dopplershift/dissertationcode

Code used in my dissertation

dissertationcode's Issues

Output file names

Output files can easily be the same across different runs, resulting in files overlapping each other. Need to:

• Check for duplicate file and append a suffix to file name if necessary
• Find a way to better distinguish outputfiles. Metadata in file isn't enough. Thoughts:
  • Radar name
  • Beamwidth
  • Model Information
  • Run start time
  • File output time
• This should, in general, be configurable. Options:
  • Straight dict-based string interpolation
  • Python 3-based string formatting, which is available in Python 2.6?

Contribute code to PyART

At least the implementation of self-consistent on top of zphi should be contributed to pyart.

Scattering shape relation

Need to add the ability to run the simulator using a different shape relation than the one default currently available. We need:

• Take as input a string that can be passed to scattering library
• Add this to the config parsing code
• Pass this down to the actual scattering calculation code
• Add the config value to the output file

Data cache lookups

Need to make a better way to query the data cache. This should eliminate the list comprehensions scattered everywhere. Also, this should be done in a way that permits us to get different sets of keys; for instance, allowing different sets of "experiments" per wavelength.

Regression coefficients

There seems to be an awful lot of variability in the KDP-A coefficients at least (as in V different from H). Also, these coefficients seem very different from the canonical 0.08 (for C band), which seems to work relatively well. Look into the regression.

Colorbar labels

Right now colorbars are labelled using a fixed function that does some munging of the field. We need a way to make this more configurable, so that the algorithms in the attenuation results appear better.

Algorithm registration

Registration of correction algorithms is still really ugly. Need to look for a way to clean this up.

Attenuation analysis

Need to start looking at streamlining comparing attenuation results between runs.