Workflow Canvas

Tutorials

  • Getting Started with Workflow Canvas
    • What is Workflow Canvas?
    • Prerequisites
    • Installation
    • Your First Pipeline
      • 1. Initialize a project
      • 2. Register a module
      • 3. Register a method
      • 4. Register a sample
      • 5. Define and run a pipeline
    • Next Steps
  • Tutorial: Authoring a Method Script
    • Authoring a Method Script
    • Step 1 — Lay out the method directory
      • Where methods live
    • Step 2 — Write the script with the wfc-client decorator
      • Install
      • The decorator surface
    • Step 3 — Understand the contract underneath
      • The contract
      • The same method, stdlib only
      • Recording metrics without the client
      • What wfc does after your script exits
    • Step 4 — Declare slots in method.yaml
    • Next steps
  • Tutorial: Registering an Environment
    • Overview
    • Why environments exist
    • Choosing a backend
    • Referencing an env from a method
    • The ephemeral-container model
    • Requesting GPUs
    • Working inside an environment
    • Docker readiness and getting runnable
    • Next steps
  • Tutorial: Writing Contracts
    • Writing contracts
    • Two tiers: module and method contracts
    • Declaring module contracts: file vs. CLI
    • Column validation: what’s inside the file
    • Worked example: from_params
    • Where contracts are enforced
    • Next steps

How-to Guides

  • Registering Modules, Methods, and Samples
    • Registering a Module
      • Command
      • Two ways to supply contracts
      • What happens in the database
    • Registering a Method
      • Command
      • What the command does (step by step)
      • save_artifact static validation (Tier-1 only)
      • Key constraints
    • Registering a Sample
      • Command
      • Prerequisites
      • What the command does
      • Important: data/samples/ is ephemeral
      • Restoring samples
    • Registering from the Canvas
    • Next Steps
  • Canvas Visual Builder
    • Overview
    • System Nodes
    • Builder Basics
    • Registry View
    • Pipeline Variables
    • History Basics
    • Next Steps
  • Run & Inspect Results
    • Run & inspect results
    • Runs and outputs
      • Where a run’s outputs live
      • How a run’s outputs are retrieved
    • Using the Canvas to inspect runs
      • Launch the Canvas
      • History tab
      • Filter runs
      • Browse lineage chains
      • Detail panel
      • Descendants view
      • Full fan-in lineage from the command line
      • Bulk export
    • Next steps
  • How-to: Sweep Parameters & Fan Out
    • Sweep parameters and fan out
    • Parameter sweeps, variants & per-sample overrides
    • Fan-out and fan-in
    • Reusing values with pipeline variables
    • Next steps

Explanation

  • Project Anatomy
    • Directory Structure
    • The Database
    • Configuration
    • Modules vs. Methods
      • Nested vs. Flat
      • Registration validations
    • The Run Workspace (.runs/)
      • What lives where
    • Next Steps
  • How a Run Executes
    • How a run executes
    • Snakemake is the orchestrator
    • The per-step unit: wfc run-step
    • The wfc ↔ method boundary: env vars + files
    • Where to go next
  • Storage & Provenance
    • Storage & Provenance
    • The content-addressed cache is the only authoritative store
    • Archiving is deferred, and indexed by the database
    • What git tracks (and what it doesn’t)
    • Sharing results across machines
    • Where to go next
  • Caching & Reproducibility
    • Caching & Reproducibility
    • Why a step re-runs (the cache key)
    • Lineage: tracing how a result was produced
      • How lineage is stored
      • Fan-in support
      • Cache-hit runs in lineage
      • Querying lineage
    • Next steps
Workflow Canvas
  • Search

© Copyright .

Built with Sphinx using a theme provided by Read the Docs.