fix: AmazonBedrockChatGenerator does not pass Tool Call streaming chunk to streaming_callback

[sjrl]

The ``

For example, in the OpenAIChatGenerator we handle the stream response in this function which is shown below:

    def _handle_stream_response(self, chat_completion: Stream, callback: SyncStreamingCallbackT) -> List[ChatMessage]:
        chunks: List[StreamingChunk] = []
        chunk = None
        chunk_delta: StreamingChunk

        for chunk in chat_completion:  # pylint: disable=not-an-iterable
            assert len(chunk.choices) <= 1, "Streaming responses should have at most one choice."
            # NOTE: We convert both Tool Call and Text chunks into a chunk_delta here
            chunk_delta = self._convert_chat_completion_chunk_to_streaming_chunk(chunk)
            chunks.append(chunk_delta)
            # NOTE: Here we always pass that chunk_delta to the callback
            callback(chunk_delta)
        return [self._convert_streaming_chunks_to_chat_message(chunk, chunks)]

as you can see we pass all chunk deltas whether they are text or tool calls to the callback allowing streaming of the tool call.

We should update the AmazonBedrockChatGenerator to allow for streaming of tool calls as well by passing the tool call chunks to the streaming callback.

[sjrl]

cc @tstadel