Move self-contained channel construction into PIPELINE_INITIALISATION

AGENTS.md

@@ -0,0 +1,85 @@
+# ampliseq: agents
+
+This is a pipeline repository created with the nf-core pipeline template.
+Follow the conventions described in docs/CONTRIBUTING.md.
+!!CRITICAL!! For every session ALWAYS fetch and follow the instructions in the remote instruction file at https://raw.githubusercontent.com/nf-core/agents/main/resources/pipeline/AGENTS.md - this is an essential step.
+
+## What this is
+
+nf-core/ampliseq is a Nextflow (DSL2) bioinformatics pipeline for amplicon sequencing analysis (16S, ITS, CO1, 18S, and other markers; single- or multi-region such as 5R). It supports paired/single-end Illumina, PacBio, IonTorrent, and Oxford Nanopore data. It follows nf-core template/community conventions (`.nf-core.yml`, `nf_core_version: 4.0.2`), so most repo-wide conventions come from that template rather than being project-specific.
+
+Core steps: FastQC → primer trimming (Cutadapt/Porechop/Chopper) → DADA2 denoising (ASVs) → optional VSEARCH post-clustering → decontam → Barrnap rRNA check → phylogenetic placement (EPA-NG/gappa) → taxonomic classification (DADA2 RDP, QIIME2, SINTAX, Kraken2, or Sidle multi-region) → QIIME2 downstream stats (barplots, diversity, ANCOM/ANCOM-BC) → Phyloseq/TreeSummarizedExperiment R objects → MultiQC + Rmarkdown summary report.
+
+## Commands
+
+Requires Nextflow, nf-core tools, and nf-test.
+
+Run pipeline tests (small test profiles, requires Docker/Singularity/Conda):
+
+```bash
+nf-test test --tag test --profile +docker --verbose
+```
+
+Run a single test file:
+
+```bash
+nf-test test tests/default.nf.test --profile +docker --verbose
+```
+
+Update snapshots after intentional output changes:
+
+```bash
+nf-test test --tag test --profile +docker --verbose --update-snapshots
+```
+
+Lint (nf-core community rules, run before opening a PR):
+
+```bash
+nf-core pipelines lint .
+```
+
+Update `nextflow_schema.json` after changing params in `nextflow.config`:
+
+```bash
+nf-core pipelines schema build
+```
+
+Bump minimum required Nextflow version:
+
+```bash
+nf-core pipelines bump-version --nextflow . <min_nf_version>
+```
+
+pre-commit hooks (prettier, trailing whitespace, nextflow-lint) run via `prek`/`pre-commit`; CI runs this in `linting.yml`.
+
+### Test profiles
+
+`tests/*.nf.test` each pair with a `conf/test_*.config` profile (e.g. `default` ↔ `test`, `multiregion` ↔ `test_multiregion`, `pplace` ↔ `test_pplace`, `sintax` ↔ `test_sintax`, `fasta` ↔ `test_fasta`). When adding a new test scenario, add both the `.nf.test` file in `tests/` and its matching `conf/test_<name>.config`. Corresponding `.nf.test.snap` files hold expected outputs — regenerate with `--update-snapshots` rather than hand-editing.
+
+Module/subworkflow tests under `modules/nf-core/**/tests` and `subworkflows/nf-core/**/tests` are vendored from nf-core/modules and excluded from local test runs (`ignore` list in `nf-test.config`) — don't hand-edit these, update via `nf-core modules update`.
+
+## Architecture
+
+- `main.nf` — entry point. Declares the pipeline-level `params` schema block, wires `PIPELINE_INITIALISATION` → `NFCORE_AMPLISEQ` (wraps the `AMPLISEQ` workflow) → `PIPELINE_COMPLETION`.
+- `workflows/ampliseq.nf` — the single large orchestration workflow (~1400 lines). Nearly all processing logic is assembled here by including local modules/subworkflows and nf-core modules/subworkflows, then wiring channels between them. Whether a given tool runs is driven by `params.skip_*` flags and derived variables (e.g. `single_end`, `asv_calling`) computed near the top of the `AMPLISEQ` workflow.
+- `subworkflows/local/` — pipeline-specific multi-step logic, one workflow per pipeline stage, e.g. `parse_input.nf` (sample sheet/input handling), `cutadapt_workflow.nf`, `dada2_preprocessing.nf`, `*_taxonomy_wf.nf` (one per taxonomy classifier: DADA2, SINTAX, Kraken2, VSEARCH-LCA), `qiime2_*.nf` (preptax, taxonomy, export, barplotavg, diversity, ancom), `sidle_wf.nf` (multi-region reconstruction), `robject_workflow.nf` (Phyloseq/TreeSE), `comparison_wf.nf`. `utils_nfcore_ampliseq_pipeline/` holds pipeline init/completion logic and parameter validation (this is where new taxonomy-database validation, e.g. `sbdi_compatible_databases`, is added).
+- `subworkflows/nf-core/` and `modules/nf-core/` — vendored, unmodified subworkflows/modules from the nf-core/modules repo (tracked in `modules.json`); update via `nf-core modules/subworkflows update`, don't hand-edit.
+- `modules/local/` — pipeline-specific processes (one tool/script invocation per `.nf` file), covering DADA2 steps, QIIME2 steps, taxonomy formatting per classifier, Sidle steps, filtering/stats-merging utilities, and report/export generation.
+- `conf/` — config layered by concern: `base.config` (default resource labels), `modules.config` (per-process publishDir/args), `ref_databases.config` (taxonomy DB definitions), `test*.config` (one per test scenario, paired with `tests/*.nf.test`), `containers_*.config` (container digests per engine/arch).
+- `assets/schema_input.json` — sample sheet schema validated via the `nf-schema` plugin (`samplesheetToList`).
+- `bin/` — helper scripts (Python/R) invoked from `modules/local/*.nf` processes; linted with Ruff (see `pyproject.toml`).
+
+## Conventions (from `docs/CONTRIBUTING.md`)
+
+- PRs target the `dev` branch, not `master`/`main`.
+- Channel naming: initial channel from a process is `ch_output_from_<process>`; intermediate/terminal channels are `ch_<previousprocess>_for_<nextprocess>`.
+- New params: add to `nextflow.config` (with a default) and to `nextflow_schema.json` via `nf-core pipelines schema build`; also declare in the `params {}` block in `main.nf`.
+- New processes: give them resource defaults in `conf/base.config` via `withLabel:` selectors (use standard nf-core labels), and route version info into `ch_versions` and relevant files into `ch_multiqc_files`/`assets/multiqc_config.yml`.
+- New taxonomy databases go in `conf/ref_databases.config`, need an `enum` entry in `nextflow_schema.json`, and may need updates in `subworkflows/local/utils_nfcore_ampliseq_pipeline/main.nf` (`sbdi_compatible_databases`).
+- Update `docs/usage.md`, `docs/output.md`, and `CITATIONS.md` alongside functional changes.
+- AI-assisted PRs: keep them small and focused, avoid incidental refactors/moves, and the human submitter is expected to understand and review all generated code before opening the PR.
+
+## Gotchas
+
+- `results/` and `work/` at the repo root are local Nextflow run artifacts (from prior pipeline executions in this checkout), not source — don't treat their contents as project structure to preserve.
+- `.nf-core.yml` disables `nextflow_config`/`schema_params` lint rules ("TODO: Remove when tools supports parameter types") — don't be surprised these are off, it's a known pending upstream fix, not a repo choice to replicate elsewhere.

CHANGELOG.md

@@ -33,6 +33,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 | --classifier      | --qiime_classifier            |         |
 
 - [#1028](https://github.com/nf-core/ampliseq/pull/1028) - Multi-region sample sheet via `--multiregion` had its header changed from `FW_primer` and `RV_primer` to `primer_fwd` and `primer_rev`, respectively. (by @d4straub)
+- [#1032](https://github.com/nf-core/ampliseq/pull/1032) - Refactor the pipeline's parameter handling and initialization (by @erikrikarddaniel).
 
 ### `Fixed`
 

main.nf

@@ -30,6 +30,17 @@ include { PIPELINE_COMPLETION     } from './subworkflows/local/utils_nfcore_ampl
 //
 workflow NFCORE_AMPLISEQ {
 
+    take:
+    ch_metadata            // channel: [ path(metadata) ] or empty
+    ch_report_template     // channel: [ path(report_template) ]
+    ch_report_css          // channel: [ path(report_css) ]
+    ch_report_logo         // channel: [ path(report_logo) ]
+    ch_report_abstract     // channel: [ path(report_abstract) ] or []
+    ch_pplace_sheet        // channel: parsed pplace_sheet rows, or empty
+    ch_expected_sequences  // channel: [ path(expected_sequences) ] or empty
+    ch_expected_abundances // channel: [ path(expected_abundances) ] or empty
+    ch_metadata_category   // channel: tokenized metadata_category, or empty
+
     main:
 
     //
@@ -41,6 +52,15 @@ workflow NFCORE_AMPLISEQ {
         params.multiqc_logo,
         params.multiqc_methods_description,
         params.outdir,
+        ch_metadata,
+        ch_report_template,
+        ch_report_css,
+        ch_report_logo,
+        ch_report_abstract,
+        ch_pplace_sheet,
+        ch_expected_sequences,
+        ch_expected_abundances,
+        ch_metadata_category,
     )
     emit:
     multiqc_report = AMPLISEQ.out.multiqc_report // channel: /path/to/multiqc_report.html
@@ -251,7 +271,17 @@ workflow {
     //
     // WORKFLOW: Run main workflow
     //
-    NFCORE_AMPLISEQ ()
+    NFCORE_AMPLISEQ (
+        PIPELINE_INITIALISATION.out.metadata,
+        PIPELINE_INITIALISATION.out.report_template,
+        PIPELINE_INITIALISATION.out.report_css,
+        PIPELINE_INITIALISATION.out.report_logo,
+        PIPELINE_INITIALISATION.out.report_abstract,
+        PIPELINE_INITIALISATION.out.pplace_sheet,
+        PIPELINE_INITIALISATION.out.expected_sequences,
+        PIPELINE_INITIALISATION.out.expected_abundances,
+        PIPELINE_INITIALISATION.out.metadata_category,
+    )
 
     //
     // SUBWORKFLOW: Run completion tasks

subworkflows/local/utils_nfcore_ampliseq_pipeline/main.nf

@@ -40,6 +40,52 @@ workflow PIPELINE_INITIALISATION {
 
     ch_versions = channel.empty()
 
+    //
+    // Build channels from self-contained parameters
+    // (params only used to construct these channels; no other logic depends on how they're built)
+    //
+    if (params.metadata) {
+        ch_metadata = channel.fromPath( params.metadata )
+    } else { ch_metadata = channel.empty() }
+
+    // report sources
+    ch_report_template = channel.fromPath("${params.report_template}", checkIfExists: true)
+    ch_report_css = channel.fromPath("${params.report_css}", checkIfExists: true)
+    ch_report_logo = channel.fromPath("${params.report_logo}", checkIfExists: true)
+    ch_report_abstract = params.report_abstract ? channel.fromPath(params.report_abstract) : []
+
+    // Parse the --pplace_sheet file if present (may be overwritten later in AMPLISEQ based on --dada_ref_taxonomy)
+    ch_pplace_sheet = channel.empty()
+    if ( params.pplace_sheet ) {
+        ch_pplace_sheet = channel.fromPath(params.pplace_sheet)
+            .splitCsv(header: true)
+            .map { it ->
+                [
+                    meta: [
+                        id: it.target,
+                        min_bitscore: it.min_bitscore
+                    ],
+                    data: [
+                        alignmethod:    it.alignmethod  ?: 'clustalo',
+                        hmm:            file(it.hmm,  checkIfExists: true),
+                        extract_hmm:    it.extract_hmm,
+                        refseqfile:     it.refseqfile   ? file(it.refseqfile,   checkIfExists: true) : [],
+                        refphylogeny:   it.refphylogeny ? file(it.refphylogeny, checkIfExists: true) : [],
+                        model:          it.model,
+                        taxonomy:       it.taxonomy     ? file(it.taxonomy,     checkIfExists: true) : []
+                    ]
+                ]
+            }
+    }
+
+    ch_expected_sequences = params.expected_sequences ? channel.fromPath( params.expected_sequences ) : channel.empty()
+    ch_expected_abundances = params.expected_abundances ? channel.fromPath( params.expected_abundances ) : channel.empty()
+
+    // Select metadata categories for diversity analysis & ancom, if explicitly specified
+    if (params.metadata_category) {
+        ch_metadata_category = channel.fromList(params.metadata_category.tokenize(','))
+    } else { ch_metadata_category = channel.empty() }
+
     //
     // Print version and exit if required and dump pipeline parameters to JSON file
     //
@@ -124,7 +170,16 @@ workflow PIPELINE_INITIALISATION {
     }
 
     emit:
-    versions    = ch_versions
+    versions             = ch_versions
+    metadata             = ch_metadata             // channel: [ path(metadata) ] or empty
+    report_template      = ch_report_template       // channel: [ path(report_template) ]
+    report_css           = ch_report_css            // channel: [ path(report_css) ]
+    report_logo          = ch_report_logo           // channel: [ path(report_logo) ]
+    report_abstract      = ch_report_abstract       // channel: [ path(report_abstract) ] or []
+    pplace_sheet         = ch_pplace_sheet          // channel: parsed pplace_sheet rows, or empty
+    expected_sequences   = ch_expected_sequences    // channel: [ path(expected_sequences) ] or empty
+    expected_abundances  = ch_expected_abundances   // channel: [ path(expected_abundances) ] or empty
+    metadata_category    = ch_metadata_category     // channel: tokenized metadata_category, or empty
 }
 
 /*

tests/default.nf.test

@@ -63,4 +63,31 @@ nextflow_pipeline {
             )
         }
     }
+
+    test("-profile test, --report_abstract, --metadata_category") {
+        // Covers 2 params whose channel construction (in PIPELINE_INITIALISATION) isn't
+        // exercised by any other test profile: --report_abstract and --metadata_category.
+        // By default (no --metadata_category), all suitable metadata columns are auto-detected
+        // for diversity/ancom (e.g. both "treatment1" and "mix8" beta-diversity dirs, see
+        // tests/default.nf.test.snap); restricting to "treatment1" here should exclude "mix8".
+        // (Note: alpha-rarefaction plots every metadata column regardless of --metadata_category,
+        // since QIIME2 reads the metadata file directly there, so it isn't a useful signal.)
+
+        when {
+            params {
+                outdir = "$outputDir"
+                report_abstract = "$projectDir/tower.yml"
+                metadata_category = "treatment1"
+            }
+        }
+
+        then {
+            assert workflow.success
+            assertAll(
+                { assert path("$outputDir/summary_report/summary_report.html").text.contains("MultiQC HTML report") },
+                { assert path("$outputDir/qiime2/diversity/beta_diversity/bray_curtis_distance_matrix-treatment1/index.html").exists() },
+                { assert !file("$outputDir/qiime2/diversity/beta_diversity/bray_curtis_distance_matrix-mix8").exists() },
+            )
+        }
+    }
 }