Volume - new gl3d trace by archmoj · Pull Request #3488 · plotly/plotly.js

archmoj · 2019-01-28T04:14:15Z

Supersedes #3479 initial prototype.
New trace type for volume data visualisation.
Here is an interactive demo

Note: now opacityscale attribute options could be applied to emphasize on maximum or minimum values as well as both extremes.

archmoj · 2019-01-28T04:27:09Z

    "gl-line3d": "^1.1.8",
    "gl-mat4": "^1.2.0",
-    "gl-mesh3d": "^2.0.7",
+    "gl-mesh3d": "git://github.com/gl-vis/gl-mesh3d.git#a361fdeae2fcaeed52efe05906f212e4b7c7bafa",


gl-vis/gl-mesh3d#19

jackparmer · 2019-01-28T04:32:00Z

Interesting view on isosurface vs volume plots by a lab at Lawrence Livermore:

https://wci.llnl.gov/simulation/computer-codes/visit/screenshots

etpinard · 2019-01-28T15:28:21Z

+        var height = this.data._Ys.length;
+        var depth = this.data._Zs.length;
+
+        var i = findNearestOnAxis(x, this.data._Xs).id;


... and "nearest" here means nearest on the "top" slice, correct?

... which I think is ok for the first iteration, but could be potentially improved in the future.

When possible, would you mind hovering on gl3d_volume4d_airflow in your browser?

jackparmer · 2019-02-05T21:35:53Z

@archmoj do you have any more screenshots of the state of the art of this trace type so far?

etpinard · 2019-04-02T16:08:59Z

+    var cond =
+    !(
        mockName === 'font-wishlist' ||
+        mockName.indexOf('gl3d_volume_') !== -1 ||


Hmm. That's less than ideal. Do you have an idea why we can't pixel-compare volume mocks?

Moreover, have you tried generating the volume baseline using orca?

I think there is one mock that caused a problem. Now with @antoinerg parallel PR in place I could fix this.
Second question: No I have not tried volume with orca.

Second question: No I have not tried volume with orca.

Cool, could you give it a try?

See

plotly.js/tasks/noci_test.sh

Lines 31 to 38 in c657348

test_image () {

$root/../orca/bin/orca.js graph \

$root/test/image/mocks/mapbox_* \

--plotly $root/build/plotly.js \

--mapbox-access-token "pk.eyJ1IjoiZXRwaW5hcmQiLCJhIjoiY2luMHIzdHE0MGFxNXVubTRxczZ2YmUxaCJ9.hwWZful0U2CQxit4ItNsiQ" \

--output-dir $root/test/image/baselines/ \

--verbose || EXIT_STATE=$?

}

for a sample orca command.

etpinard · 2019-04-02T16:10:40Z

+    }
+};
+
+function isValidScaleArray(scl) {


Very nicely done 🥇

etpinard · 2019-04-02T21:44:08Z

Looks like adding those volume tests to npm run test-image makes our CI run slower by about 1 minute:

but oh well, I'm ok with this 👌 . We should try to keep to number of volume down to a minimum though.

etpinard · 2019-04-02T21:45:07Z

@archmoj could you tell us a little bit more about how your volume renderer compare to MATLAB's https://www.mathworks.com/help/matlab/volume-visualization.html ? Thanks!

archmoj · 2019-04-02T22:03:40Z

Looks like adding those volume tests to npm run test-image makes our CI run slower by about 1 minute:

but oh well, I'm ok with this . We should try to keep to number of volume down to a minimum though.

Good observation.

…ojection type to orthographic in one mock

archmoj · 2019-04-03T01:43:44Z

Looks like adding those volume tests to npm run test-image makes our CI run slower by about 1 minute:

That was basically the result of increasing the waiting time between the batches.
Now reset to normal in 0aa25f0.

etpinard · 2019-04-03T14:27:30Z

+var Registry = require('../../registry');
+var colorscaleDefaults = require('../../components/colorscale/defaults');
+
+module.exports = function supplyIsoDefaults(traceIn, traceOut, defaultColor, layout, coerce) {


Nicely done! But I would prefer having this supplyIsoDefaults in isosurface/defaults.js.

We're trying to move away form this pattern where each defaults-related functions would get their own files. We now prefer having just one defaults.js file per trace module and have its reusable pieces exported along with the main trace supplyDefaults. function. See box/defaults.js for an example.

Done in af17adc.

etpinard · 2019-04-03T14:42:53Z

@archmoj can you add volume and isosurface to the gl3d partial bundle list: https://github.com/plotly/plotly.js/blob/master/lib/index-gl3d.js

archmoj · 2019-04-03T14:50:41Z

@archmoj can you add volume and isosurface to the gl3d partial bundle list: https://github.com/plotly/plotly.js/blob/master/lib/index-gl3d.js

Yes. Thanks for the link.
Done in 130da2f.

etpinard · 2019-04-04T13:56:40Z

This PR is looking good!

@archmoj two things:

can you confirm that orca can export your volume traces ok?
can you write down a few words about how your volume renderer compares to MATLAB's https://www.mathworks.com/help/matlab/volume-visualization.html ?

Thanks very much!

archmoj · 2019-04-04T17:06:42Z

can you write down a few words about how your volume renderer compares to MATLAB's https://www.mathworks.com/help/matlab/volume-visualization.html ?

Thanks for the link. From that link I also found this interesting link:
https://www.mathworks.com/help/matlab/visualize/overview-of-volume-visualization.html

It looks that the way matlab does this require multiple steps to create the patch. Since our isosurface was capable of generating two or multiple surfaces between iso-values, and at the same time drawing iso-caps and slices those features are accessible to the user in volume trace too.
Thinking of visualisation of wind data in 3D (x, y, z, u, v, w), one could generate an example to display opaque slices, caps from isosurface and use volume on wind speed {value = (u*u + v*v + z*z)^0.5}, then add cone or streamtube plots to draw 3D wind direction vectors (u, v, w).
Some other 'nice-to-have' features (for future work) could be adding attributes to isosurface & volume for clipping/reducing volume data similar to https://www.mathworks.com/help/matlab/ref/subvolume.html and drawing contour-slices on the plot https://www.mathworks.com/help/matlab/ref/contourslice.html.

etpinard · 2019-04-05T16:11:23Z

@archmoj the gl3d_volume mocks can't render in time (i.e. before the timeout) on my machine, making npm run test-image fail 😡

Do we have to test such large mocks in order to 🔒 down the behaviour?

archmoj · 2019-04-05T16:19:30Z

@archmoj the gl3d_volume mocks can't render in time (i.e. before the timeout) on my machine, making npm run test-image fail

Do we have to test such large mocks in order to down the behaviour?

Should we rename the mocks?
@etpinard if you are still on the branch could you please check the orca export?

etpinard · 2019-04-05T16:26:03Z

Should we rename the mocks?

No, I think we should reduce the size of their data arrays.

etpinard · 2019-04-05T16:26:27Z

if you are still on the branch could you please check the orca export?

Are you having issues running orca?

archmoj · 2019-04-05T16:28:37Z

if you are still on the branch could you please check the orca export?

Are you having issues running orca?

Yes - I needed to reinstall it. But then it didn't work : (

archmoj · 2019-04-05T16:29:47Z

Should we rename the mocks?

No, I think we should reduce the size of their data arrays.

OK. I will take care of that.

archmoj · 2019-04-05T20:35:41Z

@etpinard I was able to test orca export.
When possible please retry running the test-image with the reduced volume mocks on your machine.
Thank you very much in advance.

etpinard · 2019-04-08T14:07:55Z

@archmoj commit e23b2c3 is brilliant. Thanks very much for doing this.

I can confirm that the image tests are passing smoothly on my machine and orca exports are looking crisp!

Let's get this in 💃 💃 💃

adding volume4d new trace

3a39fd8

archmoj added status: reviewable labels Jan 28, 2019

This was referenced Jan 28, 2019

Volume3d iso #3479

Closed

Options to enable opacityscale gl-vis/gl-mesh3d#19

Merged

archmoj commented Jan 28, 2019

View reviewed changes

etpinard reviewed Jan 28, 2019

View reviewed changes

Comment thread lib/index.js Outdated

etpinard reviewed Jan 28, 2019

View reviewed changes

Comment thread src/traces/isosurface/attributes.js Outdated

etpinard reviewed Jan 28, 2019

View reviewed changes

archmoj added 17 commits January 28, 2019 16:59

opacityscale as 2d array

e3a59da

correct volume baselines

c138d9d

volume4d > volume

5dbe037

new baselines

f141868

add multi trace mock and add it to mock list

b1170a9

baseline for multiple volume traces

1a49fa2

revised the colorscale of datacube mock

e0b786c

updated baseline with revised colorscale

76dae20

bring isosurface draws to front and added a mock with mri data

fb502bd

enter mri data

c74daad

revised mri mock

ea92715

fixup test

b07aa6a

using new shader and mock with two volume traces in one cube

48909e0

correct baseline using new shader

02f171f

added new baseline

3b7320f

mock and baseline changes

1ec0941

exclude mri to run on CI

51035f9

archmoj added 2 commits February 5, 2019 19:14

removed mri baseline

7a13bde

new volume mock with aspect ratio and opacityscale

7249f8f

etpinard reviewed Apr 2, 2019

View reviewed changes

archmoj added 2 commits April 2, 2019 13:38

Merge remote-tracking branch 'origin/master' into volume4d

d35f69a

run volume image test on CI

e6b0efd

archmoj added 2 commits April 2, 2019 20:36

reset batch wait to 500

0aa25f0

fixup volume default to coerce volume attributes and dflt - switch pr…

77f7317

…ojection type to orthographic in one mock

etpinard reviewed Apr 3, 2019

View reviewed changes

archmoj added 3 commits April 3, 2019 11:13

revised isosurface default function exports

af17adc

add volume and isosurface to the gl3d partial bundle list

130da2f

revert unwanted scattergl changes in the previous commit

01a9426

reduced volume mock data

e23b2c3

bump gl-mesh3d 2.1.0

4a9987d

archmoj merged commit e5b8c71 into master Apr 8, 2019

archmoj deleted the volume4d branch April 8, 2019 14:56

	test_image () {
	$root/../orca/bin/orca.js graph \
	$root/test/image/mocks/mapbox_* \
	--plotly $root/build/plotly.js \
	--mapbox-access-token "pk.eyJ1IjoiZXRwaW5hcmQiLCJhIjoiY2luMHIzdHE0MGFxNXVubTRxczZ2YmUxaCJ9.hwWZful0U2CQxit4ItNsiQ" \
	--output-dir $root/test/image/baselines/ \
	--verbose \|\| EXIT_STATE=$?
	}

Uh oh!

Uh oh!

Conversation

archmoj commented Jan 28, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

jackparmer commented Jan 28, 2019

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

archmoj Jan 28, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

jackparmer commented Feb 5, 2019

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

etpinard commented Apr 2, 2019

Uh oh!

etpinard commented Apr 2, 2019

Uh oh!

archmoj commented Apr 2, 2019

Uh oh!

archmoj commented Apr 3, 2019

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

etpinard commented Apr 3, 2019

Uh oh!

archmoj commented Apr 3, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

etpinard commented Apr 4, 2019

Uh oh!

archmoj commented Apr 4, 2019

Uh oh!

etpinard commented Apr 5, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

archmoj commented Apr 5, 2019

Uh oh!

etpinard commented Apr 5, 2019

Uh oh!

etpinard commented Apr 5, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

archmoj commented Apr 5, 2019

Uh oh!

archmoj commented Apr 5, 2019

Uh oh!

archmoj commented Apr 5, 2019

Uh oh!

etpinard commented Apr 8, 2019

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

archmoj commented Jan 28, 2019 •

edited

Loading

archmoj Jan 28, 2019 •

edited

Loading

archmoj commented Apr 3, 2019 •

edited

Loading

etpinard commented Apr 5, 2019 •

edited

Loading

etpinard commented Apr 5, 2019 •

edited

Loading