feat: add ChatInterface - wrapper around typical chat use-case

[mhordynski]

Feature description

Goal of this task is to create a minimal wrapper around typical use-case that we develop - chat interface. ChatInterface should allow to get standardized inputs and stream text and rich elements.

Input:

• message: current message from the user
• history: previous messages in the converation
• context: anything extra coming from API (for example user info)

Output (yields):

• str: Regular text responses streamed chunk by chunk
• Reference: Source documents used to generate the answer
• LiveUpdate: Status updates during processing (tool-call, searching, etc.)
• Action: Suggested actions for the user to take (follow-up questions, write action, etc.)

Motivation

While #330 will provide a generic approach to define workflows, we still need a dead simple way of orchestrating any chat-like workload without forcing users to write code in any specific way.

Idea behind ChatInterface is to provide light-weight guidelines what input and output can be handled in such scenario and allow user to write any code using components already provided in ragbits.

Additional context

class SimpleChatImplementation(ChatInterface):
    """A simple example implementation of the ChatInterface that demonstrates different response types."""

    async def chat(
        self,
        message: str,
        history: list[Message] | None = None,
        context: dict | None = None,
    ) -> AsyncGenerator[ChatResponse, None]:

        yield self.create_live_update(
            message="Searching for relevant documents...",
        )

        references = await get_document_search().search(message)

        for reference in references:
            yield self.create_reference(reference)

        for word in "Hello my name is John Doe, I'm a software engineer and I'm looking for a new job. Please help me find a new job.".split():
            yield self.create_text_response(word + " ")
            await asyncio.sleep(0.05)