The package apogeenet2 makes predictions of a star's temperature, surface
gravity, and metallicity from spectra.
To use the pipeline after installing, run apogeenet2 on the command line. By
default, the program searches the current directory for the .fits table, uses
the current directory to search for the spectra listed in the table, then writes
its predictions to predictions.csv in the current directory.
Alternatively, the program can be used and imported as a package. Inside
apogeenet2.pipeline, the Pipeline object can be found. The method
Pipeline.predict() does most of the work, feeding data through the model and
writing predictions out.
For help and further options, run apogeenet2 --help.
To install, run pip install apogeenet2. This package requires Python 3.6 or
newer.
The paper "APOGEE Net: An expanded spectral model of both low mass and high mass stars" (Sprague et al., 2021), detailing the significance and architecture of this project, can be found here: https://arxiv.org/abs/2201.03661
The summary is that this package uses a CNN to process the spectra, obtaining previously-unseen results on high-temperature stars.
This package is unmaintained and provided as-is. However, bugs can still be filed under the "Issues" tab of the project's GitHub page