search

probcomp / GenExperimental.jl

Featherweight embedded probabilistic programming language and compositional inference programming library
MIT License
17 stars 2 forks source link

Trace visualization API for Jupyter notebooks #13

Closed marcoct closed 7 years ago

marcoct commented 7 years ago

A little API for writing interactive custom trace visualizations with Snap.svg, with the ability to drag around elements on the canvas and have the corresponding trace object be updated.

http://snapsvg.io/

Also needs to be integrated with IJulia

marcoct commented 7 years ago

I have a basic API with a D3 example already.

D3 should be the canonical rendering scheme for Jupyter notebooks. Users can use a plotting library like matplotlib if they really need to, but our examples should use D3 or other tools better designed for more general graphics and not specifically plots. Also web-standards rendering tools are preferred to Julia- or Python-specific libraries, which are less portable.

I already have D3 renderering running in the goal inference notebook, with the d3 library made available according to instructions in notebook_install.

The trace rendering API for jupyter notebooks should not and does not restrict itself to D3. It should permit e.g. WebGL Javascript renderers as well.