Alpha Complex Filtraton about cechmate

[arksch]

As this is important to my research, I will investigate this issue and might just send you a pull request in the coming days.

[ctralie]

Thank you so much for bringing this to my attention! Yes, if you could have a look and make a pull request, that would be a huge contribution. If you are on a tighter timescale, note that GUDHI also has an alpha filtration module: http://gudhi.gforge.inria.fr/doc/latest/group__alpha__complex.html GUDHI is harder to install, which is part of what spurred me to write this library, but it should work well. Regardless, I'll have a look at this as soon as I am able. Many thanks, Chris

…

On Fri, Oct 4, 2019 at 6:55 AM arksch ***@***.***> wrote: As this is important to my research, I will investigate this issue and might just send you a pull request in the coming days. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#10?email_source=notifications&email_token=AAJWDZTGEU45YGKNVV54C2DQM4OKRA5CNFSM4I5OIXYKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEALJHMQ#issuecomment-538350514>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAJWDZULVJOENHWJ7SMNZ5LQM4OKRANCNFSM4I5OIXYA> .

[arksch]

Hi Chris,
in the example I posted, Gudhi even creates filtration values of over 100. Quite odd. Maybe an issue of the miniball algorithm.
It also seems that this error does not affect the actual persistence module, as in all instances I found there are a 2-simplex and one of its faces with the same value, so it will only create ephemeral homology classes of zero persistence.
I forked and keep you posted.

Update: Gudhi does not take the square root. They actually compute the same value of 11.18 for the given triangle.