dust-compendium 0.1.0__tar.gz

dust_compendium-0.1.0/.github/workflows/ci.yml

@@ -0,0 +1,31 @@
+name: CI
+
+on:
+  push:
+    branches: [main]
+  pull_request:
+
+jobs:
+  test:
+    runs-on: ubuntu-latest
+    strategy:
+      fail-fast: false
+      matrix:
+        python-version: ["3.11", "3.12", "3.13"]
+    steps:
+      - uses: actions/checkout@v4
+      - uses: actions/setup-python@v5
+        with:
+          python-version: ${{ matrix.python-version }}
+      - run: pip install -e ".[dev,plots,docs]"
+      - run: ruff check .
+      # pre-commit is run here rather than installed as a git hook: this
+      # machine's git template points .git/hooks at the shared Galacticus
+      # hooks directory, so `pre-commit install` would overwrite it.
+      - uses: pre-commit/action@v3.0.1
+      # Hyperion is not installed here, so tests needing it or a built solver
+      # binary are skipped. The core physics has no such dependency.
+      - run: pytest -v -m "not hyperion and not solver and not reference"
+      # The documentation is built with warnings as errors, as Read the Docs
+      # does, so a broken cross-reference fails here rather than there.
+      - run: python -m sphinx -b html -W docs docs/_build/html

dust_compendium-0.1.0/.github/workflows/publish.yml

@@ -0,0 +1,41 @@
+# Publishes to PyPI when a release is created.
+#
+# Uses trusted publishing, so there is no token to keep: PyPI has to be told,
+# once, to trust this repository and this workflow. Until that is done the
+# publish step will fail, which is the right way round -- it cannot publish by
+# accident before anyone has decided it should.
+name: Publish
+
+on:
+  release:
+    types: [published]
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "3.12"
+      - run: pip install build
+      - run: python -m build
+      # Check what is about to be published before publishing it.
+      - run: pip install twine && twine check dist/*
+      - uses: actions/upload-artifact@v4
+        with:
+          name: distributions
+          path: dist/
+
+  publish:
+    needs: build
+    runs-on: ubuntu-latest
+    environment: pypi
+    permissions:
+      id-token: write
+    steps:
+      - uses: actions/download-artifact@v4
+        with:
+          name: distributions
+          path: dist/
+      - uses: pypa/gh-action-pypi-publish@release/v1

dust_compendium-0.1.0/.gitignore

@@ -0,0 +1,33 @@
+# Virtual environments
+.venv/
+venv/
+
+# Python build and cache artifacts
+__pycache__/
+*.py[cod]
+*.egg-info/
+build/
+dist/
+.pytest_cache/
+.ruff_cache/
+.coverage
+htmlcov/
+
+# Documentation build
+docs/_build/
+
+# Hyperion model inputs and outputs, and the compendium tabulations themselves.
+# These are large, reproducible from the configuration, and published to Zenodo
+# rather than tracked here. Anchored to the repository root, so that package
+# data such as src/dustcompendium/dust/data/ is not swept up with it.
+/data/
+*.hdf5
+*.rtout
+*.pbs
+*.log
+
+# Editor and OS clutter
+.DS_Store
+*~
+.idea/
+.vscode/

dust_compendium-0.1.0/.pre-commit-config.yaml

@@ -0,0 +1,22 @@
+# NOTE: do not run `pre-commit install` in this repository until the hooks
+# directory is made repository-local. `.git/hooks` here is a symlink to
+# ~/Galacticus/gitHooks, inherited from this machine's `init.templateDir`, and
+# installing into it would overwrite the shared Galacticus hooks for every
+# repository on the machine.
+repos:
+  - repo: https://github.com/astral-sh/ruff-pre-commit
+    rev: v0.16.7
+    hooks:
+      - id: ruff-check
+        args: [--fix]
+      - id: ruff-format
+
+  - repo: https://github.com/pre-commit/pre-commit-hooks
+    rev: v5.0.0
+    hooks:
+      - id: trailing-whitespace
+      - id: end-of-file-fixer
+      - id: check-yaml
+      - id: check-toml
+      - id: check-added-large-files
+        args: [--maxkb=512]

dust_compendium-0.1.0/.readthedocs.yaml

@@ -0,0 +1,18 @@
+version: 2
+
+build:
+  os: ubuntu-24.04
+  tools:
+    python: "3.12"
+
+sphinx:
+  configuration: docs/conf.py
+  fail_on_warning: true
+
+python:
+  install:
+    - method: pip
+      path: .
+      extra_requirements:
+        - docs
+        - plots

dust_compendium-0.1.0/CHANGELOG.md

@@ -0,0 +1,58 @@
+# Changelog
+
+## 0.1.0 (unreleased)
+
+First release as a package. Replaces the Perl and Python 2 scripts used for
+[Benson (2018)](https://ui.adsabs.harvard.edu/abs/2018RNAAS...2..188B), which
+are preserved verbatim in the first commit of this repository.
+
+### What it does
+
+- Galaxies are a set of named components, each of which may carry a stellar
+  profile, a dust profile, both or neither, so dust in the spheroid as well as
+  the disk is a change to a configuration rather than to any code.
+- Campaigns are described in YAML and validated on load. Any value written as a
+  range becomes an axis of the tabulation. `validate` reports what a campaign
+  would run without needing Hyperion.
+- `build`, `run` and `collect` each resume from what the last left on disk.
+  `run` has local and Slurm backends; PBS is not carried over.
+- The output format is versioned, self-describing, and reproduces the layout
+  Galacticus reads.
+- `plot` replaces nine near-identical plotting scripts, and drops the gnuplot,
+  cairolatex and pdflatex chain. `compare` measures agreement against a
+  published atlas.
+- The core — geometry, optical depths, the reader and writer, post-processing,
+  plotting and comparison — needs no Hyperion at all.
+
+### Fixed, relative to the original scripts
+
+Six defects, found by testing the ported code against numerical quadrature and
+against Ferrara et al. (1999). Two of them reached published data; see the
+documentation for which tabulations are affected.
+
+- The spheroid cell-mass antiderivative carried the wrong sign below the
+  midplane at the `R = 0`, `|z| = r_s` corner, which the compendium grids reach.
+  For a spheroid the size of the disk this put 54% of its light into two cells
+  holding 0.0003% of its mass.
+- The recorded inclination axis was always a uniform grid, even when the models
+  were solved at other angles, which the Ferrara-matched tabulations were.
+- Cell corners were combined through nested absolute values, so a cell
+  straddling the midplane was wrong and one symmetric about it came out empty.
+- The spheroid antiderivatives were unhandled on the sphere `r = r_s`, returned
+  NaN at small non-zero radius, and lost all precision near `R = r_s`.
+- `spheroidCutOff` was parsed and never applied; spheroids can now be truncated.
+- The source spectrum spanned exactly the dust's frequency range, so a photon
+  drawn at its edge aborted the solve. Hyperion exits zero when it aborts, so
+  those models were recorded as successes; solved models are now checked for the
+  SEDs they were meant to produce.
+
+### Added beyond the original
+
+- `spacing: nested`, which resolves a thin disk inside a large spheroid. The
+  published grid cannot: at a spheroid ten times the disk scale length it leaves
+  a model asked for an optical depth of one with one of zero.
+- `sampling: average`, which puts the profile's true mass in each cell. Point
+  sampling loses 7% of a Jaffe dust profile, and up to 95% in the innermost cell.
+- The extrapolation fit's residual is stored beside its coefficients, so the
+  power-law assumption can be checked rather than trusted.
+- Uncertainties propagate from both the attenuated model and the reference.

dust_compendium-0.1.0/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2018-2026 Andrew Benson
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

dust_compendium-0.1.0/PKG-INFO

@@ -0,0 +1,108 @@
+Metadata-Version: 2.5
+Name: dust-compendium
+Version: 0.1.0
+Summary: Dust attenuation tables for simple galactic geometries, computed with Hyperion
+Project-URL: Homepage, https://github.com/galacticusorg/dust-compendium
+Project-URL: Documentation, https://dust-compendium.readthedocs.io
+Project-URL: Issues, https://github.com/galacticusorg/dust-compendium/issues
+Author-email: Andrew Benson <abenson@carnegiescience.edu>
+License-Expression: MIT
+License-File: LICENSE
+Keywords: astronomy,dust,galacticus,galaxies,radiative transfer
+Classifier: Development Status :: 3 - Alpha
+Classifier: Intended Audience :: Science/Research
+Classifier: Programming Language :: Python :: 3
+Classifier: Topic :: Scientific/Engineering :: Astronomy
+Requires-Python: >=3.11
+Requires-Dist: h5py>=3.8
+Requires-Dist: numpy>=1.24
+Requires-Dist: pydantic>=2.0
+Requires-Dist: pyyaml>=6.0
+Requires-Dist: scipy>=1.10
+Requires-Dist: typer>=0.12
+Provides-Extra: dev
+Requires-Dist: pre-commit>=3.5; extra == 'dev'
+Requires-Dist: pytest-cov; extra == 'dev'
+Requires-Dist: pytest>=7.4; extra == 'dev'
+Requires-Dist: ruff==0.16.7; extra == 'dev'
+Provides-Extra: docs
+Requires-Dist: myst-parser>=2.0; extra == 'docs'
+Requires-Dist: sphinx-autodoc-typehints; extra == 'docs'
+Requires-Dist: sphinx-book-theme; extra == 'docs'
+Requires-Dist: sphinx>=7.0; extra == 'docs'
+Provides-Extra: plots
+Requires-Dist: matplotlib>=3.7; extra == 'plots'
+Description-Content-Type: text/markdown
+
+# Dust Compendium
+
+[![CI](https://github.com/galacticusorg/dust-compendium/actions/workflows/ci.yml/badge.svg)](https://github.com/galacticusorg/dust-compendium/actions/workflows/ci.yml)
+[![Documentation](https://readthedocs.org/projects/dust-compendium/badge/?version=latest)](https://dust-compendium.readthedocs.io)
+
+Dust attenuation tables for simple galactic geometries, computed by running the
+[Hyperion](http://www.hyperion-rt.org) Monte Carlo radiative transfer code over
+a grid of galaxy models.
+
+A tabulation gives the fraction of starlight escaping a galaxy as a function of
+wavelength, inclination, the optical depth of each dusty component, and the
+sizes of the components themselves. [Galacticus](https://github.com/galacticusorg/galacticus)
+reads them through its `dustAttenuationAtlasCompendium` class. The method is
+described in [Benson (2018)](https://ui.adsabs.harvard.edu/abs/2018RNAAS...2..188B).
+
+Full documentation is at [dust-compendium.readthedocs.io](https://dust-compendium.readthedocs.io).
+
+## Installation
+
+```bash
+pip install dust-compendium
+```
+
+That gives the library and the command line tool. Computing *new* tabulations
+additionally needs Hyperion, in two parts, neither of which is a dependency of
+this package:
+
+- its **Python front-end**, which has to come from source for now — the current
+  PyPI release calls `np.string_` when writing a model, which NumPy 2 removed;
+- its **Fortran solvers**, which `pip` does not install at all.
+
+See [the installation notes](https://dust-compendium.readthedocs.io/en/latest/installation.html).
+
+## Using it
+
+```bash
+dust-compendium validate configs/compendium-d03-rv3.1.yaml   # what would this run?
+dust-compendium build    configs/compendium-d03-rv3.1.yaml   # write the model inputs
+dust-compendium run      configs/compendium-d03-rv3.1.yaml   # solve them, here or on Slurm
+dust-compendium collect  configs/compendium-d03-rv3.1.yaml   # assemble the tabulation
+dust-compendium plot     attenuations.hdf5 -q curve -c opticalDepth
+dust-compendium compare  attenuations.hdf5 --atlas /path/to/atlas.hdf5
+```
+
+A campaign is described by a YAML file. Any value written as a range becomes an
+axis of the tabulation; a galaxy is a set of named components, each of which may
+carry stars, dust, both or neither. Each step reads what the one before it left
+on disk, so any of them can be interrupted and resumed.
+
+`configs/` holds the configuration for the published Draine R_V=3.1 tabulation,
+one adding dust to the spheroid, and one matched to Ferrara et al. (1999) for
+validation.
+
+## Validation
+
+Reproducing Ferrara et al. (1999) with matched grains, geometry and grids gives
+a median absolute difference from their published atlas of **0.0034** in disk
+transmission, with 99.9% of entries within 0.02 — as close as the original
+scripts got, and near the 0.005 floor set by the atlas being published to two
+decimal places.
+
+## Provenance
+
+This replaces the Perl and Python 2 scripts written for the 2018 paper, which
+are preserved verbatim in the first commit of this repository. Rewriting them
+turned up several defects, two of which reached published data; those are
+described under
+[Published datasets](https://dust-compendium.readthedocs.io/en/latest/datasets.html).
+
+## License
+
+MIT. See [LICENSE](LICENSE).

dust_compendium-0.1.0/README.md

@@ -0,0 +1,72 @@
+# Dust Compendium
+
+[![CI](https://github.com/galacticusorg/dust-compendium/actions/workflows/ci.yml/badge.svg)](https://github.com/galacticusorg/dust-compendium/actions/workflows/ci.yml)
+[![Documentation](https://readthedocs.org/projects/dust-compendium/badge/?version=latest)](https://dust-compendium.readthedocs.io)
+
+Dust attenuation tables for simple galactic geometries, computed by running the
+[Hyperion](http://www.hyperion-rt.org) Monte Carlo radiative transfer code over
+a grid of galaxy models.
+
+A tabulation gives the fraction of starlight escaping a galaxy as a function of
+wavelength, inclination, the optical depth of each dusty component, and the
+sizes of the components themselves. [Galacticus](https://github.com/galacticusorg/galacticus)
+reads them through its `dustAttenuationAtlasCompendium` class. The method is
+described in [Benson (2018)](https://ui.adsabs.harvard.edu/abs/2018RNAAS...2..188B).
+
+Full documentation is at [dust-compendium.readthedocs.io](https://dust-compendium.readthedocs.io).
+
+## Installation
+
+```bash
+pip install dust-compendium
+```
+
+That gives the library and the command line tool. Computing *new* tabulations
+additionally needs Hyperion, in two parts, neither of which is a dependency of
+this package:
+
+- its **Python front-end**, which has to come from source for now — the current
+  PyPI release calls `np.string_` when writing a model, which NumPy 2 removed;
+- its **Fortran solvers**, which `pip` does not install at all.
+
+See [the installation notes](https://dust-compendium.readthedocs.io/en/latest/installation.html).
+
+## Using it
+
+```bash
+dust-compendium validate configs/compendium-d03-rv3.1.yaml   # what would this run?
+dust-compendium build    configs/compendium-d03-rv3.1.yaml   # write the model inputs
+dust-compendium run      configs/compendium-d03-rv3.1.yaml   # solve them, here or on Slurm
+dust-compendium collect  configs/compendium-d03-rv3.1.yaml   # assemble the tabulation
+dust-compendium plot     attenuations.hdf5 -q curve -c opticalDepth
+dust-compendium compare  attenuations.hdf5 --atlas /path/to/atlas.hdf5
+```
+
+A campaign is described by a YAML file. Any value written as a range becomes an
+axis of the tabulation; a galaxy is a set of named components, each of which may
+carry stars, dust, both or neither. Each step reads what the one before it left
+on disk, so any of them can be interrupted and resumed.
+
+`configs/` holds the configuration for the published Draine R_V=3.1 tabulation,
+one adding dust to the spheroid, and one matched to Ferrara et al. (1999) for
+validation.
+
+## Validation
+
+Reproducing Ferrara et al. (1999) with matched grains, geometry and grids gives
+a median absolute difference from their published atlas of **0.0034** in disk
+transmission, with 99.9% of entries within 0.02 — as close as the original
+scripts got, and near the 0.005 floor set by the atlas being published to two
+decimal places.
+
+## Provenance
+
+This replaces the Perl and Python 2 scripts written for the 2018 paper, which
+are preserved verbatim in the first commit of this repository. Rewriting them
+turned up several defects, two of which reached published data; those are
+described under
+[Published datasets](https://dust-compendium.readthedocs.io/en/latest/datasets.html).
+
+## License
+
+MIT. See [LICENSE](LICENSE).

dust_compendium-0.1.0/configs/compendium-d03-rv3.1.yaml

@@ -0,0 +1,64 @@
+# The published compendium tabulation for Draine (2003) grains with R_V = 3.1.
+#
+# Reproduces the configuration of the original runModels.pl: an exponential disk
+# with a sech^2 vertical structure, dust following the same distribution, and a
+# Hernquist stellar spheroid tabulated over 51 sizes.
+#
+# Lengths are in units of the disk scale length. The original used a physical
+# disk scale radius of 1e-3 Mpc and noted the results do not depend on it.
+#
+# The vertical scale height is from Kregel, van der Kruit & de Grijs (2002;
+# MNRAS; 334; 646).
+label: compendium:exp:sech:Hernquist:hd0.137:hz0.137:dustD03Rv3.1
+description: >-
+  Exponential disk with sech^2 vertical structure and a Hernquist spheroid,
+  with Draine (2003) R_V = 3.1 grains.
+
+dust:
+  file: hyperion-dust-0.1.0/dust_files/d03_3.1_6.0_A.hdf5
+  description: >-
+    Weingartner & Draine (2001; ApJ; 548; 296) with R_V=3.1
+
+geometry:
+  cutOff: 10.0
+  radialCells: 100
+  verticalCells: 100
+  # `published` and `centre` reproduce the original exactly. See the notes in
+  # dustcompendium.grid and dustcompendium.model: they cannot resolve a thin
+  # disk inside a large spheroid, and `nested` with `average` should be
+  # preferred for anything new.
+  spacing: published
+  sampling: centre
+  components:
+    disk:
+      stellar:
+        profile: exponentialDisk
+        verticalStructure: sechSquared
+        scaleRadial: 1.0
+        scaleHeight: 0.137
+      dust:
+        profile: exponentialDisk
+        verticalStructure: sechSquared
+        scaleRadial: 1.0
+        scaleHeight: 0.137
+        opticalDepth:
+          minimum: 0.01
+          maximum: 1.0e4
+          count: 60
+          includeZero: true
+    spheroid:
+      stellar:
+        profile: hernquist
+        scaleRadial:
+          minimum: 1.0e-3
+          maximum: 1.0e2
+          count: 51
+        # The original parsed a spheroid cut off of 10 and then never applied
+        # it; setting it here truncates the profile as was intended.
+        truncation: 10.0
+
+tabulation:
+  wavelengths: {minimum: 0.01, maximum: 3.0, count: 250}
+  inclinations: {minimum: 0.0, maximum: 90.0, count: 46, spacing: linear}
+  photons: 100000
+  seed: -653

dust_compendium-0.1.0/configs/compendium-spheroid-dust.yaml

@@ -0,0 +1,60 @@
+# A tabulation with dust in the spheroid as well as the disk: the extension the
+# published datasets do not cover.
+#
+# Differs from the published configuration in three ways beyond the spheroid
+# dust itself, each deliberate:
+#
+#   * `nested` spacing, because the published grid cannot resolve a thin disk
+#     inside a large spheroid -- at a spheroid ten times the disk scale length
+#     it leaves a model asked for an optical depth of one with one of zero.
+#   * `average` sampling, because a point sample at cell centres loses 7% of the
+#     mass of a cuspy dust profile, and up to 95% in the innermost cell.
+#   * the spheroid truncation is applied, which the original never did.
+#
+# The spheroid's dust is given the same scale radius grid as its stars, so the
+# two are one axis rather than a grid of a size against itself.
+label: compendium:spheroidDust:exp:sech:Hernquist:hd0.137:hz0.137:dustD03Rv3.1
+description: >-
+  As the published Draine (2003) R_V = 3.1 compendium, but with dust in the
+  spheroid as well as the disk, on a grid that resolves both components.
+
+dust:
+  file: hyperion-dust-0.1.0/dust_files/d03_3.1_6.0_A.hdf5
+  description: >-
+    Weingartner & Draine (2001; ApJ; 548; 296) with R_V=3.1
+
+geometry:
+  cutOff: 10.0
+  radialCells: 100
+  verticalCells: 100
+  spacing: nested
+  sampling: average
+  components:
+    disk:
+      stellar:
+        profile: exponentialDisk
+        verticalStructure: sechSquared
+        scaleRadial: 1.0
+        scaleHeight: 0.137
+      dust:
+        profile: exponentialDisk
+        verticalStructure: sechSquared
+        scaleRadial: 1.0
+        scaleHeight: 0.137
+        opticalDepth: {minimum: 0.01, maximum: 1.0e4, count: 60, includeZero: true}
+    spheroid:
+      stellar:
+        profile: hernquist
+        scaleRadial: {minimum: 1.0e-3, maximum: 1.0e2, count: 51}
+        truncation: 10.0
+      dust:
+        profile: hernquist
+        scaleRadial: {minimum: 1.0e-3, maximum: 1.0e2, count: 51}
+        truncation: 10.0
+        opticalDepth: {minimum: 0.01, maximum: 1.0e4, count: 6, includeZero: true}
+
+tabulation:
+  wavelengths: {minimum: 0.01, maximum: 3.0, count: 250}
+  inclinations: {minimum: 0.0, maximum: 90.0, count: 46, spacing: linear}
+  photons: 100000
+  seed: -653

dust_compendium-0.1.0/configs/ferrara1999-milkyway.yaml

@@ -0,0 +1,85 @@
+# Matched to Ferrara et al. (1999; ApJS; 123; 437), for validation.
+#
+# Same grain properties, same galactic geometry, and tabulated at the same
+# wavelengths, inclinations, optical depths and spheroid sizes, so that the
+# result can be compared with their published atlas directly. That atlas ships
+# with Galacticus as
+# `datasets/static/dust/atlasFerrara2000/attenuations_MilkyWay_dustHeightRatio1.0.hdf5`,
+# in the same layout this package writes.
+#
+# Two conversions are worth stating, because the two tabulations do not use the
+# same variables. Ferrara et al. list the spheroid's *effective* radius, and the
+# scale radius is 1.16 times it, so their R_e of 0.1, 0.4 and 1.6 are the scale
+# radii below. And their wavelengths are published in Angstroms.
+#
+# An optical depth of zero is added to their eight, since every model is divided
+# by the one with no dust; it has no counterpart in the atlas and is dropped
+# when comparing.
+label: Ferrara1999:MilkyWay:hz1.0
+description: >-
+  Matched to Ferrara et al. (1999) with Milky Way grains and a dust scale height
+  equal to the stellar one.
+
+dust:
+  # Only ratios of emergent to emitted luminosity are used, and the densities
+  # are set by optical depth, so the absolute opacity cancels and the published
+  # extinction curve can be used as it stands.
+  ferrara: milkyWay
+  referenceOpacity: 1.0
+  description: >-
+    Gordon, Calzetti & Witt (1997; ApJ; 487; 625) Milky Way grains with
+    Henyey-Greenstein scattering, as used by Ferrara et al. (1999).
+
+geometry:
+  cutOff: 6.0
+  radialCells: 60
+  verticalCells: 60
+  spacing: nested
+  sampling: average
+  components:
+    disk:
+      stellar:
+        profile: exponentialDisk
+        verticalStructure: exponential
+        scaleRadial: 1.0
+        scaleHeight: 0.0875
+      dust:
+        profile: exponentialDisk
+        verticalStructure: exponential
+        scaleRadial: 1.0
+        scaleHeight: 0.0875
+        opticalDepth: [0.0, 0.1, 0.5, 1.0, 2.0, 5.0, 10.0, 20.0, 50.0]
+    spheroid:
+      stellar:
+        profile: jaffe
+        # r_s = 1.16 R_e, for R_e of 0.1, 0.4 and 1.6.
+        scaleRadial: [0.116, 0.464, 1.856]
+        # Deliberately left untruncated, against what Ferrara et al. describe.
+        #
+        # They truncate at five effective radii, which with r_s = 1.16 R_e is
+        # 4.3103 scale radii. Applying that makes agreement with their atlas
+        # *worse*, not better. Measured by running two campaigns identical but
+        # for this line, with the same seeds -- the disk arrays came out
+        # bit-for-bit identical, so the difference is the truncation alone --
+        # truncating lowers the spheroid transmission everywhere, by a median of
+        # 0.009 and up to 0.08, growing with optical depth as it removes the
+        # outer and least attenuated light. That is the size of the gap between
+        # this package and the original pipeline when both are compared with the
+        # atlas.
+        #
+        # So the published atlas behaves as though the spheroids were not
+        # truncated at 4.31 scale radii. Either the 1.16 between scale and
+        # effective radius is not the relation Ferrara et al. used, or their
+        # truncation is weaker than stated. Until that is settled, matching
+        # their atlas is the thing this configuration is for, and it is matched
+        # better without. Set a truncation here to explore it.
+        #
+        # The original wrote spheroidCutOff: 5.8, which is 5 times 1.16 rather
+        # than 5 divided by it, and then never applied it.
+
+tabulation:
+  wavelengths: [0.1250, 0.1515, 0.1775, 0.1995, 0.2215, 0.2480, 0.2895,
+                0.3605, 0.4413, 0.5512, 0.6594, 0.8059, 1.2369, 2.1578]
+  inclinations: [9.3, 22.9, 30.0, 40.0, 50.0, 60.0, 70.0, 80.0, 90.0]
+  photons: 50000
+  seed: -653