Question/request about representative document truncation

Nobody has claimed this yet.

Hi Maarten,

I'm curious about the truncation at 255 characters of representative documents in both the `OpenAI` and `TextGeneration` representation models (at least).

https://github.com/MaartenGr/BERTopic/blob/58d90bc65e95a11718e63a0834809d156dcf431d/bertopic/representation/_textgeneration.py#L143C1-L147C67

I imagine it's helpful if you are paying for an API, but if using a local text generation model e.g. as in your Llama2 google colab example, is there any other reason that the representative documents have to be cut down so hard?

Could we have some control over the truncation e.g. add a `truncate: int = 255` parameter to the representation models and then:

```
for doc in docs:
doc = doc[:truncate] if truncate else doc
to_replace += f"- {doc}\n"
```

Start with bertopic/representation/_textgeneration.py at the linked truncation logic, then locate the corresponding OpenAI representation implementation. Check how representative documents are passed into each model and whether existing tests cover their prompts. Done means both models expose configurable truncation while retaining the current default behavior.

Written by the indexing model from the issue text.