ExoGibbs 0.2.2__tar.gz → 0.2.3__tar.gz
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- {exogibbs-0.2.2 → exogibbs-0.2.3}/MANIFEST.in +1 -2
- {exogibbs-0.2.2/src/exogibbs.egg-info → exogibbs-0.2.3}/PKG-INFO +1 -1
- exogibbs-0.2.3/documents/_static/exogibbs_logo.png +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/documents/conf.py +1 -1
- exogibbs-0.2.3/documents/exogibbs/conf.py +36 -0
- exogibbs-0.2.3/documents/index.rst +47 -0
- exogibbs-0.2.3/documents/presets/ykb4.rst +196 -0
- exogibbs-0.2.3/examples/benchmark/clfac_yk.py +73 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/examples/comparisons/comparison_with_ykcode.py +27 -8
- exogibbs-0.2.3/examples/layers/layer_ykb4.py +44 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/src/exogibbs/api/chemistry.py +2 -2
- {exogibbs-0.2.2 → exogibbs-0.2.3}/src/exogibbs/api/equilibrium.py +63 -2
- exogibbs-0.2.3/src/exogibbs/exogibbs_version.py +1 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/src/exogibbs/optimize/minimize.py +61 -80
- {exogibbs-0.2.2 → exogibbs-0.2.3}/src/exogibbs/presets/ykb4.py +22 -11
- {exogibbs-0.2.2 → exogibbs-0.2.3}/src/exogibbs/thermo/stoichiometry.py +2 -2
- {exogibbs-0.2.2 → exogibbs-0.2.3/src/exogibbs.egg-info}/PKG-INFO +1 -1
- {exogibbs-0.2.2 → exogibbs-0.2.3}/src/exogibbs.egg-info/SOURCES.txt +6 -1
- {exogibbs-0.2.2 → exogibbs-0.2.3}/tests/unittests/api/chemsetup_smoke_test.py +1 -1
- {exogibbs-0.2.2 → exogibbs-0.2.3}/tests/unittests/api/equilibrium_api_test.py +2 -2
- exogibbs-0.2.3/tests/unittests/api/equilibrium_profile_test.py +61 -0
- exogibbs-0.2.2/examples/benchmark/clfac_yk.py +0 -61
- exogibbs-0.2.2/src/exogibbs/data/number_of_species.sample.list +0 -1
- exogibbs-0.2.2/src/exogibbs/exogibbs_version.py +0 -1
- {exogibbs-0.2.2 → exogibbs-0.2.3}/.github/workflows/main.yml +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/.gitignore +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/AGENTS.md +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/CLAUDE.md +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/LICENSE +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/README.md +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/documents/Makefile +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/documents/ipynb/chemical_potential.ipynb +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/documents/ipynb/tce_two_species.ipynb +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/examples/GALLERY_HEADER.rst +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/examples/comparisons/comparison_with_hcosystem.py +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/examples/comparisons/comparison_with_hsystem.py +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/examples/data/p10.txt +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/examples/data/yk.list +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/examples/numpyro/example_nuts.py +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/pyproject.toml +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/results/README.md +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/setup.cfg +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/src/exogibbs/ExoGibbs_version.py +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/src/exogibbs/__init__.py +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/src/exogibbs/api/__init__.py +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/src/exogibbs/data/gibbs_matrices.npz +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/src/exogibbs/data/janaf_raw_sample.txt +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/src/exogibbs/data/matrix_v3.dat +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/src/exogibbs/data/molecule_names.csv +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/src/exogibbs/data/testdata.dat +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/src/exogibbs/equilibrium/__init__.py +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/src/exogibbs/equilibrium/gibbs.py +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/src/exogibbs/io/__init__.py +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/src/exogibbs/io/load_data.py +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/src/exogibbs/optimize/__init__.py +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/src/exogibbs/optimize/core.py +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/src/exogibbs/optimize/vjpgibbs.py +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/src/exogibbs/presets/__init__.py +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/src/exogibbs/test/__init__.py +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/src/exogibbs/test/analytic_hcosystem.py +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/src/exogibbs/test/analytic_hsystem.py +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/src/exogibbs/test/generate_gibbs.py +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/src/exogibbs/test/reference_data.py +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/src/exogibbs/thermo/__init__.py +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/src/exogibbs/thermo/providers.py +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/src/exogibbs/utils/__init__.py +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/src/exogibbs/utils/constants.py +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/src/exogibbs.egg-info/dependency_links.txt +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/src/exogibbs.egg-info/requires.txt +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/src/exogibbs.egg-info/top_level.txt +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/tests/unittests/equilibrium/gibbs_test.py +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/tests/unittests/io/formula_matrix_reference_vs_computed_test.py +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/tests/unittests/io/load_data_test.py +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/tests/unittests/optimize/lagrange/hcosystem_test.py +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/tests/unittests/optimize/lagrange/lagrange_test.py +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/tests/unittests/optimize/lagrange/minimize_test.py +0 -0
- {exogibbs-0.2.2 → exogibbs-0.2.3}/tests/unittests/optimize/lagrange/minimize_vmap_test.py +0 -0
|
@@ -2,5 +2,4 @@ include src/exogibbs/data/janaf_raw_sample.txt
|
|
|
2
2
|
include src/exogibbs/data/testdata.dat
|
|
3
3
|
include src/exogibbs/data/matrix_v3.dat
|
|
4
4
|
include src/exogibbs/data/molecule_names.csv
|
|
5
|
-
include src/exogibbs/data/gibbs_matrices.npz
|
|
6
|
-
include src/exogibbs/data/number_of_species.sample.list
|
|
5
|
+
include src/exogibbs/data/gibbs_matrices.npz
|
|
Binary file
|
|
@@ -58,7 +58,7 @@ html_theme = "sphinx_rtd_theme"
|
|
|
58
58
|
# relative to this directory. They are copied after the builtin static files,
|
|
59
59
|
# so a file named "default.css" will overwrite the builtin "default.css".
|
|
60
60
|
html_static_path = ["_static"]
|
|
61
|
-
html_logo = "_static/
|
|
61
|
+
html_logo = "_static/exogibbs_logo.png"
|
|
62
62
|
|
|
63
63
|
# html_theme_options = {
|
|
64
64
|
# 'style_nav_header_background': '#333',
|
|
@@ -0,0 +1,36 @@
|
|
|
1
|
+
# Configuration file for the Sphinx documentation builder.
|
|
2
|
+
#
|
|
3
|
+
# For the full list of built-in configuration values, see the documentation:
|
|
4
|
+
# https://www.sphinx-doc.org/en/master/usage/configuration.html
|
|
5
|
+
|
|
6
|
+
# -- Project information -----------------------------------------------------
|
|
7
|
+
# https://www.sphinx-doc.org/en/master/usage/configuration.html#project-information
|
|
8
|
+
|
|
9
|
+
project = 'exogibbs'
|
|
10
|
+
copyright = '2025, Author'
|
|
11
|
+
author = 'Author'
|
|
12
|
+
|
|
13
|
+
# -- General configuration ---------------------------------------------------
|
|
14
|
+
# https://www.sphinx-doc.org/en/master/usage/configuration.html#general-configuration
|
|
15
|
+
|
|
16
|
+
extensions = [
|
|
17
|
+
'sphinx.ext.autodoc',
|
|
18
|
+
'sphinx.ext.viewcode',
|
|
19
|
+
'sphinx.ext.todo',
|
|
20
|
+
]
|
|
21
|
+
|
|
22
|
+
templates_path = ['_templates']
|
|
23
|
+
exclude_patterns = ['_build', 'Thumbs.db', '.DS_Store']
|
|
24
|
+
|
|
25
|
+
language = 'en'
|
|
26
|
+
|
|
27
|
+
# -- Options for HTML output -------------------------------------------------
|
|
28
|
+
# https://www.sphinx-doc.org/en/master/usage/configuration.html#options-for-html-output
|
|
29
|
+
|
|
30
|
+
html_theme = 'alabaster'
|
|
31
|
+
html_static_path = ['_static']
|
|
32
|
+
|
|
33
|
+
# -- Options for todo extension ----------------------------------------------
|
|
34
|
+
# https://www.sphinx-doc.org/en/master/usage/extensions/todo.html#configuration
|
|
35
|
+
|
|
36
|
+
todo_include_todos = True
|
|
@@ -0,0 +1,47 @@
|
|
|
1
|
+
.. exojax documentation master file, created by
|
|
2
|
+
sphinx-quickstart on Mon Jan 11 14:38:51 2021.
|
|
3
|
+
You can adapt this file completely to your liking, but it should at least
|
|
4
|
+
contain the root `toctree` directive.
|
|
5
|
+
|
|
6
|
+
ExoGibbs
|
|
7
|
+
==================================
|
|
8
|
+
|
|
9
|
+
Version 0.2
|
|
10
|
+
|
|
11
|
+
|:frog:| `DeepWiki for ExoGibbs <https://deepwiki.com/HajimeKawahara/exogibbs>`_
|
|
12
|
+
|
|
13
|
+
|
|
14
|
+
`ExoGibbs <https://github.com/HajimeKawahara/exogibbs>`_ provides an auto-differentiable thermalchemical equilibrium solver
|
|
15
|
+
based on the Gibbs free energy minimization method, powered by `JAX <https://github.com/google/jax>`_.
|
|
16
|
+
|
|
17
|
+
|
|
18
|
+
Contents
|
|
19
|
+
==================================
|
|
20
|
+
|
|
21
|
+
.. toctree::
|
|
22
|
+
:maxdepth: 2
|
|
23
|
+
:caption: PRESETS:
|
|
24
|
+
|
|
25
|
+
presets/ykb4.rst
|
|
26
|
+
|
|
27
|
+
|
|
28
|
+
.. toctree::
|
|
29
|
+
:maxdepth: 1
|
|
30
|
+
:caption: API:
|
|
31
|
+
|
|
32
|
+
exogibbs/exogibbs.rst
|
|
33
|
+
|
|
34
|
+
|
|
35
|
+
|
|
36
|
+
|
|
37
|
+
License & Attribution
|
|
38
|
+
---------------------
|
|
39
|
+
|
|
40
|
+
Copyright 2025-, Contributors
|
|
41
|
+
|
|
42
|
+
- `Hajime Kawahara <http://secondearths.sakura.ne.jp/en/index.html>`_ (@HajimeKawahara, maintainer)
|
|
43
|
+
- `Yui Kawashima <https://sites.google.com/view/yuikawashima/home>`_ (@ykawashima, the reference code author we inherit, the presets ykb4 originally comes from this code.)
|
|
44
|
+
-
|
|
45
|
+
|
|
46
|
+
ExoGibbs is free software made available under the MIT License.
|
|
47
|
+
|
|
@@ -0,0 +1,196 @@
|
|
|
1
|
+
YKB4 Preset
|
|
2
|
+
===========
|
|
3
|
+
|
|
4
|
+
Overview
|
|
5
|
+
--------
|
|
6
|
+
The ykb4 preset bundles a ready-to-use ChemicalSetup based on JANAF-like
|
|
7
|
+
Gibbs tables. It exposes a device-friendly stoichiometric matrix and a
|
|
8
|
+
JAX-differentiable h(T) interpolator for equilibrium calculations.
|
|
9
|
+
|
|
10
|
+
Quick Start
|
|
11
|
+
-----------
|
|
12
|
+
.. code-block:: python
|
|
13
|
+
|
|
14
|
+
from exogibbs.presets.ykb4 import prepare_ykb4_setup
|
|
15
|
+
from exogibbs.api.equilibrium import equilibrium
|
|
16
|
+
|
|
17
|
+
setup = prepare_ykb4_setup()
|
|
18
|
+
T, P = 1500.0, 1.0 # K, bar
|
|
19
|
+
b = setup.b_element_vector_reference # or your own jnp.array([...])
|
|
20
|
+
result = equilibrium(setup, T=T, P=P, b=b)
|
|
21
|
+
print(result.x) # mole fractions (K,)
|
|
22
|
+
|
|
23
|
+
Elements
|
|
24
|
+
--------
|
|
25
|
+
The following elements are included in this preset (including electrons as "e-"):
|
|
26
|
+
|
|
27
|
+
``C, H, He, K, N, Na, O, P, S, Ti, V, e-``
|
|
28
|
+
|
|
29
|
+
The reference elemental abundances (`b_element_vector_ref`) are taken from Asplund, Amarsi & Grevesse (2021).
|
|
30
|
+
|
|
31
|
+
Species (JANAF key)
|
|
32
|
+
-------------------
|
|
33
|
+
The species list corresponds to the ``JANAF`` column in ``molecule_names.csv``.
|
|
34
|
+
|
|
35
|
+
.. code-block:: text
|
|
36
|
+
|
|
37
|
+
000: C1
|
|
38
|
+
001: H2
|
|
39
|
+
002: He1
|
|
40
|
+
003: K1
|
|
41
|
+
004: N2
|
|
42
|
+
005: Na1
|
|
43
|
+
006: O2
|
|
44
|
+
007: Ti1
|
|
45
|
+
008: V1
|
|
46
|
+
009: P1
|
|
47
|
+
010: S1
|
|
48
|
+
011: e1-
|
|
49
|
+
012: C1H1
|
|
50
|
+
013: C1H1N1
|
|
51
|
+
014: C1H1N1O1
|
|
52
|
+
015: C1H1O1
|
|
53
|
+
016: C1H2
|
|
54
|
+
017: C1H2O1
|
|
55
|
+
018: C1H3
|
|
56
|
+
019: C1H4
|
|
57
|
+
020: C1K1N1
|
|
58
|
+
021: C1N1
|
|
59
|
+
022: C1N1Na1
|
|
60
|
+
023: C1N1O1
|
|
61
|
+
024: C1N2(CNN)
|
|
62
|
+
025: C1N2(NCN)
|
|
63
|
+
026: C1O1
|
|
64
|
+
027: C1O2
|
|
65
|
+
028: C2
|
|
66
|
+
029: C2H1
|
|
67
|
+
030: C2H2
|
|
68
|
+
031: C2H4
|
|
69
|
+
032: C2H4O1
|
|
70
|
+
033: C2K2N2
|
|
71
|
+
034: C2N1
|
|
72
|
+
035: C2N2
|
|
73
|
+
036: C2N2Na2
|
|
74
|
+
037: C2O1
|
|
75
|
+
038: C3
|
|
76
|
+
039: C3O2
|
|
77
|
+
040: C4
|
|
78
|
+
041: C4N2
|
|
79
|
+
042: C5
|
|
80
|
+
043: H1
|
|
81
|
+
044: H1K1
|
|
82
|
+
045: H1K1O1
|
|
83
|
+
046: H1N1
|
|
84
|
+
047: H1N1O1
|
|
85
|
+
048: H1N1O2(cis)
|
|
86
|
+
049: H1N1O2(trans)
|
|
87
|
+
050: H1N1O3
|
|
88
|
+
051: H1Na1
|
|
89
|
+
052: H1Na1O1
|
|
90
|
+
053: H1O1
|
|
91
|
+
054: H1O2
|
|
92
|
+
055: H2K2O2
|
|
93
|
+
056: H2N1
|
|
94
|
+
057: H2N2
|
|
95
|
+
058: H2Na2O2
|
|
96
|
+
059: H2O1
|
|
97
|
+
060: H2O2
|
|
98
|
+
061: H3N1
|
|
99
|
+
062: H4N2
|
|
100
|
+
063: K1O1
|
|
101
|
+
064: K2
|
|
102
|
+
065: N1
|
|
103
|
+
066: N1O1
|
|
104
|
+
067: N1O2
|
|
105
|
+
068: N1O3
|
|
106
|
+
069: N2O1
|
|
107
|
+
070: N2O3
|
|
108
|
+
071: N2O4
|
|
109
|
+
072: N2O5
|
|
110
|
+
073: N3
|
|
111
|
+
074: Na1O1
|
|
112
|
+
075: Na2
|
|
113
|
+
076: O1
|
|
114
|
+
077: O3
|
|
115
|
+
078: N1V1
|
|
116
|
+
079: O1Ti1
|
|
117
|
+
080: O1V1
|
|
118
|
+
081: O2Ti1
|
|
119
|
+
082: O2V1
|
|
120
|
+
083: C1H1P1
|
|
121
|
+
084: C1O1S1
|
|
122
|
+
085: C1P1
|
|
123
|
+
086: C1S1
|
|
124
|
+
087: C1S2
|
|
125
|
+
088: H1P1
|
|
126
|
+
089: H1S1
|
|
127
|
+
090: H2O4S1
|
|
128
|
+
091: H2P1
|
|
129
|
+
092: H2S1
|
|
130
|
+
093: H3P1
|
|
131
|
+
094: K2O4S1
|
|
132
|
+
095: N1P1
|
|
133
|
+
096: N1S1
|
|
134
|
+
097: Na2O4S1
|
|
135
|
+
098: O1P1
|
|
136
|
+
099: O1S1
|
|
137
|
+
100: O1S2
|
|
138
|
+
101: O2P1
|
|
139
|
+
102: O2S1
|
|
140
|
+
103: O3S1
|
|
141
|
+
104: O6P4
|
|
142
|
+
105: O10P4
|
|
143
|
+
106: P1S1
|
|
144
|
+
107: P2
|
|
145
|
+
108: P4
|
|
146
|
+
109: P4S3
|
|
147
|
+
110: S2
|
|
148
|
+
111: S3
|
|
149
|
+
112: S4
|
|
150
|
+
113: S5
|
|
151
|
+
114: S6
|
|
152
|
+
115: S7
|
|
153
|
+
116: S8
|
|
154
|
+
117: C1+
|
|
155
|
+
118: C1-
|
|
156
|
+
119: C1H1+
|
|
157
|
+
120: C1H1O1+
|
|
158
|
+
121: C1N1+
|
|
159
|
+
122: C1N1-
|
|
160
|
+
123: C1O2-
|
|
161
|
+
124: C2-
|
|
162
|
+
125: H1+
|
|
163
|
+
126: H1-
|
|
164
|
+
127: H1K1O1+
|
|
165
|
+
128: H1Na1O1+
|
|
166
|
+
129: H1O1+
|
|
167
|
+
130: H1O1-
|
|
168
|
+
131: H2+
|
|
169
|
+
132: H2-
|
|
170
|
+
133: H3O1+
|
|
171
|
+
134: He1+
|
|
172
|
+
135: K1+
|
|
173
|
+
136: K1-
|
|
174
|
+
137: K1O1-
|
|
175
|
+
138: N1+
|
|
176
|
+
139: N1-
|
|
177
|
+
140: N1O1+
|
|
178
|
+
141: N1O2-
|
|
179
|
+
142: N2+
|
|
180
|
+
143: N2-
|
|
181
|
+
144: N2O1+
|
|
182
|
+
145: Na1+
|
|
183
|
+
146: Na1-
|
|
184
|
+
147: Na1O1-
|
|
185
|
+
148: O1+
|
|
186
|
+
149: O1-
|
|
187
|
+
150: O2+
|
|
188
|
+
151: O2-
|
|
189
|
+
152: P1+
|
|
190
|
+
153: P1-
|
|
191
|
+
154: S1+
|
|
192
|
+
155: S1-
|
|
193
|
+
156: Ti1+
|
|
194
|
+
157: Ti1-
|
|
195
|
+
158: V1+
|
|
196
|
+
159: V1-
|
|
@@ -0,0 +1,73 @@
|
|
|
1
|
+
"""
|
|
2
|
+
Validation of Gibbs Minimization Against ykawashima's B4 code
|
|
3
|
+
============================================================
|
|
4
|
+
|
|
5
|
+
This example demonstrates and validates the ExoGibbs thermochemical equilibrium
|
|
6
|
+
solver against the code by ykawashima when she was at B4.
|
|
7
|
+
|
|
8
|
+
Updated to use the high-level API: exogibbs.api.equilibrium.equilibrium.
|
|
9
|
+
"""
|
|
10
|
+
from exogibbs.presets.ykb4 import prepare_ykb4_setup
|
|
11
|
+
import jax.numpy as jnp
|
|
12
|
+
from exogibbs.api.chemistry import ThermoState
|
|
13
|
+
from exogibbs.optimize.minimize import minimize_gibbs_core
|
|
14
|
+
|
|
15
|
+
from exogibbs.optimize.core import compute_ln_normalized_pressure
|
|
16
|
+
|
|
17
|
+
from jax import config
|
|
18
|
+
|
|
19
|
+
config.update("jax_enable_x64", True)
|
|
20
|
+
|
|
21
|
+
#chemical setup
|
|
22
|
+
chem = prepare_ykb4_setup()
|
|
23
|
+
|
|
24
|
+
# Thermodynamic conditions
|
|
25
|
+
temperature = 500.0 # K
|
|
26
|
+
P = 10.0 # bar
|
|
27
|
+
Pref = 1.0 # bar, reference pressure
|
|
28
|
+
ln_normalized_pressure = compute_ln_normalized_pressure(P, Pref)
|
|
29
|
+
|
|
30
|
+
# Initial guess for log number densities
|
|
31
|
+
ln_nk = jnp.zeros(chem.formula_matrix.shape[1]) # log(n_species)
|
|
32
|
+
ln_ntot = 0.0 # log(total number density)
|
|
33
|
+
|
|
34
|
+
# old reference elemental abundance b from yk's sample number densities
|
|
35
|
+
b_old_ref = jnp.array(
|
|
36
|
+
[
|
|
37
|
+
4.8774824e-04,
|
|
38
|
+
1.6749767e00,
|
|
39
|
+
1.6143440e-01,
|
|
40
|
+
2.5438149e-07,
|
|
41
|
+
1.3435642e-04,
|
|
42
|
+
3.9624806e-06,
|
|
43
|
+
9.7356725e-04,
|
|
44
|
+
5.7690579e-07,
|
|
45
|
+
3.0653933e-05,
|
|
46
|
+
1.6687756e-07,
|
|
47
|
+
1.9870969e-08,
|
|
48
|
+
0.0000000e00,
|
|
49
|
+
]
|
|
50
|
+
)
|
|
51
|
+
# ThermoState instance
|
|
52
|
+
thermo_state = ThermoState(
|
|
53
|
+
temperature=temperature,
|
|
54
|
+
ln_normalized_pressure=ln_normalized_pressure,
|
|
55
|
+
b_element_vector=b_old_ref,
|
|
56
|
+
)
|
|
57
|
+
|
|
58
|
+
# Convergence criteria
|
|
59
|
+
epsilon_crit = 1e-11
|
|
60
|
+
max_iter = 1000
|
|
61
|
+
|
|
62
|
+
ln_nk_result, _, icount = minimize_gibbs_core(
|
|
63
|
+
thermo_state,
|
|
64
|
+
ln_nk,
|
|
65
|
+
ln_ntot,
|
|
66
|
+
chem.formula_matrix,
|
|
67
|
+
chem.hvector_func,
|
|
68
|
+
epsilon_crit=epsilon_crit,
|
|
69
|
+
max_iter=max_iter,
|
|
70
|
+
)
|
|
71
|
+
nk_result = jnp.exp(ln_nk_result)
|
|
72
|
+
print("icount", icount)
|
|
73
|
+
#702 for clfac = 0.1, 347 for _cea_lambda
|
|
@@ -11,7 +11,7 @@ Updated to use the high-level API: exogibbs.api.equilibrium.equilibrium.
|
|
|
11
11
|
from exogibbs.presets.ykb4 import prepare_ykb4_setup
|
|
12
12
|
from exogibbs.api.equilibrium import equilibrium, EquilibriumOptions
|
|
13
13
|
import numpy as np
|
|
14
|
-
|
|
14
|
+
import jax.numpy as jnp
|
|
15
15
|
from jax import config
|
|
16
16
|
|
|
17
17
|
config.update("jax_enable_x64", True)
|
|
@@ -22,9 +22,28 @@ temperature = 500.0 # K
|
|
|
22
22
|
P = 10.0 # bar
|
|
23
23
|
Pref = 1.0 # bar, reference pressure
|
|
24
24
|
|
|
25
|
-
#chemical setup
|
|
25
|
+
# chemical setup
|
|
26
26
|
chem = prepare_ykb4_setup()
|
|
27
27
|
|
|
28
|
+
# old reference elemental abundance b from yk's sample number densities
|
|
29
|
+
b_old_ref = jnp.array(
|
|
30
|
+
[
|
|
31
|
+
4.8774824e-04,
|
|
32
|
+
1.6749767e00,
|
|
33
|
+
1.6143440e-01,
|
|
34
|
+
2.5438149e-07,
|
|
35
|
+
1.3435642e-04,
|
|
36
|
+
3.9624806e-06,
|
|
37
|
+
9.7356725e-04,
|
|
38
|
+
5.7690579e-07,
|
|
39
|
+
3.0653933e-05,
|
|
40
|
+
1.6687756e-07,
|
|
41
|
+
1.9870969e-08,
|
|
42
|
+
0.0000000e00,
|
|
43
|
+
]
|
|
44
|
+
)
|
|
45
|
+
|
|
46
|
+
|
|
28
47
|
##############################################################################
|
|
29
48
|
# Solve equilibrium via high-level API
|
|
30
49
|
# ------------------------------------
|
|
@@ -33,7 +52,7 @@ res = equilibrium(
|
|
|
33
52
|
chem,
|
|
34
53
|
T=temperature,
|
|
35
54
|
P=P,
|
|
36
|
-
b=
|
|
55
|
+
b=b_old_ref,
|
|
37
56
|
Pref=Pref,
|
|
38
57
|
options=opts,
|
|
39
58
|
)
|
|
@@ -43,21 +62,22 @@ nk_result = res.n
|
|
|
43
62
|
|
|
44
63
|
# load yk's results for 10 bar
|
|
45
64
|
dat = np.loadtxt("../data/p10.txt", delimiter=",")
|
|
46
|
-
mask = dat > 1.
|
|
65
|
+
mask = dat > 1.0e-14
|
|
47
66
|
mask_nk_result = nk_result[mask]
|
|
48
67
|
mask_dat = dat[mask]
|
|
49
68
|
|
|
50
|
-
res = mask_nk_result/mask_dat - 1.0
|
|
51
|
-
print(res,"diff for n>1.e-14")
|
|
69
|
+
res = mask_nk_result / mask_dat - 1.0
|
|
70
|
+
print(res, "diff for n>1.e-14")
|
|
52
71
|
assert np.max(np.abs(res)) < 0.051
|
|
53
72
|
# 8/9/2025
|
|
54
|
-
#[-0.00163185 -0.00163185 0.02571018 -0.00203837 -0.05069541 -0.00163185
|
|
73
|
+
# [-0.00163185 -0.00163185 0.02571018 -0.00203837 -0.05069541 -0.00163185
|
|
55
74
|
# -0.00481986 -0.00420364 -0.00161074 -0.00163182 -0.00163185 -0.00163183
|
|
56
75
|
# -0.00163184 -0.00163178 -0.00163185 -0.00163184]
|
|
57
76
|
|
|
58
77
|
|
|
59
78
|
ind = np.arange(len(nk_result))
|
|
60
79
|
import matplotlib.pyplot as plt
|
|
80
|
+
|
|
61
81
|
plt.plot(ind, nk_result, "+", label="ExoGibbs")
|
|
62
82
|
plt.plot(ind, dat, ".", alpha=0.5, label="yk B4 code")
|
|
63
83
|
plt.xlabel("Species Index")
|
|
@@ -66,4 +86,3 @@ plt.yscale("log")
|
|
|
66
86
|
plt.legend()
|
|
67
87
|
plt.grid()
|
|
68
88
|
plt.show()
|
|
69
|
-
|
|
@@ -0,0 +1,44 @@
|
|
|
1
|
+
"""
|
|
2
|
+
Validation of Gibbs Minimization for Layered Systems
|
|
3
|
+
===================================================================================
|
|
4
|
+
|
|
5
|
+
This example demonstrates and validates the ExoGibbs thermochemical equilibrium for the layered system
|
|
6
|
+
|
|
7
|
+
Updated to use the high-level API: exogibbs.api.equilibrium.equilibrium.
|
|
8
|
+
"""
|
|
9
|
+
|
|
10
|
+
from exogibbs.presets.ykb4 import prepare_ykb4_setup
|
|
11
|
+
from exogibbs.api.equilibrium import equilibrium_profile, EquilibriumOptions
|
|
12
|
+
import numpy as np
|
|
13
|
+
import jax.numpy as jnp
|
|
14
|
+
from jax import config
|
|
15
|
+
|
|
16
|
+
config.update("jax_enable_x64", True)
|
|
17
|
+
|
|
18
|
+
|
|
19
|
+
# Thermodynamic conditions
|
|
20
|
+
Pref = 1.0 # bar, reference pressure
|
|
21
|
+
Parr = jnp.logspace(-8, 2, 10)/Pref # bar
|
|
22
|
+
Tarr = 1500*Parr**0.1
|
|
23
|
+
|
|
24
|
+
|
|
25
|
+
#chemical setup
|
|
26
|
+
chem = prepare_ykb4_setup()
|
|
27
|
+
|
|
28
|
+
##############################################################################
|
|
29
|
+
# Solve equilibrium via high-level API
|
|
30
|
+
# ------------------------------------
|
|
31
|
+
opts = EquilibriumOptions(epsilon_crit=1e-11, max_iter=1000)
|
|
32
|
+
|
|
33
|
+
res = equilibrium_profile(
|
|
34
|
+
chem,
|
|
35
|
+
Tarr,
|
|
36
|
+
Parr,
|
|
37
|
+
chem.b_element_vector_reference,
|
|
38
|
+
Pref=Pref,
|
|
39
|
+
options=opts,
|
|
40
|
+
)
|
|
41
|
+
|
|
42
|
+
##############################################################################
|
|
43
|
+
nk_result = res.n
|
|
44
|
+
print(nk_result)
|
|
@@ -46,7 +46,7 @@ class ChemicalSetup:
|
|
|
46
46
|
hvector_func : Callable[[float|Array], Array]
|
|
47
47
|
h(T) used by the optimizer (JAX-differentiable).
|
|
48
48
|
|
|
49
|
-
|
|
49
|
+
elements : Optional[tuple[str, ...]]
|
|
50
50
|
Element symbols (E,) if available.
|
|
51
51
|
species : Optional[tuple[str, ...]]
|
|
52
52
|
Species names (K,) if available.
|
|
@@ -59,7 +59,7 @@ class ChemicalSetup:
|
|
|
59
59
|
hvector_func: Callable[[Union[float, Array]], Array]
|
|
60
60
|
|
|
61
61
|
# Optional metadata (host-side; NOT traced)
|
|
62
|
-
|
|
62
|
+
elements: Optional[Tuple[str, ...]] = None
|
|
63
63
|
species: Optional[Tuple[str, ...]] = None
|
|
64
64
|
b_element_vector_reference: Optional["np.ndarray"] = None # host-side
|
|
65
65
|
metadata: Optional[Mapping[str, str]] = None
|
|
@@ -60,7 +60,9 @@ class EquilibriumResult:
|
|
|
60
60
|
|
|
61
61
|
# Make this dataclass a JAX pytree (so vmap/jit can pass it around)
|
|
62
62
|
def tree_flatten(self):
|
|
63
|
-
|
|
63
|
+
# Avoid coercing to jnp.asarray here to keep compatibility with
|
|
64
|
+
# transformation-time abstract values (e.g., vmap/jit tracing).
|
|
65
|
+
children = (self.ln_n, self.n, self.x, self.ntot)
|
|
64
66
|
aux = (self.iterations, self.metadata)
|
|
65
67
|
return children, aux
|
|
66
68
|
|
|
@@ -148,4 +150,63 @@ def equilibrium(
|
|
|
148
150
|
ln_n=ln_n, n=n, x=x, ntot=ntot, iterations=None, metadata=None
|
|
149
151
|
)
|
|
150
152
|
|
|
151
|
-
|
|
153
|
+
def equilibrium_profile(
|
|
154
|
+
setup: ChemicalSetup,
|
|
155
|
+
T: Array,
|
|
156
|
+
P: Array,
|
|
157
|
+
b: Array,
|
|
158
|
+
*,
|
|
159
|
+
Pref: float = 1.0,
|
|
160
|
+
options: Optional[EquilibriumOptions] = None,
|
|
161
|
+
) -> EquilibriumResult:
|
|
162
|
+
"""Vectorized equilibrium along a 1D T/P profile (layers).
|
|
163
|
+
|
|
164
|
+
This computes equilibrium independently for each (T[i], P[i]) pair while
|
|
165
|
+
keeping the elemental abundances ``b`` fixed across layers.
|
|
166
|
+
|
|
167
|
+
Args:
|
|
168
|
+
setup: ChemicalSetup with formula matrix and hvector_func(T).
|
|
169
|
+
T: Temperatures, shape (N,).
|
|
170
|
+
P: Pressures, shape (N,).
|
|
171
|
+
b: Elemental abundances, shape (E,), shared across layers.
|
|
172
|
+
Pref: Reference pressure (bar).
|
|
173
|
+
options: Solver options.
|
|
174
|
+
|
|
175
|
+
Returns:
|
|
176
|
+
Batched EquilibriumResult with fields stacked over the leading dimension N:
|
|
177
|
+
- ln_n: (N, K)
|
|
178
|
+
- n: (N, K)
|
|
179
|
+
- x: (N, K)
|
|
180
|
+
- ntot: (N,)
|
|
181
|
+
"""
|
|
182
|
+
T = jnp.asarray(T)
|
|
183
|
+
P = jnp.asarray(P)
|
|
184
|
+
if T.ndim != 1 or P.ndim != 1:
|
|
185
|
+
raise ValueError("T and P must be 1D arrays of equal length.")
|
|
186
|
+
if T.shape[0] != P.shape[0]:
|
|
187
|
+
raise ValueError("T and P must have the same length.")
|
|
188
|
+
if b.ndim != 1:
|
|
189
|
+
raise ValueError("b must be a 1D array shared across layers.")
|
|
190
|
+
|
|
191
|
+
# Vectorize over T and P; keep setup and b static. Pass Pref/options as kwargs.
|
|
192
|
+
layer_fn = jax.vmap(
|
|
193
|
+
lambda Ti, Pi: equilibrium(
|
|
194
|
+
setup,
|
|
195
|
+
Ti,
|
|
196
|
+
Pi,
|
|
197
|
+
b,
|
|
198
|
+
Pref=Pref,
|
|
199
|
+
options=options,
|
|
200
|
+
),
|
|
201
|
+
in_axes=(0, 0),
|
|
202
|
+
)
|
|
203
|
+
return layer_fn(T, P)
|
|
204
|
+
|
|
205
|
+
|
|
206
|
+
__all__ = [
|
|
207
|
+
"equilibrium",
|
|
208
|
+
"equilibrium_profile",
|
|
209
|
+
"EquilibriumOptions",
|
|
210
|
+
"EquilibriumInit",
|
|
211
|
+
"EquilibriumResult",
|
|
212
|
+
]
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
__version__ = "0.2.3"
|
|
@@ -191,7 +191,9 @@ def minimize_gibbs_core(
|
|
|
191
191
|
return ln_nk, ln_tot, counter
|
|
192
192
|
|
|
193
193
|
|
|
194
|
-
|
|
194
|
+
# Only function and scalar options are non-differentiable.
|
|
195
|
+
# Array-valued args (init vectors, matrices) must NOT be in nondiff_argnums
|
|
196
|
+
# to avoid UnexpectedTracerError under vmap/jit.
|
|
195
197
|
def minimize_gibbs(
|
|
196
198
|
state: ThermoState,
|
|
197
199
|
ln_nk_init: jnp.ndarray,
|
|
@@ -215,87 +217,66 @@ def minimize_gibbs(
|
|
|
215
217
|
Returns:
|
|
216
218
|
Final log number of species vector (n_species,).
|
|
217
219
|
"""
|
|
220
|
+
# Define an inner custom_vjp function that captures static arrays
|
|
221
|
+
# (ln_nk_init, ln_ntot_init, formula_matrix) to avoid passing them as
|
|
222
|
+
# arguments, which can become JAX Tracers under vmap/jit.
|
|
223
|
+
@custom_vjp
|
|
224
|
+
def _solve(inner_state: ThermoState) -> jnp.ndarray:
|
|
225
|
+
ln_nk, _, _ = minimize_gibbs_core(
|
|
226
|
+
inner_state,
|
|
227
|
+
ln_nk_init,
|
|
228
|
+
ln_ntot_init,
|
|
229
|
+
formula_matrix,
|
|
230
|
+
hvector_func,
|
|
231
|
+
epsilon_crit,
|
|
232
|
+
max_iter,
|
|
233
|
+
)
|
|
234
|
+
return ln_nk
|
|
218
235
|
|
|
219
|
-
|
|
220
|
-
|
|
221
|
-
|
|
222
|
-
|
|
223
|
-
|
|
224
|
-
|
|
225
|
-
|
|
226
|
-
|
|
227
|
-
|
|
228
|
-
|
|
229
|
-
|
|
230
|
-
|
|
231
|
-
|
|
232
|
-
|
|
233
|
-
|
|
234
|
-
|
|
235
|
-
|
|
236
|
-
|
|
237
|
-
|
|
238
|
-
|
|
239
|
-
|
|
240
|
-
)
|
|
241
|
-
|
|
242
|
-
|
|
243
|
-
|
|
244
|
-
|
|
245
|
-
|
|
246
|
-
|
|
247
|
-
|
|
248
|
-
|
|
249
|
-
|
|
250
|
-
|
|
251
|
-
|
|
236
|
+
def _solve_fwd(inner_state: ThermoState):
|
|
237
|
+
ln_nk, ln_ntot, _ = minimize_gibbs_core(
|
|
238
|
+
inner_state,
|
|
239
|
+
ln_nk_init,
|
|
240
|
+
ln_ntot_init,
|
|
241
|
+
formula_matrix,
|
|
242
|
+
hvector_func,
|
|
243
|
+
epsilon_crit,
|
|
244
|
+
max_iter,
|
|
245
|
+
)
|
|
246
|
+
dfunc = jacrev(hvector_func)
|
|
247
|
+
hdot = dfunc(inner_state.temperature)
|
|
248
|
+
residuals = (ln_nk, hdot, inner_state.b_element_vector, ln_ntot)
|
|
249
|
+
return ln_nk, residuals
|
|
250
|
+
|
|
251
|
+
def _solve_bwd(res, g):
|
|
252
|
+
ln_nk, hdot, b_element_vector, ln_ntot = res
|
|
253
|
+
|
|
254
|
+
nk = jnp.exp(ln_nk)
|
|
255
|
+
ntot_result = jnp.exp(ln_ntot)
|
|
256
|
+
|
|
257
|
+
Bmatrix = _A_diagn_At(nk, formula_matrix)
|
|
258
|
+
c, lower = cho_factor(Bmatrix)
|
|
259
|
+
alpha = cho_solve((c, lower), formula_matrix @ g)
|
|
260
|
+
beta = cho_solve((c, lower), b_element_vector)
|
|
261
|
+
beta_dot_b_element = jnp.vdot(beta, b_element_vector)
|
|
262
|
+
|
|
263
|
+
cot_T = vjp_temperature(
|
|
264
|
+
g,
|
|
265
|
+
nk,
|
|
266
|
+
formula_matrix,
|
|
267
|
+
hdot,
|
|
268
|
+
alpha,
|
|
269
|
+
beta,
|
|
270
|
+
b_element_vector,
|
|
271
|
+
beta_dot_b_element,
|
|
272
|
+
)
|
|
273
|
+
cot_P = vjp_pressure(g, ntot_result, alpha, b_element_vector, beta_dot_b_element)
|
|
274
|
+
cot_b = vjp_elements(g, alpha, beta, b_element_vector, beta_dot_b_element)
|
|
275
|
+
return (ThermoState(jnp.asarray(cot_T), jnp.asarray(cot_P), cot_b),)
|
|
252
276
|
|
|
253
|
-
|
|
254
|
-
return ln_nk, residuals
|
|
277
|
+
_solve.defvjp(_solve_fwd, _solve_bwd)
|
|
255
278
|
|
|
279
|
+
return _solve(state)
|
|
256
280
|
|
|
257
|
-
def minimize_gibbs_bwd(
|
|
258
|
-
ln_nk_init,
|
|
259
|
-
ln_ntot_init,
|
|
260
|
-
formula_matrix,
|
|
261
|
-
hvector_func,
|
|
262
|
-
epsilon_crit,
|
|
263
|
-
max_iter,
|
|
264
|
-
res,
|
|
265
|
-
g,
|
|
266
|
-
):
|
|
267
281
|
|
|
268
|
-
|
|
269
|
-
|
|
270
|
-
# Guard against underflow and ensure strictly positive species for SPD matrix
|
|
271
|
-
# dtype = ln_nk.dtype
|
|
272
|
-
#tiny = jnp.finfo(dtype).tiny
|
|
273
|
-
# nk = jnp.maximum(jnp.exp(ln_nk), tiny)
|
|
274
|
-
# ntot_result = jnp.maximum(jnp.exp(ln_ntot), tiny)
|
|
275
|
-
nk = jnp.exp(ln_nk)
|
|
276
|
-
ntot_result = jnp.exp(ln_ntot)
|
|
277
|
-
|
|
278
|
-
# solves the linear systems
|
|
279
|
-
Bmatrix = _A_diagn_At(nk, formula_matrix)
|
|
280
|
-
|
|
281
|
-
# Add a tiny diagonal jitter to stabilize Cholesky for near-singular cases
|
|
282
|
-
# size_e = Bmatrix.shape[0]
|
|
283
|
-
# Scale jitter relative to matrix magnitude
|
|
284
|
-
#rel = jnp.maximum(jnp.linalg.norm(Bmatrix, ord=jnp.inf), 1.0)
|
|
285
|
-
#jitter = jnp.asarray(1e-12, dtype=Bmatrix.dtype) * rel
|
|
286
|
-
#Bmatrix = Bmatrix + jitter * jnp.eye(size_e, dtype=Bmatrix.dtype)
|
|
287
|
-
|
|
288
|
-
c, lower = cho_factor(Bmatrix)
|
|
289
|
-
alpha = cho_solve((c, lower), formula_matrix @ g)
|
|
290
|
-
beta = cho_solve((c, lower), b_element_vector)
|
|
291
|
-
beta_dot_b_element = jnp.vdot(beta, b_element_vector)
|
|
292
|
-
|
|
293
|
-
#VJPs
|
|
294
|
-
cot_T = vjp_temperature(g, nk, formula_matrix, hdot, alpha, beta, b_element_vector, beta_dot_b_element)
|
|
295
|
-
cot_P = vjp_pressure(g, ntot_result, alpha, b_element_vector, beta_dot_b_element)
|
|
296
|
-
cot_b = vjp_elements(g, alpha, beta, b_element_vector, beta_dot_b_element)
|
|
297
|
-
|
|
298
|
-
return (ThermoState(jnp.asarray(cot_T), jnp.asarray(cot_P), cot_b),)
|
|
299
|
-
|
|
300
|
-
|
|
301
|
-
minimize_gibbs.defvjp(minimize_gibbs_fwd, minimize_gibbs_bwd)
|
|
282
|
+
# Note: custom_vjp is defined inside minimize_gibbs to capture static arrays.
|
|
@@ -11,10 +11,10 @@ import jax.numpy as jnp
|
|
|
11
11
|
import jax
|
|
12
12
|
|
|
13
13
|
JANAF_GIBBS_MATRICES_YKB4 = "gibbs_matrices.npz"
|
|
14
|
-
NUMBER_OF_SPECIES_SAMPLE = "number_of_species.sample.list" # this is a sample file of number_of_species for testing
|
|
15
14
|
MOLNAME_YKB4 = "molecule_names.csv"
|
|
16
15
|
JANAF_NAME_KEY = "JANAF" # key for JANAF name in the molecule names file
|
|
17
16
|
|
|
17
|
+
|
|
18
18
|
def load_molname_ykb4() -> pd.DataFrame:
|
|
19
19
|
"""Load the YKB4 molecular names table.
|
|
20
20
|
|
|
@@ -30,7 +30,6 @@ def load_molname_ykb4() -> pd.DataFrame:
|
|
|
30
30
|
return df_molname
|
|
31
31
|
|
|
32
32
|
|
|
33
|
-
|
|
34
33
|
def prepare_ykb4_setup() -> ChemicalSetup:
|
|
35
34
|
"""
|
|
36
35
|
Prepare a JAX-friendly ChemicalSetup from JANAF-like Gibbs matrices.
|
|
@@ -43,16 +42,28 @@ def prepare_ykb4_setup() -> ChemicalSetup:
|
|
|
43
42
|
"""
|
|
44
43
|
# Species / formula matrix (fixed)
|
|
45
44
|
df_molname = load_molname_ykb4()
|
|
46
|
-
formula_matrix_np,
|
|
45
|
+
formula_matrix_np, elements, species = build_formula_matrix(df_molname)
|
|
47
46
|
# Keep the matrix fixed as requested, but move to device
|
|
48
47
|
formula_matrix = jnp.asarray(formula_matrix_np)
|
|
49
|
-
|
|
50
|
-
# Reference elemental abundance b from
|
|
51
|
-
|
|
52
|
-
|
|
53
|
-
|
|
54
|
-
|
|
55
|
-
|
|
48
|
+
|
|
49
|
+
# Reference elemental solar abundance b from AAG21 (from exojax.utils.zsol import nsol)
|
|
50
|
+
# AAG21 = Asplund, M., Amarsi, A. M., & Grevesse, N. 2021, arXiv:2105.01661
|
|
51
|
+
b_element_vector_ref = jnp.array(
|
|
52
|
+
[
|
|
53
|
+
2.6627135e-04,
|
|
54
|
+
9.2326087e-01,
|
|
55
|
+
7.5739846e-02,
|
|
56
|
+
1.0847369e-07,
|
|
57
|
+
6.2420098e-05,
|
|
58
|
+
1.5322316e-06,
|
|
59
|
+
4.5219361e-04,
|
|
60
|
+
2.3731458e-07,
|
|
61
|
+
1.2170949e-05,
|
|
62
|
+
8.6163716e-08,
|
|
63
|
+
7.3337216e-09,
|
|
64
|
+
0.0,
|
|
65
|
+
]
|
|
66
|
+
)
|
|
56
67
|
# Gibbs matrices -> (molecules, T_table, mu_table, grid_lens)
|
|
57
68
|
path = get_data_filepath(JANAF_GIBBS_MATRICES_YKB4)
|
|
58
69
|
gibbs_matrices = np.load(path, allow_pickle=True)["arr_0"].item()
|
|
@@ -87,7 +98,7 @@ def prepare_ykb4_setup() -> ChemicalSetup:
|
|
|
87
98
|
return ChemicalSetup(
|
|
88
99
|
formula_matrix=formula_matrix,
|
|
89
100
|
hvector_func=hvector_func_jit,
|
|
90
|
-
|
|
101
|
+
elements=tuple(elements) if elements is not None else None,
|
|
91
102
|
species=tuple(species) if species is not None else None,
|
|
92
103
|
b_element_vector_reference=b_element_vector_ref,
|
|
93
104
|
metadata={"source": "JANAF"},
|
|
@@ -171,8 +171,8 @@ def build_formula_matrix(
|
|
|
171
171
|
# Parse using the clean base (never use raw for parsing)
|
|
172
172
|
parsed_list = [parse_formula_with_charge(s) for s in species_clean]
|
|
173
173
|
|
|
174
|
-
|
|
175
|
-
elements = sorted(
|
|
174
|
+
all_elements = set().union(*parsed_list) if parsed_list else set()
|
|
175
|
+
elements = sorted(all_elements) if element_order is None else list(element_order)
|
|
176
176
|
|
|
177
177
|
A = np.zeros((len(elements), len(species_clean)), dtype=np.int64)
|
|
178
178
|
elem_index = {e: i for i, e in enumerate(elements)}
|
|
@@ -8,8 +8,12 @@ pyproject.toml
|
|
|
8
8
|
.github/workflows/main.yml
|
|
9
9
|
documents/Makefile
|
|
10
10
|
documents/conf.py
|
|
11
|
+
documents/index.rst
|
|
12
|
+
documents/_static/exogibbs_logo.png
|
|
13
|
+
documents/exogibbs/conf.py
|
|
11
14
|
documents/ipynb/chemical_potential.ipynb
|
|
12
15
|
documents/ipynb/tce_two_species.ipynb
|
|
16
|
+
documents/presets/ykb4.rst
|
|
13
17
|
examples/GALLERY_HEADER.rst
|
|
14
18
|
examples/benchmark/clfac_yk.py
|
|
15
19
|
examples/comparisons/comparison_with_hcosystem.py
|
|
@@ -17,6 +21,7 @@ examples/comparisons/comparison_with_hsystem.py
|
|
|
17
21
|
examples/comparisons/comparison_with_ykcode.py
|
|
18
22
|
examples/data/p10.txt
|
|
19
23
|
examples/data/yk.list
|
|
24
|
+
examples/layers/layer_ykb4.py
|
|
20
25
|
examples/numpyro/example_nuts.py
|
|
21
26
|
results/README.md
|
|
22
27
|
src/exogibbs/ExoGibbs_version.py
|
|
@@ -34,7 +39,6 @@ src/exogibbs/data/gibbs_matrices.npz
|
|
|
34
39
|
src/exogibbs/data/janaf_raw_sample.txt
|
|
35
40
|
src/exogibbs/data/matrix_v3.dat
|
|
36
41
|
src/exogibbs/data/molecule_names.csv
|
|
37
|
-
src/exogibbs/data/number_of_species.sample.list
|
|
38
42
|
src/exogibbs/data/testdata.dat
|
|
39
43
|
src/exogibbs/equilibrium/__init__.py
|
|
40
44
|
src/exogibbs/equilibrium/gibbs.py
|
|
@@ -58,6 +62,7 @@ src/exogibbs/utils/__init__.py
|
|
|
58
62
|
src/exogibbs/utils/constants.py
|
|
59
63
|
tests/unittests/api/chemsetup_smoke_test.py
|
|
60
64
|
tests/unittests/api/equilibrium_api_test.py
|
|
65
|
+
tests/unittests/api/equilibrium_profile_test.py
|
|
61
66
|
tests/unittests/equilibrium/gibbs_test.py
|
|
62
67
|
tests/unittests/io/formula_matrix_reference_vs_computed_test.py
|
|
63
68
|
tests/unittests/io/load_data_test.py
|
|
@@ -22,7 +22,7 @@ def test_prepare_ykb4_setup_basic():
|
|
|
22
22
|
assert E > 0 and K > 0
|
|
23
23
|
|
|
24
24
|
# optional metadata
|
|
25
|
-
assert (setup.
|
|
25
|
+
assert (setup.elements is None) or isinstance(setup.elements, tuple)
|
|
26
26
|
assert (setup.species is None) or isinstance(setup.species, tuple)
|
|
27
27
|
assert (setup.b_element_vector_reference is None) or isinstance(
|
|
28
28
|
setup.b_element_vector_reference, (np.ndarray, jnp.ndarray)
|
|
@@ -41,9 +41,9 @@ from exogibbs.api.equilibrium import EquilibriumInit, EquilibriumOptions
|
|
|
41
41
|
|
|
42
42
|
class FakeSetup:
|
|
43
43
|
"""Minimal stand-in for ChemicalSetup for interface testing."""
|
|
44
|
-
def __init__(self, A,
|
|
44
|
+
def __init__(self, A, elements=None):
|
|
45
45
|
self.formula_matrix = A
|
|
46
|
-
self.
|
|
46
|
+
self.elements = elements
|
|
47
47
|
|
|
48
48
|
def hvector_func(self, T):
|
|
49
49
|
# Shape must be (K,), but values are irrelevant for these tests
|
|
@@ -0,0 +1,61 @@
|
|
|
1
|
+
import importlib
|
|
2
|
+
import jax.numpy as jnp
|
|
3
|
+
|
|
4
|
+
import exogibbs.api.equilibrium as eqmod
|
|
5
|
+
from exogibbs.api.equilibrium import EquilibriumOptions
|
|
6
|
+
|
|
7
|
+
|
|
8
|
+
class FakeSetup:
|
|
9
|
+
"""Minimal stand-in for ChemicalSetup for vectorized API testing."""
|
|
10
|
+
|
|
11
|
+
def __init__(self, A):
|
|
12
|
+
self.formula_matrix = A
|
|
13
|
+
|
|
14
|
+
def hvector_func(self, T):
|
|
15
|
+
K = self.formula_matrix.shape[1]
|
|
16
|
+
return jnp.zeros((K,))
|
|
17
|
+
|
|
18
|
+
|
|
19
|
+
def test_equilibrium_profile_shapes_and_values(monkeypatch):
|
|
20
|
+
"""Simple happy-path test for equilibrium_profile with a stubbed minimizer.
|
|
21
|
+
|
|
22
|
+
Ensures batched shapes are correct and values match the stub behavior
|
|
23
|
+
(ln_n = 0 => n = 1, x uniform, ntot = K) across all layers.
|
|
24
|
+
"""
|
|
25
|
+
E, K, N = 2, 3, 4
|
|
26
|
+
A = jnp.array([[1, 0, 1], [0, 1, 0]], dtype=jnp.float64)
|
|
27
|
+
setup = FakeSetup(A)
|
|
28
|
+
|
|
29
|
+
def stub_minimize_gibbs(state, ln_nk0, ln_ntot0, A_in, hfunc, **kwargs):
|
|
30
|
+
# Basic sanity checks on inputs
|
|
31
|
+
assert A_in.shape == A.shape
|
|
32
|
+
assert ln_nk0.shape == (K,)
|
|
33
|
+
assert ln_ntot0.shape == ()
|
|
34
|
+
return jnp.zeros((K,), dtype=jnp.result_type(ln_nk0, A_in.dtype))
|
|
35
|
+
|
|
36
|
+
monkeypatch.setattr(
|
|
37
|
+
"exogibbs.api.equilibrium.minimize_gibbs", stub_minimize_gibbs, raising=True
|
|
38
|
+
)
|
|
39
|
+
|
|
40
|
+
# Profile inputs (N layers)
|
|
41
|
+
T = jnp.linspace(1000.0, 2000.0, N)
|
|
42
|
+
P = jnp.linspace(0.1, 1.0, N)
|
|
43
|
+
b = jnp.array([1.0, 2.0], dtype=jnp.float64)
|
|
44
|
+
|
|
45
|
+
out = eqmod.equilibrium_profile(
|
|
46
|
+
setup, T, P, b, options=EquilibriumOptions(epsilon_crit=1e-11, max_iter=50)
|
|
47
|
+
)
|
|
48
|
+
|
|
49
|
+
# Batched shapes
|
|
50
|
+
assert out.ln_n.shape == (N, K)
|
|
51
|
+
assert out.n.shape == (N, K)
|
|
52
|
+
assert out.x.shape == (N, K)
|
|
53
|
+
assert out.ntot.shape == (N,)
|
|
54
|
+
|
|
55
|
+
# Stub behavior reflected across layers
|
|
56
|
+
assert jnp.allclose(out.ln_n, 0.0)
|
|
57
|
+
assert jnp.allclose(out.n, 1.0)
|
|
58
|
+
assert jnp.allclose(out.x, jnp.ones((N, K)) / K)
|
|
59
|
+
assert jnp.allclose(out.ntot, K)
|
|
60
|
+
assert jnp.allclose(jnp.sum(out.x, axis=1), 1.0)
|
|
61
|
+
|
|
@@ -1,61 +0,0 @@
|
|
|
1
|
-
"""
|
|
2
|
-
Validation of Gibbs Minimization Against ykawashima's B4 code
|
|
3
|
-
============================================================
|
|
4
|
-
|
|
5
|
-
This example demonstrates and validates the ExoGibbs thermochemical equilibrium
|
|
6
|
-
solver against the code by ykawashima when she was at B4.
|
|
7
|
-
|
|
8
|
-
Updated to use the high-level API: exogibbs.api.equilibrium.equilibrium.
|
|
9
|
-
"""
|
|
10
|
-
from exogibbs.presets.ykb4 import prepare_ykb4_setup
|
|
11
|
-
from exogibbs.api.equilibrium import equilibrium, EquilibriumOptions
|
|
12
|
-
import numpy as np
|
|
13
|
-
import jax.numpy as jnp
|
|
14
|
-
|
|
15
|
-
from jax import config
|
|
16
|
-
|
|
17
|
-
config.update("jax_enable_x64", True)
|
|
18
|
-
|
|
19
|
-
#chemical setup
|
|
20
|
-
chem = prepare_ykb4_setup()
|
|
21
|
-
|
|
22
|
-
# Thermodynamic conditions
|
|
23
|
-
temperature = 500.0 # K
|
|
24
|
-
P = 10.0 # bar
|
|
25
|
-
Pref = 1.0 # bar, reference pressure
|
|
26
|
-
opts = EquilibriumOptions(epsilon_crit=1e-11, max_iter=1000)
|
|
27
|
-
res = equilibrium(
|
|
28
|
-
chem,
|
|
29
|
-
T=temperature,
|
|
30
|
-
P=P,
|
|
31
|
-
b=chem.b_element_vector_reference,
|
|
32
|
-
Pref=Pref,
|
|
33
|
-
options=opts,
|
|
34
|
-
)
|
|
35
|
-
nk_result = res.n
|
|
36
|
-
|
|
37
|
-
# load yk's results for 10 bar
|
|
38
|
-
dat = np.loadtxt("../data/p10.txt", delimiter=",")
|
|
39
|
-
|
|
40
|
-
mask = dat > 1.e-14
|
|
41
|
-
mask_nk_result = nk_result[mask]
|
|
42
|
-
mask_dat = dat[mask]
|
|
43
|
-
|
|
44
|
-
res = mask_nk_result/mask_dat - 1.0
|
|
45
|
-
print(res,"diff for n>1.e-14")
|
|
46
|
-
assert np.max(np.abs(res)) < 0.051
|
|
47
|
-
# 8/9/2025, 0.1
|
|
48
|
-
#[-0.00163185 -0.00163185 0.02571018 -0.00203837 -0.05069541 -0.00163185
|
|
49
|
-
# -0.00481986 -0.00420364 -0.00161074 -0.00163182 -0.00163185 -0.00163183
|
|
50
|
-
# -0.00163184 -0.00163178 -0.00163185 -0.00163184]
|
|
51
|
-
|
|
52
|
-
import matplotlib.pyplot as plt
|
|
53
|
-
plt.plot(nk_result, "+", label="ExoGibbs")
|
|
54
|
-
plt.plot(dat, ".", alpha=0.5, label="yk B4 code")
|
|
55
|
-
plt.xlabel("Species Index")
|
|
56
|
-
plt.ylabel("Number (log scale)")
|
|
57
|
-
plt.yscale("log")
|
|
58
|
-
plt.legend()
|
|
59
|
-
plt.grid()
|
|
60
|
-
plt.show()
|
|
61
|
-
|
|
@@ -1 +0,0 @@
|
|
|
1
|
-
4.8771844088134374E-004,0.83748832109661187,0.16143440365984060,2.4842019364382504E-007,6.7166286367535330E-005,3.9565195261941036E-006,4.8675855005686264E-004,1.6588401261773752E-007,1.8380646975815483E-008,5.6349288368902435E-007,3.0627107689945438E-005,0.0,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,4.9677424258960761E-010,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0
|
|
@@ -1 +0,0 @@
|
|
|
1
|
-
__version__ = "0.2.2"
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
{exogibbs-0.2.2 → exogibbs-0.2.3}/tests/unittests/io/formula_matrix_reference_vs_computed_test.py
RENAMED
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|