Nano-cat

*****
* Added the option, when constructing core atom subsets,
the use a distance matrix representing the shortest paths along the
edges of a polyhedron, rather than through space.
Enabling this option will result in more accurate ``"uniform"`` and
``"cluster"`` distributions at the cost of increased computational time.
* Updated and improved the ``"uniform"`` and ``"cluster"`` distributions.
* https://github.com/nlesc-nano/CAT/pull/65: Fixed a bug where ``uniform_idx()`` yielded the rolled,
rather than unshifted, indices.
* https://github.com/nlesc-nano/CAT/pull/64: Bug fix: the subset Schema now checks for instances of
int ``Or`` float.
* https://github.com/nlesc-nano/CAT/pull/66: Return the identity (rotation) matrix if a ``FloatingPointError`` is
encountered during the creation of rotation matrices.
This can occur if a ligand consists of a single atom.
* https://github.com/nlesc-nano/CAT/pull/66: Fixed a bug in the parsing of the mode parameter of ``distribute_idx()``;
``"uniform"`` and ``"cluster"`` will now correctly link to ``np.argmax`` and
``np.argmin`` instead of the other way around.

*****
* Added the ability to populate only a (random-ish) subset of
core anchors with ligands.

*****
* The ligand rotation check is now substantially faster:
a distance cutoff has been implemented for the construction
of distance matrices.

*****
* Added an option perform an ensemble-averaged QD activation strain
analyses in Nano-CAT_.
* Removed a number of redundant modules.
* QD optimization now properly respect the ``optional.qd.opt.use_ff`` keyword.

*****
* Minor tweaks to the default forcefield-related CP2K input files.
* Fixed a couple of bugs in the ligand dissociation workflow.
* Reworked the ligand dissociation procedure in Nano-CAT_.

*****
* Bug fix: Added a missing value to the to-be exported ASA columns.