Black Lives Matter. Please consider donating to Black Girls Code today.

Parallel coordinates plot constraint range snapping behavior

I am having an odd behavior with parallel coordinates plots that I haven’t had before. When I drag to constrain the data, the constraint range snaps to some illogical set of values. It does this even if I explicitly set the constraintrange argument for the dimension! It’s making it very difficult to explore the data. Any tips on correcting this would be greatly appreciated.

This is my code to build the plot.

fig = go.Figure()
fig.add_trace(
    go.Parcoords(
        line=dict(
            color=synergy_plot['review_encoded'],
            colorscale = [(0,DEFAULT_PLOTLY_COLORS[0]),
                          (0.5,DEFAULT_PLOTLY_COLORS[2]),
                          (1,DEFAULT_PLOTLY_COLORS[1])]
        ),
        dimensions=list([
          dict(range=[0,len(synergy_plot['review_encoded'].unique()) - 1],
            label='MD Review',
            values=synergy_plot['review_encoded'],
            tickvals=[0,1],
            ticktext=synergy_plot['review'].unique()),
          dict(range=[0,1],
            label='RNA coverage',
            tickvals=[0,0.2,0.4,0.6,0.8,1],
            values=synergy_plot['rna_coverage']),
          dict(range=[0,1],
            label='DNA coverage',
            tickvals=[0,0.2,0.4,0.6,0.8,1],
            values=synergy_plot['dna_coverage']),
          dict(range=[0,1],
            label='Core coverage',
            tickvals=[0,0.2,0.4,0.6,0.8,1],
            values=synergy_plot['core_coverage'])
        ])
    )
)
fig.update_layout(font=dict(size=24), margin=dict(l=200, r=200),
                  height=600)
fig.show()

It appears that the tickvals argument is causing this. Once I remove that, the snapping behavior stops.