Conditional rules structure

[Headonpillow]

Summary:

1. Replace the two mutually exclusive trimming rules with a single prepare_trimmed rule that branches on enabled("preprocess").
2. Declare rules unconditionally; use branching inside run/shell blocks instead of wrapping entire rules in Python conditionals.
3. Optionally split large rule groups into modules (e.g., rules/preprocess.smk, rules/phylogeny.smk) and include them conditionally to keep the DAG tidy.

---

Implementation notes / concrete refactor steps (proposed):

1. Use proper booleans in config and add a small helper

• Recommend config_templates/basic.yaml use YAML booleans (e.g., preprocess: true, phylogeny: true).
• Add an enabled(key) helper in the Snakefile to accept multiple truthy formats (strings, integers, booleans).

Example:

def enabled(key):
    v = config.get(key, False)
    if isinstance(v, str):
        return v.lower() in ("yes","y","true","1")
    return bool(v)

2. Build rule all targets dynamically

• Replace the current ad-hoc myoutput list + manual appends with a function all_targets() that returns the final target list depending on flags.

Example:

def all_targets():
    targets = [
        "results/denoising/read_count_tracking.tsv",
        "results/asv/ASVs.fa",
        "results/asv/ASVs_counts.tsv",
        "results/asv/ASVs_taxonomy.tsv",
        "results/phyloseq/starting_phyla_table.tsv",
        "results/phyloseq/prevalence_graph.png",
        "results/phyloseq/Phyloseq.RData",
        "results/phyloseq/ASVs_good.fasta",
        "results/phyloseq/plots/plot_1.tiff",
    ]
    if enabled("preprocess"):
        targets += [
            "results/multiQC/report_R1.html",
            "results/multiQC/report_R2.html",
            "results/multiQC_trimmed/report_R1.html",
            "results/multiQC_trimmed/report_R2.html",
        ]
    if enabled("phylogeny"):
        targets += [
            "results/phyloseq/ASV_alignment.mafft",
            "results/phyloseq/ASV_alignment.mafft.treefile",
        ]
    return targets

rule all:
    input: all_targets()

3. Declare rules unconditionally and branch inside rule bodies where needed

• Avoid defining whole rules inside Python if blocks. Instead define rules always and use the enabled() helper inside run/shell blocks to branch behavior.
• Example: replace the current two mutually-exclusive rules that create intermediate/trimmed/* with a single prepare_trimmed rule that either runs trim_galore or copies raw files depending on enabled("preprocess").

4. Optionally split rule modules and use conditional includes

• Put large groups of rules in rules/preprocess.smk, rules/phylogeny.smk, etc., and use if enabled("preprocess"): include: "rules/preprocess.smk". Use this only for large modules to keep the top-level Snakefile tidy.

5. Safety and housekeeping

• Remove transient print() calls from the Snakefile.
• Use mkdir -p or directory() outputs instead of ad-hoc mkdir guards.
• Add a short test checklist: create runs/demo with config.yaml toggling preprocess and phylogeny and run snakemake --configfile runs/demo/config.yaml --directory runs/demo -n to assert the DAG only contains intended targets.

Testing / Acceptance checklist for this issue

• Add config_templates/basic.yaml with boolean flags.
• Add enabled() and all_targets() to Snakefile.
• Replace ad-hoc myoutput usage with all_targets().
• Convert the two mutually-exclusive trimming rules into a single branching prepare_trimmed rule.
• Run snakemake -n with permutations of the flags and confirm the DAG and planned jobs match expectations.

Notes on scope

• This refactor touches Snakefile and possible new config_templates/basic.yaml.