Make it easier to specify a normalization method for `Chat.append_message_stream()`

[cpsievert]

Currently, if you want to teach Chat.append_message_stream() about response formats that it doesn't already know about, you have to register a new "message normalizer" strategy with this internal object.

Although convenient for developers and still potentially worth exporting (it's currently internal), it's a lot to ask for most users to learn and implement.

It'd be much easier if you could just pass a function to .append_message_stream() to grab the relevant content from each iteration of the stream.

For example, something like this (from #1610):

from shiny.ui._chat_normalize import BaseMessageNormalizer, message_normalizer_registry

class LangchainAgentResponseNormalizer(BaseMessageNormalizer):
    # For each chunk of a .append_message_stream()
    def normalize_chunk(self, chunk):
        return chunk["messages"][0].content

    def can_normalize_chunk(self, chunk):
        return "messages" in chunk and len(chunk["messages"]) > 0

    # For .append_message()
    def normalize(self, message):
        return message["messages"][0].content

    def can_normalize(self, message):
        return "messages" in message and len(message["messages"]) > 0

message_normalizer_registry.register(
    "langchain-agents", LangchainAgentResponseNormalizer()
)

could instead become something like:

chat.append_message_stream(response, lambda x: x["messages"][0].content)