I must refactor the tests to make them more rational an by removing a lot of code-repetition

Bug in line 112 in SOAPClassify.py!!!

Should address this and a test should catch this error!!!!

if x[0].nmax != i.nmax or x[0].nmax != i.lmax:

Like title, need to add a way to convert mda/ase trajectories to extended xyz

Title pretty much says it all.

The menu does not show in the read the docs documentation

Extended xyz and other formats supports the origin of the box
We should save this information:

Step to do:

• Import Origin From ase and MDA
• Export Origin to xyz and ase/MDA

A few points:

• #55
  • #72
• #56
• #57
• #58

As title:
The actual behavior for transitionMatrixFromSOAPClassification adds an additional column to the states, that is useful only if the user gives -1 as states to represent errors.
A workaround is ti print the transition matrix with [:-1,:-1]. But I think I need to make this clearer to the user

Hello
I posted an issue on the paper repos GitHub (GMPavanLab/LENS#1) and was pointed towards this GitHub.
I can see that some of the improvements have already been implemented here, but I think that introducing atom selections in the function call might improve the versatility of the code.

My suggestion is to make the following change to the function "listNeighboursAlongTrajectory":
From:

def listNeighboursAlongTrajectory(
    inputUniverse: Universe, cutOff: float, trajSlice: slice = slice(None)
) -> "list[list[AtomGroup]]":
    nnListPerFrame = []
    for ts in inputUniverse.universe.trajectory[trajSlice]:
        nnListPerAtom = []
        nnSearch = AtomNeighborSearch(inputUniverse.atoms, box=inputUniverse.dimensions)
        for atom in inputUniverse.atoms:
            nnListPerAtom.append(nnSearch.search(atom, cutOff))
        nnListPerFrame.append([at.ix for at in nnListPerAtom])
    return nnListPerFrame

To:

def listNeighboursAlongTrajectory(
    inputUniverse: Universe, cutOff: float, trajSlice: slice = slice(None), atom_sel: str = "all"
) -> "list[list[AtomGroup]]":
    nnListPerFrame = []
    for ts in inputUniverse.universe.trajectory[trajSlice]:
        nnListPerAtom = []
        nnSearch = AtomNeighborSearch(inputUniverse.select_atoms(atom_sel), box=inputUniverse.dimensions)
        for atom in inputUniverse.select_atoms(atom_sel):
            nnListPerAtom.append(nnSearch.search(atom, cutOff))
        nnListPerFrame.append([at.ix for at in nnListPerAtom])
    return nnListPerFrame

This would allow easy selections of atoms in the universe while defaulting to all atoms.

Describe the bug
A clear and concise description of what the bug is.
KeyError Traceback (most recent call last)
/nfs/scistore14//timesoap/SOAPify/Examples/timeSOAP.ipynb Cell 7 line 2
24 slide = 1
26 return nAt, timedSOAP, np.diff(timedSOAP.T, axis=-1)
---> 29 nAtoms, tSOAP, dtSOAP = getTimeSOAP(soapFileName, trajAddress)

/nfs/scistore14//timesoap/SOAPify/Examples/timeSOAP.ipynb Cell 7 line 6
4 print(f)
5 ds = f[f"/SOAP/{trajAddress}"]
----> 6 fillSettings = getSOAPSettings(ds)
7 print(fillSettings)
8 print(ds.shape)

File ~/software/soapify/lib/python3.10/site-packages/SOAPify/utils.py:345, in getSOAPSettings(fitsetData)
328 """Gets the settings of the SOAP calculation
329
330 you can feed directly this output to :func:fillSOAPVectorFromdscribe
(...)
342
343 """
344 print(fitsetData)
--> 345 lmax = fitsetData.attrs["l_max"]
346 nmax = fitsetData.attrs["n_max"]
347 symbols, atomicSlices = getSlicesFromAttrs(fitsetData.attrs)
...
File h5py/_objects.pyx:55, in h5py._objects.with_phil.wrapper()

File h5py/h5a.pyx:80, in h5py.h5a.open()

KeyError: "Can't open attribute (can't locate attribute: 'l_max')"

Additional context
I use "wget https://github.com/GMPavanLab/dynNP/releases/download/V1.0-trajectories/ico309.hdf5" and put the data file in the same directory with "trajAddress = './' .

dscribe1.2.1 returns an array of shappe (nat,nsoap) instead of shape (1,nat,nsoap) if we are analysing only 1 frame!
This breaks the saponifyWorker function

As title

As title

Describe the bug
The box size does not update within each trajChunkSize when creating the .hdf5 calling MDA2HDF5 function.
A possible workaround is to set trajChunkSize=1, thus the box update every step, but extracting the SOAP is slow.

like in the title, as usual, I need also a test to cover this

This should be the 0.1 feature:

• A way for the user to get and interact easier with trajectories out of the hdf5 files
• An instrument that reads the groups and makes some assumptions about their content, to show the user faster the content of the files
• I should not over-engineer this
• Maybe create a reader compatible with MDA

Describe the bug

In SOAPify/Examples/LENS.ipynb, tmat labels are given incorrectly when the clusters assigned by KMeans are not in order (e.g.: [C0=0, C2=2, C1=1]).
The output of calculateTransitionMatrix is a matrix with columns and rows corresponding to ordered clusters (e.g. for columns: C0=0 in col 0, C1=1 in col 1, C2=2 in col 2 ...) while the label assignment is given depending on the cluster order (C0 for col 0, C2 for col 1, C1 for col 2).
The problem is fixed by sorting the labels, from:

classifications = SOAPclassification(
    [], prepareData(classifiedFilteredLENS), [f"C{m[0]}" for m in minmax]
)

to:

classifications = SOAPclassification(
    [], prepareData(classifiedFilteredLENS), [f"C{m[0]}" for m in np.sort(minmax, axis=0)]
)

To reproduce the bug, changing the random_state parameter in KMeans (and thus the cluster assignment order) changes the exchanging probabilities.

I have compiled a very spartan documentation, this should be improved:

• #8
• #9
• #34
• #32