Quantum chemistry add-ons to Density Functional Theory

[EBB2675]

No description provided.

[coveralls]

Pull Request Test Coverage Report for Build 19224631192

Details

• 302 of 331 (91.24%) changed or added relevant lines in 7 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+0.9%) to 83.59%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
src/nomad_simulations/schema_packages/utils/libxc/build.py	15	20	75.0%
src/nomad_simulations/schema_packages/numerical_settings.py	8	14	57.14%
src/nomad_simulations/schema_packages/model_method.py	100	107	93.46%
src/nomad_simulations/schema_packages/utils/libxc/expand.py	63	74	85.14%

Totals
Change from base Build 19170712288:	0.9%
Covered Lines:	6882
Relevant Lines:	8233

---

💛 - Coveralls

[JFRudzinski]

What do you mean exactly with Point 2?

[EBB2675]

@ladinesa @ndaelman-hu @JFRudzinski please take a look at LibXC canonicalization and the new DFT architecture

I added XCComponent for a single LibXC identified kernel, and then XCFunctional as the whole XC approximation in accordance with the community jargon and theory.

As you can understand, there are multiple ways this canonicalization can take place. this is one complete implementation. I do not have any personal preferences. I am ready to adopt to whatever you suggest 👍

[JFRudzinski]

@ladinesa @ndaelman-hu @JFRudzinski please take a look at LibXC canonicalization and the new DFT architecture

I added XCComponent for a single LibXC identified kernel, and then XCFunctional as the whole XC approximation in accordance with the community jargon and theory.

As you can understand, there are multiple ways this canonicalization can take place. this is one complete implementation. I do not have any personal preferences. I am ready to adopt to whatever you suggest 👍

For me this looks very nice, but I have no insight into the domain details here. I expect @ndaelman-hu will give you more specific feedback/confirmation

[ndaelman-hu]

Like the DFT mapping additions. A few quantities that need to be added.
Also comments on the structure, suggesting some reorganization. I think that if these are addressed, you can go ahead and merge. No need for another review.

Disclaimer: didn't check the actual code implementations too much (would first need to understand the exact workflow). You could ask the copilot reviewer.

[ndaelman-hu]

Please import this.

[ndaelman-hu]

I think these are also present in expand.py, correct? Unfortunately, there isn't much of a way for testing these kinds of mappings [1]. You basically generated them twice and check for typos.

Personally, I feel that you can safely remove this test.

[1]: I guess that with a very granular format specification, you could run that to impose compliance. But I think that's overkill in this case.

[ndaelman-hu]

@EBB2675 , at this stage, I'd also request feedback from the co-pilot reviewer.

[Copilot]

Pull Request Overview

This PR adds quantum chemistry extensions to Density Functional Theory (DFT) capabilities by introducing LibXC-based XC functional normalization and additional quantum chemistry model methods. The changes include a comprehensive XC functional expansion system, new model method classes for implicit solvation, dispersion corrections, and relativistic treatments, along with extensive test coverage.

Key Changes

• Implements LibXC-based XC functional canonicalization with registry lookup and expansion capabilities
• Adds new model method classes: ImplicitSolvationModel, ExplicitDispersionModel, RelativityModel, and CoreElectronTreatment
• Refactors DFT class to use new XCFunctional/XCComponent schema with automatic Jacob's ladder derivation

Reviewed Changes

Copilot reviewed 9 out of 10 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
tests/utils/test_libxc.py	Comprehensive test suite for LibXC functional expansion and DFT normalization
tests/test_model_method.py	Extended tests for DFT class with XC component handling and parameter derivation
src/nomad_simulations/schema_packages/utils/libxc/registry.py	LibXC registry lookup functionality with label and ID-based queries
src/nomad_simulations/schema_packages/utils/libxc/expand.py	XC functional expansion logic with alias tables and fallback mechanisms
src/nomad_simulations/schema_packages/utils/libxc/build.py	XCComponent factory functions for creating components from LibXC data
src/nomad_simulations/schema_packages/utils/libxc/init.py	Module interface exposing build functions
src/nomad_simulations/schema_packages/utils/libxc/README.md	Documentation for LibXC canonicalization pipeline and supported formats
src/nomad_simulations/schema_packages/numerical_settings.py	Added FrozenCore, IntegralDecomposition, and OrbitalLocalization settings classes
src/nomad_simulations/schema_packages/model_method.py	Major refactor with new quantum chemistry model classes and updated DFT implementation

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[Copilot]

Pull Request Overview

Copilot reviewed 9 out of 10 changed files in this pull request and generated 3 comments.

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[ndaelman-hu]

Some structural remarks in the schema. These should be addressed.

Normally, I'd give this "Comment", but since no further iteration is needed and you typically follow up well with the comments, I'll add it as "Approve". Note that some older comments are also still open.

[ndaelman-hu]

@ladinesa

[JosePizarro3]

Hey, @EBB2675, I will check this later next week. It looks good, but there are a couple of sections I will move out of ModelMethodElectronic, as these are only relevant for quantum chemistry simulations.

Furthermore, I am not sure about the role here of dielectrics and so. In cond mat, these are typically something else (I am guessing, I dont know. I would like to check this, or if you know, please let me know :-) ).

[Copilot]

Pull Request Overview

Copilot reviewed 9 out of 10 changed files in this pull request and generated 1 comment.

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[Copilot]

Pull Request Overview

Copilot reviewed 9 out of 10 changed files in this pull request and generated 3 comments.

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[kubanmar]

Looks good to me.

[JosePizarro3]

All good, just some minor thing about a couple of sub-sections in ModelMethodElectronic that I would move out.

[EBB2675]

Hi @ladinesa @JosePizarro3 @JFRudzinski, @Bernadette-Mohr , @ndaelman-hu , @kubanmar ,

This PR is ready for rigorous DFT parser testing.

• Please skim utils/libxc/README.md for the canonicalization pipeline overview.
• Added extensive tests in tests/utils/test_libxc.py.
• Removed a few out-of-scope items (e.g., CoreElectronTreatment) to keep scope manageable; will handle separately.
• In ModelMethodElectronic, I dropped QC-specific SubSections as suggested by @JosePizarro3, they will instead be attached to the parent method through ModelMethod.contributions when and if needed. I added tests for these as well.
• Finally, i replaced the legacy ModelMethodElectronic.is_spin_polarized with a clearer determinant:

    determinant = Quantity(
        type=MEnum('unrestricted', 'restricted', 'restricted-open-shell'),
        description="""
        The spin-coupling form of the determinant used for the
        self-consistent field (SCF) or Kohn-Sham calculation.

        - **restricted**  (RHF/RKS): α and β electrons share the same spatial orbitals  
        - **unrestricted** (UHF/UKS): α and β orbitals are optimized independently  
        - **restricted-open-shell** (ROHF/ROKS): closed-shell core with spin-unpaired electrons
        sharing spatial orbitals in the open-shell manifold
        """,
    )

If you prefer a different name for determinant, please let me know 🏃‍♀️

[ndaelman-hu]

Hey, great work on this PR! Structurally it looks good overall.
Given how mature it is already, I just went over all my old comments. I noticed that I missed some of your replies under all the e-mail notifications.

I followed up on a few comments. I just have a few suggestions for minor tweaks.
I also opened a bunch of issues where you suggested. Feel free to add to them.

Edit: I added a bunch of our comments via VS Code's GitLens, but they appear as new messages now without their thread... My apologies for the mess.

[ndaelman-hu]

This comment has not yet been applied. It's not a big deal, but would mean 1 field less when parsing.