xmovie

A simple way of creating movies from xarray objects.

With ever increasing detail, modern scientific observations and model results
lend themselves to visualization in the form of movies.

Not only is a beautiful movie a fantastic way to wake up the crowd on a Friday
afternoon of a weeklong conference, but it can also speed up the discovery
process, since our eyes are amazing image processing devices.

This module aims to facilitate movie rendering from data object based on
xarray objects. It is already possible to
get very quick and beautiful static images from xarray, but movies can still
present a hassle.

The main aims of this module are:

• Enable quick but high quality movie frame creation from existing xarray
  objects with preset plot functions.
• Convert your static plot workflow to a movie with only a few lines of code, while maintaining all the flexibility of xarray and matplotlib.

Installation

Clone this repository with $ git clone git@github.com:jbusecke/xmovie.git and
install it from source $ python setup.py install

Presets for quick movies

Using the presets of xmovie making a movie is very easy:

from xmovie import Movie
from xmovie.presets import rotating_globe_dark

ds = xr.tutorial.open_dataset('air_temperature').load().isel(time=slice(0,150))
mov = Movie(ds)

Preview single frames in the movie with the Movie.preview function:

# preview 10th frame
mov.preview(10);
# preview 100th frame. Note the rotation
mov.preview(100);

[图片上传失败...(image-a9e882-1553128130115)][图片上传失败...(image-d2751b-1553128130115)]
ocs
and save out each frame as a picture

mov.save('.') # saves to current directory

Convert images to movies

In the commandline you can now convert the frames to a movie file using ffmpeg

$ ffmpeg -y -i "frame_%05d.png" -c:v libx264 -preset veryslow -crf 10 -pix_fmt yuv420p -framerate 20 movie.mp4

from a notebook simply use a ! before the command to execute a shell command:

# convert to movie file
! ffmpeg -y -i "frame_%05d.png" -c:v libx264 -preset veryslow -crf 10 -pix_fmt yuv420p -framerate 20 movie.mp4

Convert to gif

To convert your movie file into a gif do

# convert to gif
! ffmpeg -y -i movie.mp4 -vf palettegen palette.png
! ffmpeg -y -i movie.mp4 -i palette.png -filter_complex paletteuse -r 10 -s 480x320 movie.gif

Just clean up all the image files

!rm *.png

and enjoy your masterpiece.

[图片上传失败...(image-b2ff17-1553128130115)]

Modify plots

The preset plot-functions each have a unique set of keyword arguments for custom looks, but they all support the xarray plotting modes via the plotmethod keyword:

mov = Movie(ds, rotating_globe_dark, plotmethod='contourf', coastline=False, land=True)
mov.save('.')

[图片上传失败...(image-a23483-1553128130115)]

mov = Movie(ds, rotating_globe_dark, plotmethod='contour', coastline=False, land=True)
mov.save('.')

[图片上传失败...(image-3a10d6-1553128130115)]

ds = xr.tutorial.open_dataset('rasm').Tair

# Interpolate time for smoother animation
ds['time'].data = np.arange(len(ds['time']))
ds = ds.interp(time=np.linspace(0,10, 60))

mov = Movie(ds, rotating_globe_dark,
            cmap='RdYlBu_r',
            x='xc',
            y='yc', #accepts keyword arguments from the xarray plotting interface
            lat_start=45, # Custom keywords from `rotating_globe_dark
            lon_rotations=0.2)
mov.save('.')

[图片上传失败...(image-666045-1553128130115)]

Custom Plots

You can customize any plot based on an xarray data structure and a 'frame-dimension' (usually time, or another dimension that will evolve with time in the movie).

Take this example:

# some awesome static plot
import matplotlib.pyplot as plt
fig = plt.figure(figsize=[10,5])
tt = 40

station = dict(x=100, y=150)
ds_station = ds.sel(**station)

(ax1, ax2) = fig.subplots(ncols=2)
ds.isel(time=tt).plot(ax=ax1)
ax1.plot(station['x'], station['y'], marker='*', color='k' ,markersize=15)
ax1.text(station['x']+4, station['y']+4, 'Station', color='k' )
ds_station.isel(time=slice(0,tt)).plot(ax=ax2)

ax2.set_xlim(ds.time.min(), ds.time.max())
ax2.set_ylim(ds_station.min(), ds_station.max())

ax1.set_aspect(1)
ax1.set_facecolor('0.5')
ax2.set_aspect(0.2)
ax1.set_title('');
ax2.set_title('Data at station');
fig.subplots_adjust(wspace=0.4)

[图片上传失败...(image-c71841-1553128130115)]

All that is needed to wrap this into a function with the signature func(ds, fig, timestamp, **kwargs), where ds is an xarray Dataset or DataArray, fig is a matplotlib.figure object and timestamp is an integer which indicates the movie frame.

def custom_plotfunc(ds, fig, tt):
    station = dict(x=100, y=150)
    ds_station = ds.sel(**station)

    (ax1, ax2) = fig.subplots(ncols=2)

    # Colorlimits need to be fixed or your video is going to cause seizures.
    # This is the only modification from the code above!
    ds.isel(time=tt).plot(ax=ax1, vmin=ds.min(), vmax=ds.max(), cmap='RdBu_r')
     
    ax1.plot(station['x'], station['y'], marker='*', color='k' ,markersize=15)
    ax1.text(station['x']+4, station['y']+4, 'Station', color='k' )
    ds_station.isel(time=slice(0,tt)).plot(ax=ax2)

    ax2.set_xlim(ds.time.min(), ds.time.max())
    ax2.set_ylim(ds_station.min(), ds_station.max())

    ax1.set_aspect(1)
    ax1.set_facecolor('0.5')
    ax2.set_aspect(0.2)
    ax1.set_title('');
    ax2.set_title('Data at station');
    fig.subplots_adjust(wspace=0.4)

mov_custom = Movie(ds, custom_plotfunc)
mov_custom.preview(2)

[图片上传失败...(image-2010ed-1553128130115)]

Note: This animation looks terrible as a gif if using the suggested settings
above. Instead use ! ffmpeg -y -i {moviename}.mp4 {moviename}.gif