Skip to content

🤖 CausalIQ Workflow

The CausalIQ Workflow framework provides a comprehensive solution for designing, executing, and reproducing causal discovery experiments at scale. Using a GitHub Actions-inspired YAML syntax, it enables researchers to conduct rigorous, reproducible studies while managing complex experimental configurations.

Current Version: v0.5.0

Quick Links:

Key Features

🔄 Workflow Orchestration

GitHub Actions-inspired YAML workflows with matrix expansion for systematic parameter sweeps, multi-step sequential execution, and template variable validation.

⚡ Action Framework

Three formalised action patterns — CREATE, UPDATE, and AGGREGATE — with conservative execution that automatically skips completed work. Use --mode=force to bypass skip checks.

🔍 Filter Expressions

Logical expressions evaluated against flattened entry metadata, with template variable resolution, random() sampling, and relaxed numeric suffix matching (k, M, G, T).

💾 Workflow Caching

SQLite-based result storage with SHA-256 matrix keys, null wildcards for flexible matching, and CLI commands for cache export/import.

🔌 Plug-in Actions

Auto-discovery system for registering action providers from any CausalIQ package, with type-safe ActionInput/ActionResult interfaces built on causaliq-core.

Future Directions

  • Step output chaining: Step output references and cache restoration for resumable workflows
  • Dry and comparison runs: Runtime estimation and processing summaries
  • Discovery integration: Structure learning algorithms as workflow actions
  • CI testing: Workflow specification syntax validation
  • Distributed computing: Scalable parallel processing

Integration with Ecosystem

  • 🔍 CausalIQ Discovery (causaliq-discovery) is called by this package to perform structure learning.
  • 📊 CausalIQ Analysis (causaliq-analysis) is called by this package to perform results analysis and generate assets for research papers.
  • 🧠 CausalIQ Knowledge (causaliq-knowledge) provides LLM graph generation as a workflow action with cache integration.
  • 🔮 CausalIQ WhatIf (causaliq-whatif) is called by this package to perform causal prediction.
  • 🔄 Zenodo Synchronisation (zenodo-sync) is used by this package to download datasets and upload results.
  • 🧪 CausalIQ Research (causaliq-research) experiments are defined as CausalIQ Workflows, allowing full reproduction of experiments, results, and published paper assets.


The CausalIQ Workflow framework enables reproducible, scalable causal discovery research by providing comprehensive tools for experiment design, execution, and analysis, supporting the advancement of reliable causal inference methodologies.