app.py

import gradio as gr
from bs4 import BeautifulSoup
from markdownify import MarkdownConverter
from playwright.sync_api import sync_playwright


def md(soup, **options):
    return MarkdownConverter(**options).convert_soup(soup)


def main_fn(url: str, check: list[int], request: gr.Request):

    user_agent = request.headers["user-agent"]

    with sync_playwright() as p:

        browser = p.chromium.launch(
            args=[
                "--single-process",
                "--no-zygote",
                "--no-sandbox",
                "--disable-gpu",
                "--disable-dev-shm-usage",
                "--headless=new",
            ]
        )

        context = browser.new_context(user_agent=user_agent)
        page = context.new_page()

        response = page.goto(url=url)
        status = response.status

        content = page.content()
        title = page.title()

        browser.close()

    soup = BeautifulSoup(content, features="html.parser")

    for tag in ["script", "style"]:
        target = soup.find_all(tag)
        for t in target:
            t.clear

    body = soup.find("body")
    main = soup.find("main")

    body = md(body)

    if main:
        body = md(main, strip=check)

    return f"{title}\n======\n\n{body}"


demo = gr.Interface(
    main_fn,
    title="URL to Markdown",
    description="""<div style="width: fit-content; margin: 0 auto;">Gets HTML given by URL and converts it to Markdown.Does not support dynamically generated HTML such as React.</div>
    <div style="width: fit-content; margin: 0 auto;">URLで与えたHTMLを取得してMarkdownに変換します。Reactなどの動的に生成されるHTMLには対応していません</div>""",
    inputs=[
        gr.Text(label="URL", placeholder="https://*****"),
        gr.CheckboxGroup(
            label="Ignore tags（無視するタグ）",
            choices=["a", "img", "noscript"],
            value=["a", "img"],
        ),
    ],
    outputs=[gr.TextArea(label="Markdown", show_copy_button=True)],
    allow_flagging="never",
)

demo.launch(server_name="0.0.0.0")