add native tool functionality (e.g. google_search for Gemini)

[avergin]

Adds support for calling native tools like Google Search in Gemini models.

For example, What is the current Google stock price? will return the current stock price.

alias LangChain.Chains.LLMChain
alias LangChain.Message
alias LangChain.NativeTool

model = ChatGoogleAI.new!(%{temperature: 0, stream: false, model: "gemini-2.0-flash"})

{:ok, updated_chain} =
   %{llm: model, verbose: false, stream: false}
   |> LLMChain.new!()
   |> LLMChain.add_message(
     Message.new_user!("What is the current Google stock price?")
   )
   |> LLMChain.add_tools(NativeTool.new!(%{name: "google_search", configuration: %{}}))
   |> LLMChain.run()

I have also added a metadata field (map) to the Message struct to flexibly capture model specific responses when native tools are called. For example when google_search is called, Gemini returns groundingMetadata data, and this can be accessed from metadata.

[avergin]

For example, the metadata field in the Message struct may help with capturing the citations data that is returned from Perplexity's API response.