Allow users to pick Anthropic or OpenAI for LLM extraction

[aar2dee2]

I've added an llmOptions param in extract options for the v1/scrape endpoint.
This accepts an object with a 'model' and a 'provider' property.
For example,

    "extract": {
          "llmOptions": {"provider": "anthropic", "model": "claude-3-5-sonnet-20241022"},
       }

Demo video: https://youtu.be/BqsS9UTMD3s

I've retained logs for now. Let me know if you'd like to make changes for merging.

[aar2dee2]

I tried to not commit the formatter changes, apologies for that!

[rafaelsideguide]

Could you check #414 to see if it fits your needs?

[aar2dee2]

Could you check #414 to see if it fits your needs?

Thanks @rafaelsideguide . That PR allows for different models from OpenAI.
For anthropic, you have to use a different client, plus some differences in how the APIs are called and responses are parsed.

[rafaelsideguide]

Awesome work, @aar2dee2! Just one thing: if the env variable isn’t set and a request is made, it returns a 500 error with a response directly from Anthropic. It would be helpful if the response clarified the issue (e.g., 'Anthropic API key is not set' or 'OpenAI API key is not set' for OpenAI).

Current response:

{
  "success": false,
  "error": "(Internal server error) - Could not resolve authentication method. Expected either apiKey or authToken to be set. Or for one of the \"X-Api-Key\" or \"Authorization\" headers to be explicitly omitted - Could be due to LLM parsing issues"
}