elusight 1.0.0

EluSight - Chromatographic Decision Intelligence Framework

🔬 EluSight – Chromatographic Decision Intelligence Framework

EluSight converts chromatographic optimization outputs into explainable, trustworthy, and visual decision intelligence.

Not just what works — but why it works, how reliable it is, and what risks exist.

---

🎯 What EluSight Does

Traditional tools → “Best method = X” EluSight → “X is best because…”

• ✅ Scientific reasoning
• ✅ Trust score (0–100)
• ✅ Risk & robustness analysis
• ✅ Tradeoff insights (Pareto)
• ✅ Feature importance (SHAP)
• ✅ Automatic visualizations

---

📊 Key Outputs

🔹 Trust Score

{
  "overall_score": 87.3,
  "recommendation": "Strongly Recommend"
}

🔹 Scientific Reasoning

• Constraint satisfaction with margins
• Robustness probability
• Risk (co-elution, failure)
• Expert alignment

🔹 Visualizations (Auto-generated)

• Correlation heatmap
• Feature importance
• SHAP summary
• Partial dependence
• Parallel coordinates
• Radar chart

---

⚙️ Core Engines

Engine	Purpose
Constraint Engine	AQbD evaluation
Robustness Engine	Sensitivity analysis
Risk Engine	Failure probability
Tradeoff Engine	Pareto analysis
Explainability Engine	SHAP insights
Trust Engine	Final scoring
Visualization Engine	Auto plots

---

🔌 Supported Inputs

• Bayesian Optimization
• NSGA-II / Genetic Algorithms
• Optuna / ML pipelines
• DryLab / AQbD DoE
• Any CSV / JSON dataset

---

🚀 Installation

pip install elusight

---

📦 Quick Example

from elusight import Ingestor, TrustEngine
from elusight.constraints import ConstraintEngine

methods = Ingestor.from_csv("results.csv")

constraints = {
    "resolution": {"type": "minimum", "value": 2.0}
}

report = ConstraintEngine(constraints).evaluate(methods[0])
trust = TrustEngine().compute_trust(
    method_id=methods[0].method_id,
    constraint_report=report
)

print(trust.overall_score)

---

📁 Output Structure

elusight_outputs/
├── correlation_heatmap.png
├── feature_importance.png
├── shap_summary.png
├── parallel_coordinates.png
├── radar_chart.png
├── report.md
├── report.html
└── report.json

---

💡 Use Cases

• Pharmaceutical method development (ICH Q14)
• QC validation & transfer
• Regulatory submissions
• High-throughput screening

---

🧪 Testing

pytest tests/ -v

---

📚 Scientific Basis

• ICH Q14 – Method Development
• ICH Q9 – Risk Management
• AQbD principles
• Multi-objective optimization

---

📞 Support

• GitHub: https://github.com/SubhraC-prog/EluSight
• Issues: Bug reports & features

---

📄 License

MIT License

---

🔬 Philosophy

Scientists need confidence, evidence, and reasoning — not just results.