand is used to create the lines, texts and other items. But also the raise and lower subcommands are used, because with these the module can influence the drawing order - important to simulate a clipping rectangle around the axes. (The routine DrawMask is responsible for this - if the output format supports proper clipping areas, then a redefinition of this routine might just solve this).
Furthermore, the module uses the cget subcommand to find out the sizes of the canvas. A more mundane aspect of this is that the module currently assumes that the text is 14 pixels high and that 80 pixels in width suffice for the axis' labels. No "hook" is provided to customise this.
In summary:
Emulate the create subcommand to create all the items in the correct format
Emulate the cget subcommand for the options -width and -height to allow the correct calculation of the rectangle's position and size
Solve the problem of raising and lowering the items so that they are properly clipped, for instance by redefining the routine DrawMask.
Take care of the currently fixed text size properties
As an example of some special effects you can achieve, here is the code for a plot where the area below the data line varies in colour:
canvas .c -background white -width 400 -height 200 pack .c -fill both set s [::Plotchart::createXYPlot .c {0.0 100.0 10.0} {0.0 100.0 20.0}] $s background gradient green top-down $s dataconfig series1 -filled up -fillcolour white $s plot series1 0.0 20.0 $s plot series1 10.0 20.0 $s plot series1 30.0 50.0 $s plot series1 35.0 45.0 $s plot series1 45.0 25.0 $s plot series1 75.0 55.0 $s plot series1 100.0 55.0 $s plaintext 30.0 60.0 "Peak" south
The trick is to fill the background with a colour that changes from green at the top to white at the bottom. Then the area above the data line is filled with a white polygon. Thus the green shading varies with the height of the line.
In this version there are a lot of things that still need to be implemented:
More robust handling of incorrect calls (right now the procedures may fail when called incorrectly):
The axis drawing routines can not handle inverse axes right now.
If the user provides an invalid date/time string, the routines simply throw an error.
Plotchart has not been designed to create plots and charts that keep track of the data that are put in. This means that if an application needs to allow the user to resize the window holding the plot or chart, it must take care to redraw the complete plot.
The code below is a simple example of how to do that:
package require Plotchart grid [canvas .c -background white] -sticky news grid columnconfigure . 0 -weight 1 grid rowconfigure . 0 -weight 1 bind .c <Configure> {doResize} proc doPlot {} { # # Clean up the contents (see also the note below!) # .c delete all # # (Re)draw the bar chart # set p [::Plotchart::createBarchart .c {x y z} {0 100 10} 3] $p plot R {10 30 40} red $p plot G {30 40 60} green } proc doResize {} { global redo # # To avoid redrawing the plot many times during resizing, # cancel the callback, until the last one is left. # if { [info exists redo] } { after cancel $redo } set redo [after 50 doPlot] }
Please note: The code above will work fine for barcharts and many other types of plots, but as Plotchart keeps some private information for xy plots, more is needed in these cases. This actually requires a command "destroyPlot" to take care of such details. A next version of Plotchart will have that.
As the Plotchart package does not keep track of the data itself, rescaling an existing plot - for instance when zooming in - would have to be done by redefining the plot and redrawing the data. However, the canvas widget offers a way out by scaling and moving items, so that zooming in becomes a bit simpler.
Whether zooming is indeed useful, depends on the type of plot. Currently it is defined for XY-plots only. The method is called "rescale" and simply redraws the axes and scales and moves the data items so that they conform to the new axes. The drawback is that any symbols are scaled by the same amount. The rescale method works best for plots that only have lines, not symbols.
The method works very simply:
$p rescale {newxmin newxmax newxstep} {newymin newymax newystep}
The command plotconfig can be used to set all manner of options. The syntax is:
Set a new value for the property of a component in a particular chart or plot type or query its current value. Each argument is optional.
The type of chart or plot (see the configuration type that is mentioned for each create command). If not given or empty, a list of chart types is returned. If it is given, the properties for that particular type are used.
The component of the plot/chart: leftaxis, rightaxis, background, margin and so on. If not given or empty, a list of components is returned. If it is given, the properties for that particular component will be set for that particular type of chart.
The property of the component of the plot/chart: textcolor, thickness of the axis line, etc. If not given or empty, a list of properties is returned. If it is given, that particular property for that particular component will be set for that particular type of chart.
The new value for the property. If empty, the current value is returned. If the value is "default", the default value will be restored.
Note, that in some cases an empty value is useful. Use "none" in this case - it can be useful for colours and for formats.
Below is a more detailed list of the components and properties:
Axes come in a wide variety:
leftaxis, rightaxis, topaxis, bottomaxis for the plots with a rectangular shape.
xaxis, yaxis and zaxis are used for the 3D plots
axis, this represents the radial and tangential axes of a polar plot
All axes have the following properties:
color - the colour of the line and the tickmarks
thickness - the width of the line of the axis itself, not the tickmarks
ticklength - the length of the tickmarks in pixels. A positive value is outward, a negative value is inward.
font - the font for the labels and the text at the axis
format - the format for rendering the (numerical) labels. For the time axis it is the format for a date and time.
textcolor - the colour for the labels and the text.
The margin is important for the layout. Currently only the rectangular plots allow the margins to be set: left, right, top and bottom. The values are in pixels.
The text component is meant for any text appearing via the plaintext subcommand. The properties are: textcolor, font and anchor (positioning of the text relative to the given coordinates).
The background has two properties: outercolor, the colour outside of the actual plot, and innercolor, the colour inside the plot. (Note: only "outercolor" has now been implemented).
The legend has three properties: background, border and position. See the legend subcommand for the meaning.
See the examples in plotdemos7.tcl for it use.
For two types of plots automatic scrolling management has been implemented: timecharts and Gantt charts. The subcommands hscroll and vscroll associate (existing) scrollbars to the plot, in much the same way as for text and canvas widgets.
Once the association is made, the scrollbars are automatically updated if:
You add an item with a period wider than the current one.
You add a vertical line for a time beyond the current bounds.
You add an extra item beyond the number that was used to create the chart.
For instance:
package require Plotchart canvas .c -width 400 -height 200 scrollbar .y -orient vertical scrollbar .x -orient horizontal grid .c .y -sticky news grid .x -sticky news source plotchart.tcl set s [::Plotchart::createTimechart .c "1 january 2004" "31 december 2004" 4] $s period "Spring" "1 march 2004" "1 june 2004" green $s period "Summer" "1 june 2004" "1 september 2004" yellow $s vertline "1 jan" "1 january 2004" $s vertline "1 apr" "1 april 2004" $s vertline "1 jul" "1 july 2004" $s vertline "1 oct" "1 october 2004" $s vertline "1 jan" "1 january 2005" $s vertline "1 apr" "1 april 2005" $s vertline "1 jul" "1 july 2005" $s milestone "Longest day" "21 july 2004" $s milestone "Longest day 2" "21 july 2004" $s milestone "Longest day 3" "21 july 2004" $s milestone "Longest day 4" "21 july 2004" $s milestone "Longest day 5" "21 july 2004" $s milestone "Longest day 6" "21 july 2004" $s title "Seasons (northern hemisphere)" $s vscroll .y $s hscroll .x
The original extent of the chart is from 1 january 2004 to 31 december 2004. But because of the addition of vertical lines in 2005 and more items than was specified at the creation of the chart, both the horizontal and the vertical scrollbar will be enabled.
The command plotpack allows you to copy the contents of a plot into another canvas widget. This canvas widget does not act as a composite plot, but it can be saved as a PostScript file for instance: Note: the command simply takes a snapshot of the plots/charts as they are at that moment.
Copy the contents of the plots/charts into another widget, in a manner similar to the pack geometry manager.
The name of the canvas widget to copy the plots/charts into
The direction of the arrangement - top, left, bottom or right
List of plots/charts to be copied.
For example:
set p1 [createXYPlot ...] set p2 [createBarchart ...] ... fill the plots ... toplevel .t pack [canvas .t.c2 -width ...] # # Copy the two plots above each other in the new canvas # plotpack .t.c2 top $p1 $p2
P.M.
I have the following wishlist:
Isometric plots - allow new items to be implemented easily.
A general 3D viewer - emphasis on geometry, not a ray-tracer.
Several improvements for boxplots:
Height of the box scales with the logarithm of the number of points
Marker line to indicate a "current" value
Box drawn from quantiles
3D bars, 3D surfaces, bar charts, charts, coordinate transformations, coordinates, graphical presentation, isometric plots, pie charts, plotting, polar plots, strip charts, time charts, xy-plots
Copyright © 2007 Arjen Markus <arjenmarkus@users.sourceforge.net>