File Formats¶
Model Packs¶
YAML form¶
The main way for the user to set up models, is by using (or copying and editing) one of the default science packs provided as YAML files. These YAML files are internally converted to an astropy Table (see next section Table form).
YAML is a data serialization language. The model packs are constructed as lists of dictionaries of different feature components (eg., Line Features, Dust Features). Each list consists of dictionaries which define the feature and the parameters associated with it. For example
PAH_5.3:
kind : dust_feature
wavelength: 5.27
fwhm: 0.17918
This shows a single PAH feature and its corresponding attributes (wavelength, fwhm). The parameters specified depend on the kind of the feature and its respective functional form. Bounds (min and max) can also be set for different parameters as well as for all the parameters of a certain group.
The fit parameters for continua (starlight and dust_continuum) are temperature and tau (emission strength).
For features like line and dust_feature, the fit parameters are wavelength, fwhm and power.
For attenuation, the fit parameters are tau, model, and geometry, while absorption features have tau, wavelength, fwhm and geometry.
See Example Input for an example of YAML input model file. For a detailed account of the YAML based science packs and the relevant syntax, please follow the Wiki.
Table form¶
When a model is set up from a YAML file, or when the fit results are saved to disk, the model is described by a table compatible with astropy, called the “feature table”.
This table has one row for each model component.
Each row has a value for the name (e.g., BB3, DF10, H2 S(3), [OIV]), group, and kind columns.
The name and group columns come from the YAML file, and can be any string, as defined by the user in the YAML file from which this table was generated.
The kind column indicates one of the supported component types, and determines the function by which the component wil be represented (e.g., BlackBody1D for dust_continuum, Drude1D for dust_feature, Gaussian1D for line).
Following these three columns, there are multiple columns for the fit parameters.
These columns contain 3-tuples with the value and the bounds: [value, min, max].
The bounds min and max are masked if a parameter is fixed.
If a parameter is not relevant for a certain feature, then value will be masked.
To reduce the number of columns, some parameters have different meaning depending on the kind of the feature.
For the dust_continuum components, the fit parameters are temperature (temperature) and the “intrinsic” or “unattenuated” amplitude (tau).
For the dust_feature and (gas) line models, the fit parameters are “intrinsic” or “unattenuated” amplitude (power), center wavelength (wavelength), and full width at half maximum (fwhm).
For the attenuation model, the fit is a measure for the optical depth (tau).
The model tables can be stored as ECSV tables as these provide readable (ASCII) files with metadata support. The ECSV metadata is used to read in a model from file, after a fit result has already been saved.
The astropy version of Drude1D and Gaussian1D are used under the hood. See reference API for description of BlackBody1D and the dust attenuation model.
See example fit output for an example of a feature table/output file.
Note
IDL PAHFIT uses fractional FWHM as input/output for the Drude profiles, python PAHFIT uses FWHM as input/output for the Drude profiles.
Example Input (YAML)¶
# PAHFIT Classic Model Pack
# Implements the IDL-based PAHFIT v1.2 model
# v0.1, May 2022
#######################
# Starlight continuum #
#######################
starlight:
kind: starlight_continuum
temperature: 5000
##################
# Dust continuum #
##################
dust_cont: # Modified Blackbodies
kind: dust_continuum
temperature: [300, 200, 135, 90, 65, 50, 40, 35]
##############################
# H_2 Line Emission Features #
##############################
H2_lines:
kind: line
wavelength:
H2_S(7): 5.5115
H2_S(6): 6.1088
H2_S(5): 6.9091
H2_S(4): 8.0258
H2_S(3): 9.6649
H2_S(2): 12.2785
H2_S(1): 17.0346
H2_S(0): 28.2207
################################
# Ionic Line Emission Features #
################################
ionic_lines:
kind: line
wavelength:
'[ArII]': 6.985274
'[ArIII]': 8.99138
'[SIV]': 10.5105
'[NeII]': 12.813
'[NeIII]': 15.555
'[SIII]_18': 18.713
'[OIV]': 25.91
'[FeII]': 25.989
'[SIII]_33': 33.480
'[SiII]': 34.8152
'[FeII]_35': 35.349
#################
# Dust Features #
#################
PAH_5.3:
kind: dust_feature
wavelength: 5.27
fwhm: 0.17918
PAH_5.7:
kind: dust_feature
wavelength: 5.7
fwhm: 0.1995
PAH_6.2:
kind: dust_feature
wavelength: 6.22
fwhm: 0.1866
PAH_6.7:
kind: dust_feature
wavelength: 6.69
fwhm: 0.4683
PAH_7.7_cmp:
kind: dust_feature
features:
PAH_7.7a:
wavelength: 7.42
fwhm: 0.93492
PAH_7.7b:
wavelength: 7.6
fwhm: 0.3344
PAH_7.7c:
wavelength: 7.85
fwhm: 0.41605
PAH_8.3:
kind: dust_feature
wavelength: 8.33
fwhm: 0.4165
PAH_8.6:
kind: dust_feature
wavelength: 8.61
fwhm: 0.33579
PAH_10.7:
kind: dust_feature
wavelength: 10.68
fwhm: 0.2136
PAH_11.3_cmp:
kind: dust_feature
features:
PAH_11.3a:
wavelength: 11.23
fwhm: 0.13476
PAH_11.3b:
wavelength: 11.33
fwhm: 0.36256
PAH_12:
kind: dust_feature
wavelength: 11.99
fwhm: 0.53955
PAH_12.6_cmp:
kind: dust_feature
features:
PAH_12.6a:
wavelength: 12.62
fwhm: 0.53004
PAH_12.6b:
wavelength: 12.69
fwhm: 0.16497
PAH_13.48:
kind: dust_feature
wavelength: 13.48
fwhm: 0.5392
PAH_14.04:
kind: dust_feature
wavelength: 14.04
fwhm: 0.22464
PAH_14.19:
kind: dust_feature
wavelength: 14.19
fwhm: 0.35475
PAH_15.9:
kind: dust_feature
wavelength: 15.9
fwhm: 0.318
PAH_17_cmp:
kind: dust_feature
features:
PAH_17a:
wavelength: 16.45
fwhm: 0.2303
PAH_17b:
wavelength: 17.04
fwhm: 1.1076
PAH_17c:
wavelength: 17.37
fwhm: 0.2085
PAH_17d:
wavelength: 17.87
fwhm: 0.28592
# This dust feature, in the PAHFIT classic model, is attributed to C60 and omitted.
#PAH_18.92:
# kind: dust_feature
# wavelength: 18.92
# fwhm: 0.35948
PAH_33.1:
kind: dust_feature
wavelength: 33.1
fwhm: 1.655
##########################
# Attenuation Model #
##########################
silicate:
kind: attenuation
model: S07_attenuation
geometry: mixed
Instrument packs¶
Rationale¶
While the physical parameters are provided by the science packs, the observed emission lines will be spectroscopically unresolved for certain instruments. Therefore, the width of the lines will be determined mainly by instrumental effects. For this purpose, instrument packs have been developed, which describe the spectral resolution as a function of wavelength. Each instrument pack (usually pertaining to one observatory) contains multiple instrument models, typically one per spectral segment. See the example below. One or multiple instrument configurations need to be provided when setting up a fit, either on the command line or when a Model instance is set up (see <fit_spectrum>).
Example file¶
The user can find the instrument configurations inspecting the contents of each instrument pack.
For example, the Spitzer pack (spitzer.yaml) is shown below.
It has a hierarchical structure to define the possible instrument configurations.
To use a single segment configuration, the user needs to provide a string containing the name of the science pack, followed by the names in the hierarchy separated by dots,
e.g. "spitzer.irs.sl.1".
Because most spectroscopic obervations will combine multiple of those configurations, a list of strings or wildcard string can be passed as well.
For example, if observations were taken with the SL1, SL2, LL1, and LL2 modes, then the instrument configuration can be writen down as "spitzer.irs.*.[12]".
The asterisk will match any name at level 2, while [12] will match either ‘1’ or ‘2’ at level 3.
# ---
# PAHFIT Instrument Pack, see https://github.com/PAHFIT/pahfit/wiki/File-Formats for more information
# This is the Instrument Pack for Spitzer observations using staring
# or spectral mapping mode. For each of the four instrument
# (Short-Low = SL, Long-Low = LL, Short-High = SH, Long-High = LH,
# with LL/SL divided into order/subslits denoted 1, 2, and 3) this
# file first provides the wavelength range covered (in microns). The
# resolving power (lambda / delta_lambda, where delta_lamba is the
# FWHM of the resolution element) is represented by a polynomial of
# degree 3, and this file lists the coefficients CC of that
# polynomial, i.e.
#
# R = CC[0] + CC[1] * lambda + CC[2] * lambda^2 + CC[3] * lambda^3
#
# where lambda is expressed in microns as well.
#
# Resolution data from the IRS Instrument Handbook v5.0, Section 4.1.3
# (see Figures 4.8 and 4.9). Wave range data from the
# irs_b[0-3]_WAVECUTvN.tbl file of the CUBISM package.
irs:
sl:
'1':
range: [7.51, 14.78]
coefficients: [0.0, 8.2667]
'2':
range: [5.24, 7.6]
coefficients: [0.0, 16.5333]
'3':
range: [7.34, 8.7]
coefficients: [0.0, 8.2667]
ll:
'1':
range: [20.5, 38.5]
coefficients: [0.0, 2.9524]
'2':
range: [14.5, 21.1]
coefficients: [0.0, 5.9048]
'3':
range: [19.4, 21.65]
coefficients: [0.0, 2.9524]
sh:
range: [9.9661, 19.4386]
coefficients: [600.]
lh:
range: [19.1095, 37.1661]
coefficients: [600.]
Fit Outputs¶
The fit results are saved in the same format as the feature table. This enables the fit output file to be used as an input file to reload a model. This is potentially useful for fitting multiple spectra that are similar (e.g., JWST spectral cubes).