跳转至

sage-apps

Application examples showcasing SAGE's capabilities.

Layer: L5 (Applications)

Overview

sage-apps contains real-world application examples built with SAGE:

  • Medical Diagnosis: AI-powered medical image analysis system
  • Video Intelligence: Video content analysis pipeline
  • More applications coming soon...

Applications

Medical Diagnosis

AI-powered medical diagnosis system using MRI image analysis.

Features:

  • Diagnostic agent with image analysis
  • Report generation
  • Knowledge base integration
  • Batch processing support

Location: packages/sage-apps/src/sage/apps/medical_diagnosis/

Quick Start:

# Setup
cd packages/sage-apps/src/sage/apps/medical_diagnosis

# Download dataset (optional)
python scripts/download_lumbar_dataset.py

# Prepare data
python scripts/prepare_data.py

# Run diagnosis
python run_diagnosis.py

Components:

  • agents/: Diagnostic agents (DiagnosticAgent, ImageAnalyzer, ReportGenerator)
  • tools/: Medical knowledge base and utilities
  • config/: Agent and model configurations
  • scripts/: Data preparation scripts

Documentation:

  • See packages/sage-apps/src/sage/apps/medical_diagnosis/README.md
  • Example workflows in examples/apps/medical_diagnosis/

Video Intelligence

Video content analysis and understanding pipeline.

Features:

  • Video frame extraction
  • Object detection and tracking
  • Scene understanding
  • Event detection
  • Content summarization

Location: packages/sage-apps/src/sage/apps/video/

Quick Start:

# Run video pipeline
cd packages/sage-apps/src/sage/apps/video
python video_intelligence_pipeline.py --video input.mp4

Operators:

  • perception.py: Visual perception (object detection, face recognition)
  • analytics.py: Scene analysis and event detection
  • preprocessing.py: Video preprocessing (frame extraction, resizing)
  • formatters.py: Output formatting
  • sources.py: Video input sources
  • sinks.py: Result output

Documentation:

  • See packages/sage-apps/src/sage/apps/video/README.md

Installation

Install sage-apps:

pip install -e packages/sage-apps

Or with all dependencies:

pip install -e packages/sage-apps[all]

Running Examples

Medical Diagnosis

from sage.apps.medical_diagnosis import DiagnosticAgent

# Create diagnostic agent
agent = DiagnosticAgent(config_path="config/agent_config.yaml")

# Diagnose single case
result = agent.diagnose(
    image_path="data/case_001.npy", patient_info={"age": 45, "gender": "男"}
)

print(result.report)

Video Intelligence

from sage.kernel.api.local_environment import LocalStreamEnvironment
from sage.apps.video import VideoIntelligencePipeline

# Create pipeline
env = LocalStreamEnvironment("video_analysis")
pipeline = VideoIntelligencePipeline(env)

# Process video
results = pipeline.process("input.mp4")

Development

Adding New Applications

  1. Create application directory:
packages/sage-apps/src/sage/apps/my_app/
├── __init__.py
├── README.md
├── operators/      # Custom operators
├── agents/         # If using agents
├── config/         # Configuration files
└── scripts/        # Utility scripts

  1. Implement your application logic

  2. Add documentation and examples

  3. Add tests in packages/sage-apps/tests/my_app/

Testing

# Run all tests
cd packages/sage-apps
pytest tests/ -v

# Test specific app
pytest tests/medical_diagnosis/ -v
pytest tests/video/ -v

Architecture

Applications in sage-apps demonstrate:

  • Integration patterns: How to combine kernel, libs, and middleware
  • Best practices: Recommended ways to structure SAGE applications
  • Real-world usage: Practical examples with complete workflows

See Also

Contributing

We welcome new application examples! Please:

  1. Ensure code quality and documentation
  2. Add comprehensive tests
  3. Follow SAGE architecture guidelines
  4. Include setup instructions and dependencies

See Contributing Guide for details.