atomref-proatoms neutral radial profiles v1¶

This notebook inspects the generated v1 release artifacts. It reads radial profiles, cutoff radii, and QA tables from data/; it does not run SCF calculations and does not define the production workflow.

For the density model, see docs/theory.md. For file and column conventions, see docs/data.md.

Data contract¶

The active v1 profile datasets contain neutral atoms only. Dataset scope, method settings, radial grids, QA grids, and density cutoffs are declared in data/profile_datasets.yaml.

In [1]:

from pathlib import Path

import pandas as pd

ROOT = Path.cwd()
if not (ROOT / 'data').exists():
    ROOT = Path.cwd().parents[1]
profiles_root = ROOT / 'data' / 'profiles'
radii_root = ROOT / 'data' / 'radii'
qa_root = ROOT / 'data' / 'qa'

dataset_dirs = sorted(
    path for path in profiles_root.glob('*') if path.is_dir()
)
[
    (path.name, len(pd.read_csv(path / 'profiles.csv')))
    for path in dataset_dirs
]

from pathlib import Path import pandas as pd ROOT = Path.cwd() if not (ROOT / 'data').exists(): ROOT = Path.cwd().parents[1] profiles_root = ROOT / 'data' / 'profiles' radii_root = ROOT / 'data' / 'radii' qa_root = ROOT / 'data' / 'qa' dataset_dirs = sorted( path for path in profiles_root.glob('*') if path.is_dir() ) [ (path.name, len(pd.read_csv(path / 'profiles.csv'))) for path in dataset_dirs ]

Out[1]:

[('pbe0_sfx2c_dyallv4z_h-lr_spherical_v1', 1200),
 ('pbe0_sfx2c_x2cqzvpall_h-rn_spherical_v1', 1200)]

In [2]:

qa_summary = pd.read_csv(qa_root / 'qa_summary.csv')
qa_summary

qa_summary = pd.read_csv(qa_root / 'qa_summary.csv') qa_summary

Out[2]:

	dataset_id	basis_id	state_count	passed_count	failed_count	max_abs_electron_count_error_qa	max_rel_angular_sigma	linear_dependency_warning_count
0	pbe0_sfx2c_x2cqzvpall_h-rn_spherical_v1	x2c-QZVPall	86	86	0	2.287948e-12	2.597422e-15	0
1	pbe0_sfx2c_dyallv4z_h-lr_spherical_v1	dyall-v4z	103	103	0	2.415845e-12	2.684266e-15	55

In [3]:

radii_tables = {
    path.parent.name: pd.read_csv(path) for path in radii_root.glob('*/radii.csv')
}
pd.concat(radii_tables, names=['dataset_id']).reset_index(level=0).head()

radii_tables = { path.parent.name: pd.read_csv(path) for path in radii_root.glob('*/radii.csv') } pd.concat(radii_tables, names=['dataset_id']).reset_index(level=0).head()

Out[3]:

	dataset_id	state_id	symbol	z	charge	electron_count	multiplicity	state_category	state_role	r_iso_0.003_e_bohr3_bohr	r_iso_0.001_e_bohr3_bohr	r_iso_0.0001_e_bohr3_bohr	r_iso_0.003_e_bohr3_angstrom	r_iso_0.001_e_bohr3_angstrom	r_iso_0.0001_e_bohr3_angstrom
0	pbe0_sfx2c_dyallv4z_h-lr_spherical_v1	H_q0_mult2_hund	H	1	0	1	2	nist_ground_state	recommended	2.332423	2.902821	4.128081	1.234265	1.536107	2.184487
1	pbe0_sfx2c_dyallv4z_h-lr_spherical_v1	He_q0_mult1_hund	He	2	0	2	1	nist_ground_state	recommended	2.148013	2.535794	3.374066	1.136679	1.341884	1.785479
2	pbe0_sfx2c_dyallv4z_h-lr_spherical_v1	Li_q0_mult2_hund	Li	3	0	3	2	nist_ground_state	recommended	2.699133	4.168646	6.469639	1.428320	2.205952	3.423585
3	pbe0_sfx2c_dyallv4z_h-lr_spherical_v1	Be_q0_mult1_hund	Be	4	0	4	1	nist_ground_state	recommended	3.362242	4.129170	5.685431	1.779222	2.185063	3.008601
4	pbe0_sfx2c_dyallv4z_h-lr_spherical_v1	B_q0_mult2_hund	B	5	0	5	2	nist_ground_state	recommended	3.257091	3.914727	5.403529	1.723578	2.071584	2.859424

Interpretation notes¶

The profile tables are the canonical radial-density release artifacts. The radii tables are derived from those profiles by interpolation at the declared density cutoffs, and the QA tables summarize release-gate checks for each generated state.