Document IDs encoded server side on set, not on get

[fischer70]

Describe the bug
If a document ID contains a utf-8 character or (most) reserved characters it is percent encoded (looks like on the server side) for the document id. However, it is not on a get_document call, which means clients need to carefully encode for get_doc to work as expected.

To Reproduce

    class FakeDoc(ParentDocument):
        name: str
        _key = LexicalKey(["name"])
        aliastype: Optional[str]

    client = OpenDB()
    awfulstring = "2H₂ + O₂ ⇌ 2H₂O, R = 4.7 kΩ, ⌀ 200 mm(ish)!"
    f = FakeDoc(name=awfulstring, aliatype=awfulstring)
    client.update_document(f)
    # Fails with document not found error
    try:
        client.get_document("FakeDoc/%s" % awfulstring)
    except:
        print("Nope")
    #Succeeds:
    t = client.get_document("FakeDoc/%s" % urllib.parse.quote(awfulstring, safe=",=()'!"))
    print(t)

Expected behavior
It would be kind if the client provided encoding, so that set/get on the same docid worked. Unclear exactly what encoding is going on, I merely emulated it by trial and error. IE, "()" are not encoded but "[]" are, so minimally a helper to emulate the encoding propertly would be useful.

Error logs
Here is the encoding of the awful string above:
{'@id': 'FakeDoc/2H%E2%82%82%20%2B%20O%E2%82%82%20%E2%87%8C%202H%E2%82%82O,%20R%20=%204.7%20k%CE%A9,%20%E2%8C%80%20200%20mm(ish)!', '@type': 'FakeDoc', 'name': '2H₂ + O₂ ⇌ 2H₂O, R = 4.7 kΩ, ⌀ 200 mm(ish)!'}

System information (please complete the following information):

• OS: RHEL8 Server Ubuntu 22 client
• terminus-client-python version = 10.2.3

Additional context
Not sure if this is actually a bug or a feature request but certainly unexpected behavior.

[hoijnet]

This is an enhancement request. There could be a place for a helper to write correct IRIs, but since path segments need to be encoded and the full IRI is also an option, this is a library usage need and a documentation issue primarily, helper secondarily. It should not be implemented directly in the get_document and similar functions.

Changed from bug to enhancement request.