;+
; Procedure: plotxyz
;
; Purpose: Generates an isotropic spectrographic plot. It takes one
; 2-d array(Z) and plots it using the values in 2 1-d arrays(X,Y) to
; scale the data in Z. The X and Y axes can be any kind of data,
; most specifically the X axis need not be time. By default the
; plots, are scaled isotropically. Meaning that a unit on the x axis
; will have the same length on the screen as a unit on the z axis.
;
; The plots can be interleaved with plotxy/tplotxy plots in the same
; panel/window. Calling plotxyz with no arguments will redraw the
; entire window(including plotxy/tplotxy plots)
;
; The most significant restiction to this function is that it will
; clip any negative(or 0) data(in X,Y or Z) if you select a
; logarithmic axis.
;
; If one scaling axis is set to logarithmic and the other is not,
; but the plot is set to isotropic, 1 power unit on the logarithmic
; axis will take up the same space on screen as one normal unit on
; the normal axis. Despite the capability to perform these mixed
; x,y log plots, this is not recommended...but feel free to experiment.
;
;
;************************************
; Detailed explanation of Plot windows and panels:
; /addpanel,/noisotropic and multi=
;
; To put multiple panels in a window first call
; plotxyz with the multi keyword. It will either
; plot in which ever window is your current one, or
; create a new one if no window exists or if you
; request the use of a nonexistent window.
;
; During this first call you may want to specify things
; like wtitle,xsize,ysize,window...in addition to your normal
; plotting options, setting these windowing options will
; interfere with the creation of postscript
;
; multi specifies the plot window panel layout.
; So if you set multi='3,2' you will get 6 plots
; in your window with a layout like:
; -------
; |x x x|
; |x x x|
; -------
;
; Each panel will have dimensions x number of pixels = 1/3 *
; xsize of window and y number of pixels = 1/2 * ysize of window.
;
; Your first call should also specify the layout of your
; first panel. To add to that panel use the /overplot keyword.
;
; When you use the /add keyword the program will move on to
; the next panel within the plot window and you should add
; options to specify the layout of that panel.
;
; If you set the xmargin or ymargin keyword the margin will be
; relative to the overall size of that panel. When using the
; not using the noisotropic keyword the procedure will make
; each axis vary over the same range AND make the
; largest possible square window given the size of the panel
; and the sizes of the margins you have provided, if possible.
; In some cases when ranges are set explictly the plot must
; be rectangular.
;
; An entire plot window must be filled in sequence, if you move
; on to a new window you will not be able to go back to the
; previous window without restarting.
;
; If you call plotxyz with no arguments it will redraw the
; entire window including all panels and overplots. If you
; resize the window before calling with new arguments it
; will redraw the isotropic panels as the largest possible
; squares. This comes at a cost of storing copies of the
; commands and data you made in memory. If you need to save
; memory you can call the function with the /memsave argument,
; but then redraws will be done using hardware and window resizes
; can distort isotropic plots.
;
; NOTE TO PROGRAMMERS:
; Information about plotting for plotxyz is stored in
; the global variable !TPLOTXY, this includes
; information about the layout of the plot window
; which panel it is currently working on, and the
; sequence of commands used to generate current plot window
; so that it can regenerate the plotwindow when called
; with no arguments. This variable also stores information
; used by the plotxy function so line plots
; can be interleaved with xyz spectrographic plots.
;
; Required Inputs:
; x: 1-d array specifying the scaling/spacing of the x axis
; This array must have the same number of elements as the
; 1st dimension of Z. One can think of the x coordinates of
; the centers of the Z array data
;
;
; y: 1-d array specifying the scaling/spacing of the y axis
; This array must have the same number of elements as the
; 2nd dimension of Z. One can think of the y coordinates of
; the centers of the Z array data
;
; z: 2-d array specifying the intensity of each element, or the
; height of the z-axis. This will be represented by color
; in the 2-d plot this procedure generates.
;
; Optional keywords:
; interpolate: set this argument if you want the data to
; be interpolated between z data. This will give the
; appearance of smooth gradations, although this may not
; exist in the data. If your Z data has blanks(NaNs),
; interpolation can give inaccurate results near the blanks.
;
; noisotropic: set this argument if you don't want
; the plot to be isotropic. If this is set the plot
; will fill the entire space available to it,
; regardless of data scaling.
;
; xlog,ylog,zlog: set any of these to create logarithmic
; scaling on the appropriate axis. It is recommended,
; but not required that if you set xlog on an isotropic
; plot you also set ylog.
;
; multi: as explained above set this to a string
; indicating the desired number of columns and rows
; in your window. This string can must contain 2 numbers
; delimited by any common delimiter character or space.
; the numbers may optionally be followed by an r to
; indicate a reversal in the direction that the plots
; will be added to the window.
;
; addpanel: set this keyword to make the procedure
; move on to the next plot in window, if you have
; previously set multi. If this is not set, it will
; generate a new plot in a clear window, if this is
; set and there are no more available spaces for
; plots an error will be generated.
;
; memsave: To allow replotting of the data when the
; procedure is called with no arguments, the copies of
; the data are stored in memory. If memsave is set
; these copies will not be saved and you will be unable
; to replot with a 0 argument call.
;
; xmargin,ymargin: Set these keyword to a 2 element
; array to set extra space around the plot. Margins are
; measured proportionally(from 0.0 to 1.0) and are
; separate for each plot(not global to the entire window).
; The arrays store the [left,right] xmargin or
; [bottom,top] ymargin. Default xmargin for xyz plots is
; [.15,.15] and ymargin is [.1,.075].
;
; xrange,yrange,zrange: Set these keywords to a 2 element
; array to control the range of values to be displayed
; for each axis.
;
; title: set this to a string indicating the title at
; the top of the plot
;
; xtitle,ytitle,ztitle: set this to a string indicating
; the title of the appropriate axis.
;
; charsize: set this to a number to scale the character
; size of writing on the plot. 1.0 is the default,
; less than 1.0 decreases charsize, greater increases.
;
; WARNING setting window, xsize, ysize or wtitle will
; interfere with the creation of postscript
;
; window: specify the window in which the plots should be
; made. The default is the current window. If the window
; number does not exist one will be made
;
; xsize,ysize: Specify the number of pixels of the window you are
; plotting in. This can be done ahead by the user if they
; like, or just by stretching the window.
;
; wtitle: Specify the title for the bar at the top of the
; window as a string.
;
; noticks: set this if you do not want ticks on the plot
; (mutually exclusive with grid)
;
; grid: set this if you want a grid on your plot
; (mutually exclusive with noticks)
; Use xticks and yticks to manipulate the
; spacing of your grid
;
; markends: This keyword is deprecated. You can use
; all the normal options for plot to manipulate the
; position of the ticks on the axes.
;
; xtick_get,ytick_get: These behave exactly as the plot
; command versions, but they had to be identified explictly
; to ensure they would be passed through correctly.
;
; zticks: this acts like the normal x,y ticks option in
; idl plots. Set it to some number greater than 1 to set
; the number of tick marks of the z axis. It is available
; because draw color scale will sometimes supress all the
; tick marks on the z axis.
;
; ps_resolution: set the resolution, if you are using
; postscript (default is 150 pts/cm)
;
; You can also use many normal plot options.
;
; NOTES:
; All NaN's & INFs in the x and y axes will be removed from the data. All
; NaN's in the z data will be replaced by the minimum value.
;
; bin2d is VERY useful for preparing data for use in this routine
;
; Be very careful when manually setting the ticks. While some options like [xy]ticks
; are quite safe, others can inadvertently produce inaccurate labels as idl will sometimes
; make assumptions about positioning of axes, by rounding off. If you plan on using [xy]tickv,
; or [xy]style be careful to verify that the axis labeling is working correctly. This can best be done by
; testing on a data set where the axes are irregularly spaced and where some of the values at the axes are
; irrational.
;
;
; SEE ALSO:
; plotxy,tplotxy,thm_crib_tplotxy,thm_crib_plotxy,thm_crib_plotxyz,bin2d,
; plotxylib,plotxyvec
;
;
; $LastChangedBy: pcruce $
; $LastChangedDate: 2008-01-16 16:54:40 -0800 (Wed, 16 Jan 2008) $
; $LastChangedRevision: 2283 $
; $URL: svn+ssh://thmsvn@ambrosia.ssl.berkeley.edu/repos/ssl_general/trunk/tplot/tplotxy.pro $
;-
;helper for the data processing routine
;code to clip the x and y axes
pro pxyz_dim_clip,zp,ap,r1,r2,an,bn,znans
;the cases below determine if the data intersects with
;the xrange provided and decides how to fill the space
;between the data and the ranges
id = where(r1 le ap and r2 ge ap)
;this case handles one element x
if an eq 1 or id[0] eq -1 then begin
ap = [r1,r2]
an = 2
if id[0] ne -1 then begin
zp = [zp,zp]
znans = [znans,znans]
endif else begin
zp = dblarr(2,bn)
znans = dblarr(2,bn)
znans[*] = 1
endelse
endif else begin
;if the lower range is outside the data, pad the edge with black
if r1 lt ap[0] then begin
;decide where to place the boundary point
;this decision is essentially aesthetic
;it should determine how far the data will
;extend until it reaches the filler color
if abs(ap[1]-ap[0]) lt abs(ap[0]-r1) then begin
ap = [r1,ap[0]-abs(ap[1]-ap[0]),ap]
zp = [dblarr(2,bn),zp]
znans_fill = dblarr(2,bn)
znans_fill[*] = 1
znans = [znans_fill,znans]
an+=2
endif else begin
ap = [r1,ap]
zp = [dblarr(1,bn),zp]
znans_fill = dblarr(1,bn)
znans_fill[*] = 1
znans = [znans_fill,znans]
an+=1
endelse
;if the lower range is within the data
;clip to the data edge and interpolate to the range
;marker
endif else if r1 gt ap[0] then begin
idx = where(ap gt r1)
zp_edge = zp[(idx[0]-1):idx[0],*]
znans_edge = znans[(idx[0]-1):idx[0],*]
interp_loc = r1-ap[idx[0]-1]/(ap[idx[0]]-ap[idx[0]-1])
zp_interp = interpolate(zp_edge,interp_loc,dindgen(bn),/grid)
znans_interp = interpolate(znans_edge,interp_loc,dindgen(bn),/grid)
zp = [zp_interp,zp[idx,*]]
znans = [znans_interp,znans[idx,*]]
ap = [r1,ap[idx]]
an = n_elements(ap)
endif
;if the upper range is outside the
;data then pad the edge with black
if r2 gt ap[an-1] then begin
;decide where to place the boundary point
;this decision is essentially aesthetic
;it should determine how far the data will
;extend until it reaches the filler color
if abs(ap[an-1]-ap[an-2]) lt abs(r2 - ap[an-1]) then begin
ap = [ap,ap[an-1]+abs(ap[an-1]-ap[an-2]),r2]
zp = [zp,dblarr(2,bn)]
znans_fill = dblarr(2,bn)
znans_fill[*] = 1
znans = [znans,znans_fill]
an+=2
endif else begin
ap = [ap,r2]
zp = [zp,dblarr(1,bn)]
znans_fill = dblarr(1,bn)
znans_fill[*] = 1
znans = [znans,znans_fill]
an+=1
endelse
;if the lower range is within the data
;clip to the data edge and interpolate to the range
;marker
endif else if r2 lt ap[an-1] then begin
idx = where(ap lt r2)
n = n_elements(idx)
zp_edge = zp[idx[n-1]:(idx[n-1]+1),*]
znans_edge = znans[idx[n-1]:(idx[n-1]+1),*]
interp_loc = r2-ap[idx[n-1]]/(ap[idx[n-1]+1]-ap[idx[n-1]])
zp_interp = interpolate(zp_edge,interp_loc,dindgen(bn),/grid)
znans_interp = interpolate(znans_edge,interp_loc,dindgen(bn),/grid)
zp = [zp[idx,*],zp_interp]
znans = [znans[idx,*],znans_interp]
ap = [ap[idx],r2]
an = n_elements(ap)
endif
endelse
end
;do log processing for the x and y dims with this function
;for negative values, it replaces the values with -abs(log(value))
;all 0s are removed
pro pxyz_dim_log,zp,ap,range,znans
zdims = size(zp,/dimensions)
idx = where(ap gt 0)
if idx[0] eq -1 then begin
message,'entire logarithmic dimension is <= 0'
endif else begin
ap = ap[idx]
;may need to reset the range because low end got clipped by log
if range[0] lt range[1] then begin
range[0] = min(ap)
endif else begin
range[1] = min(ap)
endelse
ap = alog(ap)
zp = zp[idx,*]
znans = znans[idx,*]
endelse
end
pro pxyz_process_data,xp,yp,zp,xrange,yrange,zrange,xlog,ylog,zlog,znans
zdims = size(zp,/dimensions)
xn = n_elements(xp)
yn = n_elements(yp)
if size(xp,/n_dim) ne 1 || size(yp,/n_dim) ne 1 then begin
message,'x and y must be 1 dimensional'
endif
if n_elements(zdims) ne 2 then begin
message,'z must be 2 dimensional'
endif
if zdims[0] ne xn then begin
message,'dim 1 of z must equal number of elements in x'
endif
if zdims[1] ne yn then begin
message,'dim 2 of z must equal number of elements in y'
endif
;make rules to allow single elements...
;process infinities and nans
fn_x = where(finite(xp))
if fn_x[0] eq -1 then begin
message,'no finite values in x'
endif
fn_y = where(finite(yp))
if fn_y[0] eq -1 then begin
message,'no finite values in y'
endif
;jury rigging some code to deal with NaNs
;since original code just code rid of them
;NaN locations with be tracked in separate data
;structure, all modifications to the zp structure will
;also occur to the NaN structure
;(Infinities are treated like NaNs, but if we wanted to
;treat them separately the code below would be where to
;do it)
xp = double(xp[fn_x])
yp = double(yp[fn_y])
zt = zp[fn_x,*]
zp = double(zt[*,fn_y])
fn_znf = where(~finite(zp))
fn_zf = where(finite(zp))
;all locations where there are nans end up 1s others are 0
znans = dblarr(size(zp,/dimensions))
if fn_zf[0] eq -1 then begin
zp[*] = 0
znans[*] = 1
endif else begin
if fn_znf[0] ne -1 then begin
zp[fn_znf] = min(zp[fn_zf])
znans[fn_znf] = 1
endif
endelse
;sort data
xs = sort(xp)
ys = sort(yp)
xp = xp[xs]
yp = yp[ys]
zp = zp[xs,*]
zp = zp[*,ys]
znans = znans[xs,*]
znans = znans[*,ys]
;set ranges(by either lengthening or shortening)
if keyword_set(xrange) then begin
if n_elements(xrange) ne 2 then begin
message,'xrange must have exactly 2 elements if set'
endif
;the calculations below are made simpler
;if we perform them while the data is
;monotic and ascending then reverse
;according to ranges later
if xrange[0] gt xrange[1] then begin
xr1 = xrange[1]
xr2 = xrange[0]
endif else begin
xr1 = xrange[0]
xr2 = xrange[1]
endelse
pxyz_dim_clip,zp,xp,xr1,xr2,xn,yn,znans
if xrange[0] gt xrange[1] then begin
xp = reverse(xp)
zp = reverse(zp,1)
znans = reverse(znans,1)
endif
endif else begin
xrange = [min(xp),max(xp)]
endelse
;set ranges(by either lengthening or shortening)
if keyword_set(yrange) then begin
if n_elements(yrange) ne 2 then begin
message,'yrange must have exactly 2 elements if set'
endif
;the calculations below are made simpler
;if we perform them while the data is
;monotic and ascending then reverse
;according to ranges later
if yrange[0] gt yrange[1] then begin
yr1 = yrange[1]
yr2 = yrange[0]
endif else begin
yr1 = yrange[0]
yr2 = yrange[1]
endelse
;to simplfy concatenations along the y dimension
;calculations will be done with a transposed z
;this means the clipping code for
;the x and y axes is now symmetric
zp = transpose(zp)
znans = transpose(znans)
pxyz_dim_clip,zp,yp,yr1,yr2,yn,xn,znans
zp = transpose(zp)
znans = transpose(znans)
if yrange[0] gt yrange[1] then begin
yp = reverse(yp)
zp = reverse(zp,2)
znans = reverse(znans,2)
endif
endif else begin
yrange = [min(yp),max(yp)]
endelse
;handle zrange...
;clipping and reversal
;should clipped z data be filled with???
;filler or should it be max = to max???
;and min???
if keyword_set(zrange) then begin
if n_elements(zrange) ne 2 then begin
message,'zrange must have exactly 2 elements if set'
endif
;the calculations below are made simpler
;if we perform them while the data is
;monotic and ascending then reverse
;according to ranges later
if zrange[0] gt zrange[1] then begin
zr1 = zrange[1]
zr2 = zrange[0]
endif else begin
zr1 = zrange[0]
zr2 = zrange[1]
endelse
idx = where(zp lt zr1)
if idx[0] ne -1 then begin
zp[idx] = zr1
endif
idx = where(zp gt zr2)
if idx[0] ne -1 then begin
zp[idx] = zr2
endif
if zrange[0] gt zrange[1] then begin
maxz = max(zp,min=minz)
zp = minz + maxz - zp
endif
endif else begin
zrange = [min(zp),max(zp)]
endelse
;perform logarithms
if keyword_set(xlog) then begin
pxyz_dim_log,zp,xp,xrange,znans
endif
if keyword_set(ylog) then begin
zp = transpose(zp)
znans = transpose(znans)
pxyz_dim_log,zp,yp,yrange,znans
zp = transpose(zp)
znans = transpose(znans)
endif
if keyword_set(zlog) then begin
idx = where(zp le 0)
idx2 = where(zp gt 0)
if idx2[0] eq -1 then begin
message,'all values in logarthimic z dimension are <= 0'
endif
if idx[0] ne -1 then begin
message,/continue,'some values are <= 0 and are being removed'
mn = min(zp[idx2])
zp[idx] = mn
znans[idx] = 1
if zrange[0] lt zrange[1] then begin
zrange[0] = mn
endif else begin
zrange[1] = mn
endelse
endif
zp = alog(zp)
endif
end
;this function tricks the interpolate function in idl
;to generate a grid of points for output with tv
function pxyz_grid,x,y,z,pixx,pixy,interp,znans
compile_opt hidden,idl2
xi = x
yi = y
xn = n_elements(xi)
yn = n_elements(yi)
if pixx lt 1. then begin
message,'x dimension soooo small that plot is less than 1 pixel wide'
endif
if pixy lt 1. then begin
message,'y dimension soooo small that plot is less than 1 pixel high'
endif
;the proportional positions on the x-axis
xi = (xi - xi[0])/(xi[xn-1]-xi[0])
;the positions to interpolate to on the x axis
xpos = interpol(dindgen(xn),xi,dindgen(pixx)/pixx)
;the proportional positions on the y-axis
yi = (yi - yi[0])/(yi[yn-1]-yi[0])
ypos = interpol(dindgen(yn),yi,dindgen(pixy)/pixy)
if ~keyword_set(interp) then begin
xpos = round(xpos)
ypos = round(ypos)
endif
z = bytscl(z,top=247)+7
outz = floor(interpolate(z,xpos,ypos,/grid))
outnulls = interpolate(znans,round(xpos),round(ypos),/grid)
idx = where(outnulls eq 1.0)
if idx[0] ne -1 then begin
outz[idx] = !P.background
endif
;grid the points and scale them into the range of color indices
;return, floor(interpolate(z,xpos,ypos,/grid))
return,outz
end
;data must be processed as follows
;dimensions are already clipped and ordered
;pos of panel calculated
;all inputs are assigned
;values are logarithm'd(if requested)
;nans are reassigned to a default
pro pxyz_make_spec,x,y,z,pos,interp,znans,ps_resolution
compile_opt hidden,idl2
x_p_sz = pos[2]-pos[0]
y_p_sz = pos[3]-pos[1]
if(!D.name eq 'PS') then begin
ps_sz_x_cm = !D.x_size/!D.X_PX_CM
ps_sz_y_cm = !D.y_size/!D.Y_PX_CM
x_p_sz_cm = x_p_sz* ps_sz_x_cm
y_p_sz_cm = y_p_sz* ps_sz_y_cm
x_p_off_cm = pos[0]*ps_sz_x_cm
y_p_off_cm = pos[1]*ps_sz_y_cm
image = pxyz_grid(x,y,z,x_p_sz_cm*ps_resolution,y_p_sz_cm*ps_resolution,interp,znans)
tv,image,x_p_off_cm,y_p_off_cm,/CENTIMETERS,xsize=x_p_sz_cm,ysize=y_p_sz_cm
endif else begin
pxx = x_p_sz * !D.x_size
pxy = y_p_sz * !D.y_size
image = pxyz_grid(x,y,z,pxx,pxy,interp,znans)
tv,image,pos[0],pos[1],/normal
endelse
;isotropic plot will often leave blank space in the plot area
end
;makes the labels for the plot since tv doesn't
pro pxyz_make_labels,xrange,yrange,zrange,pos,xlog,$
ylog,zlog,charsize,xtitle,ytitle,ztitle,title,noticks,$
grid,markends,zticks,xtick_get,ytick_get,xstyle=xstyle,$
ystyle=ystyle,_extra=_extra
compile_opt hidden,idl2
x = dindgen(100)/99 * (xrange[1]-xrange[0]) + yrange[0]
y = dindgen(100)/99 * (yrange[1]-yrange[0]) + yrange[0]
if keyword_set(noticks) && keyword_set(grid) then begin
message,'grid and noticks are mutually exclusive'
endif
if keyword_set(noticks) then begin
ticklen=0
endif
if keyword_set(grid) then begin
ticklen = 1.0
endif
if ~keyword_set(xstyle) then begin
xstyle=1
endif
if ~keyword_set(ystyle) then begin
ystyle=1
endif
;This option ay produce unreliable results so it has been disabled
;The user should be able to perform this operation manually if needed
;if keyword_set(markends) then begin
if keyword_set(markends) then begin
dprint,'Option: markends has been deprecated, you can control the placement of ticks using all the standard commands from plot.'
endif
;keyword xstyle,ystyle=1.0 is needed to control exact positioning of the axes.
;Otherwise the spectrographic image and the ploted axes may not coincide
plot,x,y,xlog=xlog,ylog=ylog,ticklen=ticklen,/normal,/noclip,position=pos,xrange=xrange,yrange=yrange,xstyle=xstyle,ystyle=ystyle,$
charsize=charsize,/noerase,/nodata,xtitle=xtitle,ytitle=ytitle,title=title,xtick_get=xtick_get,ytick_get=ytick_get,_extra=_extra
p = pos
p[0] = pos[2] + (pos[2]-pos[0])*.05
p[2] = pos[2] + (pos[2]-pos[0])*.1
draw_color_scale,range=zrange,log=zlog,charsize=charsize,position=p,title=ztitle,yticks=zticks,brange=[7,254]
return
end
pro plotxyz,x,y,z,interpolate=interpolate,noisotropic=noisotropic,xlog=xlog,ylog=ylog,zlog=zlog,addpanel=addpanel,$
multi=multi,memsave=memsave,xmargin=xmargin,ymargin=ymargin,xrange=xrange,yrange=yrange,zrange=zrange,$
title=title,xtitle=xtitle,ytitle=ytitle,ztitle=ztitle,charsize=charsize,window=window,xsize=xsize,$
ysize=ysize,wtitle=wtitle,replot=replot,noticks=noticks,grid=grid,markends=markends,zticks=zticks,$
ps_resolution=ps_resolution,xtick_get=xtick_get,ytick_get=ytick_get,_extra=_extra
compile_opt idl2
plotxylib
if undefined(x) and undefined(y) and undefined(z) then begin
pxy_replot
return
endif
pxy_set_window,overplot,addpanel,replot,window,xsize,ysize,wtitle,multi,noisotropic,isotropic=isotropic
if ~keyword_set(ps_resolution) then begin
ps_resolution = 150
endif
;the process data function
;mutates these variables
;so these are stored for use
;in the state push...
;and to prevent mutation
;of variables in parent env
if undefined(x) || undefined(y) || undefined(z) then begin
message,'x y & z must be set'
endif
xs = double(x)
ys = double(y)
zs = double(z)
if keyword_set(xrange) then begin
xr = xrange
endif
if keyword_set(yrange) then begin
yr = yrange
endif
if keyword_set(zrange) then begin
zr = zrange
endif
;clear
if ~keyword_set(addpanel) then begin
plot,[1],[1],/nodata,color=!P.background
endif
pxyz_process_data,xs,ys,zs,xr,yr,zr,xlog,ylog,zlog,znans
pos = pxy_get_pos([min(xs),max(xs)],[min(ys),max(ys)],isotropic,xmargin,ymargin)
;isotropic calculation
;generate plot
pxyz_make_spec,xs,ys,zs,pos,interpolate,znans,ps_resolution
pxyz_make_labels,xr,yr,zr,pos,xlog,ylog,zlog,charsize,xtitle,ytitle,ztitle,title,noticks,grid,markends,zticks,xtick_get,ytick_get,_extra=_extra
;state push is called last
;to ensure that only successful
;plot operations are stored
if ~keyword_set(replot) then begin
pxy_push_state,'plotxyz',{x:x,y:y,z:z},interpolate=interpolate,noisotropic=noisotropic,xlog=xlog,ylog=ylog,zlog=zlog,$
addpanel=addpanel,multi=multi,memsave=memsave,xmargin=xmargin,ymargin=ymargin,xrange=xrange,$
yrange=yrange,zrange=zrange,title=title,xtitle=xtitle,ytitle=ytitle,ztitle=ztitle,$
charsize=charsize,noticks=noticks,grid=grid,markends=markends,zticks=zticks,ps_resolution=ps_resolution,$
xtick_get=xtick_get,ytick_get=ytick_get,_extra=_extra
endif
end