Airflow 2.3.3插件托管Great Expectations HTML页面404问题求助

问题 #

Airflow 2.3.3插件用于托管Great Expectations 0.1.5生成的HTML输出，目前能成功加载index.html，但点击页面内的任意链接（如validations/src/dag1/warning/...路径下的HTML文件）时，会跳转到Airflow 404页面。

现有插件代码 #

from airflow.plugins_manager import AirflowPlugin
from flask import Blueprint
from flask_appbuilder import expose, BaseView as AppBuilderBaseView
import os

class GreatExpectationsView(AppBuilderBaseView):

    @expose('/')
    def list(self):
        return self.render_template("index.html", content="")

v_appbuilder_view = GreatExpectationsView()

v_appbuilder_package = {
    "name": "Great Expectations",
    "category": "Browse",
    "view": v_appbuilder_view,
}

bp = Blueprint(
    "great_expectation",
    __name__,
    template_folder="/opt/great_expectations",
    static_folder='/opt/great_expectations/static',
    static_url_path='/greatexpectationsview/static'
)


class AirflowGEPlugin(AirflowPlugin):
    name = "great_expectation"
    appbuilder_views = [v_appbuilder_package]
    flask_blueprints = [bp]

Great Expectations HTML文件夹结构 #

/opt/great_expectations/
|-- expectations
|   `-- src
|       `-- dag1
|           `-- warning.html
|-- index.html
|-- static
|   |-- fonts
|   |   `-- HKGrotesk
|   |       |-- HKGrotesk-Bold.otf
|   |       |-- HKGrotesk-BoldItalic.otf
|   |       |-- HKGrotesk-Italic.otf
|   |       |-- HKGrotesk-Light.otf
|   |       |-- HKGrotesk-LightItalic.otf
|   |       |-- HKGrotesk-Medium.otf
|   |       |-- HKGrotesk-MediumItalic.otf
|   |       |-- HKGrotesk-Regular.otf
|   |       |-- HKGrotesk-SemiBold.otf
|   |       `-- HKGrotesk-SemiBoldItalic.otf
|   |-- images
|   |   |-- favicon.ico
|   |   |-- glossary_scroller.gif
|   |   |-- iterative-dev-loop.png
|   |   |-- logo-long-vector.svg
|   |   |-- logo-long.png
|   |   |-- short-logo-vector.svg
|   |   |-- short-logo.png
|   |   `-- validation_failed_unexpected_values.gif
|   `-- styles
|       |-- data_docs_custom_styles_template.css
|       `-- data_docs_default_styles.css
`-- validations
    `-- src
        `-- dag1
            `-- warning
                |-- 20220914-170251-src_dag1
                |   `-- 20220914T170251.386427Z
                |       `-- 1811b6d2fab709c0e35cab11873288dc.html
                `-- 20220914-182123-src_dag1
                    `-- 20220914T182123.138118Z
                        `-- 1811b6d2fab709c0e35cab11873288dc.html

解决方案 #

问题核心是现有插件仅处理了根路径的请求，未对GE生成的其他HTML文件路径提供路由支持。修改插件代码，添加动态路由捕获所有文件请求，并使用send_from_directory返回对应文件：

from airflow.plugins_manager import AirflowPlugin
from flask import Blueprint, send_from_directory
from flask_appbuilder import expose, BaseView as AppBuilderBaseView
import os

# 定义GE的根目录
GE_ROOT_DIR = "/opt/great_expectations"

class GreatExpectationsView(AppBuilderBaseView):

    @expose('/')
    def list(self):
        # 根路径返回index.html
        return send_from_directory(GE_ROOT_DIR, "index.html")
    
    @expose('/<path:filename>')
    def serve_file(self, filename):
        # 捕获所有子路径请求，返回对应文件
        return send_from_directory(GE_ROOT_DIR, filename)

v_appbuilder_view = GreatExpectationsView()

v_appbuilder_package = {
    "name": "Great Expectations",
    "category": "Browse",
    "view": v_appbuilder_view,
}

bp = Blueprint(
    "great_expectation",
    __name__,
    static_folder='/opt/great_expectations/static',
    static_url_path='/greatexpectationsview/static'
)


class AirflowGEPlugin(AirflowPlugin):
    name = "great_expectation"
    appbuilder_views = [v_appbuilder_package]
    flask_blueprints = [bp]

关键修改说明 #

1. 添加动态路由：新增@expose('/<path:filename>')装饰的serve_file方法，匹配所有子路径请求。
2. 改用send_from_directory：替换render_template为send_from_directory，直接返回GE目录下的静态HTML文件（GE生成的HTML不是Jinja模板，无需渲染）。
3. 移除template_folder：既然不再使用模板渲染，Blueprint中可以去掉template_folder配置。
4. 定义GE_ROOT_DIR常量：统一管理GE的根目录路径，便于维护。

额外注意事项 #

• 确保Airflow运行进程对/opt/great_expectations目录有读取权限，否则会返回权限错误。
• 静态资源（CSS、图片、字体）已通过Blueprint的static_folder和static_url_path配置处理，无需额外修改。

内容的提问来源于stack exchange，提问作者wymangr