Plotly 快速折线图 - 获取默认 colors（如何按照字典 object 的指定对线条进行着色？）

Question

I have a data frame of a multivariate time series, for which I've created a interactive plotly express plot. I'm adding vertical lines at particular locations specified by a dictionary, each line associated to one of the time series, and wish to set the line color to agree with that of the corresponding variable. In essence, in the picture below, each vertical segment can be identified with one of Fp1 or Fp2 and I want to color it as red or black accordingly:

First I plot the data frame, where X is my time series matrix, plotly.express has been imported as px, pandas imported as pd, and channels=['Fp1','Fp2']:

df=pd.DataFrame(X,columns=channels)
df['id'] = df.index
df = pd.melt(df, id_vars='id', value_vars=df.columns[:-1])
fig = px.line(df, x='id', y='value', color='variable',labels = {'id':'time (20K~100 sec)', 
'value':'millivolts','variable':'channel'},title='Patient')

Subsequent calculations yield a dictionary, HFOs, where each key corresponds is one of my two channels, and each value is a list of times, eg something of the form

HFOs={'Fp1':[500,....,10500],'Fp2':[800,...11000]}

I then created my lines and added them to the figure:

for channel,times in HFOs.items():
    for t in times:        
        fig.add_shape(type='line',yref="y",xref="x",
        x0=t,y0=df['value'].min()*1.2,x1=t,y1=df['value'].max()*1.2,line=dict(color='black', 
        width=.25))
        fig.add_trace(go.Scatter(x=[t],y= 
        [df['value'].max()*1.5],mode='text',showlegend=False))    
fig.show()

This creates the image shown above. How do I modify line=dict(color='black',width=.25) to change the color to what I want? I wish for the vertical lines at times [500,....,10500]to be blue and times [800,...11000] to be red. (Of course, in the future, there will be many more channels.)

I tried replacing 'black' with 'variable" but that, not surprisingly, just resulted in an error message. I feel there must be a very simple way to achieve my goal.

Answer 1

Cool question. There might be a better solution, but here's the one I found. Replace the code that creates the vertical lines with the following:

# fetch the colors of the traces from the figure. 
colors = [trace.line["color"] for trace in fig.data]

for inx, (channel,times) in enumerate(HFOs.items()):
    for t in times:        
        fig.add_shape(type='line',yref="y",xref="x",
        x0=t,y0=df['value'].min()*1.2,x1=t,y1=df['value'].max()*2,
                      line=dict(color=colors[inx], width=3))
fig.show()

The resulting figure looks as follows (Random data, made the vertical lines wider for visibility):

An alternative way to get the list of default colors is to use

px.colors.qualitative.Plotly , which produces a list of 10 hex color codes. My understanding is that these colors would be used for the first 10 series, and then used again for traces 11-20, etc.