RADEX is a one-dimensional non-LTE radiative transfer code base on the LAMDA molecular database.

I have written two codes to use RADEX more efficiently:

• radex_grid.py is an executable that is intended to be user-modified. It will create a grid (cube) over density, column, and temperature parameter space. It can be run in parallel using mpi4py and mpirun.

• radex_grid_opH2.py is the same as radex_grid, but allows multiple collision partners, specifically ortho and para H2. This proved useful for using the Troscompt 2009 collision rates.

• plot_grids.py is meant to plot the resulting data and/or turn it into .fits data cubes.

In order to use radex_grid with LVG, Sphere, and Planar assumptions, I recommend generating separate executables. Edit "radex.inc" included in the RADEX source code distribution (lines 34-36 in the current version as of this posting, stating "parameter (method = 1)" ) for each assumption, then make, then move bin/radex to bin/radex_sphere or bin/radex_lvg as appropriate. I then symlinked (ln -s) the executables into my /usr/local/bin directory, but you can add them to your path however you like.