Planner Results

As discussed in the Planner section, VegaFusion uses a planning phase to split an input Vega specification into a specification to run on the client and a specification to run on the server. Additionally, a communication plan is produced that specifies which signal and dataset values must be passed between the server and client to preserve interactivity. For debugging, the client spec, server spec, and communication plan are available as properties on the VegaFusionWidget instance after the chart has been displayed.

First, wrap the chart in a VegaFusionWidget and display it.

import altair as alt
from vega_datasets import data
import vegafusion as vf

source = data.seattle_weather()
brush = alt.selection(type='interval', encodings=['x'])

bars = alt.Chart().mark_bar().encode(
    opacity=alt.condition(brush, alt.OpacityValue(1), alt.OpacityValue(0.7)),

line = alt.Chart().mark_rule(color='firebrick').encode(

widget = vf.jupyter.VegaFusionWidget(alt.layer(bars, line, data=source))

Then print out the value of the following widget properties:

  • widget.spec: This is the Vega-Lite specification created by Altair

  • widget.full_vega_spec: This is the Vega specification produced by Vega-Lite

  • widget.server_vega_spec: This is the portion of the full Vega spec that was planned to run on the server (The Python kernel in this case)

  • widget.client_vega_spec: This is the portion of the full Vega spec that was planned to run on the client and be rendered by the Vega JavaScript Library.

  • widget.comm_plan: This is the specification of which signals and datasets must be transfered between the client and server in order to preserve the interactive behavior of the original specification.