plot(x, y, G=value, C=cmap) gives strange behavior for individual points

[kmaterna]

Description of the problem

In plot(), using G=value and C=cmap to interpret a color from a CMAP for a single value doesn't work in the same way that a list/array of values does. Interpreting a single value from a color map would be useful for plotting a bunch polygons with different colors (ex: for coulomb stress changes on fault segments).

• For plotting arrays of points, G returns an array of colors extracted from the cmap, as it should
• For plotting single polygons, G returns foreground or background color. wrong?
• For plotting single points, G returns 0. wrong?
  In the single-element case, plot() requires G=str, which is also counter-intuitive.

Either way, the result makes it difficult to plot custom-colored polygons with a colormap
(usually I would just extract a color for each polygon and plot each one separately, calling plot() in a loop). Is there another way to plot polygons with cmap colors?

In the example code, I'm plotting some dummy data, once as an array of points, and once in a for-loop with points and polygons in individual plot() calls. The cmap behavior is quite different.

Full code that generated the error

import pygmt
# Make CPT and Fake Data
pygmt.makecpt(C="jet",T="-1.0/1.0/0.05",H="mycpt.cpt");
xdata = [-123, -122.5, -122.0, -121.5]; 
ydata = [36.5, 37, 37.5, 38]; 
zdata = [-0.7, -0.25, 0.25, 0.7];

# Batch Calling : works
fig = pygmt.Figure()
title="+t\"Batch to CMAP - Works for Points.\""; 
fig.coast(region=[-124, -120, 35, 39], projection="M7i", B=[title,"1.0"],shorelines="1.0p,black"); 
fig.colorbar(D="jBr+w3.5i/0.2i+o5.5c/1.5c+h",C="mycpt.cpt",I="0.8",G="-1.0/1.0",B=["x"+str(0.2),"y+L\"KPa\""]); 
fig.plot(x=xdata, y=ydata, color=zdata, style="c1.8c", pen="thick,black", C="mycpt.cpt"); 
fig.savefig("Working.png");

# Single Calls Only : doesn't do the right thing
fig = pygmt.Figure()
title="+t\"Single Calls to CMAP - Odd.\""; 
fig.coast(region=[-124, -120, 35, 39],projection="M7i",shorelines="1.0p,black",B=[title,"1.0"]); 
fig.colorbar(D="jBr+w3.5i/0.2i+o5.5c/1.5c+h",C="mycpt.cpt",I="0.8",G="-1.0/1.0",B=["x"+str(0.2),"y+L\"KPa\""]); 
for i in range(len(xdata)): # just looping through the data
	fig.plot(x=xdata[i], y=ydata[i], G=str(zdata[i]), style="c1.8c", pen="thick,black", C="mycpt.cpt"); # plotting a point with associated cmap data
	xdata_long = [xdata[i]+0.2, xdata[i]+0.3, xdata[i]+0.3, xdata[i]+0.2, xdata[i]+0.2]; # building a polygon
	ydata_long = [ydata[i], ydata[i], ydata[i]+0.1, ydata[i]+0.1, ydata[i]]; # building a polygon
	fig.plot(x=xdata_long, y=ydata_long, G=str(zdata[i]), pen="thick,black", C="mycpt.cpt"); # plotting the polygon with associated cmap data
fig.savefig("Not_Working.png");

System information
Mac OSX Mojave, miniconda3 package manager, python3.6. I installed pygmt in a brand new python environment today.

[welcome]

👋 Thanks for opening your first issue here! Please make sure you filled out the template with as much detail as possible. You might also want to take a look at our contributing guidelines and code of conduct.

[weiji14]

Right, bear with me for a little bit while I get my head around this. I think this is similar to the problem we had in #376.

To colour your points in the for-loop, the G argument needs to be given a list instead of a str (which is inconsistent with the documentation). I.e. G=[zdata[i]].

Edit: To colour the polygon, you can set the Z argument to something like Z="f0.7" which would base the fill colour of the polygon on your provided cpt. I've updated the code below.

# Single Calls Only : doesn't do the right thing
fig = pygmt.Figure()
title="+t\"Single Calls to CMAP - Odd.\""; 
fig.coast(region=[-124, -120, 35, 39],projection="M7i",shorelines="1.0p,black",B=[title,"1.0"]); 
fig.colorbar(D="jBr+w3.5i/0.2i+o5.5c/1.5c+h",C="mycpt.cpt",I="0.8",G="-1.0/1.0",B=["x"+str(0.2),"y+L\"KPa\""]); 
for i in range(len(xdata)): # just looping through the data
	fig.plot(x=xdata[i], y=ydata[i], G=[zdata[i]], style="c1.8c", pen="thick,black", C="mycpt.cpt"); # plotting a point with associated cmap data
	xdata_long = [xdata[i]+0.2, xdata[i]+0.3, xdata[i]+0.3, xdata[i]+0.2, xdata[i]+0.2]; # building a polygon
	ydata_long = [ydata[i], ydata[i], ydata[i]+0.1, ydata[i]+0.1, ydata[i]]; # building a polygon
	fig.plot(x=xdata_long, y=ydata_long, pen="thick,black", C="mycpt.cpt", Z="f" + str(zdata[i])); # plotting the polygon with associated cmap data
fig.savefig("Somewhat_Working.png");

produces:

I'll need to work out how to colour a filled polygon, let me go and have lunch first :) Still, I think we might need to extend the fix in #376 to include G/color as well, or change the documentation to make it clear, or do both.

[kmaterna]

Thank you so much for your quick reply- this is exactly it! 🎉🎉🎉
I guess this is like the -Zval in GMT, but I would have had trouble figuring this one out on my own. Thanks!!

[weiji14]

Yes it does get a bit confusing, but glad it helped. The -Z syntax seems to be new in GMT 6.0.0, and it might change in GMT 6.1.0 (see GenericMappingTools/gmt/issues/3131 for context) to make colouring the fill and outline more intuitive.

I'll leave this issue open for now until we clear up the plot documentation.

[weiji14]

Documentation for plot has been revised significantly in PyGMT v0.3.0 (see https://www.pygmt.org/v0.3.0/api/generated/pygmt.Figure.plot.html) which uses GMT 6.1.1, and long alias zvalue (Z) was added in #666. It might still be a bit confusing getting to know how color (-G) and zvalue (-Z) works, but I'll close this issue now and we can open up a fresh issue later (if there's demand) to add a gallery example for using -G/-Z in plot to colour individual points.