Enhance BeanOutputConverter to trim Preamble

Every so often I'll get responses from LLMs that look like this:

I'll analyze and evaluate that for you. Here are the results
{
   "result" : "PASS"
}

Which results in BeanOutputConverter failing to parse the response.
o.s.ai.converter.BeanOutputConverter : Could not parse the given text to the desired target type:

Would it make sense to enhance BeanOutputConverters trimming logic to trim everything before the JSON payload?

Expected Behavior

When mapping a response from a ChatClient, if the response from the LLM contains a preamble before a valid JSON response, BeanOutputConverter should trim it and successfully parse the response.

Current Behavior

BeanOutputConverter fails to parse the response and throws and error.

Start at BeanOutputConverter and reproduce the issue with the sample response containing a preamble before the JSON payload. Confirm the converter can parse a valid JSON response with leading text, while preserving existing parsing behavior; the issue does not name a test file or other entry point.