First update of plotly Dash with live update

user_1875459 0
Viewed 2801

I am trying to use live update with plotly Dash (0.18.3) in Python (3.x), but I don't need to update the plot so often (once or twice a day is what I need).

I did the following so far:

import dash
import pandas as pd
import plotly.graph_objs as go
import dash.dependencies as ddp
import dash_core_components as dcc
import dash_html_components as html

def plot_figure(data):
    layout = dict(
    title="Figure w/ plotly",
    )

    fig = dict(data=data, layout=layout)
    return fig

def serve_layout():
    return html.Div(children=[
        html.Div([
            html.H1("Plotly test with Live update",
                    style={"font-family": "Helvetica",
                           "border-bottom": "1px #000000 solid"}),
            ], className='banner'),
        html.Div([dcc.Graph(id='plot')],),
        dcc.Interval(id='live-update', interval=interval),
],)

second = 1000
minute = 60
hour   = 60
day    = 24
interval = 1/2*day*hour*minute*second

app = dash.Dash()

app.layout = serve_layout

@app.callback(
    ddp.Output('plot', 'figure'),
    [],
    [],
    [ddp.Event('live-update', 'interval')])
def gen_plot():
    ind = ['a', 'b', 'c', 'd']
    df = pd.DataFrame({'one' : pd.Series([4., 3., 2., 1.], index=ind),
                       'two' : pd.Series([1., 2., 3., 4.], index=ind)})

    trace = [go.Scatter(x=df.index, y=df['one'])]
    fig   = plot_figure(trace)
    return fig

if __name__ == '__main__':
    app.run_server(debug=True)

The problem is at first nothing appears, it updates only after interval, so after half a day. I added the serve_layout function following Dash documentation to have an update when the page is loaded, but it seems to have no effect.

How could I have a first update when the page is first accessed and then updates at every interval?

plotly-dash python-3.x
1 Answers

Based on a discussion on plotly forum, I found a solution. You need to call gen_plot() from inside serve_layout to get the most recent figure. In order to call gen_plot, you’ll need to remove the decorator from it.

To call gen_plot inside serve_layout, I moved it up in the code and add figure=gen_plot() to dcc.Graph in serve_layout.

import dash
import pandas as pd
import plotly.graph_objs as go 
import dash.dependencies as ddp
import dash_core_components as dcc
import dash_html_components as html

second = 1000
minute = 60
hour   = 60
day    = 24
interval = 1/2*day*hour*minute*second

def plot_figure(data):
    layout = dict(
        title="Figure w/ plotly",
    )

    fig = dict(data=data, layout=layout)
    return fig

def gen_plot():
    ind = ['a', 'b', 'c', 'd']
    df = pd.DataFrame({'one' : pd.Series([4., 3., 2., 1.], index=ind),
                       'two' : pd.Series([1., 2., 3., 4.], index=ind)})

    trace = [go.Scatter(x=df.index, y=df['one'])]
    fig   = plot_figure(trace)
    return fig

def serve_layout():
    return html.Div(children=[
        html.Div([
            html.H1("Plotly test with Live update",
                    style={"font-family": "Helvetica", 
                           "border-bottom": "1px #000000 solid"}),
            ], className='banner'),
        html.Div([dcc.Graph(id='plot', figure=gen_plot())],),
        dcc.Interval(id='live-update', interval=interval),
],)

app = dash.Dash()

app.layout = serve_layout

app.callback(
    ddp.Output('plot', 'figure'),
    [],
    [],
    [ddp.Event('live-update', 'interval')])(gen_plot)

if __name__ == '__main__':
    app.run_server(debug=True)

Note that in my real application I also needed to do so with excerpt of text that I update with the figure. I had a gen_text function with the same decorator as my gen_plot function, I applied the same strategy and I added a children=gen_text() argument to the related html.Div; works like a charm!

user_18754590
Related