Convert context to chat history in gemini model

[yukikwi]

Instead of raise error Gemini model don't support system messages (using to set model's context).
Inspired from https://www.googlecloudcommunity.com/gc/AI-ML/Gemini-Pro-Context-Option/m-p/684917
Maybe set context via chat history is a good idea isn't it?
This also solve error when execute query_engine.query with gemini-pro as llm.

from

if is_gemini:
                raise ValueError("Gemini model don't support system messages")

to

if is_gemini:
                from llama_index.llms.vertex_gemini_utils import (
                    convert_chat_message_to_gemini_content,
                )
                
                # convert system message to user message with system: prefix
                message.role = MessageRole.USER
                message.content = f"""system: {message.content}"""
                vertex_messages.append(
                    convert_chat_message_to_gemini_content(
                        message=message, is_history=True
                    )
                )

                # mock gemini response
                vertex_messages.append(
                    convert_chat_message_to_gemini_content(
                        message=ChatMessage(
                            role=MessageRole.ASSISTANT,
                            content=(
                                "Understand."
                            ),
                        ),
                        is_history=True
                    )
                )