gexf parsing threw away attribute names and only imported attribute IDs that could change meaning anytime.
fixed this, changed the api a bit (from array of [id,value] to object)
hope this helps
improve gexf parsing, actually use key-value pairs for attributes
also parse edges inline attributes properly
I really like the change from a list of attributes to a dictionary of named attributes. This will be must easier to work with but is not backwards compatible. I think this should be incorporated; however, the note on compatibility should be flagged up to users who might be relying on the previous implementation of attributes. Great work.
thanks. i agree on the note - should i add this to the README?
I think it's fine as is and is something jacomyal could do when the fork is pulled and incorporated. I hope that will be soon. I'm going to pull your changes and make my new parseJson plugin behave the same way.
The new gexf parser used by sigma's upcoming 1.0.0 version deals with your issues and is able to extract more data from the gexf than before.
You can find an example of data as outputted by the parser there.