CausalIQ Ecosystem - Development Roadmap¶
At-a-glance view of development releases across the CausalIQ ecosystem
Last updated: August 9, 2026
🌟 Current Ecosystem Status on PyPI¶
| Project | Current Release | Project capabilities | Detailed Roadmap |
|---|---|---|---|
| causaliq (umbrella) | 0.1.0 | Ecosystem architecture and development standards defined | n/a |
| causaliq-analysis | 0.4.0 | Structural graph merging, metric evaluation and summarisation, and legacy learning traces | here |
| causaliq-core | 0.7.0 | Utility functions, graph classes (SDG, PDAG, DAG, PDG), optimal DAG, Bayesian Networks, cache infrastructure and randomised filters | here |
| causaliq-data | 0.3.0 | Data handling, score functions and independence tests | here |
| causaliq-knowledge | 0.6.0 | PDG generation and workflow integration | here |
| causaliq-repo-template | 1.0.0 | Repo template for new CausalIQ projects | n/a |
| causaliq-workflow | 0.5.0 | Workflow patterns, multi-step workflows, caches and conservative execution | here |
| zenodo-sync | 0.1.0 | Follows CausalIQ standards | tbd |
All other projects not yet released.
Milestones:
- end-May 2026: workflows and assets for draft CausalIQ LLM-assisted model-averaging paper
- end-July 2026: workflows and assets for camera-ready CausalIQ LLM-assisted model-averaging paper
- end-October 2026: Experiments and results for another paper
- end-March 2027: Full reproducibility of key published papers with assets on Zenodo
📊 Ecosystem Development Timeline¶
September 2026 - Graph evaluation and summarisation¶
| Project | Release | Status | Key Deliverables |
|---|---|---|---|
| causaliq-analysis | 0.4.0 | 🚧 Underway | Evaluation and Summarisation: structural evaluation of graphs and result summarisation |
| causaliq-discovery | 0.1.0 | 🚧 Underway | Support for HC and Tabu-Stable and some bnlearn and Tetrad algorithms |
| causaliq-analysis | 0.5.0 | 📊 Planned | Publication Assets: table and chart generation for the PGM submission |
Code migrated from legacy monolithic repo will be modified to meet CausalIQ quality standards.
CausalIQ packages will implement the CausalIQAction interface and therefore can be included in CausalIQ Workflows
September 2026 and beyond¶
| Project | Release | Status | Key Deliverables |
|---|---|---|---|
| causaliq-analysis | Non-reference Evaluation | ✨ Envisaged | Evaluation that does not require reference graphs |
| causaliq-analysis | Statistical Significance | ✨ Envisaged | Based on January Experience |
| causaliq-whatif | Foundation Inference | ✨ Envisaged | Basic PyAgrum Inference |
| causaliq-workflow | Comparison capability | ✨ Envisaged | Comparison capability |
H2 2026 - Complete Legacy Support¶
🚀 Future Vision Post 2026¶
Research Platform Features¶
- LLM Integration: Model averaging, hypothesis generation with causal reasoning
- Web Interface: Browser-based workflow designer for non-technical researchers
- Publication Support: Reproducible research outputs with automated documentation
Advanced Capabilities¶
- Workflow Marketplace: Sharing and discovering research workflow templates
- Interactive Notebooks: Jupyter integration with workflow execution
- Multi-machine Execution: Distributed workflows across compute clusters
- AI-assisted Optimisation: Automated hyperparameter and workflow tuning
- Integration Ecosystem: Plugins for major research tools and platforms
Performance & Scale¶
- High-performance Computing: GPU acceleration for large-scale structure learning
- Distributed Storage: Cloud-native knowledge bases with global accessibility
- Real-time Analytics: Live causal discovery with streaming data sources
- Enterprise Features: Security, compliance, and enterprise-grade deployment
This roadmap is updated weekly and reflects current development priorities. For detailed project-specific roadmaps, see individual project documentation.