v0.1.3-alpha

Release Highlights
Critical Production Fixes (PR #10):
Fixed read-only filesystem errors (42% code reduction)
Fixed user scripts on noexec filesystems
Fixed package installation mismatches
Customer validated on XT-5 hardware
Major Documentation (PR #9):
QUICKSTART.md, WORKFLOWS.md, FAQ.md added
README restructured (64% size reduction)
6+ comprehensive guides
check-prerequisites validation script
NPU/Model Zoo (PR #8):
RKNNLite compatibility layer
rknn_model_zoo examples work out-of-box
YOLOX 93% accuracy validated
Breaking Change: Requires BrightSign OS 9.1.79.3+

v0.1.2-alpha

What's Changed

• Fix librknnrt by @scottrfrancis in #7

New Contributors

• @scottrfrancis made their first contribution in #7

Full Changelog: v0.1.1-alpha...v0.1.2-alpha

v0.1.1-alpha

Release Title: BrightSign Python CV Extension v0.1.1 - Enhanced Testing &
Package Optimization

Tag: v0.1.1-alpha

Release Notes:

BrightSign Python CV Extension v0.1.1 - ALPHA

🎉 Major Improvements

This release significantly enhances the Python CV development environment
with comprehensive testing capabilities and optimized package management.

🚀 Key Features

• 96.6% Package Compatibility - Complete CV/ML/AI development platform
• Hybrid Package Strategy - SDK-built core packages + runtime-installed
  extensions
• Comprehensive Testing Suite - Automated import testing for all
  packages
• Enhanced Documentation - Clear package availability guidelines

📦 Package Management Optimization

SDK-Built Packages (Always Available)

Core packages built into the extension and immediately available:

• OpenCV - Computer vision library
• pandas 2.0.3 - Data analysis (major upgrade from 1.3.5)
• Pillow - Image processing
• NetworkX - Graph analysis
• ImageIO - Image I/O operations
• psutil, tqdm, jinja2 - Essential utilities

Runtime-Installed Packages (After User Init)

Extended packages installed via pip during initialization:

• PyTorch 2.4.1 + torchvision 0.19.1 - Deep learning framework
• scikit-image 0.21.0 - Advanced image processing
• SciPy 1.10.1 - Scientific computing
• matplotlib 3.7.5 - Visualization
• RKNN Toolkit Lite2 2.3.2 - BrightSign NPU acceleration
• 50+ additional scientific packages

🔧 New Testing Infrastructure

Comprehensive Import Testing

• New Script: sh/test_python_imports with multiple testing modes
• SDK Package Testing - Verify all built-in packages
• Runtime Package Testing - Check pip-installed packages
• Integration Testing - Validate CV/ML workflows
• Verbose Output - Detailed debugging information

Usage Examples

# Basic testing
source sh/setup_python_env
sh/test_python_imports

# Detailed output
sh/test_python_imports --verbose

# Test specific categories
sh/test_python_imports --sdk-only
sh/test_python_imports --runtime-only

📚 Documentation Enhancements

- Package Availability Guide - Clear categorization of available packages
- Import Examples - Ready-to-use code snippets
- Testing Instructions - Comprehensive validation procedures
- Version Information - Specific package versions included

🛠️ Technical Improvements

Build Optimization

- Streamlined SDK Build - Removed complex scikit-image dependencies
- Faster Build Times - Eliminated problematic pythran/beniget chains
- Reliable Core Packages - pandas now built into SDK for guaranteed
availability

Code Quality

- Cleanup - Removed orphaned recipe files
- Better Error Handling - Enhanced user initialization scripts
- Comprehensive Testing - Multi-level validation approach

🎯 Target Applications

Perfect for enterprise edge CV applications:
- Audience Analytics - Real-time people counting and demographics
- Interactive Displays - Gesture recognition and user interaction
- Retail Analytics - Customer behavior analysis
- Security Applications - Motion detection and monitoring
- Industrial Automation - Quality control and inspection

🏗️ System Requirements

Development Host

- Architecture: x86_64 (Intel/AMD) - Apple Silicon incompatible
- Memory: 16GB+ RAM recommended
- Storage: 50GB+ free space
- Tools: Docker, git, cmake

Target Device

- Players: BrightSign Series 5 (XT-5, Firebird, LS-5)
- Firmware: 9.1.52+ required
- Configuration: Unsecured, SSH enabled

📈 Performance Highlights

- Complete ML Stack - PyTorch + scikit-image + OpenCV + RKNN
- Hardware Acceleration - NPU support for ~10x inference speedup
- Production Ready - Persistent extension architecture
- Robust Testing - Comprehensive validation suite

🔄 Upgrade Notes

This release maintains backward compatibility while significantly
improving:
- Package Reliability - Core packages now SDK-built
- Testing Capabilities - New comprehensive test suite
- Documentation Quality - Clear usage guidelines

🐛 Bug Fixes & Stability

- Resolved: Complex scikit-image build dependencies
- Enhanced: Error handling in initialization scripts
- Improved: Package import reliability
- Optimized: Build process efficiency

🎊 What's Next

- Performance Optimization - Further build time improvements
- Extended Package Support - Additional ML libraries
- Enhanced Examples - More comprehensive demos
- Platform Expansion - Additional BrightSign player support

---
Installation: Download the extension package and follow the deployment
guide in README.md

Support: Report issues at
https://github.com/brightsign/python-cv-dev-extension/issues