Problem linking plots when returned from funtion

Community Support
1

Hello,

I have a problem when I try to return a scatter plot from a function. I plot 3 scatters plots in jupyter notebook and I write specifically the three in one cell I have no issue to have the selection synchronised among all of them.

But when I write every figure as the output of a function and then I try to use the gridplot I have the result with the three plots but not the linked behaviour. How to force the synchronisation?

Function

def volcanoplot(df, x, y, category_a, category_b, title='Volcano Plot', logy=True):
    df = df.copy().dropna()
    if logy==True:
        df['log_y'] = np.log10(df[y]) * -1
        y = 'log_y'

    #Legend
    df[category_a] = df[category_a].astype(str)
    df[category_b] = df[category_b].astype(str)

    Col = ['True', 'False']
    Mar = ['circle_x', 'triangle']

    tooltips = """
    <div style="width:200px;">
    <b>@mgi_symbol</b> </br>
    $x </br>
    @description
    </div>
    """
    
    TOOLS = "crosshair,pan,wheel_zoom,box_zoom,reset,box_select,lasso_select"


    source = ColumnDataSource(df)
    p = figure(title=title, tools=TOOLS, tooltips=tooltips, height= 500, width=500, output_backend="webgl")
    p.scatter(x=x, y=y,
             source=source,
             size=10,
             fill_alpha=0.4,
             legend=category_a,
             marker=factor_mark(category_b, Mar, Col),
             color=factor_cmap(category_a, 'Category10_3', Col))

    p.legend.title = category_a
    p.yaxis.axis_label = '-log10\(FDR\)'
    p.xaxis.axis_label = 'log2 FC'


    limit_x_a = Span(location=-1,
                    dimension='height', line_color='red',
                    line_dash='dashed', line_width=1)

    p.add_layout(limit_x_a)

    limit_x_b = Span(location=1,
                    dimension='height', line_color='red',
                    line_dash='dashed', line_width=1)

    p.add_layout(limit_x_b)

    limit_y = Span(location=np.log10(0.01) * -1,
                    dimension='width', line_color='red',
                    line_dash='dashed', line_width=1)

    p.add_layout(limit_y)

    return p

Cell.

v1 = volcanoplot(sakers_df, 'a', 'b', 'c', 'd')
v2 = volcanoplot(sakers_df, 'a', 'b', 'c', 'd')
v3 = volcanoplot(sakers_df, 'a', 'b', 'c', 'd')

v = gridplot([[v1, v2, v3]])

show(v)

Thanks!

2

Linked selection in Bokeh is expressed by sharing a data source between different glyphs. You will have to change the structure of your code to create a single CDS up front, and then pass that same CDS to be used by all the function calls.

1 Like
3

Thanks so much!