3 develop testing guidelines - #5

- column_standardise: map CSV headers to ABCD schema terms (WP6)
- biodiv_annotate: wrap BiodivPortal Annotator REST API (WP2)
- taxonomy_classify: resolve taxon names to GBIF/WFO IDs (WP3)
- outlier_detect: statistical + LLM quality detection (WP4)
- provisional_concepts: register unresolved terms (WP5)
- rdf_transform: tabular data to RDF triples (WP7)
- collect_reports: aggregate per-step quality reports

All modules are stub implementations ready for WP teams to replace.

- main.nf: entry point accepting --input CSV and --mapping_schema
- workflows/biodivpipeline.nf: orchestrates all 7 modules
  - Step 1: Column standardisation (WP6)
  - Step 2: Parallel annotation (WP2) + taxonomy (WP3) + outlier detection (WP4)
  - Step 3: Provisional concept registration (WP5, after WP2)
  - Step 4: RDF transformation (WP7, combines all upstream outputs)
  - Step 5: Report collection
- nextflow.config: add input/mapping_schema params
- nextflow_schema.json: update input parameter validation for CSV data

- Remove FASTQC/MultiQC module references from workflow
- Replace samplesheet parsing with direct CSV file channel
- Remove genome/igenomes/fasta parameter references
- Update PIPELINE_INITIALISATION to validate CSV input
- Replace validateInputSamplesheet with validateInputParameters
- Update modules.config: per-module publishDir for our 7 modules
- Update test.config: point to test_data/sample.csv

- test_data/sample.csv: 109 records sampled from BELEGE_aus_D (every 1000th row)
- modules/stubs/: FastAPI mock service for end-to-end testing
  - Dockerfile + stub_service.py (POST /process, GET /health)
- assets/default_mapping.jsonld: stub DwC RDF mapping schema
  - Maps BGBM CSV columns to Darwin Core terms

Pipeline runs end-to-end with: nextflow run main.nf -profile test

Replace nf-core genomics boilerplate with project-specific content:
- README.md: project summary, Mermaid diagrams, quick start, output tree
- docs/usage.md: CSV input format, parameters, sequence diagram
- docs/output.md: all 18 output files across 7 directories
- docs/CONTRIBUTING.md: branch model, conventional commits, module interface
- docs/README.md: documentation index with tech stack table

Remove unused template files:
- modules/nf-core/fastqc, modules/nf-core/multiqc (genomics QC)
- assets: samplesheet.csv, schema_input.json, multiqc_config.yml, logos
- conf: igenomes configs, container lock files, test_full.config
- tower.yml, methods_description_template.yml

Clean code references:
- nextflow.config: remove igenomes include, test_full profile
- modules.json: remove fastqc/multiqc entries
- main.nf, subworkflow: remove multiqc_report channel
- subworkflow: remove samplesheetToList import, FastQC/MultiQC citations

WP1 owns the integration backbone, not module internals. Strip stub scripts
to minimal valid-output emitters and remove pre-claimed domain content:

- Delete modules/stubs/ (FastAPI HTTP service contradicts Pattern A)
- Replace 6 module main.nf stubs with minimal emitters
- Strip default_mapping.jsonld to empty placeholder
- Replace biodiv/wpX-*:latest container refs with placeholder image
- Remove WP-to-module assignments from workflow comments
- Drop BGBM example from input description

Module teams (WP2-WP7) own real implementations, container images, and
domain content. WP1 ships interface contracts.

- README.md: remove module diagram, dataset claims, project-structure
  module descriptions, credits section
- docs/output.md: reduce to directory layout + Nextflow pipeline_info
- docs/README.md: shrink to TOC + tech stack
- docs/usage.md: remove dataset specifics, column listing, unzip step,
  sequence diagram
- docs/CONTRIBUTING.md: remove project contacts, dataset record count
- nextflow.config: empty manifest.contributors placeholder
- .nf-core.yml: clear author placeholder
- CITATIONS.md: drop tools not used in this pipeline