"IndexError: This vertex is missing, maybe hidden by a duplicate or another heavier point. " when using AlphaDTMFiltration

[Lxion0553]

Hello! Thanks for your great work!
When using the AlphaDTMFiltration in DTM_filtrations.py, an IndexError: "This vertex is missing, maybe hidden by a duplicate or another heavier point." was reported. May I ask what is the cause of this error? How can I fixed it.

[mglisse]

The error message says "maybe hidden by a duplicate". Did you check if you have 2 times the same point in your point set?

[mglisse]

@VincentRouvreau

TDA-tutorial/DTM_filtrations.py

Lines 159 to 163 in 3ab7c33

	alpha_complex = gudhi.AlphaComplex(points=X)
	st_alpha = alpha_complex.create_simplex_tree()
	Y = np.array([alpha_complex.get_point(i) for i in range(N_tot)])
	distances = euclidean_distances(Y) #computes the pairwise distances
	DTM_values = DTM(X,Y,m) #/!\ in 3D, gudhi.AlphaComplex may change the ordering of the points

The comprehension with get_point was done to be robust against the ordering changes we used to do in 3d, but we haven't done that for a long time now. On the other hand, it isn't robust to duplicated points. Should we just use Y=X?

[VincentRouvreau]

I was able to reproduce it by doing:

X = SampleOnNecklace(N_obs, N_out, is_plot = True)
X[1,:] = X[0,:]  # To force 2 points to be equal
# ...
simplex_tree_DTM_alpha = AlphaDTMFiltration(X, m, p, dimension_max)

# File alpha_complex.pyx:119, in gudhi.alpha_complex.AlphaComplex.get_point()
# IndexError: This vertex is missing, maybe hidden by a duplicate or another heavier point.

[VincentRouvreau]

The comprehension with get_point was done to be robust against the ordering changes we used to do in 3d, but we haven't done that for a long time now. On the other hand, it isn't robust to duplicated points. Should we just use Y=X?

Your proposal works fine and I think this is the good way to handle it

[Lxion0553]

@mglisse

The error message says "maybe hidden by a duplicate". Did you check if you have 2 times the same point in your point set?

It seems that there exist no duplicated points in my dataset, maybe two points that are too close can cause this error also?

[mglisse]

It seems that there exist no duplicated points in my dataset, maybe two points that are too close can cause this error also?

I don't think it can, but it is going to be hard to say much more without a complete reproducible testcase.
Anyway, does the change in #73 help you?

[Lxion0553]

I don't think it can, but it is going to be hard to say much more without a complete reproducible testcase. Anyway, does the change in #73 help you?

It does help! Thank you!