Update scalars on existing mesh

jbusecke commented 1 year ago

📰 Custom Issue

I recently produced this and am wondering if I could improve performance with these sorts of high resolution climate outputs. The movie shows two 'animated' features: a) evolving scalars (time dependent velocity and salinity fields) with time b) A camera move

I wrote the model out in a fashing like this (pseudocode):

import geovista as gv

ds = load_xarray_dataset
frames = range(n)
camera_path = define_camera_path(frames)

p = gv.GeoPlotter()
p.open_movie(f'something.mp4')

for frame in frames:
    # select timestep
    da = ds.sel(time=frame).YOUR_VARIABLE

   # create new mesh for each timestep (point a)
    mesh = gv.Transform.from_2d(da.lon, da.lat, data=da.data)
    mesh = mesh.threshold()
    actor = p.add_mesh(mesh)

   # Animate camera position (point b)
    p.camera_position = camera_path.points[frame,:]

    # write movie frame and remove actor
    p.write_frame()
    p.remove_actor(actor)

p.close()

This seems super inefficent to me. I actually do not need to redraw the mesh, but instead the whole point a) is currently a change in the scalar values. I saw that there is a possibility to update just the scalars in pyvista. I wonder if there is an internal/external function that would allow me to 'map' my n dimensional array of scalar data with geovista to be in the proper format to use pyvista.Plotter.update_scalars() instead of removing/rebuilding the mesh.

bjlittle commented 1 year ago

Hey @jbusecke!

Great to hear from you.

Okay, I can see what you're doing, and your intuition is right ...

Given your example above, you should be able to do "something" like this:

import geovista as gv

ds = load_xarray_dataset
frames = range(n)
camera_path = define_camera_path(frames)

p = gv.GeoPlotter()
p.open_movie(f'something.mp4')

# create the mesh with no data payload
mesh = gv.Transform.from_2d(da.lon, da.lat)

# this is the name of the active scalar that contains your frame data
# which can be anything, let's go with ...
name = "data-frame"

for frame in frames:
    # select timestep
    da = ds.sel(time=frame).YOUR_VARIABLE

    # set the active scalar with the data payload for the frame
    mesh[name] = da.data

    # threshold the mesh - this can change the mesh geometry each frame,
    # so create a separate "frame" mesh
    frame = mesh.threshold()

    # this will add the named "data-frame" mesh to the plotter or replace it
    actor = p.add_mesh(frame, name=name)

    # Animate camera position (point b)
    p.camera_position = camera_path.points[frame, :]

    # write movie frame
    p.write_frame()

p.close()

The above assumes that the relationship between the data and the original mesh geometry doesn't change.

Give it a whirl and let me know if that works. It's difficult to give solid advice without the data to play with, so I may hunt around for some at this end to confirm a working pattern.

Also, you may have already seen this, but this may be of help.

HTH

🍻

github-actions[bot] commented 8 months ago

In order to maintain a backlog of relevant issues, we automatically label them as stale after 180 days of inactivity.

If this issue is still important to you, then please comment on this issue and the stale label will be removed.

Otherwise this issue will be automatically closed in 28 days time.