line style settings for temporally linked profiles should be located inside the Spectral Profile Sources panel (maybe a bug)

[Andreas Janz]

To just plot the data points of a temporally linked profile, you have to change the line style in the Spectral View visualization settings:

‌

This is really counter intuitive. The line style should be right next to the color setting in the Spectral Profile Sources panel:

‌

I’m not sure, but maybe the current implementation could be concidered as buggy, because right next to the line style, there is another color setting, which is not used:

‌

[Benjamin Jakimow]

There are two kind of profiles:
- profile candidates eventually to be added to the speclib, e.g. by clicking on [+]
- profiles in the spectral library

The color defined in the Spectral Profile Source panel is used for these candidates, which are replaced by the next candidate if not added to the speclib (by [+])

The color defined in the Spectral Profile Plot Settings is used for the profiles selected by the given profile view.

Using different colors for candidates and finally added profiles is similar to what is known from the digitalization of geometries. It allows to better distinguish these two profile types and to make the final decision if candidate(s) should be added to the speclib or not.

[Andreas Janz]

Hi @Benjamin Jakimow , not sure if you got the point of this issue:

The question is: why is the line style of a temporally linked profile not here, right next to it’s color?:

‌

[Andreas Janz]

I still would say that having to change the style of temporal profiles in the Vis-Settings of the Spectral View is kind of a bug.

‌

[Benjamin Jakimow]

addresses #905 : temporary profile candidate gets color from profile source panel

→ <<cset 83274557bcd7>>

[Benjamin Jakimow]

(except a small bug, which is fixed with https://bitbucket.org/hu-geomatics/enmap-box/pull-requests/31)

changed type to enhancement because the requested functionality is implemented
we might discuss if colors specifically for such profile candidates should be defined somewhere else. However, so far it makes much sense to me doing it in the profile source panel, where source, sampling type and further profile attribute are defined as well

‌

[Andreas Janz]

we might discuss if colors specifically for such profile candidates should be defined somewhere else. However, so far it makes much sense to me doing it in the profile source panel,

No, we don’t need to discuss that . The SpectralProfileSources panel is the right place to define the color!

The issue is about the LINE STYLE of linked pixel profiles, which should be right next to the COLOR here:

‌

[Andreas Janz]

The “bug” would be, that the line style of a linked pixel profile can be changed here:

‌

Why would that make sense?

What has that to do with the temporally linked pixel profiles?

I mean, I can have arbitrary many visualizations. That should have nothing to do with the line style of a linked pixel profile:

‌

[Andreas Janz]

It is I bug, right?

If not, please let’s make it a critical enhancement to be sure to implement this for v3.10.

[Benjamin Jakimow]

Squashed 'enmapbox/qgispluginsupport/' changes from 697d895ef..8604ddec4

8604ddec4 flaked files f04bccfaf spectralprocessingdialog.py writes & reads spectral setting to/from temporary files 94a92b761 spectralprocessingdialog.py write spectral setting to temporary files 95fbb2ea1 Merge branch 'write_spectral_setting' 0171ee8da python-tests.yml added releast-3_22 to test matrix cf607f050 added workaround for missing QgsFieldCalculator 9b2199131 added workaround QgsProcessingParametersGenerator.Flags() 2d9c72c6b hides QgsFieldCalculator if not available 8886405dd added repaint trigger added typechecking benchmark a60f4d101 added repaint trigger added typechecking benchmark 56afe2b3d enhances SpectralSettings added release-3_22 to test routines Signed-off-by: jakimowb [e-mail] dccdef480 Spectral Processing Dialog got static window title, as algo name is shown in separate widget fixed XML read error cb6eb54b5 added QgsFieldCalculator 3b0199fd4 refactoring spectral profile plots included spectral processing dialog 03ab35255 WIP refactoring profile plot model towards using QStandardItems 0f6b2f813 addresses #1 - added SelectMapLayerDialog b1761414c addresses #1 - added LayerRendererVisualization + unit test - flaked repo 59e6a783f addresses #b902032 addresses Issue #875 improved spectral profile plotting added workaround for wavelength definition at provider level 121dcec80 added test case with different pixel resolutions resolves #853 ddd03c777 addresses #853 (QPS updates) - added test_PixelScaleExtentMapTool - re-implemented zoom-to-native resolution from QGIS\src\app\qgisapp.cpp in maptools.py f74e1a3da qgisinfo.py works without start_app speclibrasterdataprovider initialized during init.py phase 4259415eb addresses #905 : temporary profile candidate gets color from profile source panel a236b916f qps pyqtgraph - try-except itemChange

git-subtree-dir: enmapbox/qgispluginsupport git-subtree-split: 8604ddec47b6d4add389005ec6bc3a09a8bde3fd

→ <<cset 03db28f8ba8e>>