HTTPChannel basic authentication

Scripting Issues
1

I need to issue HTTPS GETs from within my Panda3d application to a site that uses basic authentication. This is my first time using HTTPSClient and HTTPSChannel. I haven’t found how to get them to successfully handle basic authentication. When I run the following code, I’m expecting the returned data to be the requested document. Instead, I receive a login page.

The login page response indicates that the server did not recognize an acceptable Authorization header such as “Authorization: Basic userid:password”. It has returned a login page for me to interactively login.

The site returns the expected document when I access it using the python requests HTTPS library and the Postman application, but not when using HTTPClient+HTTPChannel. What am I missing in setting up basic authentication?

server = 'ncs01.case.edu'
url = 'https://' + server + '/webacs/api/v4/data/AccessPointDetails.json?.full=true&.maxResults=1000&.nocount=true'
client = HTTPClient()
client.add_direct_host(server)
client.set_verify_ssl(HTTPClient.VS_no_verify)
user, password = mylib.credentials(server, None)  # get my default credentials for server
user_password = HTTPClient.base64_encode(f"{user}:{password}")
client.set_username(server, '', user_password)
chan: HTTPChannel = client.make_channel(False)
chan.beginConnectTo(url)   # non-blocking direct connection
while chan.run():
    time.sleep(0.1)
print(f"is_connection_ready={chan.is_connection_ready()}")
chan.beginGetDocument(url)  # sends non-blocking GET?
while chan.run():  # turns False with operation is complete
    time.sleep(0.1)
if not chan.is_valid():
    raise ValueError(f"invalid url")
status = chan.get_status_code()
if status == 200:  # success
    stream: 'ISocketStream' = chan.open_read_body()
    data = bytes()
    reader = StreamReader(stream, True)
    while True:
        s = reader.extract_bytes(100000)
        print(f"extracted {len(s)} bytes")
        if len(s) == 0:  # stream EOF?
            if stream.is_closed():  # document end?
                print(f"stream is_closed")  # Yes
                break  # end of data
            else:  # No. More to come
                time.sleep(0.1)
        else:  # has data
            data += s  # aggregate segment into data
    # Received whole document
    print(f"Received {len(data)} bytes")
    ss = str(data, 'UTF-8')
    print(ss)
else:
    print(f"get_status_code={status}")
2

Just specify “user:password”, Panda will do the base64 step, I think it’s currently being doubly encoded.

I also suggest leaving the server argument blank to be sure that Panda will pick the chosen username for this request.

Note that 100 ms is a long time to sleep while handling the request. If you want to do it synchronously, you can just use getDocument instead of beginGetDocument.

3

Thanks. 100 msec is just for this test. In the Panda application, I’ll test the .run() once per frame in a task.

I changed the client.set_username line to

client.set_username('', '', f"{user}:{password}")

However, the server still returns the login page.

4

Could it be that the server is not returning a 401 error with the WWW-Authenticate header? I think HTTPChannel expects that to be sent in order to retry with authentication.

You can always set the header yourself using channel.send_extra_header.

5

I inserted the send_extra_header as below. Same results…

user, password = mylib.credentials(server, None)  # get my default credentials for server
client.set_username('', '', f"{user}:{password}")
chan: HTTPChannel = client.make_channel(False)
chan.send_extra_header('Authorization', f"Basic {user}:{password}")
chan.beginConnectTo(url)   # non-blocking direct connection
6

In another trial, instead of putting the send_extra_header just prior to the beginConnect(url), I tried putting the send_extra_header just prior to the beginGetDocument. This causes chan.is_valid() to return false.

server = 'ncs01.case.edu'
url = 'https://' + server + '/webacs/api/v4/data/AccessPointDetails.json?.full=true&.maxResults=1000&.nocount=true'
client = HTTPClient()
client.add_direct_host(server)
client.set_verify_ssl(HTTPClient.VS_no_verify)
user, password = mylib.credentials(server, None)  # get my default credentials for server
client.set_username('', '', f"{user}:{password}")
chan: HTTPChannel = client.make_channel(False)
chan.beginConnectTo(url)   # non-blocking direct connection
while chan.run():
    time.sleep(0.1)
print(f"is_connection_ready={chan.is_connection_ready()}")
chan.send_extra_header('Authorization', f"Basic {user}:{password}")
chan.beginGetDocument(url)  # sends non-blocking GET?
while chan.run():  # turns False with operation is complete
    time.sleep(0.1)
if not chan.is_valid():
    raise ValueError(f"invalid url")
...
7

If you pass it as custom header then you should base64 it, since Panda isn’t handling that for you then. It should be done before begin_get_document.

8

I’m just shooting in the dark for what send_extra_header needs given its (key, value) signature. I tried the three versions below, with no success.

...
chan: HTTPChannel = client.make_channel(False)
chan.beginConnectTo(url)   # non-blocking direct connection
while chan.run():
    time.sleep(0.1)
print(f"is_connection_ready={chan.is_connection_ready()}")
# chan.send_extra_header('Authorization', HTTPClient.base64_encode(f"Basic {user}:{password}"))  # does nothing
# chan.send_extra_header('Authorization', HTTPClient.base64_encode(f"{user}:{password}"))  # does nothing
chan.send_extra_header(HTTPClient.base64_encode('Authorization'),
                       HTTPClient.base64_encode(f"Basic {user}:{password}"))  # --> invalid URL
chan.beginGetDocument(url)  # sends non-blocking GET?
while chan.run():  # turns False with operation is complete
    time.sleep(0.1)
if not chan.is_valid():
    raise ValueError(f"invalid url")
...
9

This is what you would need:

chan.send_extra_header('Authorization', "Basic " + HTTPClient.base64_encode(f"{user}:{password}"))
10

It works! thanks!

1 Like