unsafe_update_figure Chart Customization

To customize a chart in a way that is not directly supported by Deephaven Plotly Express (dx), use the unsafe_update_figure parameter.
\\ Every dx chart is backed by a Plotly Figure. This object gets passed to unsafe_update_figure where it can be modified directly. See the Plotly Figure Reference docs for details on available Figure properties.

Warning

Update figure is marked “unsafe” because some modifications can entirely break your figure, and care must be taken.
\\ dx maps Table columns to an index of a trace within Figure.data which will break if the trace order changes. Do not remove traces. Add new traces at the end of the list.

unsafe_update_figure accepts a function that takes a Plotly Figure object as input and optionally returns a modified Figure object. If a Figure is not returned, it is assumed that the input Figure has been modified in place.

Examples

Bar Line

Add a line to bars in a bar plot with update_traces.

import deephaven.plot.express as dx

tips = dx.data.tips()


def update(figure):
    # Add a gray line to the bars
    figure.update_traces(marker_line_width=3, marker_line_color="gray")


bar_lined_plot = dx.bar(tips, x="Day", unsafe_update_figure=update)

Unsafe Update Figure Example

Legend Location

Change the location of the legend to the bottom of the plot by updating the layout.

import deephaven.plot.express as dx

tips = dx.data.tips()


def update(figure):
    # Update the layout to move the legend to the bottom
    # y is negative to move the legend outside the plot area
    figure.update_layout(
        legend=dict(orientation="h", yanchor="bottom", y=-0.3, xanchor="left", x=0.3)
    )


legend_bottom_plot = dx.scatter(
    tips, x="TotalBill", y="Tip", color="Day", unsafe_update_figure=update
)

Vertical Line

Add a vertical line to a plot with add_vline.

import deephaven.plot.express as dx

tips = dx.data.tips()


def update(figure):
    # Add a dashed orange vertical line at x=20
    figure.add_vline(x=20, line_width=3, line_dash="dash", line_color="orange")


scatter_vline_plot = dx.scatter(
    tips, x="TotalBill", y="Tip", unsafe_update_figure=update
)

Between Line Fill

Fill the area between lines in a line plot with fill="tonexty".

import deephaven.plot.express as dx

my_table = dx.data.stocks()

# subset data for just DOG transactions and add upper and lower bounds
dog_prices = my_table.where("Sym = `DOG`").update_view(
    ["UpperPrice = Price + 5", "LowerPrice = Price - 5"]
)


def update(figure):
    # tonexty fills the area between the trace and the previous trace in the list
    figure.update_traces(
        fill="tonexty", fillcolor="rgba(123,67,0,0.3)", selector={"name": "LowerPrice"}
    )
    figure.update_traces(
        fill="tonexty", fillcolor="rgba(123,67,0,0.3)", selector={"name": "Price"}
    )


# Order matters for y since the fill is between the trace and the previous trace in the list
filled_line_plot = dx.line(
    dog_prices,
    x="Timestamp",
    y=["UpperPrice", "Price", "LowerPrice"],
    unsafe_update_figure=update,
)