如何在Python Plotly时间轴中添加可打开本地Word文档的点击链接？

我在Jupyter环境用Python 3，通过以下代码用Plotly生成PDF时间轴：

import plotly.express as px
import pandas as pd

source = pd.DataFrame([
    {"filename": 'blablabla', "start": '1795-01-01', "end": '1810-01-01'},
    {"filename": "Sense and Sensibility","start": '1792-01-01', "end": '1797-01-01'},
    {"filename": "Call the Midwife", "start": '1957-01-01', "end": '1967-02-01'}, 
])

source['start'] = pd.to_datetime(source['start'])
source['end'] = pd.to_datetime(source['end'])
 
fig = px.timeline(source.sort_values('start'),
                  x_start="start",
                  x_end="end",
                  y="filename",
                  text="filename",
                  color_discrete_sequence=["tan"])
fig.show()

fig.write_image("C://Users/mytimeline.pdf")

生成和导出PDF都正常，但想把文件名改成可点击的超链接，点击打开本地Word文档。尝试用HTML <a> 标签替换文件名：

{"filename": '<a href=".1CL.docx">Call the Midwife</a>', "start": '1957-01-01', "end": '1967-02-01'}

遇到两个问题：

• Jupyter预览时，文本显示为蓝色链接，但点击会打开新标签页报错；用完整本地路径（C://Users...1CL.docx）时，连新标签页都不会打开。
• 导出PDF后，链接样式消失，也无法点击。

1. Jupyter预览中实现可点击本地链接 #

Plotly的text参数默认不解析HTML链接的本地协议，浏览器对file://协议有安全限制，直接点击会被拦截。需通过自定义JavaScript事件+customdata存储路径实现：

import plotly.express as px
import pandas as pd
from IPython.display import display, HTML

# 新增本地路径列，空路径对应无需链接的项
source = pd.DataFrame([
    {"filename": 'blablabla', "start": '1795-01-01', "end": '1810-01-01', "local_path": ""},
    {"filename": "Sense and Sensibility","start": '1792-01-01', "end": '1797-01-01', "local_path": ""},
    {"filename": "Call the Midwife", "start": '1957-01-01', "end": '1967-02-01', "local_path": "C://Users/xxx/1CL.docx"}, 
])

source['start'] = pd.to_datetime(source['start'])
source['end'] = pd.to_datetime(source['end'])
 
fig = px.timeline(source.sort_values('start'),
                  x_start="start",
                  x_end="end",
                  y="filename",
                  text="filename",
                  color_discrete_sequence=["tan"])

# 给图形trace绑定本地路径数据
fig.data[0].customdata = source['local_path'].tolist()

# 添加点击事件监听，仅在Jupyter环境生效
display(HTML('''
<script>
document.addEventListener('DOMContentLoaded', function() {
    let plotDiv = document.querySelector('.plotly-graph-div');
    plotDiv.on('plotly_click', function(data) {
        let localPath = data.points[0].customdata;
        if (localPath) {
            // 转换路径格式并调用浏览器打开
            let formattedPath = 'file:///' + localPath.replace(/\\\\/g, '/');
            window.open(formattedPath, '_blank');
        }
    });
});
</script>
'''))

fig.show()

注意：Chrome等浏览器默认禁止网页打开本地文件，需启动时添加参数解除限制：

• Windows：chrome.exe --allow-file-access-from-files
• macOS：open /Applications/Google\ Chrome.app --args --allow-file-access-from-files

2. 导出PDF时保留可点击本地链接 #

Plotly的write_image依赖Kaleido，不支持导出HTML链接到PDF。需先导出为HTML，再用weasyprint转PDF：

步骤1：安装依赖

pip install weasyprint

步骤2：导出带链接的Plotly HTML

import plotly.express as px
import pandas as pd

source = pd.DataFrame([
    {"filename": 'blablabla', "start": '1795-01-01', "end": '1810-01-01', "local_path": ""},
    {"filename": "Sense and Sensibility","start": '1792-01-01', "end": '1797-01-01', "local_path": ""},
    {"filename": "Call the Midwife", "start": '1957-01-01', "end": '1967-02-01', "local_path": "C://Users/xxx/1CL.docx"}, 
])

source['start'] = pd.to_datetime(source['start'])
source['end'] = pd.to_datetime(source['end'])
 
fig = px.timeline(source.sort_values('start'),
                  x_start="start",
                  x_end="end",
                  y="filename",
                  color_discrete_sequence=["tan"])

# 手动修改y轴标签为带本地链接的HTML
fig.update_yaxes(
    tickmode='array',
    tickvals=source['filename'],
    ticktext=[
        f'<a href="file:///{row["local_path"].replace("\\\\", "/")}">{row["filename"]}</a>' 
        if row["local_path"] else row["filename"] 
        for _, row in source.iterrows()
    ]
)

# 导出HTML文件
fig.write_html("timeline_with_links.html")

步骤3：HTML转PDF

from weasyprint import HTML

HTML("timeline_with_links.html").write_pdf("C://Users/mytimeline_with_links.pdf")

导出的PDF中，文件名会显示为可点击链接，点击后可打开对应本地Word文档（需PDF阅读器支持本地链接跳转）。

内容的提问来源于stack exchange，提问作者Hugo