Lily02 — Agentic Ocean Intelligence Workstation (v2.0)

A premium, agentic scientific workstation for advanced oceanographic telemetry analysis. Fuses AutoGluon forensic ML models, local LLM orchestration (LM Studio), and high-speed LangChain integrations into a "Pure Sea" minimalist research environment.

🌊 New in v2.0: The "Pure Sea" Evolution

• Lily02 CLI (Open Source): Cross-platform terminal interface for scientific intelligence (Windows, Linux, iOS via iSH).
• GPT-OSS-120B Integration: Premier local model support for complex scientific reasoning and forensic analysis.
• Local Installation Portal: Integrated setup hub for deploying the hyperpipeline on private architecture.
• Search Dock Action Menus: A dynamic "Plus" (+) toggle system for scientific presets and AutoML controls.
• Auto-ML Deep Insights: Integrated AutoGluon statistical forensics (Zap icon) for deep-data pattern recognition.

Project Structure

argoland/
├── hyperpipeline/               # Decoupled Reasoning Engine (OSS Core)
├── lily02_cli/                  # Terminal Command Interface [NEW]
├── lily02_frontend/             # React (Vite) Workstation UI
│   ├── src/pages/
│   │   └── LocalInstallation.jsx # Deployment Hub [NEW]
├── lily02_backend/              # Unified FastAPI API (Port 8000)
├── scripts/                     # Operational & Recovery scripts
├── pyproject.toml               # Python Packaging Spec [NEW]
└── README.md

🚀 Local Installation (CLI)

Lily02 is now available as a pip-installable agentic package for Windows, Linux, and iOS.

1. Install the CLI Package

pip install .

2. Configure the reasoning core Set your model in new.env:

LOCAL_AGENT_MODEL="gpt-oss-120b"

3. Run the CLI

# General query
lily02 ask "Analyze heat trends in the North Atlantic"

# Enable deep Auto-ML forensics
lily02 ask "Scan for BGC anomalies" --automl

🖥️ Workstation Quick Start

1. Initialize Environment

pip install -r requirements.txt
cd lily02_frontend && npm install && npm run build

2. Launch Unified Services

python recovery.py

Access the workstation UI locally at http://localhost:8000.

High-Performance Architecture

Layer	System	Capability
Interface	React v18 / CLI	"Pure Sea" UX + Premium Terminal Diagnostics
Logic	ParentOrchestrator	Multi-step reasoning with gpt-oss-120b
Data	Supabase/pgvector	Sub-second RAG retrieval
Forensics	AutoGluon Tabular	Real-time statistical anomaly detection
Resilience	Recovery.py	Automated port clearing and service stabilization

---

© 2026 Lily02 Scientific Mission Control. Open Source Agentic Research.