fire_an

code and scripts for finding main halos, making histograms and 2D projected maps, and analysing those products in FIRE simulations

'installation':

This is an in-development set of scripts, not a polished package. Many of the scripts are in flux, especially those in queuerun and makeplots, and I do have a tendency to do development on the main branch (especially if it's unlikely to break something important). Because I'm updating this a lot, it's simpler to just use git pulls instead of running an installation every time. However, I have provided a condaenv_firep1.yml file that includes the python packages required to run these scripts. It can be used to set up a conda environment where this code should run.

Note that a few functions requires additional packages, such as Andrew Wetzel's FIRE analysis code. I haven't included these here, since these are no longer the defaults I use.

Note that the main projection code and ion balance calculations call some compiled C code. I don't have that publicly available right now, because I did not write the SPH projection code myself (I only made small modifications), and I have not gotten around to asking the authors for permission.

setup:

The utils directory should contain a file called opts_locs.py. This file simply lists a number of files and directories needed to run the code, such as where to search for simulation outputs, and where to store halo data. Some example files opts_locs_<system>.py are provided. Note that the kernel_list, desngb, solar_abunds_ea, and Zsun_ea variables do not need to be changed. They're basically just stored here because this started off as a parameter file.

Storing files as opts_locs_<system>.py and copying those to opts_locs.py locally is recommended, as it provides a back-up for the setups on github, without constantly overwriting local locations. (utils/opts_locs.py is included in .gitignore for this reason.)

running the scripts:

To run, for example, run.py, options are:

• run from the directory above fire_an as

  python -m fire_an/queuerun/run -3
  

  (This will raise an error because run.py only takes postive integer arguments, but it also won't accidentally start an hour-long process on a login node.)
• or add the directory above fire_an (e.g. foo) to PYTHONPATH:

  export PYTHONPATH="/path/to/foo/:${PYTHONPATH}"
  

  and then, from anywhere, run:

  python -m fire_an.queuerun.run -3
  

  or specify an absolute or relative path to run.py from your current directory:

  python /path/to/fire_an/queuerun/run.py -3
  

This is basically an issue of the different scripts in different directories needing to be able to find each other. Running from the directory above fire_an adds that directory to PYTHONPATH, like adding it to PYTHONPATH explicitly does.

main functions:

TODO

warning:

This repo is mainly for me to sync up in-progress work. Much of this code is untested, half-finished projects I abandoned, or work in progress. Also, I have a tendency to do work and testing on the main branch.

Note that tests are usually of the 'plot and see if it makes sense' form, not functions that will return a boolean value.