SAGE Agentic Framework
Independent package for agentic AI capabilities: tool selection, planning, workflows, and agent coordination
🎯 Overview
sage-agentic provides a comprehensive framework for building agentic AI systems with:
- Tool Selection: Multiple strategies (keyword, embedding, hybrid, DFS-DT, Gorilla)
- Planning Algorithms: ReAct, Tree of Thoughts (ToT), hierarchical planning
- Workflow Management: Workflow orchestration and optimization
- Agent Coordination: Multi-agent collaboration and registry
- SIAS: Sample-Importance-Aware Selection for tool/trajectory curation
- 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]
�� Quick Start
Tool Selection
from sage_agentic.agents.action.tool_selection import HybridToolSelector
# Create selector
selector = HybridToolSelector(embedder=your_embedder)
# Select tools
tools = selector.select(
query="search for research papers",
available_tools=all_tools,
k=3
)
Planning
from 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_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)
📚 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. Tool Selection (agents/action/tool_selection/)
Tool selection strategies:
- Keyword Selector: Rule-based keyword matching
- Embedding Selector: Semantic similarity-based selection
- Hybrid Selector: Combined keyword + embedding approach
- DFS-DT Selector: Decision tree-based selection
- Gorilla Adapter: Gorilla-style tool retrieval
3. SIAS (sias/)
Sample-Importance-Aware Selection for:
- Tool selection optimization
- Trajectory curation
- Continual learning with core-set selection
4. Evaluation (eval/)
Agent evaluation capabilities:
- Metrics tracking
- Determinism testing
- Telemetry and monitoring
5. Interfaces & Registry (interface/, registry/)
Unified interfaces and registration system for:
- Planners
- Tool selectors
- Workflows
- Agents
🔧 Architecture
sage_agentic/
├── agents/ # Agent implementations
│ ├── action/ # Action and tool selection
│ ├── planning/ # Planning algorithms
│ └── intent/ # Intent detection
├── workflow/ # Workflow orchestration
├── sias/ # Sample-Importance-Aware Selection
├── reasoning/ # Reasoning capabilities
├── eval/ # Evaluation tools
├── interface/ # Protocol definitions
├── interface/ # Protocols, registries, schemas
└── registry/ # Component registry
🎓 Use Cases
- Multi-Agent Systems: Build coordinated multi-agent workflows
- Tool-Augmented LLMs: Select and use external tools intelligently
- Hierarchical Planning: Decompose complex tasks into subtasks
- Adaptive Systems: Use SIAS for intelligent sample selection
- Research: Experiment with different planning and selection strategies
🔗 Integration with SAGE
This package is part of the SAGE ecosystem but can be used independently:
# Standalone usage
from sage_agentic import ReActPlanner, HybridToolSelector
# With SAGE (if installed)
from sage.libs.agentic import ReActPlanner # Compatibility layer
📖 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
Originally part of the SAGE framework, now maintained as an independent package for broader community use.
📧 Contact
- Team: IntelliStream Team
- Email: shuhao_zhang@hust.edu.cn
- GitHub: https://github.com/intellistream
Metadata
Release files for isage-agentic 0.1.0.5
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| isage_agentic-0.1.0.5.tar.gz | 120.3 kB | Without provenance | Details for isage_agentic-0.1.0.5.tar.gz |
Built distributions (wheels)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| isage_agentic-0.1.0.5-py3-none-any.whl | Python 3 | none | any | Details for isage_agentic-0.1.0.5-py3-none-any.whl |
| isage_agentic-0.1.0.5-py2.py3-none-any.whl | Python 2, Python 3 | none | any | Details for isage_agentic-0.1.0.5-py2.py3-none-any.whl |
Total release size: 577.2 kB
Release files / isage_agentic-0.1.0.5.tar.gz
| Download URL | isage_agentic-0.1.0.5.tar.gz |
|---|---|
| Size | 120.3 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
b81925266bfd8f36fe509f4eea61b06f562a94411167b07b0014db314a74cf9c
|
|
BLAKE2b-256 checksum How to use checksums |
57b628470caa30128e8da2b184e33db898839b0e91bdf8ded9e8e7ced8495c81
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.11.14
|
Release files / isage_agentic-0.1.0.5-py3-none-any.whl
| Download URL | isage_agentic-0.1.0.5-py3-none-any.whl |
|---|---|
| Size | 302.0 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
1e617302af297d242b26ad66a936749baaf57225c337841af3aa9a9b720a0ab3
|
|
BLAKE2b-256 checksum How to use checksums |
16af112e692cad43a0cebafef5d490afa4a4226121966444a95b4d4d2dc5f6a1
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.13.11
|
Release files / isage_agentic-0.1.0.5-py2.py3-none-any.whl
| Download URL | isage_agentic-0.1.0.5-py2.py3-none-any.whl |
|---|---|
| Size | 155.0 kB |
| Tags | Python 2 Python 3 |
|
SHA-256 checksum How to use checksums |
2f137c9c597d7d4635522c1e9ef19b07e7f14cf44fece2903f095c8472ac33df
|
|
BLAKE2b-256 checksum How to use checksums |
97bc3d521c0e2170bbe42101b8b194857b9d0cd8cd4246d7cd79cfcf731278b5
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.11.14
|
Security for isage-agentic 0.1.0.5
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.