Request: ChatVertexAI preserving grounding_metadata

[jamesbraza]

Feature request

As of langchain==0.0.348 in ChatVertexAI here:

1. vertexai.language_models.ChatSession.send_message returns a vertexai.language_models.MultiCandidateTextGenerationResponse response
2. ChatVertexAI throws out the response's grounding_metadata, only preserving the text

The request is to not discard this useful metadata

Motivation

I am trying to use the response.grounding_metadata.citations in my own code

Your contribution

I will open a PR to make this:

generations = [
    ChatGeneration(
        message=AIMessage(content=r.text),
        generation_info=r.raw_prediction_response,
    )
    for r in response.candidates
]

So the raw response remains accessible

[dosubot]

🤖

Thank you for your interest in contributing to LangChain! It's great to see your motivation to enhance the functionality of ChatVertexAI by preserving the grounding_metadata. Your proposed solution looks promising, and I encourage you to go ahead and open a PR with your changes. Your contribution will definitely add value to the framework and benefit other users as well. Keep up the good work, and thank you for your contribution to the LangChain community!

---

This response is meant to be useful and save you time. It is not meant to be a precise solution, but rather a starting point for your own research.

Help me be more useful! Please leave a 👍 if this is helpful and 👎 if it is irrelevant.

If you want to continue the conversation, start your reply with @dosu-bot.