用 matplotlib 库将数据集 links = [ {"source": "category1", "target": "category2", "value": 10}, {"source": "category2", "target": "category3", "value": 15}, {"source": "category3", "target": "category4", "value": 20}, {"source": "category5", "target": "category6", "value": 25}, ]绘制sankey图的程序

matplotlib库也可以绘制Sankey图。下面是一个示例程序，根据提供的数据集绘制Sankey图。

import matplotlib.pyplot as plt
from matplotlib.sankey import Sankey

# 数据集
links = [
    {"source": "category1", "target": "category2", "value": 10},
    {"source": "category2", "target": "category3", "value": 15},
    {"source": "category3", "target": "category4", "value": 20},
    {"source": "category5", "target": "category6", "value": 25},
]

# 定义节点和链接
node_labels = []
node_indices = {}
flows = []

for link in links:
    source = link['source']
    target = link['target']
    value = link['value']
    
    if source not in node_indices:
        node_indices[source] = len(node_labels)
        node_labels.append(source)
    if target not in node_indices:
        node_indices[target] = len(node_labels)
        node_labels.append(target)
    
    flows.append(value)
    
# 创建Sankey图
fig, ax = plt.subplots()
sankey = Sankey(ax=ax, scale=0.1, offset=0.3)
sankey.add(flows=flows, 
           labels=node_labels, 
           orientations=[1, -1, 0, 1],
           pathlengths=[0.05, 0.05, 0.2, 0.2],
           trunklength=1.5,
           facecolor='#37c959',
           edgecolor='black',
           lw=1)
sankey.finish()

# 显示图像
plt.show()

程序中使用matplotlib库中的Sankey类绘制图像，需要定义节点和链接的信息。首先遍历数据集，根据数据集中的源节点和目标节点构建节点列表，同时记录每个节点的索引。接着根据链接信息构建流量列表。最后使用Sankey类绘制图形。

Sankey类的构造函数需要传入画布和缩放比例等参数。调用add方法添加流量和节点标签等信息，可以设置流量的方向和长度等属性。最后调用finish方法完成Sankey图的绘制。

程序运行后会在浏览器中打开一个新的网页，展示绘制的Sankey图。