isage-agentic 0.1.0.3

SAGE Agentic Framework - Agent framework, planning, tool selection, and workflow

SAGE Agentic Framework

Independent package for agentic AI capabilities: planning, workflows, and agent coordination

📢 Note: Tool selection algorithms have been moved to sage-agentic-tooluse for focused development.

🎯 Overview

sage-agentic provides a comprehensive framework for building agentic AI systems with:

• Planning Algorithms: ReAct, Tree of Thoughts (ToT), hierarchical planning
• Workflow Management: Workflow orchestration and optimization
• Agent Coordination: Multi-agent collaboration and registry
• Reasoning: Advanced reasoning capabilities and timing decisions

📦 Installation

# Basic installation
pip install isage-agentic

# With LLM support
pip install isage-agentic[llm]

# Development installation
pip install isage-agentic[dev]

🔧 Tool Selection (Moved to sage-agentic-tooluse)

Tool selection algorithms are now in a separate package:

• Repository: https://github.com/intellistream/sage-agentic-tooluse
• Install: pip install isage-agentic-tooluse
• Import: from sage_libs.sage_agentic_tooluse import ...

# Tool selection - use sage-agentic-tooluse package
from sage_libs.sage_agentic_tooluse import (
    KeywordToolSelector,
    EmbeddingToolSelector,
    HybridToolSelector,
    DFSDTToolSelector,
    GorillaAdapter,
)

Why separate tool selection?

• Focused development by dedicated team
• Rapid iteration with independent versioning
• Can be used outside SAGE ecosystem

🚀 Quick Start

Planning

from sage_libs.sage_agentic.agents.planning import ReActPlanner

# Create planner
planner = ReActPlanner(llm=your_llm_client)

# Generate plan
plan = planner.plan(
    task="Analyze this document and summarize key findings",
    context={"document": doc_content}
)

Workflow Management

from sage_libs.sage_agentic.workflow import WorkflowEngine

# Create workflow
workflow = WorkflowEngine()

# Register and execute workflows
workflow.register("data_pipeline", pipeline_config)
result = workflow.execute("data_pipeline", inputs=data)

Intent Recognition

from sage_libs.sage_agentic.agents.intent import IntentClassifier

# Create intent classifier
classifier = IntentClassifier()

# Classify user intent
intent = classifier.classify("Show me the sales report for last month")

📚 Key Components

1. Planning (agents/planning/)

Planning algorithms and strategies:

• ToT (Tree of Thoughts): Multi-path reasoning with backtracking
• ReAct: Reasoning + Acting interleaved execution
• Hierarchical Planner: Hierarchical task decomposition
• Dependency Graph: Task dependency management
• Timing Decider: Execution timing optimization

2. Workflow (workflow/, workflows/)

Workflow orchestration capabilities:

• Workflow Engine: Execute multi-step workflows
• Workflow Nodes: Define workflow components
• Workflow Edges: Connect workflow steps
• Optimization: Workflow optimization strategies

3. Reasoning (reasoning/)

Advanced reasoning capabilities:

• Chain of Thought: Step-by-step reasoning
• Reflection: Self-evaluation and correction
• Meta-reasoning: Reasoning about reasoning processes

4. Evaluation (eval/)

Agent evaluation capabilities:

• Metrics tracking
• Determinism testing
• Telemetry and monitoring

5. Interfaces & Registry (interface/, interfaces/, registry/)

Unified interfaces and registration system for:

• Planners
• Workflows
• Agents
• Intent classifiers

🔧 Architecture

sage_libs/sage_agentic/
├── agents/                 # Agent implementations
│   ├── planning/          # Planning algorithms (ReAct, ToT, etc.)
│   ├── intent/            # Intent detection and classification
│   ├── bots/              # Bot implementations
│   ├── runtime/           # Runtime execution
│   └── profile/           # Agent profiles
├── workflow/              # Workflow orchestration
├── workflows/             # Workflow implementations
├── reasoning/             # Reasoning capabilities
├── eval/                  # Evaluation tools
├── interface/             # Protocol definitions
├── interfaces/            # Interface implementations
└── registry/              # Component registry

🎓 Use Cases

1. Multi-Agent Systems: Build coordinated multi-agent workflows
2. Complex Task Planning: Decompose tasks with hierarchical planning
3. Adaptive Workflows: Dynamic workflow execution with reasoning
4. Intent-Driven Systems: Classify and route based on user intent
5. Research: Experiment with different planning strategies

🔗 Integration with SAGE

This package is part of the SAGE ecosystem but can be used independently:

# Standalone usage
from sage_libs.sage_agentic import ReActPlanner

# With SAGE interface layer (if installed)
from sage.libs.agentic import ReActPlanner

Related Packages

• sage-agentic-tooluse: Tool selection algorithms
• sage-agentic-tooluse-benchmark: Tool selection evaluation

📖 Documentation

• Repository: https://github.com/intellistream/sage-agentic
• SAGE Documentation: https://intellistream.github.io/SAGE-Pub/
• Issues: https://github.com/intellistream/sage-agentic/issues

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

📄 License

MIT License - see LICENSE file for details.

🙏 Acknowledgments

Part of the SAGE ecosystem for stream analytics and generative AI.

📧 Contact

• Team: IntelliStream Team
• Email: [e-mail]
• GitHub: https://github.com/intellistream

---

Part of the SAGE ecosystem - Stream Analytics for Generative AI Engines