go.Marker update problem

[agahkarakuzu]

Hi All,

When defined in a conditional statement, go.Marker fails to update marker attributes as described. To have a different color (than other markers in trace 3), I had to comment out the respective line.

if len(args) > 9:
        trace4 = go.Scatter(
                          x= args[9],
                          y= args[10],
                          mode='markers',
                          #marker=go.Marker(color='rgb(107,174,214)'),
                          xaxis='x1',
                          yaxis='y2',
                          name= 'Outliers',
                          text= args[11],
                          hoverinfo= 'all'
                          )
    else:
        trace4 = []

Am I not allowed to do such thing in the first place, or am I missing something?

Best,
Agah

[cldougl]

Hey @agahkarakuzu can you attach a simple and reproducible example and we'll take a look (you should be able to conditionally set marker color)

[agahkarakuzu]

Hi @cldougl , of course. Here is the link to the plot. Outliers does not appear as I define in the script:

def corPlotOutlier(*args):
    xData = args[0]
    yData = args[1]
    colorPoint = args[2]
    colorFill = args[3]
    dataName = args[4]
    xRefName = args[5]
    hvr = args[6]
    elX = args[7]
    elY = args[8]
    
    slope, intercept, r_value, p_value, std_err = stats.linregress(xData,yData)
    line = slope*xData+intercept
    if intercept<0:
        aNot = 'r: ' + str(round(r_value,2)) + ' | ' + 'p: ' + str(round(p_value,2)) + '<br>' + dataName  + ' | ' + 'y=' + str(round(slope,2)) + 'x'  + str(round(intercept,2))
    if intercept>0:
        aNot = 'r: ' + str(round(r_value,2)) + ' | ' + 'p: ' + str(round(p_value,2)) + '<br>' + dataName  + ' | '+  'y=' + str(round(slope,2)) + 'x+'  + str(round(intercept,2))


    
    trace1 = go.Scatter(
                      x= xData,
                      y= yData,
                      mode='markers',
                      marker=go.Marker(color=colorPoint, size = '10',line=dict(width=2, color = 'rgb(256,256,256)')),
                      opacity=0.7,
                      xaxis='x1',
                      yaxis='y2',
                      name= dataName,
                      text= hvr,
                      hoverinfo= 'all'
                      )

    trace2 = go.Scatter(
                      x=xData,
                      y=line,
                      mode='lines',
                      xaxis='x1',
                      yaxis='y2',
                      marker=go.Marker(color= colorPoint),
                      name=dataName + ' vs ' + xRefName,
                      hoverinfo= 'none'
                      )
    trace3 = go.Scatter(
                      x= elX,
                      y= elY,
                      fill= 'toself',
                      fillcolor = colorFill,
                      xaxis='x1',
                      yaxis='y2',
                      marker=go.Marker(color= 'rgba(0,0,0,0.001)'),
                      hoverinfo= 'none'
                      )
    if len(args) > 9:
        trace4 = go.Scatter(
                          x= args[9],
                          y= args[10],
                          mode='markers',
                          marker=go.Marker(color='rgb(107,174,214)', size = '10',line=dict(width=2, color = 'rgb(107,174,214)')),
                          xaxis='x1',
                          yaxis='y2',
                          name= 'Outliers',
                          text= args[11],
                          hoverinfo= 'all'
                          )
    else:
        trace4 = []
    
    return trace1, trace2, trace3, trace4, aNot

[cldougl]

Thanks - looks like that plot is private. Could you attach a simple reproducible example with data that we can run?

[agahkarakuzu]

@cldougl Here is the link to the online executable jupyter notebook, the whole thing is reproducible there. The problem that I encounter occurs for the last figures (4.1 and 4.2).

P.S. You need to execute previous (hidden) cells prior to the last one.

Thank you!

[agahkarakuzu]

@cldougl By the way, plot was not private but I deleted that to make some space for my free account :)

[cldougl]

Thanks- I'm not able to run through that whole notebook (typically it's easiest to determine the bug with a simplified example that's based on dummy data rather than a user dataset).

Modifying the code you pasted above, you should be able to run:

import plotly.offline as off
import plotly.graph_objs as go

args = [0, 1, 2, 3, 4, 5]
args_T4 = [0, 1, 2, 3, 4, 5, 6, 7]

def generate_traces(args):
    trace0 = go.Scatter(x= [args[0]],
                        y= [args[1]],
                        mode='markers',
                        marker=go.Marker(color='blue', size=8))
    trace1 = go.Scatter(x= [args[2]],
                        y= [args[3]],
                        mode='markers',
                        marker=go.Marker(color='blue', size=8))
    trace2 = go.Scatter(x= [args[4]],
                        y= [args[5]],
                        mode='markers',
                        marker=go.Marker(color='blue', size=8))
    if len(args) > 6:
        trace3 = go.Scatter(x= [args[6]],
                            y= [args[7]],
                            mode='markers',
                            marker=go.Marker(color='pink', size=8))
    else:
        trace3 = {}
    
    data = [trace0, trace1, trace2, trace3]
    
    return data

data_T4 = generate_traces(args_T4)
off.plot(data_T4)

Here is a notebook comparing the 2 dummy datasets: https://plot.ly/~chelsea_lyn/16834

If you're still having an issue with your example feel free to check out and post in our community forum!

[agahkarakuzu]

@cldougl Thank you so much! Although brackets around args[] did not solved my issue, casting the trace data in the conditional statement similar to the remaining did the trick.

My take home message from this issue: Data type should be uniform among traces to be appended :)

I wish us Montrealers a warm weekend (I guess this would be the best wish)!