Fly

An interactive geophysical flow visualizer for Python.

fly provides an easy way to visualize both scalar and vector fields in real, interactive, 3-D spherical coordinates. Since it is written in Python, you can do all data processing as you are used to, and just pass your data to fly as NumPy arrays. And thanks to Glumpy's powerful rendering capacities, fly makes full use of your GPU to give you that FPS goodness.

See fly in action on Vimeo

But why?

Geophysicists usually visualize data as static two-dimensional plots on maps. This approach is well-suited to produce high-quality plots that convey detailed information to other scientists. However, those plots have two big practical disadvantages:

• Exploring a new data set feels clunky. Matplotlib is not built for speed, and is awkward to use interactively.
• Every two-dimensional map projection introduces distortions of some kind - when plotting vectors or streamlines, one has to be very careful not to skew the angles and magnitudes of the vectors at high latitudes.

On top of that, there is one psychological disadvantage: After hours of analysis, one tends to forget the dynamical nature of the ocean and the atmosphere. Wouldn't it be nice to see your data move, for once?

Installation

Most of the heavy lifting is done by Glumpy, but to use fly, you will have to install some additional Python dependencies. First up, note that fly only supports Python 3, and only works with Glumpy's recent master (not the release version found on PyPI).

If you have Python 3 and pip installed, you may e.g. run

pip3 install -r requirements.txt --user
pip3 install git+git://github.com/glumpy/glumpy.git --user

to install all dependencies. Alternatively, you should at least install the following packages:

• pillow
• NumPy
• SciPy
• Matplotlib
• Cython
• pyopengl
• Glumpy (from GitHub, e.g., via pip install git+git://github.com/glumpy/glumpy.git)

Usage

All code for fly is contained in a single source file, fly.py. This file contains the Fly class, which handles the visualization for you. All you need to do is to write an interface that reads your data, and pass it to the Fly class as NumPy arrays:

from fly import Fly

# create Fly instance
fly = Fly(flow_field=velocity)

# apply settings
fly.rotate = True
fly.markers = [
  ["Copenhagen", "Heidelberg"],
  [55.6761, 49.3988],
  [12.5683, 8.6724]
]
# etc

# do setup
fly.setup()

# run animation
fly.run()

Please note that all input fields must lie on a regular longitude-latitude grid with bounds (-180,180) and (-90,90). Make sure to interpolate accordingly before passing data to fly.

A (sloppy) interface script for a particular use case (interpreting Veros output) is located in the fly_veros.py file in this repository. It implements a command line interface that can e.g. be used like this:

python3 fly_veros.py veros_data.nc --flow-variables u v --shading variable --shading-variable temp --depth -10 --time -1

All available settings are class attributes of the Fly class, and documented in fly.py.