To carry over this conversation from twitter: if d3py is going to move towards a "idiom+serving" model, should the idioms be imported directly?
Right now I'm manually moving new Vincent builds into the d3py codebase, as to avoid an install dependency. But, it might be more Pythonic/easier to integrate if we have explicit imports of our different idiom libraries.
That would also let other folks build idioms and link them to d3py, via a standard set of API hooks (that would need to be sorted out). Maybe something like "d3py will process the pandas data into a JSON-like structure, and your idiom needs to parse it".
To carry over this conversation from twitter: if d3py is going to move towards a "idiom+serving" model, should the idioms be imported directly?
Right now I'm manually moving new Vincent builds into the d3py codebase, as to avoid an install dependency. But, it might be more Pythonic/easier to integrate if we have explicit imports of our different idiom libraries.
That would also let other folks build idioms and link them to d3py, via a standard set of API hooks (that would need to be sorted out). Maybe something like "d3py will process the pandas data into a JSON-like structure, and your idiom needs to parse it".