Only update some traces in graph

Dash Python
1

Hi,

I have a graph with a lot of data, then I add three horizontal lines to view a kind of threshold which can be adjusted by changing the input value of an input box.

The graph itself looks like on the attached picture

Timeline
Timeline1413×387 32.5 KB

When I update to the input value to adjust the thresholds(which is just a line between two datasets), the whole graph have to be updated which takes a couple of seconds.

Is it possible to only update the three lines so the rest of the data stays unchanged and doesn’t need to be update?

2

yes it is possible.
Just give visibility= True,
or Visibility = legendonly

Just look at the plotly documentation https://plot.ly/javascript/reference/#scatter-visible

3

I don’t think this answers my question…
Because all 4 legends should be visible, I just only want to update 3 of them. The 4’th should somehow be static.

4

Any updates on this? I have a vertical bar I want to update like OP but I want the data to stay static (most of the time). The performance is horrific since I update the whole figure every time. Can I update the single trace and prevent the performance hit associated with drawing 6000 points?

5

I have a temporary solution. I found this custom dash component: mydcc from https://github.com/jimmybow/mydcc. He has developed a ChangeTrace dash component to update plotly graph trace (only for graph with one trace).
You can install it with:

pip install mydcc

But, the problem is if you have multiples traces in the graph that component won’t work properly. I changed the code of him a little bit to support multiple traces. But, my version needs all the traces in a list to be updated (traces with new data and traces with old data):

You can try it and see if this temporary solution is suitable for you. The good thing with this is that you don’t need to re-draw the graph again, just all the data in the graph. It will be much faster. In the future a could try a way to set a option to delete a specific trace and then, change its content something like:

fig['data][trace to modify]=new data

2 Likes
6

Seems somewhat basic, is there really no way to update only some traces in a figure supported officially?
Is there a way to up vote a request for such a feature?

7

There is the extend data property,

8

Not exactly what I wanted but it helped, I ended up using State as I found here https://community.plotly.com/t/extend-or-append-data-instead-of-update/8898
Here is a working code example

import dash
from dash import dcc
from dash import html
from dash.dependencies import Input, Output, State
import plotly.graph_objects as go


if __name__ == '__main__':
    app = dash.Dash()
    app.layout = html.Div([dcc.Graph(id='graph'),
                           dcc.Input(id='in', value=7, type='number')])

    @app.callback(
        Output('graph', 'figure'),
        Input('in', 'value'),
        State('graph', 'figure'))
    def update_figure(v, state):
        if state is None:
            fig = go.Figure()
            x = list(range(10))
            y = list(range(5, 25, 2))
            z = [v] * len(x)
            fig.add_trace(go.Scatter(x=x, y=y, mode='lines', name='const'))
            fig.add_trace(go.Scatter(x=x, y=z, mode='lines', name='var'))
            return fig
        offset = [elem['name'] for elem in state['data']].index('var')
        x = state['data'][offset]['x']
        state['data'][offset]['y'] = [v] * len(x)
        return state

    app.run_server()

I think this is a common need so I hope it helps

3 Likes
9

This is great! I used this approach to sync a playhead (vertical line) on my graph with an audio file played via dash_player.

@callback(
    Output('graph-content', 'figure'),
    Input('audio-player', 'currentTime'),
    State('graph-content', 'figure')
)
def update_graph(current_time: float | int, figure: dict | None):
    if figure is None:
        # Used to add figure elements on the first callback only
        first_fig = px.scatter(points, x='time', y=['SPEAKER_00', 'SPEAKER_01'])
        first_fig.add_vline(x=current_time, line_width=3, name='playhead')

        return first_fig

    # If the scatter plot has been drawn, only update the vertical line
    figure['layout']['shapes'][0]['x0'] = current_time
    figure['layout']['shapes'][0]['x1'] = current_time

    return figure```
10

partial callbacks Patch do the trick.