wf2wf – Universal Workflow-Format Converter

wf2wf is a format-agnostic converter: any supported engine → Intermediate Representation (IR) → any other engine. The core library handles: 

• Snakemake • Nextflow • CWL • HTCondor/DAGMan • WDL • Galaxy • bundled BioCompute Objects

graph TD;
  A[Snakemake] -->|import| IR((IR));
  B[DAGMan]   --> IR;
  C[CWL]      --> IR;
  D[Nextflow] --> IR;
  E[WDL]      --> IR;
  F[Galaxy]   --> IR;
  IR -->|export| GA2[Galaxy];
  IR --> SMK2[Snakemake];
  IR --> CWL2[CWL];
  IR --> DAG2[DAGMan];
  IR --> NF2[Nextflow];
  IR --> WDL2[WDL];

Loading

wf2wf is called from the command line as:

# Convert Snakemake → DAGMan and auto-generate Markdown report
wf2wf convert -i pipeline.smk -o pipeline.dag --auto-env build --interactive --report-md

---

📋 Table of Contents

• Features
• Installation
• Quick CLI Tour
• Commands
• Examples
• Contributing
• Support
• License
• Acknowledgements

---

✨ Features

• 🔄 Universal Conversion – Any supported engine → IR → any other engine with a single command.
• 🧬 Loss-Mapping & Round-Trip Fidelity – Structured loss reports (*.loss.json) and automatic reinjection guarantee nothing disappears silently.
• 🐳 Automated Environment Builds – Optional Conda-to-OCI pipeline (micromamba → conda-pack → buildx/buildah) with digest-pinned image references, SBOMs and Apptainer conversion.
• ⚖ Regulatory & Provenance Support – BioCompute Object generation, schema validation and side-car provenance for FDA submissions.
• 🧪 Aiming for Quality – high test coverage, semantic versioning, graceful degradation when optional external tools are missing.
• 🔧 Smart Configuration Analysis – Automatic detection and warnings for missing resource requirements, containers, error handling, and file transfer modes when converting between shared filesystem and distributed computing workflows.
• 💬 Interactive Mode – Guided prompts to help users address configuration gaps and optimize workflows for target execution environments.

Information-loss workflow

wf2wf should record every field the target engine cannot express:

⚠ Conversion losses: 2 (lost), 1 (lost again), 7 (reapplied)

• lost – field dropped in this conversion
• lost again – it was already lost by a previous exporter
• reapplied – successfully restored from a side-car when converting back to a richer format

Use --fail-on-loss to abort if any lost/lost again entries remain.

Configuration Analysis

When converting between different workflow execution environments, wf2wf automatically detects potential issues:

## Configuration Analysis

### Potential Issues for Distributed Computing

* **Memory**: 2 tasks without explicit memory requirements
* **Containers**: 3 tasks without container/conda specifications
* **Error Handling**: 3 tasks without retry specifications
* **File Transfer**: 6 files with auto-detected transfer modes

**Recommendations:**
* Add explicit resource requirements for all tasks
* Specify container images or conda environments for environment isolation
* Configure retry policies for fault tolerance
* Review file transfer modes for distributed execution

Use --interactive to get guided prompts for addressing these issues automatically.

---

📦 Installation

# PyPI (recommended)
pip install wf2wf

# or conda-forge (once feedstock is merged)
conda install -c conda-forge wf2wf

Development install:

git clone https://github.com/csmcal/wf2wf.git && cd wf2wf
pip install -e .[dev]
pre-commit install
pytest -q

---

🚀 Quick CLI Tour

# Convert Snakemake → DAGMan and build digest-pinned images
wf2wf convert -i Snakefile -o pipeline.dag --auto-env build --push-registry ghcr.io/myorg --report-md --interactive

# Convert CWL → Nextflow, abort on any information loss
wf2wf convert -i analysis.cwl -o main.nf --out-format nextflow --fail-on-loss

# Validate a workflow and its loss side-car
wf2wf validate pipeline.dag

Interactive prompts (--interactive) use y/n/always/quit; loss prompts appear only for warn/error severities.

---

🛠 Commands Overview

Command	Purpose
wf2wf convert	Convert workflows between formats (all conversions go via the IR)
wf2wf validate	Validate a workflow file or a .loss.json side-car
wf2wf info	Pretty-print summary statistics of a workflow
wf2wf bco sign	Sign a BioCompute Object and generate provenance attestation
wf2wf bco package	Bundle a BCO and its artefacts (e.g. eSTAR ZIP)

Each command accepts --help for full usage details.

Auto-detection matrix

Extension	Format
.cwl	CWL
.dag	DAGMan
.ga	Galaxy
.json	IR (JSON)
.nf	Nextflow
.smk	Snakemake
.wdl	WDL
.yaml, .yml	IR (YAML)

---

🔬 Examples

Example 1 – Snakemake → DAGMan with automatic environment build

wf2wf convert -i Snakefile \
              -o pipeline.dag \
              --out-format dagman \
              --auto-env build --push-registry ghcr.io/myorg

Example 2 – CWL → Nextflow round-trip

# CWL → IR → Nextflow
airflow_cwl="workflow.cwl"
wf2wf convert -i "$airflow_cwl" -o main.nf --out-format nextflow

# … do some edits …
# Nextflow → IR → CWL (should restore metadata)
wf2wf convert -i main.nf -o roundtrip.cwl --out-format cwl

Example 3 – WDL → CWL with loss checking

wf2wf convert -i assembly.wdl -o assembly.cwl --out-format cwl --fail-on-loss

More recipes live in the examples/ directory.

---

🔄 Workflow Conversion Differences

When converting between different workflow execution environments, several key differences need to be addressed:

Shared Filesystem vs Distributed Computing

Shared Filesystem Workflows (Snakemake, CWL, Nextflow):

• Assume all files are accessible on a shared filesystem
• Often have minimal resource specifications
• Rely on system-wide software or conda environments
• Basic error handling and retry mechanisms

Distributed Computing Workflows (HTCondor/DAGMan):

• Require explicit file transfer specifications
• Need explicit resource allocation (CPU, memory, disk)
• Require container specifications for environment isolation
• Benefit from sophisticated retry policies and error handling

Key Conversion Challenges

Challenge	Shared → Distributed	Distributed → Shared
File Transfer	Add transfer_input_files/transfer_output_files	Remove transfer specifications
Resources	Add request_cpus, request_memory, request_disk	Convert to engine-specific resource formats
Containers	Specify Docker/Singularity images	Map to conda environments or system packages
Error Handling	Add retry policies and error strategies	Convert to engine-specific error handling
Scatter/Gather	Expand to explicit job definitions	Map to wildcards or engine-specific parallelization

Interactive Configuration Assistance

Use --interactive mode to get guided assistance:

# Interactive conversion with configuration prompts
wf2wf convert -i Snakefile -o workflow.dag --interactive

# Example prompts you'll see:
# Found 3 tasks without explicit resource requirements. 
# Distributed systems require explicit resource allocation. 
# Add default resource specifications? (y)es/(n)o/(a)lways/(q)uit: y

Automatic Configuration Analysis

The conversion report includes detailed analysis:

## Configuration Analysis

### Potential Issues for Distributed Computing

* **Memory**: 2 tasks without explicit memory requirements
* **Containers**: 3 tasks without container/conda specifications  
* **Error Handling**: 3 tasks without retry specifications
* **File Transfer**: 6 files with auto-detected transfer modes

**Recommendations:**
* Add explicit resource requirements for all tasks
* Specify container images or conda environments for environment isolation
* Configure retry policies for fault tolerance
* Review file transfer modes for distributed execution

See File Transfer Handling for detailed information about file transfer modes and best practices.

---

🤝 Contributing

1. Fork the repository and create a feature branch (git checkout -b feature/amazing-feature)
2. Add tests for new functionality
3. Ensure all tests pass (pytest -q)
4. Open a Pull Request – GitHub Actions will run the test matrix automatically

Please read CONTRIBUTING.md for the full guidelines.

🧪 Testing

Run the comprehensive test suite:

# Run all tests
python -m pytest tests/ -v

# Run specific test categories
python -m pytest tests/test_conversions.py::TestConversions::test_linear_workflow_conversion -v

# Run with coverage
python -m pytest tests/ --cov=wf2wf --cov-report=html

---

📞 Support

• 📖 Documentation – The docs/ folder contains a growing knowledge-base, rendered on ReadTheDocs.
• 🐛 Issues – Found a bug or missing feature? Open an issue.
• 💬 Discussions – General questions and ideas live in GitHub Discussions.

---

📄 License

wf2wf is licensed under the MIT License – see the LICENSE file for details.

---

🙏 Acknowledgments

• CHTC - The Center for High Throughput Computing for testing and feedback
• The OpenAI and Anthropic-powered coding assistants whose suggestions accelerated feature implementation
• Cursor - Interactive IDE used for pair-programming and AI-assisted development

---

Bridging workflow ecosystems.