Open nwlandry opened 5 months ago
An initial idea that I have is that each dataset has a type: "collection", "simplicial-complex", "hypergraph", or "dihypergraph", for example. If the dataset type is a collection, then the json contains metadata about the entire collection as well as information for accessing each dataset in the collection. This could be recursive; you could have collections of collections, for example. I picture the output of load_xgi_data(<collection>)
being a dictionary or perhaps we define a data structure so that we can easily store collection-level data.
If the output was a dictionary, it might look something like this:
{
"collection-data": {"name": "hyperbard"},
"datasets": [Hypergraph1, Hypergraph2,...]
}
This tweet references a paper which contains a collection of datasets, where each dataset is the co-occurence of characters in a Shakespeare play. It should be possible to store a collection of datasets in a good way on XGI-DATA.