Enumerations
============

This page documents the high-level enums exposed by the Python bindings.

ExecutionSpace
--------------
Determines where computations execute:

* ``ExecutionSpace.Host`` — run on CPU.
* ``ExecutionSpace.Device`` — run on GPU (requires CUDA/HIP build).

Example::

   integrator = gxc.Integrator(gxc.ExecutionSpace.Host, mol, basis, grid)

GridQuality
-----------
Predefined grid quality levels (increasing accuracy → increasing cost):

* ``GridQuality.GM3`` — Treutler–Ahlrichs Grid Model 3: 35 radial, 110 angular points.
* ``GridQuality.GM5`` — Treutler–Ahlrichs Grid Model 5: 50 radial, 302 angular points.
* ``GridQuality.FineGrid`` — 75 radial, 302 angular points.
* ``GridQuality.UltraFineGrid`` — 99 radial, 590 angular points.
* ``GridQuality.SuperFineGrid`` — 175 (H-He) / 250 (Li-...) radial, 974 angular points.

RadialQuad
----------
Radial quadrature schemes for atomic grids:

* ``RadialQuad.MuraKnowles`` — Mura–Knowles Log3 quadrature.
* ``RadialQuad.MurrayHandyLaming`` — Murray–Handy–Laming quadrature.
* ``RadialQuad.TreutlerAhlrichs`` — Treutler–Ahlrichs quadrature.

PruningScheme
-------------
Grid pruning strategies to remove redundant angular points:

* ``PruningScheme.Unpruned`` — no pruning (slowest).
* ``PruningScheme.Robust`` — Psi4's robust pruning scheme (recommended).
* ``PruningScheme.Treutler`` — Treutler–Ahlrichs pruning.

Usage note
----------
These enums are passed to constructors like :class:`gauxc.Molgrid` to
control grid construction and evaluation behavior. See :doc:`molgrid`.