Don't raise errors when attempting to parse "accessory" text

[caufieldjh]

LLM completions frequently include bits of accessory text like "Here are the fields you requested" etc etc.
This happens even when the prompt includes explicit instructions to provide only a given set of results.

OntoGPT attempts to parse these because it doesn't know they aren't part of the split data, resulting in errors like:

ERROR:root:Cannot find slot for here_is_the_text_split_into_the_specified_field in Here is the text split into the specified fields:

This doesn't impact the overall results but it's not entirely an error, and could be ignored.