Refactor: Make generate_final_output() data-driven

[cmkobel]

Summary

generate_final_output() in workflow/Snakefile (lines ~323–388) uses three similar expand() blocks with conditional logic based on genome count N:

• N≥1: singular analyses
• N≥2: pairwise comparisons
• N≥3: phylogenetics

Each block manually lists rules and their expand patterns. Adding a new rule requires finding the right block and duplicating the expand boilerplate.

Proposed approach

Use a data-driven approach — a list of (rule_output_pattern, min_N) tuples — and a single loop that expands all rules where N >= min_N:

_analyses = [
    ("results/{sample}/annotation/...", 1),
    ("results/batch/panaroo/...", 2),
    ("results/batch/fasttree/...", 3),
]

def generate_final_output(N):
    return [expand(pattern, ...) for pattern, min_n in _analyses if N >= min_n]

This makes it trivial to add/remove analyses and see the N-threshold for each at a glance.

Files to touch

• workflow/Snakefile (lines ~323–388)