[BUG] cursor.rotate should rotate all vector properties

[lohedges]

Currently the cursor.rotate function only rotates the coordinates property, when it should really rotate all vector properties, i.e. velocity too. At present this can be achieved by calling the function multiple times, using the map option to remap coordinates to velocities, e.g.:

# Here m is a rotation matrix.

# First the coordinates.
cursor = cursor.rotate(matrix=m)

# Now the velocities.
cursor = cursor.rotate(matrix=m, map={"coordinates": "velocity"})

[lohedges]

Actually, the above approach doesn't work because, unless the center keyword argument is used, the coordinates property is used to compute the center-of-mass to use for the rotation, e.g. here.

[lohedges]

This is relevant to this feature branch that I'm working on, which naively uses the workaround in the original post. For now I can just compute the center of mass before calling the function and pass it in.

[lohedges]

Which also makes me realise that I'll need to be careful to choose the correct center when applying the rotation, since this won't be (necessarily) the same as what is used for the box vectors themselves which would be rotated about (0, 0, 0).

[lohedges]

I think I've now got this working (using a workaround) on my feature_reduce branch. I've also added a unit test to check that the rotation is performed correctly and that the coordinates property is updated.

The two things that I'm not sure of are:

• For the vector properties, I aim performing the rotation about the box origin, which can be specified by the user (defaulting to (0,0,0)). It would be nice to be able to determine this automatically.
• Following the rotation of the box, the z component of the third lattice vector needs to be flipped if it is negative. I'm not sure how this should impact the properties.

[chryswoods]

I've been thinking about this - am I right that the centre of rotation for each velocity vector would be the origin (0,0,0)? This is because velocities are all vectors that represent change in coordinates - their origin represents the "no change - atom not moving", and so they all have an origin at (0,0,0) regardless of where the atom is located or where the centre of the box sits (effectively a velocity vector is always in a frame of reference centered on its atom).

So, if an atom's velocity is (1,0,0), then rotating the atom by 90 degrees about z would change the velocity to (0,1,0), i.e. rotated the velocity vector about its origin. And this is the same regardless of the position of the atom or the choice of centre of box.

[chryswoods]

I'm also not sure how the above fix worked, because velocity should be of type AtomVelocities, which cannot be rotated by that function (it only works on AtomCoords).

For example;

>>> import sire as sr
>>> # Creating some velocities by running MD
>>> mols = sr.load(sr.expand(sr.tutorial_url, "ala.top", "ala.traj"), silent=True)
>>> mols = mols.minimisation().run().commit()
>>> d = mols.dynamics(timestep="4fs", temperature="25oC")
>>> d.run("10ps")
>>> mols = d.commit()
>>> mols[0].property("velocity")
SireMol::AtomVelocities( size=22
0: ( -14.35 Å ps-1, -13.7221 Å ps-1, -3.16664 Å ps-1 )
1: ( 3.83649 Å ps-1, 0.782261 Å ps-1, -1.09576 Å ps-1 )
2: ( -1.63303 Å ps-1, 30.2243 Å ps-1, 19.0436 Å ps-1 )
3: ( -6.70623 Å ps-1, -15.6312 Å ps-1, -14.4375 Å ps-1 )
4: ( -1.49417 Å ps-1, 5.25292 Å ps-1, 0.825426 Å ps-1 )
...
17: ( 11.5151 Å ps-1, 26.6377 Å ps-1, -27.9718 Å ps-1 )
18: ( 0.576416 Å ps-1, 1.62035 Å ps-1, 2.71101 Å ps-1 )
19: ( -3.38965 Å ps-1, 7.1897 Å ps-1, 8.87909 Å ps-1 )
20: ( -8.95596 Å ps-1, -22.0357 Å ps-1, -5.77348 Å ps-1 )
21: ( 0.297291 Å ps-1, 0.673374 Å ps-1, 19.2852 Å ps-1 )
)
>>> c = mols[0].cursor()
>>> c.rotate(5*sr.units.degrees, map={"coordinates":"velocity"})
TypeError: SireError::invalid_cast: Cannot cast from an object of class "SireMol::AtomVelocities" to an object of class "SireMol::AtomCoords". (call 
sire.error.get_last_error_details() for more info)

[lohedges]

Yes, I think (0, 0, 0) is correct and is what my current workaround uses.

[chryswoods]

This commit is the beginning of adding in the option of rotating velocities.