Update to HyperNetX 2

[ffl096]

TopoNetX is currently only compatible with HyperNetX 1.2, which was released more than two years ago. The latest version of HyperNetX is 2.3. We should update our supported version.

See #410 for what tests are currently failing.

[colltoaction]

Hi folks! Let me introduce @brendapraggastis from PNNL, we work in the Hypergraph Interchange Format (HIF) project together. I believe it would be beneficial to start collaborating and this issue is a great place to kickstart the conversation. Cheers!

[brendapraggastis]

Hello TopoNetX! I only recently discovered your library and it looks like you have many cool features. We would like to have you join the HIF project, allowing hypergraphs to be shared across our libraries. Please reach out to Martin to see how to make that happen!

…

__________________________________ Brenda Praggastis, PhD Senior Data Scientist and Mathematician Math, Stats & Data Science AI & Data Analytics Pacific Northwest National Laboratory 1100 Dexter Ave N, Suite 500 Seattle, WA 98109 USA (206) 528-3409 (voice only) From: Martin Coll ***@***.***> Date: Monday, December 16, 2024 at 5:41 AM To: pyt-team/TopoNetX ***@***.***> Cc: Praggastis, Brenda ***@***.***>, Mention ***@***.***> Subject: Re: [pyt-team/TopoNetX] Update to HyperNetX 2 (Issue #411) Check twice before you click! This email originated from outside PNNL. Hi folks! Let me introduce @brendapraggastis<https://github.com/brendapraggastis> from PNNL, we work in the Hypergraph Interchange Format (HIF) project together. I believe it would be beneficial to start collaborating and this issue is a great place to kickstart the conversation. Cheers! — Reply to this email directly, view it on GitHub<#411 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AJPW7D43NZSZBKGD426LR7T2F3KBLAVCNFSM6AAAAABTLDJYKCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDKNBVGY3DGNZYGE>. You are receiving this because you were mentioned.Message ID: ***@***.***>

[mhajij]

@brendapraggastis please outline your plan of contribution. Are you planning to add a hypergraph class to toponetx? or a support for HyperNetX inside every class?

So far the support for HyperNetX has been limited. Essentially every simplicial, cellular complex is a hypergraph so we have a function that supports the conversation of these complexes to hypergraph objects in HyperNetX. I would suggest we make this integration more complete.

[mhajij]

@colltoaction
@brendapraggastis feel free to start creating a series of actionable issues and we can work together on them, sounds like a plan ?

[colltoaction]

@mhajij sounds good to me! As far as I can see there are two uses of HNX in the codebase:

TopoNetX/toponetx/classes/simplicial_complex.py

Line 1744 in e9ecbeb

def to_hypergraph(self) -> Hypergraph:

TopoNetX/toponetx/classes/cell_complex.py

Line 2193 in e9ecbeb

def to_hypergraph(self):

The idea moving forward is to use a JSON interchange format, HIF, so that not only HNX but other participating libraries (HypergraphX, XGI, ...) can interoperate as well. As far as I can see you can then remove your explicit dependency on the HNX package.

So far you have implemented these functions that export a hypergraph, and now you can take advantage of HIF to access to new datasets yourselves. A good test would be to replicate the results in any of the tutorials but using TopoNetX, e.g:

https://github.com/pszufe/HIF-standard/blob/main/tutorials/hnx.ipynb

The existing integrations were straightforward to implement so I wouldn't want to create unnecessary tickets. Maybe one to implement to_hif and from_hif, and another one to send a PR to our repo with a working notebook. @mhajij what do you think?

A second step in our collaboration should look at enabling new network types, many available in your package. The HIF format currently defines three possible network types: directed, undirected and ASC.