Repository navigation
Time sliding with plume model enables is very slow #183
Description
It seems that when sliding time, the plume model is calculated for every step of the slider. Can we change to a computation at mouse release?
{{ message }}
It seems that when sliding time, the plume model is calculated for every step of the slider. Can we change to a computation at mouse release?
Activity
Peter9192 commentedon Dec 5, 2025
Computation at mouse release is not nice because you can use the slider to see the evolution of the profiles, which you'll lose then.
Instead I tried precomputing everything in #190, however, that is very slow and also very data intensive. On the flipside, it does lead to cleaner code, and you can easily get the axis limits across the full time range, so the axes don't change while sliding (#186)
Alternatively, we can reduce the vertical resolution of the profiles (currently 1m, we could make it 10 or 20 m for example). That would make compute faster and less data, but I'm not sure if it's physically okay for strong gradients. Can you provide some insight @Chiil?
With 10m resolution profile/plume calculation on slider is probably okay, or at least much better. Pre-calculation is still necessary to calculate the plot bounds, so now looking into data compression e.g. https://en.wikipedia.org/wiki/Ramer%E2%80%93Douglas%E2%80%93Peucker_algorithm