The documentation of forces in the simulation is not up to date: https://aortiza.github.io/icenumerics/OtherNotebooks/ForcesOfTheSimulation.html.

We corrected the documentation and created a new checking script in: https://github.com/crodriga/PROJECT_CairoLattice/blob/master/Scripts/Script-6.0_CheckingHillDefinition.ipynb

The function v.trj_to_vertices() is not working properly, it seems that in the new merge some function gets lost and is not included in the new version. Hence, the match between trajectories (.lammpstrj files) and vertices now is unavailable.

I left a jupyter notebook and a .lammpstrj file in Drive located at:

DegenerateIce>BidisperseSystems>Issue Github

https://drive.google.com/drive/u/0/folders/10MQ73BFr2lgPcI3b1sf0w26xDzczg5I8

@crodriga found that get_ice_trj doesn't connect the colloids to the correct trap when there are several types of traps. This likely happens also when there are several types of colloids.

We need to assign a prefactor to the energy to obtain energy in proper units instead of 1/nm^3

with pint 0.11 the following error appears; to solve it try
conda install -c conda-forge pint=0.9

TypeError Traceback (most recent call last)
in
1 col.run_simulation()
----> 2 col.load_simulation()

~\Desktop\Tutorial\icenumerics\icenumerics\colloidalice.py in load_simulation(self, sl)
372 self.trj = self.sim.load(read_trj = True, sl = sl)
373 self.frames = self.trj.index.get_level_values("frame").unique()
--> 374 self.set_state_from_frame(frame = -1)
375
376 def set_state_from_frame(self, frame):

~\Desktop\Tutorial\icenumerics\icenumerics\colloidalice.py in set_state_from_frame(self, frame)
380 for i,c in enumerate(self):
381 c.colloid = self.trj.loc[idx[frame,i+1],["x","y","z"]].values*ureg.um - c.center
--> 382 c.direction = c.direction * np.sign(np.dot(c.colloid,c.direction))
383
384 return self

TypeError: operand type(s) all returned NotImplemented from array_ufunc(<ufunc 'sign'>, 'call', <Quantity(5.296799659729004, 'micrometer')>): 'Quantity'

Creating a spin ice object, using sp.create_lattice() only periodic boundary conditions define the "unit_cell_center".

The "unit_cell_center" is defined within an if statement in the square.py file.

Occasionally, at the end of a simulation, a colloid can be found at the exact center of the trap. That results in a null direction for the colloid.
This situation raises an error when passing colloids to vertices in "vertices.py".

In the multitrap branch, traps could have different lengths and center positions, thus traps are not in the same lattice and do not have the same trap separation. Due to that, vertices in the last dataframe are not properly classified.

I have patch the issue by chaenging a bit the function from_neighbors_get_nearest_neighbors() in line 85 of vertices.py file:

def from_neighbors_get_nearest_neighbors(NeighborPairs):
    # This function takes a list of Delaunay Neighbor Pairs and returns only those which are close to the minimum distance.
    NeighborPairs['Distance']=np.around(NeighborPairs['Distance'],decimals=4)
    #print(NeighborPairs['Distance'])
    #NeighborPairs = NeighborPairs[NeighborPairs['Distance']<=np.min(NeighborPairs['Distance'])*1.1]
    NeighborPairs = NeighborPairs[NeighborPairs['Distance']<=21.2132*1.1]
    
    return NeighborPairs

I just printed the Neighborpairs['Distance'] and manually choose the one that maps the inner vertices.

In get_ice_trj, when particles jump from one side to the other, they are assigned an incorrect trap. In fact, it appears that two particles are assigned to a single trap.

To fix this, distances could be calculated in a periodic space.

This will also be useful for periodic boundaries.

Using the Ordering option in the Spins.HoneycombSpinIce function calls one of the functions in the file HoneycombSpinIceGeometry.py.

However, the name of the functions and the resulting states do not match. Could you assign the correct names @Anne-le-cunuder? It could also be useful to add some documentation in the functions Spins.HoneycombSpinIce() and HoneycombSpinIceCalculateGeometry() to briefly explain what are the options of ordered states that we have, and how should they look.

docstrings on most functions are missing or incomplete. For a guide on how to write docstrings we can see the gist https://gist.github.com/aortiza/43b16f5f609a396c29ad0cd745a511ff