Scattermode=group does not autoscale yaxis

I'm plotting some data with a very elaborate box plot series, and attempting to overlay connecting median lines between each box in each series. I can do this successfully, but due the density of information, I need to stagger each x value using boxmode=group and scattermode=group. There are no issues with boxmode, but scattermode causes the lower y min to be set to 0. When scattermode is disabled, everything scales nicely, and when no boxes are drawn, the same issues persists. Using subplots and ensuring shared axes doesn't work, nor does updating axes with autorange.

Code:
`

fig = go.Figure()

point_visibility = point_checked
point_legend = False
box_visibility = True
match box_checked:
    case False:
        box_visibility = False
        point_legend = True

for col in dcr.columns[1:]:
    if box_checked:
        fig.add_trace(go.Box(
            x=dcr['SOC'],
            y=dcr[col],
            name=col,
            legendgroup=col,
            visible=box_visibility,
        ))
    if point_checked:
        fig.add_trace(go.Scatter(
            x=medians['SOC'],
            y=medians[col],
            name=col,
            legendgroup=col,
            showlegend=point_legend,
            visible=point_visibility,
            mode='lines',
        ))

fig.update_layout(
    yaxis_title='DCIR (Ohms)',
    xaxis_title='SOC',
    boxmode='group',
    scattermode='group',  # figure out how to align this to the boxes
)
fig.update_yaxes(autorange=True)  # has no effect

`

Not scaling:

Disabling scattermode=group (No longer align to boxes):

What it should look like:

Start by reproducing the supplied box-and-scatter example in plotly.py, comparing y-axis autorange with and without scattermode='group'. Trace the scattermode grouping and autorange entry points to identify where the lower bound becomes zero, then add a regression test covering grouped scatter traces with box traces and verify that autorange matches the plotted data.