Parameter Files

So far every model has been assembled by hand in Python. That is great for exploration, but science workflows benefit from reproducibility: being able to re-run or share an exact configuration without distributing a notebook. TauREx .par files are the answer — a simple INI-style format where each section corresponds to one Python component.

The mapping is direct and transparent:

• [Temperature] → temperature profile

• [Pressure] → pressure grid

• [Chemistry] → gas mixture

• [Model] → forward-model geometry and contributions

The same model we built manually in Transmission spectrum basics, Emission spectrum basics, and Cloud models can be expressed in a dozen lines of text. More information about the full parameter-file syntax is here, and global settings (data paths, log level) are here.

Data Note

This notebook uses the opacity and CIA files set up in Setup and opacity data. TauREx provides the software to work with these datasets; the files themselves are third-party products from ExoMol and HITRAN.

from pathlib import Path
import os
import sys

for candidate in (Path.cwd(), *Path.cwd().parents):
    if (candidate / '_shared.py').exists():
        sys.path.insert(0, str(candidate))
        break
    nested = candidate / 'examples' / 'notebooks'
    if (nested / '_shared.py').exists():
        sys.path.insert(0, str(nested))
        break

from _shared import CIA_DIR, TMP_DIR, XSEC_DIR, ensure_opacity_data

ensure_opacity_data(download=False)
output_path = Path(os.path.relpath(TMP_DIR / 'input.par', Path.cwd()))
xsec_path = os.path.relpath(XSEC_DIR, Path.cwd())
cia_path = os.path.relpath(CIA_DIR, Path.cwd())

print('Opacity data located and ready.')
print(f'Parameter file will be written to: {output_path}')

Opacity data located and ready.
Parameter file will be written to: ../../examples/tmp/input.par

input_par = f'''[Global]
xsec_path = {xsec_path}
cia_path = {cia_path}

[Chemistry]
chemistry_type = taurex
fill_gases = H2,He
ratio = 0.1741

    [[H2O]]
    gas_type = constant
    mix_ratio = 1e-3

[Temperature]
profile_type = isothermal
T = 2000

[Pressure]
profile_type = Simple
atm_min_pressure = 1e-5
atm_max_pressure = 1e5
nlayers = 100

[Planet]
planet_type = Simple
planet_mass = 0.74
planet_radius = 1.38

[Star]
star_type = blackbody
temperature = 6117
radius = 1.16

[Model]
model_type = transmission

    [[Absorption]]
'''

output_path.parent.mkdir(parents=True, exist_ok=True)
output_path.write_text(input_par)
print(input_par)

[Global]
xsec_path = ../../examples/tmp/xsec
cia_path = ../../examples/tmp/cia

[Chemistry]
chemistry_type = taurex
fill_gases = H2,He
ratio = 0.1741

    [[H2O]]
    gas_type = constant
    mix_ratio = 1e-3

[Temperature]
profile_type = isothermal
T = 2000

[Pressure]
profile_type = Simple
atm_min_pressure = 1e-5
atm_max_pressure = 1e5
nlayers = 100

[Planet]
planet_type = Simple
planet_mass = 0.74
planet_radius = 1.38

[Star]
star_type = blackbody
temperature = 6117
radius = 1.16

[Model]
model_type = transmission

    [[Absorption]]

Loading the Parameter File

ParameterParser reads a .par file and reconstructs the full model hierarchy — exactly the same objects that would be created by hand in Python. This round-trip confirms that the declarative and programmatic approaches are equivalent.

Full syntax details, dynamic-header behaviour, and required section names are documented here.

from taurex.parameter import ParameterParser

parser = ParameterParser()
parser.read(str(output_path))
parser.setup_globals()
model = parser.generate_appropriate_model()

print(type(model).__name__)
print('Contributions:', [c.name for c in model.contribution_list])

TransmissionModel
Contributions: ['Absorption']

print('Temperature object:', type(model.temperature).__name__)
print('Chemistry object:', type(model.chemistry).__name__)
print('Pressure object:', type(model.pressure).__name__)

Temperature object: Isothermal
Chemistry object: TaurexChemistry
Pressure object: SimplePressureProfile