The user's interactions are now taken into account (and dynamically)
when the "center" field is edited (for a sector).

The user no longer edit the bounding box fields for the sector mask,
because this is not a well-defined operation (there is a lot of
different new sectors that correspond to a given field edit, and it's
not possible to decide which one the user meant (nor it is possible for
the user to mean one at all, in fact)).

Thus the user can enter new value on the fields, but upon leaving them /
validating, they will be reversed to their previous values.

- Solves the bug where the sector would stop being scaled by dragging a
corner and the bug where the sector's bounding box and real position
would sometimes stop corresponding to one another by changing strict
equalities of potentially very small numbers to checking their absolute
differences is less than a certain threshold.

- Manages negative values for field input, especially for angles

On the mask tab of the instrument view, when the user inputs a value in
one the field, the view is now updated on validation of this value (i.e.
when the user clicks outside the field or hits Enter), whereas before,
the user needed to hover over the instrument view to see the changes
take effect, which was illogical.

Also, when there is a correction on the value entered (for example a
negative value for a positive-only field), the value corrected is
immediatly showed, instead of waiting for the user to deselect-reselect
the shape.

- Add a proper shortcut
- Fix the bounding box not following the sector when moved, but updating
it
- Set new endAngle at creation
- Fix a bug where the topLeft corner drag was mismanaged due to a bad
variable
- Better check for angles and radii values

It is always installed with setuptools so setuptools can assume it will always be there.

28927 ill d7 signed two theta

Clarify rotations in documentation

Suppress C++17 deprecation warnings on inheriting std::iterator

Fix performance tests

Add notebook back to macOS bundle

Add apply to all button to plot axes properties

Sliceviewer sub-plots follow image scaling

Minor dgsplanner fixes

Support rotations and translations in LoadSampleShape

Add reverse reverse colourmap for sliceviewer

Newer versions of visual studio now show
a warning if a custom build doesn't produce
all of the outputs specified by the build
step.

TBB has not yet been fixed to avoid this warning so we have no
choice but to suppress it

According to the Nexus standard, if the offset is not present, it implies
the offset is and absolute timestamp, which is relative to the start of
Unix epoch (https://manual.nexusformat.org/classes/base_classes/NXlog.html)

So instead fo throwing an error, we use unix epoch.

New test cases have been added for molecular calculations with no
periodic boundary conditions. Compared to the already-tested periodic
systems, these calculations yield lattice vectors of zero and
have 3N-6 vibrational modes. This tests some different paths in the
parsing logic.

Based on a user report and sample files. Two main differences were
identified which caused the parser to trip up on a calculation without
periodic boundary conditions:

- No lattice vectors are printed in the output file, but the parser
  requires this data to exist, as does Abins' data format.

  - Here we apply the same convention as the Gaussian parser, setting
    lattice vectors to a 3x3 zero matrix

- The lowest 6 vibrational modes, which theoretically correspond to
  translation and rotation, are discarded from the output data. This
  breaks an assumption that there are 3N vibrational modes (where N is
  number of atoms). Abins in general does _not_ make this assumption,
  so it can be fixed in the parser.

  - Now the number of atoms is obtained from the coordinates data, and
    a check is made that the number of modes is either 3N or 3N-6. There
    _might_ be an issue with dimers (which have 3 translational modes,
    two rotational modes and a non-zero stretching mode); need to get
    access to Materials Studio and test this.

Some error messages were also cleaned up to remove a confusing "bytes"
type signature.

This commit adds a checkbox to sliceviewer that reverses the colourmap. because of this it also removes the _r reverse maps form the combo box.

re: #28915