Common LCAO NLDF Infrastructure =============================== The molecular NLDF implementation is divided between PySCF adapters and a host-independent linear-combination-of-atomic-orbitals (LCAO) layer. The classes on this page define the atom-centered grid indexing, auxiliary Gaussian convolutions, and grid/basis interpolation used by the PySCF forward and adjoint paths. Atom-centered grid indexing --------------------------- .. py:module:: ciderpress.dft.grids_indexer .. py:class:: AtomicGridsIndexer ``AtomicGridsIndexer`` relates PySCF's sorted and screened quadrature points to atom, radial shell, angular point, and spherical-harmonic indices. Energy, potential, and grid-response operations share this mapping. Auxiliary Gaussian convolutions ------------------------------- .. py:module:: ciderpress.dft.lcao_convolutions .. py:class:: ATCBasis .. py:class:: ConvolutionCollection .. py:class:: ConvolutionCollectionK ``ATCBasis`` stores the atom-centered auxiliary Gaussian basis and its angular layout. ``ConvolutionCollection`` builds the matrices that convolve the source expansion for version-I/J features; ``ConvolutionCollectionK`` supplies the corresponding version-K operations. Grid/basis interpolation ------------------------ .. py:module:: ciderpress.dft.lcao_interpolation .. py:class:: LCAOInterpolator .. py:class:: LCAOInterpolatorDirect The interpolators transfer fields between the atom-centered auxiliary basis and molecular quadrature points. Their forward and backward methods form the pair used for NLDF values and potentials. The direct variant treats the on-site contribution explicitly and is used by the packaged molecular NLDF path. NLDF orchestration ------------------ .. py:module:: ciderpress.dft.lcao_nldf_generator .. py:class:: LCAONLDFGenerator ``LCAONLDFGenerator`` combines an exponent-interpolation plan, convolution collection, and LCAO interpolator. The PySCF-specific :class:`~ciderpress.pyscf.nldf_convolutions.PyscfNLDFGenerator` constructs these objects from a molecule and its CIDER grid. See :doc:`../../theory/nldf_numerical` for the mathematical description and :doc:`../pyscf/numerical` for the PySCF numerical implementation details.