Add the so wished option to plot spheres in psxyz. #8844

joa-quim · 2025-12-21T23:17:51Z

They aren't really spheres but circles in 3D plots.

This is mostly a Claude.ai creation but a one that needed a lot of baby-sitting development.

Claude picked -SQ for this option and I honestly couldn't think of a better alternative, though -Sq is a decorated line and -SQ has nothing to do with it. Ideas for a better syntax are welcome. The following is the summary prepared by Claude


● Summary of -SQ Sphere Symbol Option

  Basic Syntax:
  -SQ<size>[c|i|p][+a<azimuth>][+e<elevation>][+f][+n]

  Description:
  The -SQ option plots a 3-D sphere symbol with diameter <size>. The sphere is rendered with a radial gradient shading from white at the light source position to the fill color (specified by -G) at the opposite side, creating a realistic 3-D appearance.

  Size Units:
  - Default unit: cm (centimeters)
  - Supported units: c (cm), i (inches), p (points)

  Modifiers:
  - +a - Set light source azimuth angle [default: 0°, from the right]
  - +e - Set light source elevation angle [default: 90°, perpendicular to viewing plane]
  - +f - Use flat/constant fill color (disables gradient shading)
  - +n - Draw outline only with no fill

  Fill and Outline:
  - Fill color: Specified by -G [default: black if not specified]
  - Outline: Controlled by -W (pen color, width, style)
  - Without -G and without +n, the sphere defaults to black fill

The +a and +e are not working correctly and I'm not even sure that we should have them. Opinions?

A low imagination example (used in the learning work) is:

echo 3 3 3 | gmt plot3d -R0/6/0/6/0/6 -JX10c -JZ4c -p135/45 -SQ2c -W0.2p,red -Ba -Bz -png lixo

They aren't really spheres but circles in 3D plots. This is mostly a Claude.ai creation but a one that needed a lot of baby-sitting development.

Esteban82

Great feature!!!

It worked on my pc.

Just some comments.

Is it needed to include this feature in psxy.c?
The simulate shading does not affect the outline, right?

joa-quim · 2025-12-22T01:19:52Z

I wondered about psxy.c too but this is a (pseudo) 3D feature so don't see the interest of it in there where we can do almost all of this with circles (except gradient).

The outline is controlled by -W so no shadings in it (and this was a BIG fight with Claude to make it do that).

Esteban82 · 2025-12-22T16:59:20Z

I wondered about psxy.c too but this is a (pseudo) 3D feature so don't see the interest of it in there where we can do almost all of this with circles (except gradient).

The outline is controlled by -W so no shadings in it (and this was a BIG fight with Claude to make it do that).

Ok. Fine by me.

About the letter (-SQ) I think that the capital P would be better. The lowercase p is for points.

joa-quim · 2025-12-22T20:07:08Z

Yes, that's a good idea.

Esteban82 · 2025-12-22T23:24:56Z

Ok. I will I modify the code.

Esteban82 · 2025-12-23T00:37:57Z

Looks fine. Do you forget to update the psxy.c file?

Esteban82 · 2025-12-23T00:38:10Z

Looks fine. But, do you forget to update the psxy.c file?

Esteban82 · 2025-12-23T00:40:17Z

src/gmt_init.c

 	char symbol_type, txt_a[GMT_LEN256] = {""}, txt_b[GMT_LEN256] = {""}, txt_c[GMT_LEN256] = {""}, txt_d[GMT_LEN256] = {""};
 	char text_cp[GMT_LEN256] = {""}, diameter[GMT_LEN32] = {""}, *c = NULL;
-	static char *allowed_symbols[2] = {"~=-+AaBbCcDdEefGgHhIiJjMmNnpqRrSsTtVvWwxy", "=-+AabCcDdEefGgHhIiJjMmNnOopqRrSsTtUuVvWwxy"};
+	static char *allowed_symbols[2] = {"~=-+AaBbCcDdEefGgHhIiJjMmNnpQqRrSsTtVvWwxy", "=-+AabCcDdEefGgHhIiJjMmNnOopQqRrSsTtUuVvWwxy"};


I am not sure, but should I replace here the Q with P?

I think so.

joa-quim · 2025-12-23T00:46:43Z

Don't do it. There are more things to change. And psxy ill not receive this option

joa-quim · 2025-12-23T00:51:54Z

@seisman can you please add the test/psxyz/sphere.sh PS to DVC?

seisman · 2025-12-25T09:04:25Z

I've added the baseline image to DVC, but currently it has a vertical axis on the left, should it be a 3D basemap instead?

github-actions · 2025-12-25T09:09:53Z

Summary of changed images

This is an auto-generated report of images that have changed on the DVC remote

Status	Path
modified	test/baseline/psxyz/

Image diff(s)

Details

Added images

Modified images

Path	Old	New

Report last updated at commit e544ec1

src/psxyz.c

seisman · 2025-12-25T09:11:21Z

src/psxyz.c

+	GMT_Usage (API, 3, "+a Set light source azimuth [0, from the right].");
+	GMT_Usage (API, 3, "+e Set light source elevation [90, perpendicular to viewing plane].");


Does +a/+e really work? I tried to plot a sphere with -Sc1c+a90+e10 and expect to see the white light near the north edge, but it gives me:

Co-authored-by: Dongdong Tian <seisman.info@gmail.com>

joa-quim · 2025-12-25T12:11:37Z

The vertical bar in the example actually reveals a bug in the implementation. Modern and Classic modes are not giving the same thing. Whilst modern mode works:

echo 3 3 3 | gmt psxyz -R0/6/0/6/0/6 -JX10c -JZ4c -p135/45 -SP2c -Gblue -W0.2p,red -Ba -Bz -png lixo

the classic somehow screw the axis and shows only that vertical bar.

echo 3 3 3 | gmt psxyz -R0/6/0/6/0/6 -JX10c -JZ4c -p135/45 -SP2c -Gblue -W0.2p,red -Ba -Bz -P > lixo.ps

joa-quim · 2025-12-25T12:14:20Z

Does +a/+e really work?

No. I mentioned that in my first post. And asked opinions if it really worth to exist as option. I still think yes but meanwhile other things came up and I left them in the state they are.

seisman · 2025-12-25T12:27:35Z

Does +a/+e really work?

No. I mentioned that in my first post. And asked opinions if it really worth to exist as option. I still think yes but meanwhile other things came up and I left them in the state they are.

I think they're useful and should be implemented.

joa-quim · 2025-12-25T16:15:15Z

I updated the test. Turns out that the vertical bar issue is independent of the changes in this PR. See #8845

joa-quim · 2025-12-25T16:28:26Z

I tried to plot a sphere with -SP1c+a90+e10 and expect to see the white light near the north edge,

Is it? I'm not even so sure on what to expect. I would say that the expectation would be to have the bright point at 45º lat when looking, for example from -p180/45, and that because we also have to take into account the elevation of the perspective view. This means that illumination should only be allowed to come from the illuminated hemisphere (plus some tolerance for the same reason that we still have day light even after sunset).

Oddly the PS seem right but conversion to raster flips spheres upside-down.

joa-quim · 2025-12-25T20:39:29Z

Fixed (I think) the +e and +a options. But a very strange thing decided joining the party. The psconvert part flips the spheres upsisde-down. This is flipped but the PS is correct (and the PDF too).

echo 3 3 3 | build\src\gmt psxyz -R0/6/0/6/0/6 -JX10c -JZ4c -p180/45 -SP4c+e90+a180 -Gblue -W0.2p,red -Ba -Bza -png lixo

Esteban82 · 2025-12-26T01:26:25Z

echo 3 3 3 | build\src\gmt psxyz -R0/6/0/6/0/6 -JX10c -JZ4c -p180/45 -SP4c+e90+a180 -Gblue -W0.2p,red -Ba -Bza -png lixo

I got the same result.

joa-quim · 2025-12-26T11:45:03Z

I will merge this PR as is since this flipping issue seems even a Ghostscript thing. Will try to investigate it later.

Add the so wished option to plot spheres in psxyz.

7c1d2e8

They aren't really spheres but circles in 3D plots. This is mostly a Claude.ai creation but a one that needed a lot of baby-sitting development.

joa-quim requested a review from a team December 21, 2025 23:18

Esteban82 approved these changes Dec 22, 2025

View reviewed changes

Esteban82 added enhancement Improving an existing feature add-changelog Add PR to the changelog labels Dec 22, 2025

Change -SQ to -SP for sphere plots. Add a sphere.sh test

ed7409f

Esteban82 reviewed Dec 23, 2025

View reviewed changes

joa-quim and others added 2 commits December 23, 2025 01:31

Merge branch 'master' into PS-spheres

757c470

Add the baseline image sphere.ps to dvc

d64744e

seisman reviewed Dec 25, 2025

View reviewed changes

Update src/psxyz.c

ed1f289

Co-authored-by: Dongdong Tian <seisman.info@gmail.com>

Fixes and updates to this test.

3cc5a9f

Make the -SP+e & +a work better.

e544ec1

Oddly the PS seem right but conversion to raster flips spheres upside-down.

joa-quim merged commit 65f2d39 into master Dec 26, 2025
10 of 14 checks passed

joa-quim deleted the PS-spheres branch December 26, 2025 11:45

		GMT_Usage (API, 3, "+a Set light source azimuth [0, from the right].");
		GMT_Usage (API, 3, "+e Set light source elevation [90, perpendicular to viewing plane].");

Add the so wished option to plot spheres in psxyz. #8844

Add the so wished option to plot spheres in psxyz. #8844

Conversation

joa-quim commented Dec 21, 2025

Uh oh!

Esteban82 left a comment

Choose a reason for hiding this comment

Uh oh!

joa-quim commented Dec 22, 2025

Uh oh!

Esteban82 commented Dec 22, 2025

Uh oh!

joa-quim commented Dec 22, 2025

Uh oh!

Esteban82 commented Dec 22, 2025

Uh oh!

Esteban82 commented Dec 23, 2025

Uh oh!

Esteban82 commented Dec 23, 2025

Uh oh!

Esteban82 Dec 23, 2025

Choose a reason for hiding this comment

Uh oh!

seisman Dec 25, 2025

Choose a reason for hiding this comment

Uh oh!

joa-quim commented Dec 23, 2025

Uh oh!

joa-quim commented Dec 23, 2025

Uh oh!

seisman commented Dec 25, 2025

Uh oh!

github-actions bot commented Dec 25, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary of changed images

Image diff(s)

Added images

Modified images

Uh oh!

Uh oh!

seisman Dec 25, 2025

Choose a reason for hiding this comment

Uh oh!

joa-quim commented Dec 25, 2025

Uh oh!

joa-quim commented Dec 25, 2025

Uh oh!

seisman commented Dec 25, 2025

Uh oh!

joa-quim commented Dec 25, 2025

Uh oh!

joa-quim commented Dec 25, 2025

Uh oh!

joa-quim commented Dec 25, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Esteban82 commented Dec 26, 2025

Uh oh!

joa-quim commented Dec 26, 2025

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

github-actions bot commented Dec 25, 2025 •

edited

Loading

joa-quim commented Dec 25, 2025 •

edited

Loading