Function parse_proxy

src/websockets/proxy.py:48–95

Parse and validate a proxy. Args: proxy: proxy. Returns: Parsed proxy. Raises: InvalidProxy: If ``proxy`` isn't a valid proxy.

(proxy: str)

def parse_proxy(proxy: str) -> Proxy:
    """
    Parse and validate a proxy.

    Args:
        proxy: proxy.

    Returns:
        Parsed proxy.

    Raises:
        InvalidProxy: If ``proxy`` isn't a valid proxy.

    """
    parsed = urllib.parse.urlparse(proxy)
    if parsed.scheme not in ["socks5h", "socks5", "socks4a", "socks4", "https", "http"]:
        raise InvalidProxy(proxy, f"scheme {parsed.scheme} isn't supported")
    if parsed.hostname is None:
        raise InvalidProxy(proxy, "hostname isn't provided")
    if parsed.path not in ["", "/"]:
        raise InvalidProxy(proxy, "path is meaningless")
    if parsed.query != "":
        raise InvalidProxy(proxy, "query is meaningless")
    if parsed.fragment != "":
        raise InvalidProxy(proxy, "fragment is meaningless")

    scheme = parsed.scheme
    host = parsed.hostname
    port = parsed.port or (443 if parsed.scheme == "https" else 80)
    username = parsed.username
    password = parsed.password
    # urllib.parse.urlparse accepts URLs with a username but without a
    # password. This doesn't make sense for HTTP Basic Auth credentials.
    if username is not None and password is None:
        raise InvalidProxy(proxy, "username provided without password")

    try:
        proxy.encode("ascii")
    except UnicodeEncodeError:
        # Input contains non-ASCII characters.
        # It must be an IRI. Convert it to a URI.
        host = host.encode("idna").decode()
        if username is not None:
            assert password is not None
            username = urllib.parse.quote(username, safe=DELIMS)
            password = urllib.parse.quote(password, safe=DELIMS)

    return Proxy(scheme, host, port, username, password)


def get_proxy(uri: WebSocketURI) -> str | None: