Add atom centered basis set #132

EBB2675 · 2024-09-25T08:41:56Z

Fixes #130

coveralls · 2024-09-25T12:06:59Z

Pull Request Test Coverage Report for Build 12278300834

Details

69 of 110 (62.73%) changed or added relevant lines in 4 files are covered.
1 unchanged line in 1 file lost coverage.
Overall coverage decreased (-0.8%) to 79.376%

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
src/nomad_simulations/schema_packages/numerical_settings.py	22	27	81.48%
src/nomad_simulations/schema_packages/basis_set.py	16	29	55.17%
src/nomad_simulations/schema_packages/model_method.py	29	52	55.77%

Files with Coverage Reduction	New Missed Lines	%
src/nomad_simulations/schema_packages/numerical_settings.py	1	76.07%

Totals
Change from base Build 11890140918:	-0.8%
Covered Lines:	2136
Relevant Lines:	2691

💛 - Coveralls

ndaelman-hu · 2024-09-30T08:50:20Z

src/nomad_simulations/schema_packages/basis_set.py

+        type=np.float32,
+        shape=['n_primitive'],
+        description="""
+        List of contraction coefficients corresponding to the exponents.


I would elaborate a bit more here, keeping also non-experts in mind.

ndaelman-hu · 2024-09-30T08:52:22Z

src/nomad_simulations/schema_packages/basis_set.py

+    main_basis_set = Quantity(
+        type=str,
+        description="""
+        Name of the main basis set.
+        """,
+    )


What's the distinction between main and aux? If it's method routines, you can use BasisSetContainer to distinguish them.

ndaelman-hu · 2024-09-30T08:54:40Z

src/nomad_simulations/schema_packages/basis_set.py

+        type=AtomsState,
+        shape=['*'],
+        description="""
+        References to the `AtomsState` sections that define the atoms this basis set applies to.


Connected to my comment in the issue, you might also want a map to a ModelSystem.
For simplicity sake, I'd try to use only 1 type of linker when possible. Convenience can be added in other ways.

ndaelman-hu · 2024-09-30T08:55:23Z

src/nomad_simulations/schema_packages/basis_set.py

+    aux_c_basis_set = Quantity(
+        type=str,
+        description="""
+        AuxC type of basis set.


I think it would also be good to elaborate on the description here.

ndaelman-hu · 2024-09-30T08:55:58Z

src/nomad_simulations/schema_packages/basis_set.py

+    function_type = Quantity(
+        type=MEnum('S', 'P', 'D', 'F', 'G', 'H', 'I', 'J'),
+        description="""
+        the angular momentum of the shell to be added.
+        """,
+    )


Looks more like something for AtomsState. @JosePizarro3

I think @EBB2675 wants to store the angular momentum here simply. I'd say if a string is enough, go ahead. Tho I will keep consistent with notation: small letters type=MEnum('s', 'p', 'd', 'f')

Now, are there cases where one needs to go to h, i, j? I must admit g "might" be ok in some excited state calculations, but the others... I am not sure.

Tho I will keep consistent with notation: small letters type=MEnum('s', 'p', 'd', 'f')

Shouldn't we use capital letters for the total momentum?

@EBB2675 Is this orbital angular momentum? Could you specify in the description, pls?

Ah sorry, if it is total ang mom yes, capital letters is good. Tho we need to decide where this kind of information should be living in a more consistent way.

We can leave that discussion out of this pr.

It's the type of basis functions. Let me improve all the Quantity and class descriptions, it will be clearer

ndaelman-hu · 2024-09-30T08:58:01Z

src/nomad_simulations/schema_packages/basis_set.py



 class AtomCenteredBasisSet(BasisSetComponent):
    """
    Defines an atom-centered basis set.
    """

+    main_basis_set = Quantity(
+        type=str,


I think str is fine for now, we'll likely want to constrain the choice here down the line.

…basis-set

ndaelman-hu · 2024-10-01T11:45:54Z

The basis set template works as follows:
BasisSetContainer bundles all the information per property (e.g. density, wavefunction, auxiliary grids, etc.).* Different properties go together in the list ModelMethod.numerical_settings.
The actual basis set data are stored under BasisSetContainer.basis_set_components.

In the case of LAPW, this would mean:

ModelMethod.numerical_settings = [
  ...,
  density: BasisSetContainer
  └ basis_set_components
      ├── 1 x plane-wave basis set
      └── n x muffin-tin regions -> references `AtomsState`
          └── l_max x l-channels
              ├── orbitals
              └── local orbitals,
  auxialary density: BasisSetContainer
  └ basis_set_components
      ├── 1 x plane-wave basis set
      └── m x muffin-tin regions -> references `AtomsState`
          └── l_max' x l-channels
              ├── orbitals
              └── local orbitals
]

@EBB2675 I'm curious what you think about splitting the various atom-centered auxiliary basis sets up like this.

*: Since we don't have an electronic structure yet, I Ieft it under a TODO. In the old schema, I used a str, which is fine in the beginning.

JFRudzinski · 2024-10-04T10:37:09Z

src/nomad_simulations/schema_packages/basis_set.py

+    """
+
+    basis_set_data = Quantity(
+        type=JSON,  # Use JSON to store basis set information, including atom references


@EBB2675 did you see my comment about this? it came up in my GH notifications, but I am not sure where it was stored. The comment was about using a repeating sub-section here instead of JSON

@JFRudzinski I have just seen it (but only in the notification E-mail)

Alright, im on it 👍

…ction coefficients

JFRudzinski

Just a few small comments, I will come by to discuss...

JFRudzinski · 2024-11-21T14:25:28Z

src/nomad_simulations/schema_packages/numerical_settings.py

+        """,
+    )
+
+    integration_thresh = Quantity(


I would opt for integration_threshold, you are not really saving much with the abbreviation

JFRudzinski · 2024-11-21T14:29:39Z

src/nomad_simulations/schema_packages/numerical_settings.py

        """,
    )

-    weights = Quantity(
+    integration_rule = Quantity(


I would switch to Enum now and add a few most common integration rules with references in the description using the table like:

integrator_type = Quantity( type=MEnum( 'brownian', 'conjugant_gradient', 'langevin_goga', 'langevin_schneider', 'leap_frog', 'rRESPA_multitimescale', 'velocity_verlet', 'langevin_leap_frog', ), shape=[], description=""" Name of the integrator. Allowed values are: | Integrator Name | Description | | ---------------------- | ----------------------------------------- | | `"langevin_goga"` | N. Goga, A. J. Rzepiela, A. H. de Vries, S. J. Marrink, and H. J. C. Berendsen, [J. Chem. Theory Comput. **8**, 3637 (2012)] (https://doi.org/10.1021/ct3000876) | | `"langevin_schneider"` | T. Schneider and E. Stoll, [Phys. Rev. B **17**, 1302](https://doi.org/10.1103/PhysRevB.17.1302) | | `"leap_frog"` | R.W. Hockney, S.P. Goel, and J. Eastwood, [J. Comp. Phys. **14**, 148 (1974)](https://doi.org/10.1016/0021-9991(74)90010-2) | | `"velocity_verlet"` | W.C. Swope, H.C. Andersen, P.H. Berens, and K.R. Wilson, [J. Chem. Phys. **76**, 637 (1982)](https://doi.org/10.1063/1.442716) | | `"rRESPA_multitimescale"` | M. Tuckerman, B. J. Berne, and G. J. Martyna [J. Chem. Phys. **97**, 1990 (1992)](https://doi.org/10.1063/1.463137) | | `"langevin_leap_frog"` | J.A. Izaguirre, C.R. Sweet, and V.S. Pande [Pac Symp Biocomput. **15**, 240-251 (2010)](https://doi.org/10.1142/9789814295291_0026) | """, )

Since these are very established mathematical methods, I might just put a wikipedia link or something more easily accessible along with a more persistent reference

JFRudzinski · 2024-11-21T14:37:20Z

src/nomad_simulations/schema_packages/numerical_settings.py

-        Weight of each point. A value smaller than 1, typically indicates a symmetry operation that was
-        applied to the mesh. This quantity is equivalent to `multiplicities`:
+        Accuracy threshold for integration grid.
+        GRIDTHR in Molpro.


This is interesting...maybe making this into a table of aliases/related quantities would be useful as a general format in the description

""" Related code-specific quantities: | Quantity | Program | Relation | ---------------------- | ----------------------------------------- | | GRIDTHR | Molpro | = integration_thresh | BFCut | Orca | = integration_thresh """,

JFRudzinski · 2024-11-21T14:42:23Z

src/nomad_simulations/schema_packages/basis_set.py

+            'cabs',  # complementary auxiliary basis set
+        ),
+        description="""
+        The role of the basis set.


I think you should add a short description for each of these within a table, to make sure the usage is clear

JFRudzinski · 2024-11-21T14:47:19Z

src/nomad_simulations/schema_packages/numerical_settings.py

+        """,
+    )
+
+    pruning = Quantity(


I'm wondering how general this is, i.e., how understandable or useful it is outside of the context of your particular methods.

…ined orbital types

EBB2675 added 3 commits September 24, 2024 14:26

add main and auxiliary basis set quantities to AtomCenteredBasisSet

d436e95

formatting

022c1fa

Merge branch 'develop' into coupled_cluster

bf1f6f4

EBB2675 added the new feature New feature or request label Sep 25, 2024

add a draft for AtomCenteredFunction

af88624

EBB2675 added 2 commits September 25, 2024 15:09

add a quantity for number of primitives

0a9d98b

add atoms_state reference to AtomCenteredBasisSet

a204ec4

ndaelman-hu assigned EBB2675 Sep 30, 2024

ndaelman-hu reviewed Sep 30, 2024

View reviewed changes

Merge remote-tracking branch 'origin/develop' into 130-atom-centered-…

a325753

…basis-set

assign JSON format to basis set

8866ab7

JFRudzinski reviewed Oct 4, 2024

View reviewed changes

EBB2675 added 16 commits October 8, 2024 11:38

add type and auxiliary_type quantities to AtomCenteredBasisSet

c7b078a

reformatted basis_set.py

1208e88

add NAO and point charges to basis set types

fbc13f9

add cECPs and pointcharges to AtomCenteredBasisSet

ddab789

fix point charge Quantity type

8c6ae1a

a bit of a cleanup

c94e995

merge develop

a007e1e

move GTOIntegralDecomposition to NumericalSettings

62de243

Merge branch 'develop' into 130-atom-centered-basis-set

c12adac

merge develop

52d6b9a

add Mesh, NumericalIntegration and MolecularHamiltonianSubTerms

646430c

minor adjustments to Mesh and NumericalIntegration

cca109f

add integration_thresh and weight_cutoff to NumericalIntegration

42f4f38

check whether n_primitive matches the lengths of exponents and contra…

0f92eb9

…ction coefficients

add tests for AtomCenteredBasisSet and AtomCenteredFunction

e8fb5ef

add tests for Mesh and NumericalIntegration

4826f0c

EBB2675 added 3 commits November 20, 2024 16:56

modify Mesh

c5141ab

MEnum for MolecularHamiltonianContributions

23d7615

remove contributions

9ef13ea

JFRudzinski approved these changes Nov 21, 2024

View reviewed changes

EBB2675 added 7 commits November 28, 2024 16:44

add a normalizer function for the AtomCenteredFunction to handle comb…

80c8b64

…ined orbital types

fix test_basis_set.py

4ac1076

add OrbitalLocalization to numerical_settings.py

10e782c

add method to LocalCorrelation

5c15e97

add total_charge and total_spin to ModelSystem

2fd6398

add a simple HF class

c9d6118

a placeholder for MO and LCAO

816c7ba

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add atom centered basis set #132

Add atom centered basis set #132

EBB2675 commented Sep 25, 2024 •

edited

Loading

coveralls commented Sep 25, 2024 •

edited

Loading

ndaelman-hu Sep 30, 2024

ndaelman-hu Sep 30, 2024

ndaelman-hu Sep 30, 2024

ndaelman-hu Sep 30, 2024

ndaelman-hu Sep 30, 2024

JosePizarro3 Oct 1, 2024

ndaelman-hu Oct 1, 2024

ndaelman-hu Oct 1, 2024

JosePizarro3 Oct 1, 2024

EBB2675 Oct 1, 2024

ndaelman-hu Sep 30, 2024

ndaelman-hu commented Oct 1, 2024

JFRudzinski Oct 4, 2024

EBB2675 Oct 4, 2024

JFRudzinski left a comment

JFRudzinski Nov 21, 2024

JFRudzinski Nov 21, 2024

JFRudzinski Nov 21, 2024

JFRudzinski Nov 21, 2024

JFRudzinski Nov 21, 2024

Add atom centered basis set #132

Are you sure you want to change the base?

Add atom centered basis set #132

Conversation

EBB2675 commented Sep 25, 2024 • edited Loading

coveralls commented Sep 25, 2024 • edited Loading

Pull Request Test Coverage Report for Build 12278300834

Details

💛 - Coveralls

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

ndaelman-hu commented Oct 1, 2024

Choose a reason for hiding this comment

Choose a reason for hiding this comment

JFRudzinski left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

EBB2675 commented Sep 25, 2024 •

edited

Loading

coveralls commented Sep 25, 2024 •

edited

Loading