v 0.6.0 documentation missing for local homology?

[catanzaromj]

I might be missing something obvious here. Is the documentation for v0.6.0 available yet? The links in the release (tutorial: https://giotto-ai.github.io/gtda-docs/0.6.0/notebooks/local_homology.html and example:https://giotto-ai.github.io/gtda-docs/0.6.0/notebooks/local_hom_NLP_disambiguation.html) both give 404 errors. In particular, I have my own hard-coded versions of local homology and I'd like to compare them to the newly integrated versions. The notebook in the repo renders fine.

Since this is a discussion, what "version" of local homology is this implementing, i.e., is there a corresponding paper or journal article?

[ulupo]

Sorry for the late reply @catanzaromj, I've been on vacation. Indeed, the docs for v0.6.0 have not been updated yet. I need a bit of help with this, but perhaps @wreise can be of some assistance.

Concerning the version of local homology, this is work of @jacobbamberger and @matteocao, and they are in a better position to explain the relation with other known versions.

[wreise]

I am working on the docs, but it will take some time!
If you are looking for a html version of the executed notebooks, here it is: local_hom_notebooks.zip
As for the docs of the classes, for now, I can only refer you to the docstrings. The documentation of the class is generated from those and the Theory section does not contain local homology yet anyway.

[catanzaromj]

@ulupo I hope you enjoyed your vacation. Thanks for letting me know.

@wreise Thanks for sharing. The explanation given at the bottom of the local homology notebook in the attached zip gives exactly what I was looking for.

@jacobbamberger @matteocao FWIW I was computing the topology of the annular regions without doing any coning, but then looking for features within this annular shell that persist longer than the thickness of the region. This particular idea was formalized in Geometric anomaly detection in data by Stolz et al, where they argue this idea as a proxy for coning off.