Modules for fitting FIR line ratios to infer physical conditions

Requirements:

1. numpy
2. scipy.spatial (for Delaunay triangulation)
3. matplotlib
4. astropy or pyfits
5. (optional) scikit-learn (for kdtrees)

Usage:

• See demo/fit_ngc6822.py

• Because of how the fitting is done, errors should never be set to 0. even if the line value is zero and/or the line is masked. This will result in dividing by zero and a NAN for the whole calculation.

Description:

• demo/fit_ngc6822 an example of how to fit the data and store and produce output for a given region

• pdrfit/pdrfit.py Has a method to fit data for a single pixel

• pdrfit/pdrmodel.py has classes for reading and storing holding Kaufmann model grids and for generating model predictions of observables given physical conditions.

• pdrfit/io.fits Methods for reading observational data into numpy structured arrays

• pdrfit/pdrplot.py methods to produce pretty plots and to produce point estimates from posterior grids

• pdrfit/modelgrid.py generic classes for storing and interpolating model grids. Typically should be left alone.