search

mfitzp / pathomx

Workflow based scientific analysis built on Python
https://www.mfitzp.com/tools/pathomx/
GNU General Public License v3.0
91 stars 30 forks source link

Spectra & PCA Scores Plot broken graphics #27

Open popher opened 9 years ago

popher commented 9 years ago

In release v3.0.0.rc5 for OS X, there seems to be some missing/broken graphics.

For example, when using the example thp1_1d_nmrlab_metabolab.csv dataset provided, the spectra plot differently (worse) than in an earlier release.

screenshot 2014-10-22 09 55 42

Also it appears that the markers for a PCA scores plot are missing, but based on the shapes of the circle outlines for the clusters, it must be plotting it correctly behind-the-scenes.

screenshot 2014-10-22 09 55 29

mfitzp commented 9 years ago

The difference in colours might be due to it already assigning the original colours to another class group definition (have you previously loaded another dataset?). If you go to Appearance -> Line & Marker Styles can you delete the other colour definitions? If you do that it should revert back to the 'normal' colours.

The missing markers in the PCA plot are probably caused by the same thing, but of course there should be no situation where markers end up invisible! I'll look into the auto-assignment of marker colours and see where it is going wrong.

popher commented 9 years ago

I had loaded a previous dataset, and restarting Pathomx resolved the issue (with spectra/scores showing as they should).

For what its worth, Appearance > Line & Marker Styles didnt seem to do anything (no pop up options or other discernible difference).

mfitzp commented 9 years ago

Great, I think ideally the assigned colours should be cleared when datasets are removed from the workspace but I'll have think about best to implement this.

There was a bug with the Styles window in rc5 but if you download the latest http://download.pathomx.org/Pathomx-latest.dmg it should be fixed. Clicking on Appearance > Line & Marker Styles should give you a little window where you can add/edit/remove colour settings for your plots. I'll update the documentation to mention this more explicitly.