gist:

      Copyright (c) 1996, 1997, The Regents of the University of
      California.  All rights reserved.  See Legal.htm for full
      text and disclaimer.

     This is the help file for the Python Gist Graphics Module. Gist
     is a portable graphics package for scientific applications. It
     can produce interactive graphics in the X-window and Macintosh
     environments (the Python module currently supports only X11), and
     it can produce file output conforming to ANSI standard CGM or
     standard Postscript.

     Gist was developed by David H. Munro <munro@icf.llnl.gov> at
     Lawrence Livermore National Laboratory, as part of his Yorick
     scientific interpreter. Gist is distributed as part of Yorick, and
     is available at the following sites:

       wuarchive.wustl.edu: /languages/yorick/yorick-1.2.tar.gz
       sunsite.unc.edu: /pub/languages/yorick/yorick-1.2.tar.gz
       sunsite.unc.edu: /pub/Linux/apps/math/matrix/yorick-1.2.tar.gz
       netlib.att.com: /netlib/env/yorick-1.2.tar.gz
       netlib2.cs.utk.edu: /env/yorick-1.2.tar.gz

     Much of the code in the Python Gist C extension module was
     adapted from similar code in Yorick, and this help file also
     originated there. I am greatly indebted to Dave for his prior work.
     Questions about this module should be directed to me, Lee Busby,
     <busby1@llnl.gov>.

/*--------------------------------------------------------------------------*/
/* Control functions */

window:
window( [n] [, display = "host:server.screen", dpi=100/75, wait=0/1,
                       private=0/1, hcp="hcp_filename", dump=0/1,
                       legends=1/0, style="style_sheet_filename" ] )
     select window N as the current graphics output window.  N may
     range from 0 to 7, inclusive.  Each graphics window corresponds to
     an X window, and optionally has its own associated hardcopy file.
     If N is omitted, it defaults to the current coordinate system.

     The X window will appear on your default display at 75 dpi, unless
     you specify the display and/or dpi keywords.  A dpi=100 X window
     is larger than a dpi=75 X window; both represent the same thing
     on paper.  Use display="" to create a graphics window which has
     no associated X window (you should do this if you want to make
     plots in a non-interactive batch mode).

     By default, an X window will attempt to use shared colors, which
     permits several Pygist graphics windows (including windows from
     multiple instances of Python) to use a common palette.  You can
     force an X window to post its own colormap (set its colormap
     attribute) with the private=1 keyword.  You will most likely have
     to fiddle with your window manager to understand how it handles
     colormap focus if you do this.  Use private=0 to return to shared
     colors.

     By default, Python will not wait for the X window to become visible;
     code which creates a new window, then plots a series of frames to
     that window should use wait=1 to assure that all frames are actually
     plotted.

     By default, a graphics window does NOT have a hardcopy file
     of its own -- any request for hardcopy are directed to the
     default hardcopy file, so hardcopy output from any window goes
     to a single file.  By specifying the hcp keyword, however, a
     hardcopy file unique to this window will be created.  If the
     "hcp_filename" ends in ".ps", the hardcopy file will be a PostScript
     file; otherwise, hardcopy files are in binary CGM format.  Use
     hcp="" to revert to the default hardcopy file (closing the window
     specific file, if any).  The legends keyword, if present, controls
     whether the curve legends are (legends=1, the default) or are not
     (legends=0) dumped to the hardcopy file.  The dump keyword, if
     present, controls whether all colors are converted to a gray scale
     (dump=0, the default), or the current palette is dumped at the
     beginning of each page of hardcopy output.  (The legends keyword
     applies to all pictures dumped to hardcopy from this graphics
     window.  The dump keyword applies only to the specific hardcopy
     file defined using the hcp keyword -- use the dump keyword in the
     hcp_file command to get the same effect in the default hardcopy
     file.)

     If both display="" and hcp="", the graphics window will be
     entirely eliminated.

     The style keyword, if present, specifies the name of a Gist style
     sheet file; the default is "work.gs".  The style sheet determines
     the number and location of coordinate systems, tick and label styles,
     and the like.  Other choices include "axes.gs", "boxed.gs",
     "work2.gs", and "boxed2.gs".

     Window(...) returns the current window number.
   SEE ALSO: plsys, hcp_file, fma, hcp, redraw, palette, animate, plg,
             winkill, gridxy

winkill:
winkill( [n] )
     Delete the current graphics window, or graphics window N (0-7).

current_window:
n = current_window()
     Return the number of the current graphics window, or -1 if none.

hcp_file:
hcp_file( [filename] [, dump=0/1] )
     Sets the default hardcopy file to FILENAME.  If FILENAME ends with
     ".ps", the file will be a PostScript file, otherwise it will be a
     binary CGM file.  By default, the hardcopy file name will be
     "Aa00.cgm", or "Ab00.cgm" if that exists, or "Ac00.cgm" if both
     exist, and so on.  The default hardcopy file gets hardcopy from all
     graphics windows which do not have their own specific hardcopy file
     (see the window command).  If the dump keyword is present and non-zero,
     the current palette will be dumped at the beginning of each frame
     of the default hardcopy file.  With dump=0, the default behavior of
     converting all colors to a gray scale is restored.
   SEE ALSO: window, fma, hcp, plg

hcp_finish:
filename = hcp_finish( [n] )
     Close the current hardcopy file and return the filename.
     If N is specified, close the hcp file associated with window N
     and return its name; use hcp_finish(-1) to close the default
     hardcopy file.
   SEE ALSO: window, fma, hcp, hcp_out, plg

hcp_out: **** NOT YET IMPLEMENTED ****
hcp_out( [n] [,keep = 0/1] )
     Finishes the current hardcopy file and sends it to the printer.
     If N is specified, prints the hcp file associated with window N;
     use hcp_out(-1) to print the default hardcopy file.
     Unless the KEEP keyword is supplied and non-zero, the file will
     be deleted after it is processed by gist and sent to lpr.
   SEE ALSO: window, fma, hcp, hcp_finish, plg

eps:
eps(name)
     Write the picture in the current graphics window to the Encapsulated
     PostScript file NAME+".epsi" (i.e.- the suffix .epsi is added to NAME).
     The eps function requires the ps2epsi utility which comes with the
     project GNU Ghostscript program.  Any hardcopy file associated with
     the current window is first closed, but the default hardcopy file is
     unaffected.  As a side effect, legends are turned off and color table
     dumping is turned on for the current window.
     The environment variable PS2EPSI_FORMAT contains the format for the
     command to start the ps2epsi program.
   SEE ALSO: window, fma, hcp, hcp_finish, plg

fma:
fma()
     Frame advance the current graphics window.  The current picture
     remains displayed in the associated X window until the next element
     is actually plotted.
   SEE ALSO: window, hcp, animate, plg

hcp:
hcpon:
hcpoff:
hcp(), or hcpon(), or hcpoff()
     The hcp command sends the picture displayed in the current graphics
     window to the hardcopy file.  (The name of the default hardcopy file
     can be specified using hcp_file; each individual graphics window may
     have its own hardcopy file as specified by the window command.)
     The hcpon command causes every fma (frame advance) command to do
     and implicit hcp, so that every frame is sent to the hardcopy file.
     The hcpoff command reverts to the default "demand only" mode.
   SEE ALSO: window, fma, plg

redraw:
redraw()
     Redraw the X window associated with the current graphics window.
   SEE ALSO: window, fma, hcp, plg

palette:
palette( filename )
or palette( source_window_number )
or palette( red, green, blue, ntsc=1/0 )
or palette( red, green, blue, gray )
or palette( red, green, blue, query=1 )
or palette( red, green, blue, gray, query=1 )
     Set (or retrieve with query=1) the palette for the current
     graphics window.  The FILENAME is the name of a Gist palette file;
     the standard palettes are "earth.gp", "stern.gp", "rainbow.gp",
     "heat.gp", "gray.gp", and "yarg.gp".  Use the maxcolors keyword
     in the pldefault command to put an upper limit on the number of
     colors which will be read from the palette in FILENAME.

     In the second form, the palette for the current window is copied
     from the SOURCE_WINDOW_NUMBER.  If the X colormap for the window is
     private, there will still be two separate X colormaps for the two
     windows, but they will have the same color values.

     In the third form, RED, GREEN, and BLUE are 1-D arrays of the same
     length specifying the palette you wish to install; the values
     should vary between 0 and 255, and your palette should have no
     more than 240 colors.  If ntsc=0, monochrome devices (such as most
     laser printers) will use the average brightness to translate your
     colors into gray; otherwise, the NTSC (television) averaging will
     be used (.30*RED+.59*GREEN+.11*BLUE).  Alternatively, you can specify
     GRAY explicitly.

     Ordinarily, the palette is not dumped to a hardcopy file
     (color hardcopy is still rare and expensive), but you can
     force the palette to dump using the window() or hcp_file() commands.

     See the dump= keyword for the hcp_file() and window() commands if you
     are having trouble getting color in your hardcopy files.

   SEE ALSO: window, fma, hcp, pldefault, plg

animate:
animate()
or animate( 0/1 )
     Without any arguments, toggle animation mode; with argument 0,
     turn off animation mode; with argument 1 turn on animation mode.
     In animation mode, the X window associated with a graphics window
     is actually an offscreen pixmap which is bit-blitted onscreen
     when an fma() command is issued.  This is confusing unless you are
     actually trying to make a movie, but results in smoother animation
     if you are.  Generally, you should turn animation on, run your movie,
     then turn it off.
   SEE ALSO: window, fma, plg

plsys:
plsys( n )
     Set the current coordinate system to number N in the current
     graphics window.  If N equals 0, subsequent elements will be
     plotted in absolute NDC coordinates outside of any coordinate
     system.  The default style sheet "work.gs" defines only a single
     coordinate system, so the only other choice is N equal 1.  You
     can make up your own style sheet (using a text editor) which
     defines multiple coordinate systems.  You need to do this if
     you want to display four plots side by side on a single page,
     for example.  The standard style sheets "work2.gs" and "boxed2.gs"
     define two overlayed coordinate systems with the first labeled
     to the right of the plot and the second labeled to the left of
     the plot.  When using overlayed coordinate systems, it is your
     responsibility to ensure that the x-axis limits in the two
     systems are identical.
   SEE ALSO: window, limits, plg

/*--------------------------------------------------------------------------*/
/* Plotting functions (output primitives) */

plg:
plg( y [, x] )
     Plot a graph of Y versus X.  Y and X must be 1-D arrays of equal
     length; if X is omitted, it defaults to [1, 2, ..., 1+range(len(Y))].
     The following keywords are legal (each has a separate help entry):
   KEYWORDS: legend, hide
	     type, width, color, closed, smooth
             marks, marker, mspace, mphase
	     rays, arrowl, arroww, rspace, rphase
   SEE ALSO: plg, plm, plc, plv, plf, pli, plt, pldj, plfp
             limits, logxy, ylimits, fma, hcp

plm:
plm( y, x, boundary=0/1, inhibit=0/1/2 )
or plm( y, x, ireg, boundary=0/1, inhibit=0/1/2 )
or plm( boundary=0/1, inhibit=0/1/2 )
     Plot a mesh of Y versus X.  Y and X must be 2-D arrays with equal
     dimensions.  If present, IREG must be a 2-D region number array
     for the mesh, with the same dimensions as X and Y.  The values of
     IREG should be positive region numbers, and zero for zones which do
     not exist.  The first row and column of IREG never correspond to any
     zone, and should always be zero.  The default IREG is 1 everywhere
     else.  If present, the BOUNDARY keyword determines whether the
     entire mesh is to be plotted (boundary=0, the default), or just the
     boundary of the selected region (boundary=1).  If present, the
     INHIBIT keyword causes the (X(,j),Y(,j)) lines to not be plotted
     (inhibit=1), or the (X(i,),Y(i,)) lines to not be plotted (inhibit=2).
     By default (inhibit=0), mesh lines in both logical directions are
     plotted.
     The Y, X, and IREG arguments may all be omitted to default to the
     mesh set by the most recent plmesh call.
     The following keywords are legal (each has a separate help entry):
   KEYWORDS: legend, hide
	     type, width, color
	     region
   SEE ALSO: plg, plm, plc, plv, plf, pli, plt, pldj, plfp, plmesh
             limits, logxy, ylimits, fma, hcp

plmesh:
plmesh( y, x, ireg, triangle=tri_array )
or plmesh()
     Set the default mesh for subsequent plm, plc, plv, and plf calls.
     In the second form, deletes the default mesh (until you do this,
     or switch to a new default mesh, the default mesh arrays persist and
     take up space in memory).  The Y, X, and IREG arrays should all be
     the same shape; Y and X will be converted to double, and IREG will
     be converted to int.  If IREG is omitted, it defaults to IREG(1,)=
     IREG(,1)= 0, IREG(2:,2:)=1; that is, region number 1 is the whole
     mesh.  The triangulation array TRI_ARRAY is used by plc; the
     correspondence between TRI_ARRAY indices and zone indices is the
     same as for IREG, and its default value is all zero.
     The IREG or TRI_ARRAY arguments may be supplied without Y and X
     to change the region numbering or triangulation for a given set of
     mesh coordinates.  However, a default Y and X must already have been
     defined if you do this.
     If Y is supplied, X must be supplied, and vice-versa.
   SEE ALSO: plm, plc, plv, plf, plfp

plc:
plc( z, y, x, levs=z_values )
or plc( z, y, x, ireg, levs=z_values )
or plc( z, levs=z_values )
     Plot contours of Z on the mesh Y versus X.  Y, X, and IREG are
     as for plm.  The Z array must have the same shape as Y and X.
     The function being contoured takes the value Z at each point
     (X,Y) -- that is, the Z array is presumed to be point-centered.
     The Y, X, and IREG arguments may all be omitted to default to the
     mesh set by the most recent plmesh call.
     The LEVS keyword is a list of the values of Z at which you want
     contour curves.  The default is eight contours spanning the
     range of Z.
     The following keywords are legal (each has a separate help entry):
   KEYWORDS: legend, hide
	     type, width, color, smooth
             marks, marker, mspace, mphase
	     smooth, triangle, region
   SEE ALSO: plg, plm, plc, plv, plf, pli, plt, pldj, plfp, plmesh
             limits, logxy, ylimits, fma, hcp

plv:
plv( vy, vx, y, x, scale=dt )
or plv( vy, vx, y, x, ireg, scale=dt )
or plv( vy, vx, scale=dt )
     Plot a vector field (VX,VY) on the mesh (X,Y).  Y, X, and IREG are
     as for plm.  The VY and VX arrays must have the same shape as Y and X.
     The Y, X, and IREG arguments may all be omitted to default to the
     mesh set by the most recent plmesh call.
     The SCALE keyword is the conversion factor from the units of
     (VX,VY) to the units of (X,Y) -- a time interval if (VX,VY) is a velocity
     and (X,Y) is a position -- which determines the length of the
     vector "darts" plotted at the (X,Y) points.  If omitted, SCALE is
     chosen so that the longest ray arrows have a length comparable
     to a "typical" zone size.
     You can use the scalem keyword in pledit to make adjustments to the
     SCALE factor computed by default.
     The following keywords are legal (each has a separate help entry):
   KEYWORDS: legend, hide
	     type, width, color, smooth
             marks, marker, mspace, mphase
	     triangle, region
   SEE ALSO: plg, plm, plc, plv, plf, pli, plt, pldj, plfp, plmesh, pledit,
             limits, logxy, ylimits, fma, hcp

plf:
plf( z, y, x )
or plf( z, y, x, ireg )
or plf( z )
     Plot a filled mesh Y versus X.  Y, X, and IREG are as for plm.
     The Z array must have the same shape as Y and X, or one smaller
     in both dimensions.  If Z is of type char, it is used "as is",
     otherwise it is linearly scaled to fill the current palette, as
     with the bytscl function.
     (See the bytscl function for explanation of top, cmin, cmax.)
     The mesh is drawn with each zone in the color derived from the Z
     function and the current palette; thus Z is interpreted as a
     zone-centered array.
     The Y, X, and IREG arguments may all be omitted to default to the
     mesh set by the most recent plmesh call.
     A solid edge can optionally be drawn around each zone by setting
     the EDGES keyword non-zero.  ECOLOR and EWIDTH determine the edge
     color and width.  The mesh is drawn zone by zone in order from
     IREG(2+imax) to IREG(jmax*imax) (the latter is IREG(imax,jmax)),
     so you can achieve 3D effects by arranging for this order to
     coincide with back-to-front order.  If Z is nil, the mesh zones
     are filled with the background color, which you can use to
     produce 3D wire frames.
     The following keywords are legal (each has a separate help entry):
   KEYWORDS: legend, hide
	     region, top, cmin, cmax, edges, ecolor, ewidth
   SEE ALSO: plg, plm, plc, plv, plf, pli, plt, pldj, plfp, plmesh,
             limits, logxy, ylimits, fma, hcp, palette, bytscl, histeq_scale
 */

plfc:
plfc (z, y, x, ireg, contours = 20, colors = None, region = 0,
 triangle = None, scale = "lin")
      fills contours of Z on the mesh Y versus X.  Y, X, and IREG are
      as for plm.  The Z array must have the same shape as Y and X.
      The function being contoured takes the value Z at each point
      (X, Y) -- that is, the Z array is presumed to be point-centered.
 
      The CONTOURS keyword can be an integer specifying the number of
      contours desired, or a list of the values of Z at which you want
      contour curves.  These curves divide the mesh into len(CONTOURS+1)
      regions, each of which is filled with a solid color.  If CONTOURS is
      None or not given, 20 "nice" equally spaced level values spanning the
      range of Z are selected.
 
      If you specify CONTOURS, you may also specify COLORS, an array of
      color numbers (Python typecode 'b', integers between 0 and the
      length of the current palette - 1, normally 199) of length
      len(CONTOURS)+1. If you do not specify them, equally
      spaced colors are chosen.

      If CONTOURS is an integer, SCALE expresses how contour levels
      are determined.  SCALE may be "lin", "log", or "normal"
      specifying linearly, logarithmically, or normally spaced
      contours. Note that unlike Yorick's plfc, this routine does
      not use spann to compute its contours. Neither, apparently,
      does plc, which uses a third algorithm which matches neither
      the one we use nor the one spann uses. So if you plot filled
      contours and then plot contour lines, the contours will in
      general not coincide exactly.

      Note that you may use spann to calculate your contour levels
      if you wish.
 
      The following keywords are legal (each has a separate help entry):
    KEYWORDS: triangle, region
    SEE ALSO: plg, plm, plc, plv, plf, pli, plt, pldj, plfp, plmesh
              color_bar, spann, contour, limits, logxy, range, fma, hcp

plfp:
plfp( z, y, x, n )
     Plot a list of filled polygons Y versus X, with colors Z.
     The N array is a 1D list of lengths (number of corners) of the
     polygons; the 1D colors array Z has the same length as N.  The
     X and Y arrays have length equal to the sum of all dimensions
     of N.
     The Z array must have the same shape as Y and X.  If Z is of
     type char, it is used "as is", otherwise it is linearly scaled
     to fill the current palette, as with the bytscl function.
     (See the bytscl function for explanation of top, cmin, cmax.)
     The following keywords are legal (each has a separate help entry):
   KEYWORDS: legend, hide, top, cmin, cmax
   SEE ALSO: plg, plm, plc, plv, plf, pli, plt, pldj
             limits, logxy, ylimits, fma, hcp
 */

pli:
pli( z )
or pli( z, x1, y1 )
or pli( z, x0, y0, x1, y1 )
     Plot the image Z as a cell array -- an array of equal rectangular
     cells colored according to the 2-D array Z.  The first dimension
     of Z is plotted along x, the second dimension is along y.
     If Z is of type char, it is used "as is", otherwise it is linearly
     scaled to fill the current palette, as with the bytscl function.
     (See the bytscl function for explanation of top, cmin, cmax.)
     If X1 and Y1 are given, they represent the coordinates of the
     upper right corner of the image.  If X0, and Y0 are given, they
     represent the coordinates of the lower left corner, which is at
     (0,0) by default.  If only the Z array is given, each cell will be
     a 1x1 unit square, with the lower left corner of the image at (0,0).
     The following keywords are legal (each has a separate help entry):
   KEYWORDS: legend, hide, top, cmin, cmax
   SEE ALSO: plg, plm, plc, plv, plf, pli, plt, pldj, plfp,
             limits, logxy, ylimits, fma, hcp, palette, bytscl, histeq_scale

pldj:
pldj( x0, y0, x1, y1 )
     Plot disjoint lines from (X0,Y0) to (X1,Y1).  X0, Y0, X1, and Y1
     may have any dimensionality, but all must have the same number of
     elements.
     The following keywords are legal (each has a separate help entry):
   KEYWORDS: legend, hide
	     type, width, color
   SEE ALSO: plg, plm, plc, plv, plf, pli, plt, pldj, plfp
             limits, logxy, ylimits, fma, hcp

plt:
plt( text, x, y, tosys=0/1 )
     Plot TEXT (a string) at the point (X,Y).  The exact relationship
     between the point (X,Y) and the TEXT is determined by the
     justify keyword.  TEXT may contain newline ("\n") characters
     to output multiple lines of text with a single call.  The
     coordinates (X,Y) are NDC coordinates (outside of any coordinate
     system) unless the tosys keyword is present and non-zero, in
     which case the TEXT will be placed in the current coordinate
     system.  However, the character height is NEVER affected by the
     scale of the coordinate system to which the text belongs.
     Note that the pledit command takes dx and/or dy keywords to
     adjust the position of existing text elements.
     The following keywords are legal (each has a separate help entry):
   KEYWORDS: legend, hide
	     color, font, height, opaque, path, justify
   SEE ALSO: plg, plm, plc, plv, plf, pli, plt, pldj, plfp, pledit
             limits, ylimits, fma, hcp, pltitle

pltitle:
pltitle( title )
     Plot TITLE centered above the coordinate system for any of the
     standard Gist styles.  You will need to customize this for other
     plot styles.

/*--------------------------------------------------------------------------*/
/* Plot limits and log/linear scaling */

limits:
old_limits = limits()
or old_limits = limits( xmin [, xmax, ymin, ymax,]
     [ square=0/1, nice=0/1, restrict=0/1 ] )
or limits( old_limits )

     In the first form, restore all four plot limits to extreme values,
     and save the previous limits in the tuple old_limits.

     In the second form, set the plot limits in the current coordinate
     system to XMIN, XMAX, YMIN, YMAX, which may each be a number to fix
     the corresponding limit to a specified value, or the string "e"
     to make the corresponding limit take on the extreme value of the
     currently displayed data. Arguments may be omitted from the right
     end only. (But see ``ylimits'' to set limits on the y-axis.)

     If present, the square keyword determines whether limits marked as
     extreme values will be adjusted to force the x and y scales to be
     equal (square=1) or not (square=0, the default). If present, the
     nice keyword determines whether limits will be adjusted to nice
     values (nice=1) or not (nice=0, the default). There is a subtlety
     in the meaning of "extreme value" when one or both of the limits
     on the OPPOSITE axis have fixed values -- does the "extreme value"
     of the data include points which will not be plotted because their
     other coordinate lies outside the fixed limit on the opposite axis
     (restrict=0, the default), or not (restrict=1)?

     Limits() always returns a tuple of 4 doubles and an integer;
     OLD_LIMITS[0:3] are the previous xmin, xmax, ymin, and ymax, and
     OLD_LIMITS[4] is a set of flags indicating extreme values and the
     square, nice, restrict, and log flags. This tuple can be saved and
     passed back to limits() in a future call to restore the limits to a
     previous state.

     In an X window, the limits may also be adjusted interactively with
     the mouse. Drag left to zoom in and pan (click left to zoom in on a
     point without moving it), drag middle to pan, and click (and drag)
     right to zoom out (and pan). If you click just above or below the
     plot, these operations will be restricted to the x-axis; if you
     click just to the left or right, the operations are restricted to
     the y-axis. A shift-left click, drag, and release will expand the
     box you dragged over to fill the plot (other popular software zooms
     with this paradigm). If the rubber band box is not visible with
     shift-left zooming, try shift-middle or shift-right for alternate
     XOR masks. Such mouse-set limits are equivalent to a limits command
     specifying all four limits EXCEPT that the unzoom command can
     revert to the limits before a series of mouse zooms and pans.

     The limits you set using the limits or ylimits functions carry over
     to the next plot -- that is, an fma operation does NOT reset the
     limits to extreme values.

   SEE ALSO: plsys, ylimits, logxy, zoom_factor, unzoom, plg

ylimits:
ylimits(ymin, ymax)
     Set the y-axis plot limits in the current coordinate system to
     YMIN, YMAX, which may each be a number to fix the corresponding
     limit to a specified value, or the string "e" to make the
     corresponding limit take on the extreme value of the currently
     displayed data. Arguments may be omitted only from the right. Use
     limits( xmin, xmax ) to accomplish the same function for the x-axis
     plot limits.  Note that the corresponding Yorick function for
     ylimits is ``range'' - since this word is a Python built-in function,
     I've changed the name to avoid the collision.
   SEE ALSO: plsys, limits, logxy, plg

logxy:
logxy( xflag, yflag )
     Sets the linear/log axis scaling flags for the current coordinate
     system. XFLAG and YFLAG may be 0 to select linear scaling, or 1 to
     select log scaling. YFLAG may be omitted (but not XFLAG).
   SEE ALSO: plsys, limits, ylimits, plg, gridxy

gridxy:
gridxy( flag )
or gridxy( xflag, yflag )
     Turns on or off grid lines according to FLAG.  In the first form, both
     the x and y axes are affected.  In the second form, XFLAG and YFLAG
     may differ to have different grid options for the two axes.  In either
     case, a FLAG value of 0 means no grid lines (the default), a value of
     1 means grid lines at all major ticks (the level of ticks which get
     grid lines can be set in the style sheet), and a FLAG value of 2 means
     that the coordinate origin only will get a grid line.  In styles with
     multiple coordinate systems, only the current coordinate system is
     affected.  The keywords can be used to affect the style of the grid
     lines.

     You can also turn the ticks off entirely.  (You might want to do this
     to plot your own custom set of tick marks when the automatic tick
     generating machinery will never give the ticks you want.  For example
     a latitude axis in degrees might reasonably be labeled "0, 30, 60,
     90", but the automatic machinery considers 3 an "ugly" number - only
     1, 2, and 5 are "pretty" - and cannot make the required scale.  In
     this case, you can turn off the automatic ticks and labels, and use
     plsys, pldj, and plt to generate your own.)
     To fiddle with the tick flags in this general manner, set the
     0x200 bit of FLAG (or XFLAG or YFLAG), and "or-in" the 0x1ff bits
     however you wish.  The meaning of the various flags is described
     in the ``work.gs'' Gist style sheet.  Additionally, you can use the
     0x400 bit to turn on or off the frame drawn around the viewport.
     Here are some examples:
        gridxy(0x233)          work.gs default setting
        gridxy(0, 0x200)       like work.gs, but no y-axis ticks or labels
        gridxy(0, 0x231)       like work.gs, but no y-axis ticks on right
        gridxy(0x62b)          boxed.gs default setting

   KEYWORDS: color, type, width
   SEE ALSO: window, plsys, limits, ylimits, logxy

zoom_factor:
zoom_factor( factor )
     Set the zoom factor for mouse-click zoom in and zoom out operations.
     The default FACTOR is 1.5; FACTOR should always be greater than 1.0.
   SEE ALSO: limits, ylimits, unzoom, plg

unzoom:
unzoom()
     Restore limits to their values before zoom and pan operations
     performed interactively using the mouse.
     Use    old_limits = limits()
            ...
	    limits( old_limits )
     to save and restore plot limits generally.
   SEE ALSO: limits, ylimits, zoom_factor, plg

/*--------------------------------------------------------------------------*/
/* Keywords for plotting functions */

legend:
legend = "text destined for the legend"
     Set the legend for a plot. There are no default legends in Python
     Gist. Legends are never plotted to the X window; use the plq
     command to see them interactively. Legends will appear in hardcopy
     output unless they have been explicitly turned off.
   PLOTTING COMMANDS: plg, plm, plc, plv, plf, pli, plt, pldj
   SEE ALSO: hide

hide:
hide = 0/1
     Set the visibility of a plotted element.  The default is hide=0,
     which means that the element will be visible.  Use hide=1 to remove
     the element from the plot (but not from the display list).
   PLOTTING COMMANDS: plg, plm, plc, plv, plf, pli, plt, pldj
   SEE ALSO: legend

type:
type = <line type value>
     Select line type. Valid values are the strings "solid", "dash",
     "dot", "dashdot", "dashdotdot", and "none". The "none" value causes
     the line to be plotted as a polymarker. The type value may also be
     a number; 0 is "none", 1 is "solid", 2 is "dash", 3 is "dot", 4 is
     "dashdot", and 5 is "dashdotdot".
   PLOTTING COMMANDS: plg, plm, plc, pldj
   SEE ALSO: width, color, marks, marker, rays, closed, smooth

width:
width = <floating point value>
     Select line width.  Valid values are positive floating point numbers
     giving the line thickness relative to the default line width of one
     half point, which is width = 1.0.
   PLOTTING COMMANDS: plg, plm, plc, pldj, plv (only if hollow=1)
   SEE ALSO: type, color, marks, marker, rays, closed, smooth

color:
color = <color value>
     Select line or text color.  Valid values are the strings "bg", "fg",
     "black", "white", "red", "green", "blue", "cyan", "magenta", "yellow",
     or a 0-origin index into the current palette.  The default is "fg".
     Negative numbers may be used instead of the strings: -1 is bg
     (background), -2 is fg (foreground), -3 is black, -4 is white,
     -5 is red, -6 is green, -7 is blue, -8 is cyan, -9 is magenta, and
     -10 is yellow.
   PLOTTING COMMANDS: plg, plm, plc, pldj, plt
   SEE ALSO: type, width, marks, marker, mcolor, rays, closed, smooth

marks:
marks = 0/1
     Select unadorned lines (marks=0), or lines with occasional markers
     (marks=1).  Ignored if type is "none" (indicating polymarkers instead
     of occasional markers).  The spacing and phase of the occasional
     markers can be altered using the mspace and mphase keywords; the
     character used to make the mark can be altered using the marker
     keyword.
   PLOTTING COMMANDS: plg, plc
   SEE ALSO: type, width, color, marker, rays, mspace, mphase, msize, mcolor

marker:
marker = <character or integer value>
     Select the character used for occasional markers along a polyline,
     or for the polymarker if type = "none".  The special values
     '\1', '\2', '\3', '\4', and '\5' stand for point, plus, asterisk,
     circle, and cross, which are prettier than text characters on output
     to some devices.  The default marker is the next available capital
     letter, 'A', 'B', ..., 'Z'.
   PLOTTING COMMANDS: plg, plc
   SEE ALSO: type, width, color, marks, rays, mspace, mphase, msize, mcolor

mspace:
mphase:
msize:
mcolor:
mspace = <float value>
or mphase = <float value>
or msize =  <float value>
or mcolor =  <color value>
     Select the spacing, phase, and size of occasional markers placed
     along polylines.  The msize also selects polymarker size if type
     is "none".  The spacing and phase are in NDC units (0.0013 NDC
     equals 1.0 point); the default mspace is 0.16, and the default
     mphase is 0.14, but mphase is automatically incremented for
     successive curves on a single plot.  The msize is in relative
     units, with the default msize of 1.0 representing 10 points.
     The mcolor keyword is the same as the color keyword, but controls
     the marker color instead of the line color.  Setting the color
     automatically sets the mcolor to the same value, so you only
     need to use mcolor if you want the markers for a curve to be a
     different color than the curve itself.
   PLOTTING COMMANDS: plg, plc
   SEE ALSO: type, width, color, marks, marker, rays

rays:
rays = 0/1
     Select unadorned lines (rays=0), or lines with occasional ray
     arrows (rays=1).  Ignored if type is "none".  The spacing and phase
     of the occasional arrows can be altered using the rspace and rphase
     keywords; the shape of the arrowhead can be modified using the
     arroww and arrowl keywords.
   PLOTTING COMMANDS: plg, plc
   SEE ALSO: type, width, color, marker, marks, rspace, rphase
             arroww, arrowl

rspace:
rphase:
arroww:
arrowl:
rspace = <float value>
or rphase = <float value>
or arroww = <float value>
or arrowl = <float value>
     Select the spacing, phase, and size of occasional ray arrows
     placed along polylines.  The spacing and phase are in NDC units
     (0.0013 NDC equals 1.0 point); the default rspace is 0.13, and
     the default rphase is 0.11375, but rphase is automatically
     incremented for successive curves on a single plot.
     The arrowhead width, arroww, and arrowhead length, arrowl are
     in relative units, defaulting to 1.0, which translates to an
     arrowhead 10 points long and 4 points in half-width.
   PLOTTING COMMANDS: plg
   SEE ALSO: type, width, color, marks, marker, rays

closed:
smooth:
closed = 0/1
or smooth = 0/1/2/3/4
     Select closed curves (closed=1) or default open curves (closed=0),
     or Bezier smoothing (smooth>0) or default piecewise linear curves
     (smooth=0).  The value of smooth can be 1, 2, 3, or 4 to get
     successively more smoothing.  Only the Bezier control points are
     plotted to an X window; the actual Bezier curves will show up in
     PostScript hardcopy files.  Closed curves join correctly, which
     becomes more noticeable for wide lines; non-solid closed curves
     may look bad because the dashing pattern may be incommensurate
     with the length of the curve.
   PLOTTING COMMANDS: plg, plc (smooth only)
   SEE ALSO: type, width, color, marks, marker, rays

font:
height:
opaque:
path:
justify:
font = <font value>
height = <float value>
opaque = 0/1
path = 0/1
justify = (see text description)
     Select text properties.  The font can be any of the strings
     "courier", "times", "helvetica" (the default), "symbol", or
     "schoolbook".  Append "B" for boldface and "I" for italic, so
     "courierB" is boldface Courier, "timesI" is Times italic, and
     "helveticaBI" is bold italic (oblique) Helvetica.  Your X server
     should have the Adobe fonts (available free from the MIT X
     distribution tapes) for all these fonts, preferably at both 75
     and 100 dpi.  Occasionally, a PostScript printer will not be
     equipped for some fonts; often New Century Schoolbook is missing.
     The font keyword may also be an integer: 0 is Courier, 4 is Times,
     8 is Helvetica, 12 is Symbol, 16 is New Century Schoolbook, and
     you add 1 to get boldface and/or 2 to get italic (or oblique).

     The height is the font size in points; 14.0 is the default.
     X windows only has 8, 10, 12, 14, 18, and 24 point fonts, so
     don't stray from these sizes if you want what you see on the
     screen to be a reasonably close match to what will be printed.

     By default, opaque=0 and text is transparent.  Set opaque=1 to
     white-out a box before drawing the text.  The default path
     (path=0) is left-to-right text; set path=1 for top-to-bottom text.

     The default text justification, justify="NN" is normal is both
     the horizontal and vertical directions.  Other possibilities
     are "L", "C", or "R" for the first character, meaning left,
     center, and right horizontal justification, and "T", "C", "H",
     "A", or "B", meaning top, capline, half, baseline, and bottom
     vertical justification.  The normal justification "NN" is equivalent
     to "LA" if path=0, and to "CT" if path=1.  Common values are
     "LA", "CA", and "RA" for garden variety left, center, and right
     justified text, with the y coordinate at the baseline of the
     last line in the string presented to plt.  The characters labeling
     the right axis of a plot are "RH", so that the y value of the
     text will match the y value of the corresponding tick.  Similarly,
     the characters labeling the bottom axis of a plot are "CT".
     The justification may also be a number, horizontal+vertical,
     where horizontal is 0 for "N", 1 for "L", 2 for "C", or 3 for "R",
     and vertical is 0 for "N", 4 for "T", 8 for "C", 12 for "H",
     16 for "A", or 20 for "B".

   PLOTTING COMMANDS: plt
   SEE ALSO: color

region:
region = <region number>
     Select the part of mesh to consider.  The region should match one
     of the numbers in the IREG array.  Putting region=0 (the default)
     means to plot the entire mesh, that is, everything EXCEPT region
     zero (non-existent zones).  Any other number means to plot only
     the specified region number; region=3 would plot region 3 only.
   PLOTTING COMMANDS: plm, plc, plv, plf

triangle:
triangle = <triangulation array>
     Set the triangulation array for a contour plot.  The triangulation
     array must be the same shape as the IREG (region number) array, and
     the correspondence between mesh zones and indices is the same as
     for IREG.  The triangulation array is used to resolve the ambiguity
     in saddle zones, in which the function Z being contoured has two
     diagonally opposite corners high, and the other two corners low.
     The triangulation array element for a zone is 0 if the algorithm is
     to choose a triangulation, based on the curvature of the first
     contour to enter the zone.  If zone (i,j) is to be triangulated
     from point (i-1,j-1) to point (i,j), then TRIANGLE(i,j)=1,
     while if it is to be triangulated from (i-1,j) to (i,j-1), then
     TRIANGLE(i,j)=-1.  Contours will never cross this "triangulation
     line".
     You should rarely need to fiddle with the triangulation array;
     it is a hedge for dealing with pathological cases.
   PLOTTING COMMANDS: plc

hollow:
aspect:
hollow = 0/1
aspect = <float value>
     Set the appearance of the "darts" of a vector field plot.  The
     default darts, hollow=0, are filled; use hollow=1 to get just the
     dart outlines.  The default is aspect=0.125; aspect is the ratio
     of the half-width to the length of the darts.  Use the color
     keyword to control the color of the darts.
   PLOTTING COMMANDS: plv
   SEE ALSO: color

edges:
ecolor:
ewidth:
edges = 0/1
ecolor = <color value>
ewidth = <float value>
     set the appearance of the zone edges in a filled mesh plot (plf).
     By default, edges=0, and the zone edges are not plotted.  If
     edges=1, a solid line is drawn around each zone after it is
     filled; the edge color and width are given by ecolor and ewidth,
     which are "fg" and 1.0 by default.
   PLOTTING COMMANDS: plf
   SEE ALSO: color, width

/*--------------------------------------------------------------------------*/
/* Inquiry and editing functions */

plq:
plq()
or plq( n_element )
or plq( n_element, n_contour )
or legend_list = plq() **** RETURN VALUE NOT YET IMPLEMENTED ****
or properties = plq(n_element, n_contour)
     Called as a subroutine, prints the list of legends for the current
     coordinate system (with an "(H)" to mark hidden elements), or prints
     a list of current properties of element N_ELEMENT (such as line type,
     width, font, etc.), or of contour number N_CONTOUR of element number
     N_ELEMENT (which must be contours generated using the plc command).
     Elements and contours are both numbered starting with one; hidden
     elements or contours are included in this numbering.

     The plq function always operates on the current coordinate system
     in the current graphics window; use window and plsys to change these.
   SEE ALSO: window, plsys, pledit, pldefault, plg

pledit:
pledit( key1=value1, key2=value2, ... )
or pledit( n_element, key1=value1, key2=value2, ... )
or pledit( n_element, n_contour, key1=value1, key2=value2, ... )
     Changes some property of element number N_ELEMENT (and contour
     number N_CONTOUR of that element).  If N_ELEMENT and N_CONTOUR are
     omitted, the default is the most recently added element, or the
     element specified in the most recent plq query command.

     The keywords can be any of the keywords that apply to the current
     element.  These are:
       plg:  color, type, width,
             marks, mcolor, marker, msize, mspace, mphase,
	     rays, rspace, rphase, arrowl, arroww,
	     closed, smooth
       pldj: color, type, width
       plt:  color, font, height, path, justify, opaque
       plm:  region, boundary, inhibit, color, type, width
       plf:  region
       plv:  region, color, hollow, width, aspect, scale
       plc:  region, color, type, width,
             marks, mcolor, marker, msize, mspace, mphase
	     smooth, levs
     (For contours, if you aren't talking about a particular N_CONTOUR,
      any changes will affect ALL the contours.)

     A plv (vector field) element can also take the scalem
     keyword to multiply all vector lengths by a specified factor.

     A plt (text) element can also take the dx and/or dy
     keywords to adjust the text position by (dx,dy).

   SEE ALSO: window, plsys, plq, pldefault, plg

plwf:
plwf (z, y = None, x = None, fill = None, shade = 0, edges = 1,
   ecolor =  None, ewidth = None, cull = None, scale = None, cmax = None,
   clear = 1)
     plots a 3-D wire frame of the given Z array, which must have the
     same dimensions as the mesh (X, Y).  If X and Y are not given, they
     default to the first and second indices of Z, respectively.
     The drawing order of the zones is determined by a simple "painter's
     algorithm", which works fairly well if the mesh is reasonably near
     rectilinear, but can fail even then if the viewpoint is chosen to
     produce extreme fisheye perspective effects.  Look at the resulting
     plot carefully to be sure the algorithm has correctly rendered the
     model in each case.

   KEYWORDS: FILL   -- optional colors to use (default is to make zones
                       have background color), same dimension options as
                       for z argument to plf function
             SHADE  -- set non-zero to compute shading from current
                       3D lighting sources
             EDGES  -- default is 1 (draw edges), but if you provide fill
                       colors, you may set to 0 to supress the edges
             ECOLOR, EWIDTH  -- color and width of edges
             CULL   -- default is 1 (cull back surfaces), but if you want
                       to see the "underside" of the model, set to 0
             SCALE  -- by default, Z is scaled to "reasonable" maximum
                       and minimum values related to the scale of (X,Y).
                       This keyword alters the default scaling factor, in
                       the sense that scale=2.0 will produce twice the
                       Z-relief of the default scale=1.0.
             CMAX   -- the AMBIENT keyword in light3 can be used to
                       control how dark the darkest surface is; use this
                       to control how light the lightest surface is
                       the lightwf routine can change this parameter
                       interactively

   SEE ALSO: lightwf, plm, plf, orient3, light3, fma3, window3

pldefault:
pldefault( key1=value1, key2=value2, ... )
     Set default values for the various properties of graphical elements.

     The keywords can be most of the keywords that can be passed to the
     plotting commands:
       plg:  color, type, width,
             marks, mcolor, msize, mspace, mphase,
	     rays, rspace, rphase, arrowl, arroww
       pldj: color, type, width
       plt:  color, font, height, path, justify, opaque
       plm:  color, type, width
       plv:  color, hollow, width, aspect
       plc:  color, type, width,
             marks, mcolor, marker, msize, mspace, mphase
       plf:  edges, ecolor, ewidth

     The initial default values are:
       color="fg", type="solid", width=1.0 (1/2 point),
       marks=1, mcolor="fg", msize=1.0 (10 points),
          mspace=0.16, mphase=0.14,
       rays=0, arrowl=1.0 (10 points), arroww=1.0 (4 points),
          rspace=0.13, rphase=0.11375,
       font="helvetica", height=12.0, path=0, justify="NN", opaque=0,
       hollow= 0, aspect=0.125,
       edges=0, ecolor="fg", ewidth=1.0 (1/2 point)

     Additional default keywords are:
       dpi, style, legends  (see window command)
       palette              (to set default filename as in palette command)
       maxcolors            (default 200)

   SEE ALSO: window, plsys, plq, pledit, plg

/*--------------------------------------------------------------------------*/
/* Miscellany */

bytscl:
bytscl(z)
or bytscl(z, top=max_byte, cmin=lower_cutoff, cmax=upper_cutoff)
     Returns a char array of the same shape as Z, with values linearly
     scaled to the range 0 to one less than the current palette size.
     If MAX_BYTE is specified, the scaled values will run from 0 to
     MAX_BYTE instead.
     If LOWER_CUTOFF and/or UPPER_CUTOFF are specified, Z values outside
     this range are mapped to the cutoff value; otherwise the linear
     scaling maps the extreme values of Z to 0 and MAX_BYTE.
   SEE ALSO: plf, pli, histeq_scale

contour:
[nc, yc, xc] = contour (level, z, y, x [, ireg] [, triangle = <vals>]
   [, region = num])
     returns the points on the contour curve that would have been
     plotted by plc.  Z, Y, X, and IREG are as for plc, and the
     triangle= and region= keywords are accepted and have the same
     meaning as for plc.  Unlike plc, the triangle array is an output
     as well as an input to contour; if supplied it may be modified
     to reflect any triangulations which were performed by contour.

     either:
     LEVEL is a scalar z value to return the points at that contour
     level.  All such points lie on edges of the mesh.  If a contour
     curve closes, the final point is the same as the initial point
     (i.e.- that point is included twice in the returned list).

     or:
     LEVEL is a pair of z values [z0,z1] to return the points of
     a set of polygons which outline the regions between the two
     contour levels.  These will include points on the mesh boundary
     which lie between the levels, in addition to the edge points
     for both levels.  The polygons are closed, simply connected,
     and will not contain more than about 4000 points (larger polygons
     are split into pieces with a few points repeated where the pieces
     join).

     YC and XC are the output points on the curve(s), or None if there
     are no points. The return value NC is a list of the lengths of
     the polygons/polylines returned in (XC,YC), or None if there are
     none.  len(XC) == len(YC) == sum(NC).  For the level pair
     case, YC, XC, and NC are ready to be used as inputs to plfp.

   KEYWORDS: triangle, region
   SEE ALSO: plc, plfp

mesh_loc:
mesh_loc(y0, x0)
or mesh_loc(y0, x0, y, x)
or mesh_loc(y0, x0, y, x, ireg)
     Returns the zone index (=i+imax*(j-1)) of the zone of the mesh
     (X,Y) (with optional region number array IREG) containing the
     point (X0,Y0).  If (X0,Y0) lies outside the mesh, returns 0.
     Thus, eg- ireg(mesh_loc(x0, y0, y, x, ireg)) is the region number of
     the region containing (x0,y0).  If no mesh specified, uses default.
     X0 and Y0 may be arrays as long as they are conformable.
   SEE ALSO: plmesh, moush, mouse

mouse:
result = mouse(system, style, prompt)
     Displays a PROMPT, then waits for a mouse button to be pressed,
     then released.  Returns tuple of length eleven:
       result= [x_pressed, y_pressed, x_released, y_released,
                xndc_pressed, yndc_pressed, xndc_released, yndc_released,
	        system, button, modifiers]

     If SYSTEM>=0, the first four coordinate values will be relative to
     that coordinate system.
     For SYSTEM<0, the first four coordinate values will be relative to
     the coordinate system under the mouse when the button was pressed.
     The second four coordinates are always normalized device coordinates,
     which start at (0,0) in the lower left corner of the 8.5x11 sheet of
     paper the picture will be printed on, with 0.0013 NDC unit being
     1/72.27 inch (1.0 point).  Look in the style sheet for the location
     of the viewport in NDC coordinates (see the style keyword).

     If STYLE is 0, there will be no visual cues that the mouse
     command has been called; this is intended for a simple click.
     If STYLE is 1, a rubber band box will be drawn; if STYLE is 2,
     a rubber band line will be drawn.  These disappear when the
     button is released.

     Clicking a second button before releasing the first cancels the
     mouse function, which will then return nil.
     Ordinary text input also cancels the mouse function, which again
     returns nil.

     The left button reverses forground for background (by XOR) in
     order to draw the rubber band (if any).  The middle and right
     buttons use other masks, in case the rubber band is not visible
     with the left button.

     result[8] is the coordinate system in which the first four
     coordinates are to be interpreted.
     result[9] is the button which was pressed, 1 for left, 2
     for middle, and 3 for right (4 and 5 are also possible).
     result[10] is a mask representing the modifier keys which
     were pressed during the operation: 1 for shift, 2 for shift lock,
     4 for control, 8 for mod1 (alt or meta), 16 for mod2, 32 for mod3,
     64 for mod4, and 128 for mod5.

   SEE ALSO: moush

moush:
moush(y, x, ireg)
or moush(y, x)
or moush()
     Returns the 1-origin zone index for the point clicked in
     for the default mesh, or for the mesh (X,Y) (region array IREG).

pause:
pause( milliseconds )
     Pause for the specified number of milliseconds of wall clock
     time, or until input arrives from the keyboard.
     This is intended for use in creating animated sequences.

/*--------------------------------------------------------------------------*/

histeq_scale: **** NOT YET IMPLEMENTED ****
histeq_scale(z, top=top_value, cmin=cmin, cmax=cmax)
     Returns a byte-scaled version of the array Z having the property
     that each byte occurs with equal frequency (Z is histogram
     equalized).  The result bytes range from 0 to TOP_VALUE, which
     defaults to one less than the size of the current palette (or
     255 if no pli, plf, or palette command has yet been issued).

     If non-nil CMIN and/or CMAX is supplied, values of Z beyond these
     cutoffs are not included in the frequency counts.

   SEE ALSO: bytscl, plf, pli

spann:
spann (zmin, zmax, n = 8, fudge = 0, force = 0)
      return no more than N equally spaced "nice" numbers between
      ZMIN and ZMAX.
      Note that in general spann may not supply the number of
      values that you asked for. To force it to do so, set
      keyword FORCE to nonzero.
    SEE ALSO: span, spanl, plc, plfc

lightwf:
lightwf (cmax) 
     Sets the CMAX parameter interactively, assuming the current
     3D display list contains the result of a previous plwf call.
     This changes the color of the brightest surface in the picture.
     The darkest surface color can be controlled using the AMBIENT
     keyword to light3.

   SEE ALSO: plwf, light3

orient3:
orient3 ( [phi = val1, theta = val2] )
   Set the orientation of the object to (PHI, THETA). Orientations
   are a subset of the possible rotation matrices in which the z axis
   of the object appears vertical on the screen (that is, the object
   z axis projects onto the viewer y axis). The THETA angle is the
   angle from the viewer y axis to the object z axis, positive if
   the object z axis is tilted towards you (toward viewer +z). PHI is
   zero when the object x axis coincides with the viewer x axis. If
   neither PHI nor THETA is specified, PHI defaults to - pi / 4 and
   THETA defaults to pi / 6. If only PHI is specified, THETA remains
   unchanged, unless the current THETA is near pi / 2, in which case
   THETA returns to pi / 6, or unless the current orientation does
   not have a vertical z axis, in which case THETA returns to its
   default. If only THETA is specified, PHI retains its current value.
   Unlike rot3, orient3 is not a cumulative operation.

   SEE ALSO: rot3, mov3, aim3, save3, restore3, light3

light3:
light3 (ambient=a_level,
                diffuse=d_level,
                specular=s_level,
                spower=n,
                sdir=xyz)
     Sets lighting properties for 3D shading effects.
     A surface will be shaded according to its to its orientation
     relative to the viewing direction.

     The ambient level A_LEVEL is a light level (arbitrary units)
     that is added to every surface independent of its orientation.

     The diffuse level D_LEVEL is a light level which is proportional
     to cos(theta), where theta is the angle between the surface
     normal and the viewing direction, so that surfaces directly
     facing the viewer are bright, while surfaces viewed edge on are
     unlit (and surfaces facing away, if drawn, are shaded as if they
     faced the viewer).

     The specular level S_LEVEL is a light level proportional to a high
     power spower=N of 1+cos(alpha), where alpha is the angle between
     the specular reflection angle and the viewing direction.  The light
     source for the calculation of alpha lies in the direction XYZ (a
     3 element vector) in the viewer's coordinate system at infinite
     distance.  You can have ns light sources by making S_LEVEL, N, and
     XYZ (or any combination) be vectors of length ns (3-by-ns in the
     case of XYZ).  (See source code for specular_hook function
     definition if powers of 1+cos(alpha) aren't good enough for you.)

     With no arguments, return to the default lighting.

   EXAMPLES:
     light3 ( diffuse=.1, specular=1., sdir=[0,0,-1])
       (dramatic "tail lighting" effect)
     light3 ( diffuse=.5, specular=1., sdir=[1,.5,1])
       (classic "over your right shoulder" lighting)
     light3 ( ambient=.1,diffuse=.1,specular=1.,
             sdir=[[0,0,-1],[1,.5,1]],spower=[4,2])
       (two light sources combining previous effects)

   SEE ALSO: rot3, save3, restore3

window3:
window3 ( [n,] , dump = 0, hcp = None)
   initialize style="nobox.gs" window for 3D graphics. dump = 1
   to dump the palette to the hardcopy file (if any), for color
   plots; use hcp to specify the filename for hardcopy.

rot3:
rot3 (xa = 0., ya = 0., za = 0.)
   rotate the current 3D plot by XA about viewer's x-axis,
   YA about viewer's y-axis, and ZA about viewer's z-axis.

   SEE ALSO: orient3, mov3, aim3, setz3, undo3, save3, restore3, light3

save3:
view = save3 ( )
     Save the current 3D viewing transformation and lighting.
     Actually, this doesn't save anything; it returns a copy
     of the current 3D viewing transformation and lighting, so
     that the user can put it aside somewhere.

   SEE ALSO: restore3, rot3, mov3, aim3, light3

restore3:
restore3 ( view )
   Restore a previously saved 3D viewing transformation and lighting.
   If view is missing, rotate object to viewer's coordinate system.

   SEE ALSO: restore3, rot3, mov3, aim3, light3

aim3:
aim3 ( xa = 0., ya = 0., za = 0. )
   move the current 3D plot to put the point (XA, YA, ZA) in object
   coordinates at the point (0, 0, 0) -- the aim point -- in the
   viewer's coordinates. If any of the XA, YA, or ZA is None or
   missing, it defaults to zero.

   SEE ALSO: mov3, rot3, orient3, setz3, undo3, save3, restore3, light3

mov3:
mov3 ( xa = 0., ya = 0., za = 0. )
   move the current 3D plot by XA along the viewer's x axis,
   YA along the viewer's y axis, and ZA along the viewer's z axis.

   SEE ALSO: rot3, orient3, setz3, undo3, save3, restore3, light3

setz3:
setz3 ( zc = None )
   Set the camera position to z = ZC (x = y = 0) in the viewer's coordinate
   system. If zc is None, set the camera to infinity (default).

   SEE ALSO: rot3, orient3, undo3, save3, restore3, light3

undo3:
undo3 (n = 1)
     Undo the effects of the last N (default 1) rot3, orient3, mov3, aim3,
     setz3, or light3 commands.

get3_light:
get3_light (xyz [, nxyz])
     return 3D lighting for polygons with vertices XYZ.  If NXYZ is
     specified, XYZ should be sum(nxyz)-by-3, with NXYZ being the
     list of numbers of vertices for each polygon (as for the plfp
     function).  If NXYZ is not specified, XYZ should be a quadrilateral
     mesh, ni-by-nj-by-3 (as for the plf function).  In the first case,
     the return value is len (NXYZ) long; in the second case, the
     return value is (ni-1)-by-(nj-1).

     The parameters of the lighting calculation are set by the
     light3 function.

     SEE ALSO: light3, set3_object, get3_normal, get3_centroid

set3_object:
set3_object (fnc, arg)
or set3_object (fnc, [arg1, arg2,...])
     set up to trigger a call to draw3, adding a call to the
     3D display list of the form:

        DRAWING_FUNCTION ( [ARG1, ARG2, ...]))

     When draw3 calls DRAWING_FUNCTION, the external variable _draw3
     will be non-zero, so DRAWING_FUNCTION can be written like this:

     def drawing_function(arg) :
 
       if (_draw3) :
          arg1= arg [0]
          arg1= arg [1]
          ...
          ...<calls to get3_xy, sort3d, get3_light, etc.>...
          ...<calls to graphics functions plfp, plf, etc.>...
          return
 
       ...<verify args>...
       ...<do orientation and lighting independent calcs>...
       set3_object (drawing_function, [arg1,arg2,...])
 
   SEE ALSO: get3_xy, get3_light, sort3d

sort3d:
sort3d(z, npolys)
     given Z and NPOLYS, with len(Z)==sum(npolys), return
     a 2-element list [LIST, VLIST] such that Z[VLIST] and NPOLYS[LIST] are
     sorted from smallest average Z to largest average Z, where
     the averages are taken over the clusters of length NPOLYS.
     Within each cluster (polygon), the cyclic order of Z[VLIST]
     remains unchanged, but the absolute order may change.

     This sorting order produces correct or nearly correct order
     for a plfp command to make a plot involving hidden or partially
     hidden surfaces in three dimensions.  It works best when the
     polys form a set of disjoint closed, convex surfaces, and when
     the surface normal changes only very little between neighboring
     polys.  (If the latter condition holds, then even if sort3d
     mis-orders two neighboring polys, their colors will be very
     nearly the same, and the mistake won't be noticeable.)  A truly
     correct 3D sorting routine is impossible, since there may be no
     rendering order which produces correct surface hiding (some polys
     may need to be split into pieces in order to do that).  There
     are more nearly correct algorithms than this, but they are much
     slower.
   SEE ALSO: get3_xy, plfp

draw3:
def draw3 (called_as_idler = 0, lims = None)
   Draw the current 3d display list.
   (Ordinarily triggered automatically when the drawing changes.)

get3_normal:
get3_normal (xyz [, nxyz])
     return 3D normals for polygons with vertices XYZ.  If NXYZ is
     specified, XYZ should be sum(nxyz)-by-3, with NXYZ being the
     list of numbers of vertices for each polygon (as for the plfp
     function).  If NXYZ is not specified, XYZ should be a quadrilateral
     mesh, ni-by-nj-by-3 (as for the plf function).  In the first case,
     the return value is len(NXYZ)-by-3; in the second case, the
     return value is (ni-1)-by-(nj-1)-by-3.

     The normals are constructed from the cross product of the lines
     joining the midpoints of two edges which as nearly quarter the
     polygon as possible (the medians for a quadrilateral).  No check
     is made that these not be parallel; the returned "normal" is
     [0,0,0] in that case.  Also, if the polygon vertices are not
     coplanar, the "normal" has no precisely definable meaning.

     SEE ALSO: get3_centroid, get3_light

get3_centroid:
get3_centroid (xyz [, nxyz])
     return 3D centroids for polygons with vertices XYZ.  If NXYZ is
     specified, XYZ should be sum(nxyz)-by-3, with NXYZ being the
     list of numbers of vertices for each polygon (as for the plfp
     function).  If NXYZ is not specified, XYZ should be a quadrilateral
     mesh, ni-by-nj-by-3 (as for the plf function).  In the first case,
     the return value is len(NXYZ) in length; in the second case, the
     return value is (ni-1)-by-(nj-1)-by-3.

     The centroids are constructed as the mean value of all vertices
     of each polygon.

     SEE ALSO: get3_normal, get3_light

get3_xy:
get3_xy (xyz [, flg])
     Given anything-by-3 coordinates XYZ, return X and Y in viewer's
     coordinate system (set by rot3, mov3, orient3, etc.).  If the
     second argument is present and non-zero, also return Z (for use
     in sort3d or get3_light, for example).  If the camera position
     has been set to a finite distance with setz3, the returned
     coordinates will be tangents of angles for a perspective
     drawing (and Z will be scaled by 1/zc).
     Unlike the Yorick version, this function returns a 3-by-anything
     array of coordinates.
     Actually, what it returns is a 3-by-anything python array, whose
     0th element is the x array, whose 1th element is the y array, and
     whose 2th element is the z array if asked for.
     I believe that x, y, and z can be either 1d or 2d, so this
     routine is written in two cases.

gnomon:
gnomon ( [onoff] [, chr = ["X", "Y", "Z"]])
     Toggle the gnomon display. If on is present and non-zero,
     turn on the gnomon. If zero, turn it off.
 
     The gnomon shows the X, Y, and Z axis directions in the
     object coordinate system. The directions are labeled.
     The gnomon is always infinitely far behind the object
     (away from the camera).
 
     There is a mirror-through-the-screen-plane ambiguity in the
     display which is resolved in two ways: (1) the (X, Y, Z)
     coordinate system is right-handed, and (2) If the tip of an
     axis projects into the screen, its label is drawn in opposite
     polarity to the other text in the screen.

     CHR specifies the axis labels.

spin3:
spin3 (nframes = 30, axis = array ([-1, 1, 0],  Float), tlimit = 60.,
   dtmin = 0.0, bracket_time = array ([2., 2.],  Float), lims = None,
   timing = 0, angle = 2. * pi)
     Spin the current 3D display list about AXIS over NFRAMES.  Keywords
     TLIMIT the total time allowed for the movie in seconds (default 60),
     DTMIN the minimum allowed interframe time in seconds (default 0.0),
     BRACKET_TIME (as for movie function in movie.py), TIMING = 1 if
     you want timing measured and printed out, 0 if not.

   SEE ALSO: rot3

color_bar:
color_bar (minz, maxz, split = 0)
   color_bar (minz, maxz) plots a color bar to the right of the plot square
   labelled by the z values from minz to maxz.

   plf (z, y, x)
   color_bar (z (min, min), z (max, max))

   or
   plf (z, y, x, cmin = MINZ, cmax = MAXZ)
   color_bar (MINZ, MAXZ)

   are typical usage.

   The SPLIT keyword should be nonzero for a split palette.

movie:
 movie (draw_frame, time_limit = 120., min_interframe = 0.0,
   bracket_time = array ([2., 2.], Float ), lims = None, timing = 0)
     runs a movie based on the given DRAW_FRAME function.  The movie
     stops after a total elapsed time of TIME_LIMIT seconds, which
     defaults to 60 (one minute), or when the DRAW_FRAME function
     returns zero.

     note: All but the first argument are keyword arguments, with
     defaults as shown.

     def draw_frame(i) :
       # Input argument i is the frame number.
       # draw_frame should return non-zero if there are more
       # frames in this movie.  A zero return will stop the
       # movie.
       # draw_frame must NOT include any fma command if the
       # making_movie variable is set (movie sets this variable
       # before calling draw_frame)

     If MIN_INTERFRAME is specified, a pause will be added as
     necessary to slow down the movie.  MIN_INTERFRAME is a time
     in seconds (default 0).

     The keyword BRACKET_TIME (again a time in seconds) can be
     used to adjust the duration of the pauses after the first
     and last frames.  It may also be a two element array [beg, end].
     If the pause at the end is greater than five seconds, you will
     be prompted to explain that hitting <RETURN> will abort the final
     pause. Well, the Python version does not have this capability.

     TIMING = 1 enables a timing printout for your movie.

     If every frame of your movie has the same limits, use the
     limits function to fix the limits before you call movie.

   BUG:  If you hit <RETURN> to start a movie early, it will not
         pause at the end of the movie at all.  You probably should
         not use long initial pauses.

movie_stats:
movie_stats ( [timing])
     prints statistics from the last movie command, or from the
     command which produced TIMING.  TIMING is the contents of the
     movie_timing external variable after the movie command completes.

plane3:
plane3 (normal, point)
     NORMAL and POINT are sequences of length 3.
     returns [nx,ny,nz,pp] for the specified plane.

mesh3:
m = mesh3 (x, y, z)
     or m = mesh3 (x, y, z, funcs = [f1, f2,...])
     or m = mesh3 (xyz, funcs = [f1, f2,...])
     or m = mesh3 (nxnynz, dxdydz, x0y0z0, funcs = [f1, f2,...])

     make mesh3 argument for slice3, xyz3, getv3, etc., functions.
     X, Y, and Z are each 3D coordinate arrays.  The optional F1, F2,
     etc. are 3D arrays of function values (e.g. density, temperature)
     which have one less value along each dimension than the coordinate
     arrays.  The "index" of each zone in the returned mesh3 is
     the index in these cell-centered Fi arrays, so every index from
     one through the total number of cells indicates one real cell.
     The Fi arrays can also have the same dimensions as X, Y, or Z
     in order to represent point-centered quantities.

     If X has four dimensions and the length of the first is 3, then
     it is interpreted as XYZ (which is the quantity actually stored
     in the returned cell list).

     If X is a vector of 3 integers, it is interpreted as [nx,ny,nz]
     of a uniform 3D mesh, and the second and third arguments are
     [dx,dy,dz] and [x0,y0,z0] respectively.  (DXDYDZ represent the
     size of the entire mesh, not the size of one cell, and NXNYNZ are
     the number of cells, not the number of points.)

     Added by ZCM 1/13/97: if x, y, and z are one-dimensional of
     the same length and if the keyword verts exists and yields
     an NCELLS by 8 integer array, then we have an unstructured
     rectangular mesh, and the subscripts of cell i's vertices
     are verts[i, 0:8].

     other sorts of meshes are possible -- a mesh which lives
     in a binary file is an obvious example -- which would need
     different workers for xyz3, getv3, getc3, and iterator3
     iterator3_rect may be more general than the other three.

slice3:
[NVERTS, XYZVERTS, color] = slice3 (m3, fslice, nverts, xyzverts)
     or [NVERTS, XYZVERTS, color] =
        slice3(m3, fslice, nverts, xyzverts, fcolor, node = 0)
     or [NVERTS, XYZVERTS, color] =
        slice3(m3, fslice, nverts, xyzverts, fcolor, 1, node = 0)

     slice the 3D mesh M3 using the slicing function FSLICE, returning
     the list [NVERTS, XYZVERTS, color].  Note that it is impossible to
     pass arguments as addresses, as yorick does in this routine.
     NVERTS is the number of vertices in each polygon of the slice, and
     XYZVERTS is the 3-by-sum(NVERTS) list of polygon vertices.  If the
     FCOLOR argument is present, the values of that coloring function on
     the polygons are returned as the value of the slice3 function
     (len (color_values) == len (NVERTS) == number of polygons).

     If the slice function FSLICE is a function, it should be of the
     form:
        func fslice(m3, chunk)
     returning a list of function values on the specified chunk of the
     mesh m3.  The format of chunk depends on the type of m3 mesh, so
     you should use only the other mesh functions xyz3 and getv3 which
     take m3 and chunk as arguments.  The return value of fslice should
     have the same dimensions as the return value of getv3; the return
     value of xyz3 has an additional first dimension of length 3.
     N. B. (ZCM 2/24/97) I have eliminated the globals iso_index
     and _value, so for isosurface_slicer only, the call must be
     of the form fslice (m3, chunk, iso_index, _value).
        Likewise, I have eliminated normal and projection, so
     for plane_slicer only, we do fslice (m3, chunk, normal, projection).

     If FSLICE is a list of 4 numbers, it is taken as a slicing plane
     of the form returned by plane3.

     If FSLICE is a single integer, the slice will be an isosurface for
     the FSLICEth variable associated with the mesh M3.  In this case,
     the keyword VALUE must also be present, representing the value
     of that variable on the isosurface.

     If FCOLOR is None, slice3 returns None.  If you want to color the
     polygons in a manner that depends only on their vertex coordinates
     (e.g.- by a 3D shading calculation), use this mode.

     If FCOLOR is a function, it should be of the form:
        func fcolor(m3, cells, l, u, fsl, fsu, ihist)
     returning a list of function values on the specified cells of the
     mesh m3.  The cells argument will be the list of cell indices in
     m3 at which values are to be returned.  l, u, fsl, fsu, and ihist
     are interpolation coefficients which can be used to interpolate
     from vertex centered values to the required cell centered values,
     ignoring the cells argument.  See getc3 source code.
     The return values should always have cells.shape.

     If FCOLOR is a single integer, the slice will be an isosurface for
     the FCOLORth variable associated with the mesh M3.

     If the optional argument after FCOLOR is non-nil and non-zero,
     then the FCOLOR function is called with only two arguments:
        func fcolor(m3, cells)

     The keyword argument NODE, if present and nonzero, is a signal
        to return node-centered values rather than cell-centered
        values. (ZCM 4/16/97)

slice3mesh:
     slice3mesh (z, color = None, smooth = 0)
     slice3mesh (nxny, dxdy, x0y0, z, color = None, smooth = 0)
     slice3mesh (x, y, z, color = None, smooth = 0)
    
     slice3mesh returns a triple [nverts, xyzverts, color]
      nverts is no_cells long and the ith entry tells how many
         vertices the ith cell has.
      xyzverts is sum (nverts) by 3 and gives the vertex
         coordinates of the cells in order.
      color, if present, is len (nverts) long and contains
         a color value for each cell in the mesh if smooth == 0;
         sum (nverts) long and contains a color value for each
         node in the mesh if smooth == 1.
     There are a number of ways to call slice3mesh:
        slice3mesh (z, color = None, smooth = 0)
     z is a two dimensional array of function values, assumed
     to be on a uniform mesh nx by ny nodes (assuming z is nx by ny)
     nx being the number of nodes in the x direction, ny the number
     in the y direction.
     color, if specified, is either an nx - 1 by ny - 1 array
     of cell-centered values by which the surface is to
     be colored, or an nx by ny array of vertex-
     centered values, which will be averaged over each
     cell to give cell-centered values if smooth == 0, or
     returned as a node-centered array sum (nverts) long if
     smooth == 1.
        slice3mesh (nxny, dxdy, x0y0, z, color = None, smooth = 0)
     In this case, slice3mesh accepts the specification for
     a regular 2d mesh: nxny is the number of cells in the
     x direction and the y direction (i. e., its two components
     are nx - 1 and ny - 1, nx by ny being the node size;
     x0y0 are the initial values of x and y; and dxdy are the
     increments in the two directions. z is the height of a
     surface above the xy plane and must be dimensioned nx by ny. 
     color, if specified, is as above.
       slice3mesh (x, y, z, color = None, smooth = 0)
     z is as above, an nx by ny array of function values
     on a mesh of the same dimensions. There are two choices
     for x and y: they can both be one-dimensional, dimensioned
     nx and ny respectively, in which case they represent a
     mesh whose edges are parallel to the axes; or else they
     can both be nx by ny, in which case they represent a
     general quadrilateral mesh.
     color, if specified, is as above.

iterator3:
iterator3 (m3 , chunk = None, clist = None)
   The iterator3 function combines three distinct operations:
   (1) If only the M3 argument is given, return the initial
       chunk of the mesh.  The chunk will be no more than
       chunk3_limit cells of the mesh.
   (2) If only M3 and CHUNK are given, return the next CHUNK,
       or [] if there are no more chunks.
   (3) If M3, CHUNK, and CLIST are all specified, return the
       absolute cell index list corresponding to the index list
       CLIST of the cells in the CHUNK.
       Do not increment the chunk in this case.
  
   The form of the CHUNK argument and return value for cases (1)
   and (2) is not specified, but it must be recognized by the
   xyz3 and getv3 functions which go along with this iterator3.
   (For case (3), CLIST and the return value are both ordinary
    index lists.)
   
iterator3_rect:
   This is the iterator3 function for a regular rectangular mesh.

iterator3_irreg:
   Does the same thing as iterator3_rect only for an irregular
   rectangular mesh. It simply splits a large mesh into smaller
   parts. Whether this is necessary I am not sure.
   Certainly it makes it easier in the irregular case to handle
   the four different types of cells separately.
   if clist is present, in the irregular case it is already
   the list of absolute cell indices, so it is simply returned.
   This and other routines to do with irregular meshes return a
   chunk which is a 2-list. The first item delimits the chunk;
   the second gives a list of corresponding cell numbers.

getv3:
   getv3(i, m3, chunk)

     return vertex values of the Ith function attached to 3D mesh M3
     for cells in the specified CHUNK.  The CHUNK may be a list of
     cell indices, in which case getv3 returns a 2x2x2x(CHUNK.shape)
     list of vertex coordinates.  CHUNK may also be a mesh-specific data
     structure used in the slice3 routine, in which case getv3 may
     return a (ni)x(nj)x(nk) array of vertex values.  For meshes which
     are logically rectangular or consist of several rectangular
     patches, this is up to 8 times less data, with a concomitant
     performance advantage.  Use getv3 when writing slicing functions
     for slice3.

getv3_rect:
     the getv3 function for regular rectangular meshes.

getv3_irreg:
   for an irregular mesh, returns a 3-list whose elements are:
   (1) the function values for the ith function on the vertices of the
   given chunk. (The function values must have the same dimension
   as the coordinates; there is no attempt to convert zone-centered
   values to vertex-centered values.)
   (2) an array of relative cell numbers within the list of cells
   of this type.
   (3) a number that can be added to these relative numbers to gives
   the absolute cell numbers for correct access to their coordinates
   and function values.

getc3:
getc3(i, m3, chunk)
         or getc3(i, m3, clist, l, u, fsl, fsu, cells)

     return cell values of the Ith function attached to 3D mesh M3
     for cells in the specified CHUNK.  The CHUNK may be a list of
     cell indices, in which case getc3 returns a (CHUNK.shape)
     list of vertex coordinates.  CHUNK may also be a mesh-specific data
     structure used in the slice3 routine, in which case getc3 may
     return a (ni)x(nj)x(nk) array of vertex values.  There is no
     savings in the amount of data for such a CHUNK, but the gather
     operation is cheaper than a general list of cell indices.
     Use getc3 when writing colorng functions for slice3.

     If CHUNK is a CLIST, the additional arguments L, U, FSL, and FSU
     are vertex index lists which override the CLIST if the Ith attached
     function is defined on mesh vertices.  L and U are index lists into
     the (CLIST.shape)x2x2x2 vertex value array, say vva, and FSL
     and FSU are corresponding interpolation coefficients; the zone
     centered value is computed as a weighted average of involving these
     coefficients.  The CELLS argument is required by histogram to do
     the averaging.  See the source code for details.
     By default, this conversion (if necessary) is done by averaging
     the eight vertex-centered values.

getc3_rect:
     The getc3 function for a regular rectangular mesh.

getc3_irreg:
     Returns the same type of data structure as getc3_rect,
     but from an irregular rectangular mesh.
      m3 [1] is a 2-list; m3[1] [0] is an array whose ith element
         is an array of coordinate indices for the ith cell,
         or a list of up to four such arrays.
         m3 [1] [1] is the 3 by nverts array of coordinates.
      m3 [2] is a list of arrays of vertex-centered or cell-centered
         data.
     chunk may be a list, in which case chunk [0] is a 2-sequence
      representing a range of cell indices; or it may be a one-dimensional
      array, in which case it is a nonconsecutive set of cell indices.
      It is guaranteed that all cells indexed by the chunk are the
      same type.

plzcont:
plzcont (nverts, xyzverts, contours = 8, scale = "lin", clear = 1,
   edges = 0, color = None, cmin = None, cmax = None, split = 0)
     Plot filled z contours on the specified surface. NVERTS and
     XYZVERTS arrays specify the polygons for the surface being
     drawn. CONTOURS can be one of the following:
        N, an integer: Plot N contours (therefore, N+1 colored
        components of the surface)
        CVALS, a vector of floats: draw the contours at the
        specified levels.
     SCALE can be "lin", "log", or "normal" specifying the
     contour scale. (Only applicable if contours = N, of course).
     If CLEAR = 1, clear the display list first.
     If EDGES = 1, plot the edges.
     The algorithm is to apply slice2x repeatedly to the surface.
     If color == None, then bytscl the palette into N + 1 colors
     and send each of the slices to pl3tree with the appropriate color.
     If color == "bg", will plot only the edges.
     If CMIN is given, use it instead of the minimum z actually
     being plotted. If CMAX is given, use it instead of the maximum
     z actually being plotted.
     If SPLIT is nonzero, then colors are computed based only on the
     first half of the palette (i. e., a split palette is assumed).

pl4cont:
pl4cont (nverts, xyzverts, values, contours = 8, scale = "lin", clear = 1,
   edges = 0, color = None, cmin = None, cmax = None, split = 0)
     Plot filled z contours on the specified surface. VALUES is
     a cell-centered array the same length as SUM (NVERTS) whose
     contours will be drawn. NVERTS and
     XYZVERTS arrays specify the polygons for the surface being
     drawn. CONTOURS can be one of the following:
        N, an integer: Plot N contours (therefore, N+1 colored
        components of the surface)
        CVALS, a vector of floats: draw the contours at the
        specified levels.
     SCALE can be "lin", "log", or "normal" specifying the
     contour scale. (Only applicable if contours = N, of course).
     If CLEAR == 1, clear the display list first.
     If EDGES == 1, plot the edges.
     The algorithm is to apply slice2x repeatedly to the surface.
     If color == None, then bytscl the palette into N + 1 colors
     and send each of the slices to pl3tree with the appropriate color.
     If color == "bg", will plot only the edges.
     If CMIN is given, use it instead of the minimum z actually
     being plotted. If CMAX is given, use it instead of the maximum
     z actually being plotted.
     If SPLIT is nonzero, then colors are computed based only on the
     first half of the palette (i. e., a split palette is assumed).

slice2x:
[nvf, xyzvf, colorf, nvb, xyzvb, colorb] =
   slice2x (plane, nverts, xyzverts, values = None)
     Slice a polygon list, returning in nvf and xyzvf only those
     polygons or parts of polygons on the positive side of PLANE,
     and in nvb and xyzvb only those polygons or parts of polygons
     on the negative side of PLANE.
     If PLANE is a scalar real, then VALUES must be a function
     defined on the vertices of the mesh, and the mesh will
     be sliced where the function has that value.
     The NVERTS, XYZVERTS, and VALUES arrays have the meanings
     of the return values from the slice3 function. It is legal
     to omit the VALUES argument (e.g.- if there is no fcolor
     function).
     In order to plot two intersecting slices, one could
     slice (for example) the horizontal plane twice (slice2x) -
     first with the plane of the vertical slice, then with minus
     that same plane.  Then, plot first the back part of the
     slice, then the vertical slice, then the front part of the
     horizontal slice.  Of course, the vertical plane could
     be the one to be sliced, and "back" and "front" vary
     depending on the view point, but the general idea always
     works.

slice2:
[nvf, xyzvf, colorf] = slice2 ((plane, nverts, xyzverts, values = None)
     Slice a polygon list, returning in nvf and xyzvf only those
     polygons or parts of polygons on the positive side of PLANE.
     If PLANE is a scalar real, then VALUES must be a function
     defined on the vertices of the mesh, and the mesh will
     be sliced where the function has that value.
     The NVERTS, XYZVERTS, and VALUES arrays have the meanings
     of the return values from the slice3 function. It is legal
     to omit the VALUES argument (e.g.- if there is no fcolor
     function).
     In order to plot two intersecting slices, one could
     slice (for example) the horizontal plane twice (slice2x) -
     first with the plane of the vertical slice, then with minus
     that same plane.  Then, plot first the back part of the
     slice, then the vertical slice, then the front part of the
     horizontal slice.  Of course, the vertical plane could
     be the one to be sliced, and "back" and "front" vary
     depending on the view point, but the general idea always
     works.

pl3surf:
pl3surf(nverts, xyzverts = None, values = None, cmin = None, cmax = None,
        lim = None, edges = 0)
     Perform simple 3D rendering of an object created by slice3
     (possibly followed by slice2).  NVERTS and XYZVERTS are polygon
     lists as returned by slice3, so XYZVERTS is sum(NVERTS)-by-3,
     where NVERTS is a list of the number of vertices in each polygon.
     If present, the VALUES should have the same length as NVERTS;
     they are used to color the polygon.  If VALUES is not specified,
     the 3D lighting calculation set up using the light3 function
     will be carried out.  Keywords CMIN and CMAX as for plf, pli,
     or plfp are also accepted.  (If you do not supply VALUES, you
     probably want to use the AMBIENT keyword to light3 instead of
     CMIN here, but CMAX may still be useful.)
     EDGES should be set nonzero if you wish the mesh lines plotted
     on the surface.

pl3tree:
pl3tree (nverts, xyzverts = None, values = None, plane = None,
         cmin = None, cmax = None, split = 1, edges = 0)
     Add the polygon list specified by NVERTS (number of vertices in
     each polygon) and XYZVERTS (3-by-sum(NVERTS) vertex coordinates)
     to the currently displayed b-tree.  If VALUES is specified, it
     must have the same dimension as NVERTS, and represents the color
     of each polygon.  (ZCM 7/18/97) Or, if VALUES == "bg" ("background")
     Then each polygon will be filled with the background color,
     giving just a wire frame. If VALUES is not specified, the polygons
     are assumed to form an isosurface which will be shaded by the
     current 3D lighting model; the isosurfaces are at the leaves of
     the b-tree, sliced by all of the planes.  If PLANE is specified,
     the XYZVERTS must all lie in that plane, and that plane becomes
     a new slicing plane in the b-tree.

     Each leaf of the b-tree consists of a set of sliced isosurfaces.
     A node of the b-tree consists of some polygons in one of the
     planes, a b-tree or leaf entirely on one side of that plane, and
     a b-tree or leaf on the other side.  The first plane you add
     becomes the root node, slicing any existing leaf in half.  When
     you add an isosurface, it propagates down the tree, getting
     sliced at each node, until its pieces reach the existing leaves,
     to which they are added.  When you add a plane, it also propagates
     down the tree, getting sliced at each node, until its pieces
     reach the leaves, which it slices, becoming the nodes closest to
     the leaves.

     This structure is relatively easy to plot, since from any
     viewpoint, a node can always be plotted in the order from one
     side, then the plane, then the other side.
     This routine assumes a "split palette"; the colors for the
     VALUES will be scaled to fit from color 0 to color 99, while
     the colors from the shading calculation will be scaled to fit
     from color 100 to color 199.  (If VALUES is specified as an
     unsigned char (Python typecode 'b') array, however, it will
     be used without scaling.) You may specifiy a CMIN or CMAX
     keyword to affect the scaling; CMIN is ignored if VALUES is
     not specified (use the AMBIENT keyword from light3 for that
     case).

     (ZCM 4/23/97) Add the SPLIT keyword. This will determine
     whether or not to split the palette (half to the isosurfaces
     for shading and the other half to plane sections for contouring).

split_palette:
split_palette ( [ <string>])
     split the current palette or the specified palette into two
     parts; colors 0 to 99 will be a compressed version of the
     original, while colors 100 to 199 will be a gray scale.

split_bytscl:
split_bytscl (x, upper, cmin = None, cmax = None)
     as bytscl function, but scale to the lower half of a split
     palette (0-99, normally the color scale) if the second parameter
     is zero or nil, or the upper half (100-199, normally the gray
     scale) if the second parameter is non-zero.

xyz3:
xyz3 (m3, chunk)
     return vertex coordinates for CHUNK of 3D mesh M3.  The CHUNK
     may be a list of cell indices, in which case xyz3 returns a
     (CHUNK.shape)x3x2x2x2 list of vertex coordinates.  CHUNK may
     also be a mesh-specific data structure used in the slice3
     routine, in which case xyz3 may return a 3x(ni)x(nj)x(nk)
     array of vertex coordinates.  For meshes which are logically
     rectangular or consist of several rectangular patches, this
     is up to 8 times less data, with a concomitant performance
     advantage.  Use xyz3 when writing slicing functions or coloring
     functions for slice3.

xyz3_rect:
     The xyz3 function for a regular, rectangular mesh.

xyz3_irreg:
     The xyz3 function for an irregular, unstructured mesh.

xyz3_unif:
     The xyz3 function for a uniform mesh.

to_corners3:
to_corners3(list, nj, nk)
     convert an array of cell indices in an (ni-1)-by-(NJ-1)-by-(NK-1)
     logically rectangular grid of cells into the list of
     len(LIST)-by-2-by-2-by-2 cell corner indices in the
     corresponding ni-by-NJ-by-NK array of vertices.
     Note that this computation in Yorick gives an absolute offset
     for each cell quantity in the grid. In Yorick it is legal to
     index a multidimensional array with an absolute offset. In
     Python it is not. However, an array can be flattened if
     necessary.
     Other changes from Yorick were necessitated by row-major
     order and 0-origin indices, and of course the lack of
     Yorick array facilities.