implement custom API

[YuiKasagi]

Since I would like to use the FeH line list by Hargreaves et al. (2010), I implemented a custom API by referencing radis.api.dbmanager and exojax.spec.api.

• The line list contains many unassigned lines, where only the line center wavenumber and intensity are provided. For missing values such as the Einstein A coefficient and upper state J, I assigned either constants or computed estimates.
• Additionally, the line list does not include a partition function. To address this, I used the partition function from ExoMol (via MoLLIST). Specifically, I first load the MdbExoMol for FeH, and then overwrite certain attributes using MdbHargreaves.

Please see the sample code documents/userguide/customapi.ipynb for its usage.
https://github.com/YuiKasagi/exojax/blob/api_mdb_external/documents/userguide/customapi.ipynb

[Copilot]

Pull Request Overview

Adds a custom molecular database API for the FeH line list by Hargreaves et al. (2010), including utilities to fetch, parse, and convert it into ExoMol‐compatible format.

• Introduces MdbHargreaves in customapi.py (with fetching, parsing, intensity/A‐coefficient estimation, and activation logic)
• Adds HargreavesDatabaseManager under dbmanager.py and a test harness (mock_mdbHargreaves) in emulate_mdb.py
• Provides basic unit tests for set_wavenum and line‐strength sums in customapi_test.py

Reviewed Changes

Copilot reviewed 6 out of 6 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
tests/unittests/spec/api/customapi_test.py	New tests for set_wavenum and Hargreaves line strengths
src/exojax/test/emulate_mdb.py	Added mock_mdbHargreaves and related import
src/exojax/spec/qstate.py	Added branch_to_number helper
src/exojax/spec/dbmanager.py	New CustomDatabaseManager and HargreavesDatabaseManager
src/exojax/spec/customapi.py	Implementation of MdbHargreaves and related utilities

Comments suppressed due to low confidence (4)

src/exojax/spec/customapi.py:102

• The database name string is not formatted; it literally contains {molecule}. Use an f-string (e.g. f"Hargreaves2010-{molecule}") or format before passing.

super().__init__( name="Hargreaves2010-{molecule}",

tests/unittests/spec/api/customapi_test.py:24

• There are no unit tests covering convert_to_exomol logic, NaN handling for jlower or A‐coefficient estimation paths. Consider adding tests for those branches.

def test_line_strength_hargreaves():

src/exojax/spec/qstate.py:36

• [nitpick] This error message is generic; consider including the invalid values or their indices to aid debugging, e.g. f"Invalid branch values at positions {invalid_indices}: {branch_str[invalid_indices]}".

raise ValueError("Invalid or NaN branch values found.")

src/exojax/test/emulate_mdb.py:120

• The code uses os.getcwd() and os.path.exists but os is not imported. Add import os at the top of the file.

target_dir = os.getcwd() + "/" + molecule

[YuiKasagi]

I made the following modifications in response to @HajimeKawahara's comments:

• Created data/opacity/FeH_Hargreaves2010.csv (265KB)
• Added the reference in both data/README and the header of the csv file
• Modified the code to load the CSV file by default instead of downloading the line list from the web
• Added comments at the end of customapi.ipynb explaining that the default line strength scaling factor is set to 1/3, as referenced in papers on the Sonora model

[YuiKasagi]

• Updated MdbHargreaves.activate_with_exomol so that it now extends the line list by default, returning a combination of the ExoMol and Hargreaves line lists.

[HajimeKawahara]

Thank you!