Fig.add_vline not updating with slider callback

Dash Python
1

I have a dash plot that includes two distributions and I want a vertical line to indicate the mean of the first distribution. I’d like all three pieces to be updated with a slider callback.

This plotting function successfully gets called by a callback initiated by a slider in my app. dist1, dist2, and dist1_mean all get updated, but regardless of the slider position, the vertical line stays fixed at its initial value. It’s like the add_vline isn’t getting updated with the callback even though a new fig object is presumably being created each time.

I’ve confirmed dist1_mean is changing depending on the filter and slider value.

def plot_distribution_slider(self, df, filter_ind, slider_ind):
        filter = (#filter_ind#)

        dist1 = df.loc[filter, 'dist1_cols'].iloc[slider_ind]
        dist2 = df.loc[filter, 'dist2_cols'].iloc[slider_ind]

        dist1_mean = df.loc[filter, 'dist1_mean_col'].iloc[slider_ind]

        x_axes = list(range(#, #))

        fig = go.Figure()

        # Distribution 1
        fig.add_trace(
            go.Scatter(
                x=x_axes,
                y=dist1,
                name='dist1'
            )
        )

        # Distribution 2
        fig.add_trace(
            go.Scatter(
                x=x_axes,
                y=dist2,
                name='dist2'
            )
        )

        fig.update_layout(
             ####
        )

        fig.add_vline(
            x=dist1_mean,
        )

return fig

My callback, fwiw:

@app.callback(
    Output('distributions', 'figure'),
    Input('all-data-graph', 'clickData'),
    Input('time-slider', 'drag_value'))
def update_distribution(click, slider_ind):
    filter_ind = df.loc[df.index == click['points'][0]['x'], 'runs'].values[0]
    return Plot().plot_distribution_slider(df, filter_ind, slider_ind)

I’ve had this same issue in the past with add_vline and thought maybe it was just some issue with my current environment, but it’s still not working. I fixed it previously by manually adding a line shape. But this is a verbose work-around. Id’ prefer to get this seemingly very simple feature working.

Here was my previous solution, for those having the same issue:

        fig_std.update_layout(
            ###

            shapes=[
            dict(
                type= 'line',
                yref= 'paper', y0= 0, y1= 1,
                xref= 'x', 
                x0= dist_mean,
                x1= dist_mean,
                line=dict(
                            color="Blue",
                            width=2,
                            dash="dashdot",
                    )
            ),

            )
       ])

dash==1.20.0
dash-core-components==1.16.0
dash-html-components==1.1.3
dash-renderer==1.9.1
dash-table==4.11.3
plotly==4.14.3

Edit:
I seem to have isolated the bug. When I fix the y-axis, the vertical line doesn’t get updated. If I then adjust the plot by zooming in etc, the vertical line gets updated. Here is the culprit:

        fig.update_layout(
            yaxis=dict(
                range=[-1, 50]
            ),
        )

Regardless of whether I fix the yaxis before or after setting the vertical line, it’s failing to update.