Context length exceeded when evaluating large generated repositories

[gin111-hub]

Hi,

Thanks for the great project.

I'm using this framework to evaluate repositories generated by an LLM. During the evaluation process, the generated repository is sent to the evaluator model through the API.

However, for larger repositories, I encounter a context length exceeded error because the repository content exceeds the model's context window.

I'm wondering:

• Have you encountered this issue before?
• Is there a recommended way to evaluate repositories that are larger than the model's context limit?
• Does the framework support any chunking, summarization, file selection, or other strategies to handle large repositories?

The error seems to occur when the repository content is passed to the evaluator model.

Any suggestions would be appreciated.

Thanks!