Solution 1:

You can use fig.update_traces(hovertemplate) with the following setup:

fig.update_traces(hovertemplate = 'hover_data_0=%{customdata[0]}<br>label=%{customdata[1]}<extra></extra>')

Plot 1:

enter image description here

Complete code:

import plotly.express as px
import plotly.graph_objects as go
import pandas as pd

df=(pd.DataFrame({'label_color':['white','white','cyan','cyan','cyan','cyan','white','white'],
                  'label_quality':['white','white','red','green','green','red','white','white'],
'label':['foo','foo','foo','foo','foo','foo','foo','foo']}))

df['color_value'] = 1
df['quality_value'] = 1

fig = px.bar(df, y=['color_value','quality_value'],
             x=[1]*len(df),
             orientation='h',
             barmode='group',
             template='plotly_white', hover_data=[df.index.values,df.label])

fig.data[0]['marker']['color'] = df['label_color'].tolist()
fig.data[1]['marker']['color'] = df['label_quality'].tolist()
fig.update_traces(marker_line_color='rgb(8,48,107)')
fig.update_layout(showlegend=False, yaxis_title='foo', xaxis_title='')
fig.update_traces(hovertemplate = 'hover_data_0=%{customdata[0]}<br>label=%{customdata[1]}<extra></extra>')
fig.show()