Is it possible to model something like this with pyg or pygTemporal?

[DogitoErgoSum]

Where a circle is a node and different colors represent different node types.

[rusty1s]

As outlined in #4323, this can be achieved by a list of events:

(src=0, dst=1, time=t1)
(src=0, dst=2, time=t2)
(src=1, dst=2, time=t2)

[DogitoErgoSum]

i am not following.
As an example, if orange type has a feature size of 5 and green a feature size of 3, i m not sure how would one define what type they belong or its feature size with the above, like one does with heterodata.

[rusty1s]

Yes, the feature set of different types/messages currently need to have the same dimensionality in TemporalData (or you will need to work with padding features to same length). Sorry :(

[DogitoErgoSum]

Ty for your answer. I will model the sequence has one heterogeneous graph with time has as a feature.

[rusty1s]

That should work as well :)