如何在Plotly中为X轴元素添加两组100%堆叠柱状图并归一化Y轴？

我用Python的Plotly生成了分组柱状图，目前已满足大部分需求，但最后需要将其改为100%堆叠样式，使Y轴最大值为1，所有柱状图对齐。请问在Plotly中如何实现该需求？

原生成图表代码：

fig = go.Figure()
categories = grouped_values2['portfolio_category'].unique()

category_colors = ["#0070C0", "#FFC000", "#AD1457", "#00A651", "#DB4437", "#6E3159", "#F4A000", "#00788C", "#5FB3C3"]

for i, category in enumerate(categories):
  num1 = random.randint(0, 30)
  category_data = grouped_values2[grouped_values['portfolio_category'] == category]
  
  fig.add_trace(
    go.Bar(
      x=category_data['month'],
      y=category_data['importancia_realizado'],
      name=f'R - {category}',
      marker_color=category_colors[i],
      offsetgroup=0
    )
  )
    
  fig.add_trace(
    go.Bar(
      x=category_data['month'],
      y=category_data['importancia_referencia'],
      name=f'P - {category}',
      marker_color=category_colors[i],
      offsetgroup=1
    )
  )

  fig.update_layout(
    barmode='group',
    title='IMPORTÂNCIA PROJETADA (P) vs IMPORTANCIA REALIZADA',
    plot_bgcolor="rgba(0, 0, 0, 0)",
    xaxis_title='Mês',
    yaxis_title='Importância'
  )

fig.show()

对应数据：

month,month_number,portfolio_category,gmv_mtd,importancia_referencia,importancia_realizado
Apr,4,CAT1,3242989.05,0.230,0.188
Apr,4,CAT2,3490670.33,0.190,0.202
Apr,4,CAT3,2970437.35,0.170,0.172
Apr,4,CAT4,6310704.34,0.160,0.365
Apr,4,CAT5,1046692.36,0.120,0.610
Apr,4,CAT6,216612.53,0.130,0.130
Jun,6,CAT1,4449799.95,0.210,0.274
Jun,6,CAT2,5209480.79,0.240,0.321
Jun,6,CAT3,2478664.66,0.160,0.153
Jun,6,CAT4,1318476.99,0.130,0.810
Jun,6,CAT5,2349817.17,0.130,0.145
Jun,6,CAT6,425692.50,0.130,0.260
May,5,CAT1,4057372.58,0.210,0.284
May,5,CAT2,3504235.05,0.250,0.245
May,5,CAT3,2863510.23,0.170,0.201
May,5,CAT4,900040.46,0.130,0.630
May,5,CAT5,2564349.59,0.130,0.180
May,5,CAT6,385128.35,0.120,0.270

要实现100%堆叠的分组柱状图，核心是调整Plotly的柱状图模式和归一化规则，关键修改如下：

• 将barmode从group改为stack，同时添加barnorm='percent'参数，让每个分组内的柱状图按百分比堆叠，总和自动为100%（对应Y轴数值1）。
• 修正原代码中数据筛选的变量名错误（grouped_values改为grouped_values2），移除无用的random.randint代码。
• 显式设置Y轴范围为[0, 1]，确保所有柱状图高度对齐统一。

修改后的完整代码：

import plotly.graph_objects as go
import pandas as pd
from io import StringIO

# 加载并整理数据
data = """month,month_number,portfolio_category,gmv_mtd,importancia_referencia,importancia_realizado
Apr,4,CAT1,3242989.05,0.230,0.188
Apr,4,CAT2,3490670.33,0.190,0.202
Apr,4,CAT3,2970437.35,0.170,0.172
Apr,4,CAT4,6310704.34,0.160,0.365
Apr,4,CAT5,1046692.36,0.120,0.610
Apr,4,CAT6,216612.53,0.130,0.130
Jun,6,CAT1,4449799.95,0.210,0.274
Jun,6,CAT2,5209480.79,0.240,0.321
Jun,6,CAT3,2478664.66,0.160,0.153
Jun,6,CAT4,1318476.99,0.130,0.810
Jun,6,CAT5,2349817.17,0.130,0.145
Jun,6,CAT6,425692.50,0.130,0.260
May,5,CAT1,4057372.58,0.210,0.284
May,5,CAT2,3504235.05,0.250,0.245
May,5,CAT3,2863510.23,0.170,0.201
May,5,CAT4,900040.46,0.130,0.630
May,5,CAT5,2564349.59,0.130,0.180
May,5,CAT6,385128.35,0.120,0.270"""

grouped_values2 = pd.read_csv(StringIO(data))

fig = go.Figure()
categories = grouped_values2['portfolio_category'].unique()

category_colors = ["#0070C0", "#FFC000", "#AD1457", "#00A651", "#DB4437", "#6E3159", "#F4A000", "#00788C", "#5FB3C3"]

for i, category in enumerate(categories):
    category_data = grouped_values2[grouped_values2['portfolio_category'] == category]
    
    fig.add_trace(
        go.Bar(
            x=category_data['month'],
            y=category_data['importancia_realizado'],
            name=f'R - {category}',
            marker_color=category_colors[i],
            offsetgroup=0
        )
    )
      
    fig.add_trace(
        go.Bar(
            x=category_data['month'],
            y=category_data['importancia_referencia'],
            name=f'P - {category}',
            marker_color=category_colors[i],
            offsetgroup=1
        )
    )

fig.update_layout(
    barmode='stack',
    barnorm='percent',
    title='IMPORTÂNCIA PROJETADA (P) vs IMPORTANCIA REALIZADA',
    plot_bgcolor="rgba(0, 0, 0, 0)",
    xaxis_title='Mês',
    yaxis_title='Importância',
    yaxis_range=[0, 1]
)

fig.show()

修改后效果：每个月份会显示两个高度为1的堆叠柱，分别对应「实际完成重要性（R）」和「计划重要性（P）」的类别占比堆叠，所有柱状图完全对齐到Y轴最大值1。

内容的提问来源于stack exchange，提问作者Marcos Dias