Function select_similar_code_corpus 

pub fn select_similar_code_corpus(
    functions: &[SimilarCodeSelectionInput<'_>],
    limits: SimilarCodeLimits,
) -> Result<SimilarCodeCorpusSelection, SimilarCodeError>

Select a deterministic fair corpus before provider inference.

In-scope functions receive deterministic priority, while both the scoped and background partitions use normalized source occurrence plus full source digest for fair selection. The subset is returned in stable occurrence order.