feat: ai assistant

[RaunoT]

A few things still need working out

1. Caching doesn't seem to work properly, as I'm getting hit with limit errors after just a few requests.

 ⨯ GoogleGenerativeAIFetchError: [GoogleGenerativeAI Error]: Error fetching from https://generativelanguage.googleapis.com/v1beta/cachedContents: [429 Too Many Requests] TotalCachedContentStorageTokensPerModelFreeTier limit exceeded for model gemini-1.5-flash: limit=1000000, requested=1251312

{
  status: 429,
  statusText: 'Too Many Requests',
  errorDetails: undefined
}

2. The maximum token count for Gemini cache is 1,000,000 but we also need to consider the minimum, which is 32,768. Should we just send histories with the request message every time in those cases? What are the best practices here?

⨯ GoogleGenerativeAIFetchError: [GoogleGenerativeAI Error]: Error fetching from https://generativelanguage.googleapis.com/v1beta/cachedContents: [400 Bad Request] Cached content is too small. total_token_count=6690, min_total_token_count=32768

3. The history.csv file isn't always being properly overridden when saving Chat settings.