radynpy 0.3.3

Analysis tools for Radyn in Python

This is a very early development release of RadynPy, a nascent suite of tools to allow analysis of Radyn simulations to be performed in Python.

If you don’t know what Radyn is, then this package is probably not aimed at you!

Currently the only class that is designed to be used is RadynData (and LazyRadynData), which can be used to load the .cdf files created by Radyn.

This module exposes the functions: index_convention, var_info, load_vars, and lazy_load_all.

• Calling index_convention prints the indexing convention used in var_info.

• var_info takes a str or list of str and prints the meaning of the associated variables, along with their dimensionality in the indexing convention used. var_info can also take ‘*’ as an argument, for which it will return the info on all variables.

• load_vars takes the path to a CDF file and a list of the required variables which are then loaded into an instance of RadynData and can now be accesed as .varName rather than needing to go via the strings. If the filename is in the FCHROMA format then the heating parameters will be parsed from the string. If the filename is simply radyn_out.cdf then the filename isn’t scanned for parameters. If the filename is a different format then you may need to pass parseFilenameParams=False to load_vars. load_vars can also take ‘*’ as an argument, for which it will load all variables and timesteps into memory.

• lazy_load_all loads all of the variables in the CDF in a lazy manner, meaning that they are only loaded into memory when called for the first time. This has the unfortunate side effect of needing to keep a handle to the CDF file open, so ideally the close() method should be called when you are done with the file, and not opening thousands (!) of CDF files simulataneously. For just a few files the call to close() isn’t too important as the class attempts to clean up after itself as much as possible. This mode is very good for easy exploration of CDF files as the requested variables do not need to be known a priori, making it easier to use in the REPL, without thehuge memory requirements of loading all of the variables from a file. lazy_load_all behaves in the same way as load_vars with respect to parsing the filename parameters, and also accepts the parseFilenameParams keyword.

I expect that lazy_load_all will become the prevailing mode of use, but in scripts where a large number of simulations (e.g. the grid) are being processed then load_vars will probably work out slightly more efficient.

This documentation will be fleshed out more in time, this is but a very early alpha release.