Explore and visualize your WDL (Workflow Description Language) workflows

pip install -e .

Or using uv:

uv sync

Run with Docker:

# Create a directory with your user permissions
mkdir output

# Generate documentation
docker run  --user $(id -u):$(id -g) --rm \
            -v /path/to/wdl-project:/data \
            -v $(pwd)/output:/output \
            taniguti/wdlatlas:0.1.1 generate /data -o /output

# Generate graph
docker run  --user $(id -u):$(id -g) --rm \
            -v /path/to/wdl-project:/data \
            -v $(pwd)/output:/output \
            taniguti/wdlatlas:0.1.1 graph /data/workflow.wdl -o /output/graph.md

WDL Atlas provides two main commands:

Generate complete HTML documentation for your WDL project:

wdlatlas generate /path/to/wdl-project -o docs/

Generate a Mermaid diagram for a specific workflow:

wdlatlas graph workflow.wdl -o workflow_graph.md

• 📚 Complete HTML documentation with navigation
• 📊 Interactive workflow diagrams (Mermaid)
• 🐳 Docker image inventory
• 🔗 Cross-references between workflows and tasks
• 📦 External dependency tracking

⚠️ Important: The generated HTML documentation uses ES6 modules, which require a web server to work properly due to CORS restrictions.

You cannot simply open index.html directly in your browser (using file:// protocol). Instead, use one of these options:

python serve_docs.py
# or specify a different port
python serve_docs.py 8080

This will automatically:

• Start a local HTTP server
• Open your browser at the correct URL
• Serve the documentation from the docs/ directory

cd docs
python -m http.server 8000

Then open: http://localhost:8000

Modern browsers enforce strict CORS (Cross-Origin Resource Sharing) policies for ES6 modules. When opening files directly (file:// protocol), the browser blocks module imports for security reasons. A local HTTP server provides the necessary CORS headers to allow module loading.

Run tests:

uv run pytest -v --cov=src tests --cov-report=term
# or 'make test'

src/
├── cli/                 # Command-line interface
├── domain/              # Domain models
├── application/         # Use cases
└── infrastructure/      # Parsers, renderers, repositories