Requests-HTML 0.10.0: Configure the Chromium Proxy for render()

In Requests-HTML 0.10.0, session.get() uses Requests, while response.html.render() starts Pyppeteer and normally navigates Chromium to the page URL again. A successful first response does not establish the browser's route.

For an existing compatible legacy environment using an unauthenticated HTTP proxy:

from requests_html import HTMLSession

proxy_url = 'http://proxy.example:8080'
session = HTMLSession(browser_args=[
    '--proxy-server=' + proxy_url,
])
try:
    response = session.get(
        'https://www.example.org/',
        proxies={'https': proxy_url},
        timeout=20,
    )
    response.html.render(timeout=20)
    rendered = response.html.html
finally:
    session.close()

For the public-page example, the HTTP proxy list provides candidates for proxy_url. browser_args is the session's public constructor input. A Requests proxies dictionary does not populate those Chromium arguments.

Requests' socks5h:// spelling also cannot simply be copied into Chromium. Chromium accepts socks5://, resolves SOCKS5 destination names through the proxy, and does not support SOCKS5 username/password authentication.

render(reload=False) supplies the existing HTML to the browser instead of navigating to the page URL. Scripts and subresources can still make requests. It is useful when deliberately rendering that response, but it does not turn the Requests connection into the browser's transport.