PyEchelle

PyEchelle is a simulation tool, to generate realistic 2D spectra, in particular cross-dispersed echelle spectra. However, it is not limited to echelle spectrographs, but allows simulating arbitrary spectra for any fiber-fed or slit spectrograph, where a model file is available. Optical aberrations are treated accurately, the simulated spectra include photon and read-out noise.

PyEchelle uses numba for implementing fast Python-based simulation code. It also comes with CUDA support for major speed improvements.

Example usage

You can use PyEchelle directly from the console:

pyechelle --spectrograph MaroonX --fiber 2-4 --sources Phoenix --phoenix_t_eff 3500 -t 10 --rv 100 -o mdwarf.fit

If you rather script in python, you can do the same as above with the following python script:

from pyechelle.simulator import Simulator
from pyechelle.sources import Phoenix
from pyechelle.spectrograph import ZEMAX

sim = Simulator(ZEMAX("MaroonX"))
sim.set_fibers([2, 3, 4])
sim.set_sources(Phoenix(t_eff=3500))
sim.set_exposure_time(10.)
sim.set_radial_velocities(100.)
sim.set_output('mdwarf.fits', overwrite=True)
sim.run()

Both times, a PHOENIX M-dwarf spectrum with the given stellar parameters, and a RV shift of 100m/s for the MAROON-X spectrograph is simulated.

The output is a 2D raw frame (.fits) and will look similar to:

Check out the Documentation for more examples.

Pyechelle is the successor of Echelle++ which has a similar functionality but was written in C++. This package was rewritten in python for better maintainability, easier package distribution and for smoother cross-platform development.

Installation

As simple as

pip install pyechelle

Check out the Documentation for alternative installation instruction.

Usage

See

pyechelle -h

for all available command line options.

See Documentation for more examples.

Model and output locations

PyEchelle stores downloaded spectrograph models and default outputs in platform-specific user directories (via platformdirs) instead of writing into the package installation folder.

• Model lookup order: user data models dir, user cache models dir, current working directory, legacy package models folder.
• Model download targets: user data models dir first, then user cache models dir.
• Default simulation output path (if -o/--output or Simulator.set_output(path=...) is not set): user data outputs dir.

You can inspect these paths directly:

from pyechelle.paths import model_data_dir, model_cache_dir, default_output_dir

print(model_data_dir())
print(model_cache_dir())
print(default_output_dir())

Migration note: existing models in the current working directory or legacy package models directory are still detected. For new setups, storing models in the user data directory is recommended.

Concept:

The basic idea is that any spectrograph can be modelled with a set of wavelength-dependent transformation matrices and point spread functions which describe the spectrographs' optics:

First, wavelength-dependent affine transformation matrices are extracted from the ZEMAX model of the spectrograph. As the underlying geometric transformations (scaling, rotation, shearing, translation) vary smoothly across an echelle order, these matrices can be interpolated for any intermediate wavelength.

Second, a wavelength-dependent point spread functions (PSFs) is applied on the transformed slit images to properly account for optical aberrations. Again, the PSF is only slowly varying across an echelle order, allowing for interpolation at intermediate wavelength.

Both, the matrices and the PSFs have to be extracted from ZEMAX only once. It is therefore possible to simulate spectra without access to ZEMAX

Citation

Please cite this paper if you find this work useful in your research.