You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
At present any attempt to plot using the MSlice CLI on a system without a display (or running virtual desktop) results in a Qt error Could not connect to display or Could not connect to X server.
Whilst it is not possible to run Qt graphical components "headless" it is possible to run Matplotlib headless. Thus when not using the customised plot windows, it should be possible to use MSlice headless using the 'mslice' projection in standard Matplotlib.
However, currently the overloaded errorbar and pcolormesh methods in the MSlice projections explicitly assume that they are being invoked on the customised MSlice plot windows (which needs Qt). Thus they fail in headless environments.
For use in the ISIS autoreduction system to plot figures of the autoreduced data for a quick glance, it would be good to have MSlice able to run headless. At the moment a work-around is to extract the underlying Mantid raw_ws and to use the 'mantid' projection in Matplotlib to run the plot instead.
The text was updated successfully, but these errors were encountered:
At present any attempt to plot using the MSlice CLI on a system without a display (or running virtual desktop) results in a Qt error
Could not connect to display
orCould not connect to X server
.Whilst it is not possible to run Qt graphical components "headless" it is possible to run Matplotlib headless. Thus when not using the customised plot windows, it should be possible to use MSlice headless using the
'mslice'
projection in standard Matplotlib.However, currently the overloaded
errorbar
andpcolormesh
methods in the MSlice projections explicitly assume that they are being invoked on the customised MSlice plot windows (which needs Qt). Thus they fail in headless environments.For use in the ISIS autoreduction system to plot figures of the autoreduced data for a quick glance, it would be good to have MSlice able to run headless. At the moment a work-around is to extract the underlying Mantid
raw_ws
and to use the'mantid'
projection in Matplotlib to run the plot instead.The text was updated successfully, but these errors were encountered: