refactor: bundle process_chromosome_batch() parameters into config structs

[edmundmiller]

Context

process_chromosome_batch() in src/rna/dupradar/counting.rs currently takes 19 positional parameters. This was flagged during review of #84 — the addition of skip_dup_check brought it from 18 to 19.

Problem

• Hard to read and maintain — callers must get the positional order exactly right
• Easy to accidentally swap parameters of the same type (e.g. multiple &str, bool, usize args)
• Each new feature tends to add another parameter, compounding the problem

Suggested approach

Bundle related parameters into config structs, e.g.:

struct BatchConfig<'a> {
    bam_path: &'a str,
    gtf_path: &'a str,
    stranded: Strandedness,
    paired: bool,
    skip_dup_check: bool,
    // ...
}

struct BatchResources<'a> {
    index: &'a SpatialIndex,
    interner: &'a GeneIdInterner,
    qualimap_index: Option<&'a QualimapIndex>,
    gene_to_biotype: &'a [u16],
    // ...
}

This would make call sites self-documenting and make future parameter additions non-breaking.

Priority

Low — the current code works fine, this is a maintainability improvement.