A real grab-bag of tools for spectroscopic data analysis and modelling.

Source code is available at https://github.com/aheays/spectr. The repository can be cloned locally with git clone --depth=1 https://github.com/aheays/spectr.git or a zip archive downloaded from GitHub.

Version 3.11 or above will be necessary.

It is necessary to have gfortran and lapack available on the system before installation to enable fortran extensions.

The necessary linux packages are

• Debian / Ubuntu: gfortran python3-dev liblapack-dev
• Arch: gcc-gfortran lapack

The extension can be build in the spectr/ subdirectory with the following command.

f2py3 -c --quiet -llapack --f90flags="-Wall -ffree-line-length-none" --opt="-O3" fortran_tools.f90

Under ubuntu, installing the package qt5-default seem to be enough to get matplotlib to draw windowed figures.

The following shell commands should get things going in a linux environment if the external dependencies are met.

• Get external dependencies. On Ubuntu: sudo apt update && sudo apt install gfortran python3-dev liblapack-dev qt5-default
• Clone spectr: git clone --depth=1 https://github.com/aheays/spectr.git
• Create a virtual environment using python: python -m venv env
• Activate the virtual environment: source env/bin/activate
• Build and install: pip install ./spectr

Test by importing spectr and trying to plot something

source env/bin/activate
echo "from spectr.env import *" | python
qplot absorption/data/2021_11_30_bcgr.0

Examples of using spectr are provided spectr_examples

A command to import all submodules and many common functions directly into the working namespace is from spectr.env import *. Otherwise only the needed submodules can be imported, e.g., import spectr.spectrum

The optimiser.Optimiser class is used to conveniently construct model objects with parameters that can be fit to experimental data. The real-number input arguments of most methods of objects base-classed on Optimiser can be marked for optimisation by replacing their values with a optimiser.Parameter object. This has the abbreviated definition:

P(value=float,
  vary=True|False,
  step=float,
  uncertainty=float,
  bounds=(float,float))

Only the first argument is required. For example, x=P(2,True,1e-5,bounds=(0,100)) defines a parameter x that will be varied from an initial value of 2 but constrained to the range 0 to 100. When computing the finite-difference approximation to the linear dependence of model error on x a step size of \num{e-5} will be used. The fitting uncertainty unc will be set automatically after optimisation. Multiple Optimiser objects can be combined in a hierarchy, so that multiple spectra can be fit at once to optimise a common parameter, for example a temperature-dependence coefficient fit to spectra at multiple temperatures.

TBD

This is a command line programming for making line plots, e.g., qplot datafile, or qplot -h for a list of options.

Some examples scripts are provided in the repository https://github.com/aheays/spectr_examples

Conveniently import all submodules.

Storage, manipulation, and plotting of tabular data. Allows for the recursive calculation of derived quantities

Functions for performing common mathematical and scripting tasks.

Functions for plotting built on matplotlib.

Unit conversion, species name conversion, and various conversion formulae.

General class for conveniently and hierarchically building numerical models with optimisable parameters.

Classes for analysing atmospheric photochemistry.

Dataset subclasses for storing atomic and molecular line data.

Dataset subclasses for storing atomic and molecular level data.

Interact with output files of Bruker OPUS spectroscopic acquisition and analysis software.

Interface to internal spectroscopic and chemistry database.

Calculation of diatomic level energies from potential-energy curves.

Exception used to internally communicate failure conditions.

Access HITRAN spectroscopic data with hapy.

Simulate individual and groups of spectra lines of various shapes.

Functions for manipulating atomic and molecular quantum numbers.

Classes for manipulating and modelling of experimental spectroscopic datea.

Functions for computing thermochemical equilibrium with ggchem.

Classes for simulating diatomic levels and lines defined by effective Hamiltonians.

Various fortran functions and subroutines.

Currently the o/p/q Λ-doubling is handled with effective (S,Λ)-dependent forumulae. Instead implement the last three terms of Eq. 18 of brown1979 into _get_linearH() .

When comparing thismodel with pgopher, everything works find except the sign of the interactions a³Π(v=12)~D¹Δ(v=1), a³Π(v=12)~d³Δ(v=5), and a³Π(v=12)~d³Δ(v=6) needs to be reversed. There is a phase error between these interactions and others.

##rafals draft 2021-06-24
## 
## crossing states
upper_13C18O.add_level('A¹Π(v=1)',Tv=66175.53765,Bv=1.43761743,Dv=6.11179e-06,Hv=-22.39e-12,)
upper_13C18O.add_level('D¹Δ(v=1)',Tv=66442.5076,Bv=1.12,Dv=5.79e-6,Hv=-0.22e-12,)
upper_13C18O.add_level('I¹Σ⁻(v=2)',Tv=66595.57091,Bv=1.1146473,Dv=5.68e-6,Hv=2.25e-12,)
upper_13C18O.add_level('d³Δ(v=6)',Tv=66956.97424,Bv=1.09416857,Dv=5.31e-6,Hv=-0.60e-12,Av=-16.097,ADv=-9.17e-5,λv=0.94,γv=0.76e-2,)
upper_13C18O.add_level('e³Σ⁻(v=3)',Tv=66811.0988,Bv=1.1126549,Dv=5.55e-6,Hv=-1.50e-12,λv=0.5278,)
# ## non-crossing states
upper_13C18O.add_level('d³Δ(v=5)',Tv=65949.55,Bv=1.11,Dv=5.33e-6,Hv=-0.60e-12,Av=-15.91,ADv=-9.17e-5,λv=0.85,γv=0.69e-2,)
upper_13C18O.add_level('e³Σ⁻(v=2)',Tv=65802.44,Bv=1.13,Dv=5.58e-6,Hv=-1.50e-12,λv=0.54,)
upper_13C18O.add_level('I¹Σ⁻(v=1)',Tv=65593.17,Bv=1.13,Dv=5.67e-6,Hv=2.25e-12,)
upper_13C18O.add_level('a′³Σ⁺(v=10)',Tv=66066.95,Bv=1.07,Dv=5.17e-6,Hv=-0.30e-12,)
upper_13C18O.add_level('a′³Σ⁺(v=11)',Tv=67037.79,Bv=1.05,Dv=5.16e-6,Hv=-0.30e-12,λv=-108.84e-2,γv=-0.50e-2,)
upper_13C18O.add_level('a³Π(v=12)',Tv=66355.00,Bv=1.32,Dv=5.67e-6,Av=36.97,ADv=-20.58e-5,λv=-0.49e-2,γv=0.33e-2,ov=0.64,pv=2.73e-3,qv=2.95e-5,)
# ## interactions with crossing states
upper_13C18O.add_coupling('A¹Π(v=1)','D¹Δ(v=1)',ξv=-6.1688e-2),
upper_13C18O.add_coupling('A¹Π(v=1)','I¹Σ⁻(v=2)',ξv=7.630e-2)
upper_13C18O.add_coupling('A¹Π(v=1)','d³Δ(v=6)',ηv=18.0838)
upper_13C18O.add_coupling('A¹Π(v=1)','e³Σ⁻(v=3)',ηv=-5.4206)# ## interactions with non-crossing states
upper_13C18O.add_coupling('A¹Π(v=1)','d³Δ(v=5)',ηv=15.57)
upper_13C18O.add_coupling('A¹Π(v=1)','e³Σ⁻(v=2)',ηv=14.05)
upper_13C18O.add_coupling('A¹Π(v=1)','I¹Σ⁻(v=1)',ξv=9.89e-2)
upper_13C18O.add_coupling('A¹Π(v=1)','a′³Σ⁺(v=10)',ηv=-5.29)
upper_13C18O.add_coupling('A¹Π(v=1)','a′³Σ⁺(v=11)',ηv=3.836)
## interactions not including A
upper_13C18O.add_coupling('a³Π(v=12)','I¹Σ⁻(v=2)',ηv=-7.604)
# upper_13C18O.add_coupling('a³Π(v=12)','D¹Δ(v=1)',ηv=-7.955)
# upper_13C18O.add_coupling('a³Π(v=12)','d³Δ(v=5)',ηv=-38.48,ξv=7e-2)
# upper_13C18O.add_coupling('a³Π(v=12)','d³Δ(v=6)',ηv=26.31,ξv=5.80e-2)
upper_13C18O.add_coupling('a³Π(v=12)','D¹Δ(v=1)',ηv=7.955)
upper_13C18O.add_coupling('a³Π(v=12)','d³Δ(v=5)',ηv=38.48,ξv=-7e-2)
upper_13C18O.add_coupling('a³Π(v=12)','d³Δ(v=6)',ηv=-26.31,ξv=-5.80e-2)
upper_13C18O.add_coupling('a³Π(v=12)','e³Σ⁻(v=2)',ηv=5.09,ξv=1.00e-2)
upper_13C18O.add_coupling('a³Π(v=12)','e³Σ⁻(v=3)',ηv=8.24,ξv=1.60e-2)