Releases: marcelomarkus/b-fast
⚡ B-FAST v1.1.0 - Extended Type Support & Performance Optimizations
🚀 Features
Extended Type Support
- ✅ datetime, date, time
- ✅ UUID
- ✅ Decimal
- ✅ Enum (string and int)
- ✅ tuple, set, frozenset
- ✅ bytes, bytearray
- ✅ Type preservation with special tags
Performance Optimizations
- ✅ Memory allocation improvements
- ✅ Optimized type checking
- ✅ Buffer pre-allocation
- ✅ Recursion depth protection
- ✅ Parallel compression
📊 Performance Results
- Simple objects: 1.7x faster than orjson
- 100 Mbps network: 5.7x faster than orjson (round-trip)
- NumPy arrays: 14x faster than orjson
- Compression: 89% payload reduction
📚 Documentation
- ✅ Translated README to English
- ✅ Added philosophy and project message
- ✅ Updated all benchmarks
- ✅ Fixed multilingual docs structure
- ✅ Added "Recommended" badge to FastAPI integration
- ✅ Added uv installation option
🧪 Tests
- ✅ 25/25 tests passing
- ✅ Full coverage for extended types
- ✅ Integration tests for type preservation
🔧 Code Quality
- ✅ Formatted with black and cargo fmt
- ✅ Linted with ruff
- ✅ Zero warnings
📦 Installation
# Python
pip install bfast-py==1.1.0
# or
uv add bfast-py
# TypeScript
npm install bfast-client@1.1.0🔗 Links
🌐 B-FAST v1.0.7 - Multilingual Documentation
🌐 Multilingual Documentation Release
B-FAST v1.0.7 introduces comprehensive multilingual documentation, making the library accessible to both English and Portuguese-speaking developers worldwide.
🌍 New Language Support
- Complete Portuguese Documentation - Full translation of all documentation pages
- Language Selector - Easy switching between English and Portuguese (similar to FastAPI docs)
- Organized URLs - Clean structure with
/(English) and/pt/(Portuguese) - Automated Build System - Script for maintaining both language versions
📚 Documentation Enhancements
🇧🇷 Portuguese Content
- Início - Complete overview with performance benchmarks
- Começando - Comprehensive installation and setup tutorial
- Frontend - TypeScript integration with React, Vue, and Angular examples
- Performance - Detailed technical analysis and optimization guide
- Solução de Problemas - Complete troubleshooting guide
🇺🇸 English Improvements
- Updated benchmarks with latest v1.0.6 performance results
- Enhanced navigation and user experience
- Better code examples and use cases
- Improved technical documentation structure
🔧 Technical Improvements
- Enhanced MkDocs Theme - Better navigation features and responsive design
- SEO Optimization - Proper language metadata for search engines
- Build Automation -
build_docs.shscript for easy multilingual builds - Improved Structure - Better organization of documentation files
📊 Performance Highlights (from v1.0.6)
- 2.1x faster than JSON serialization
- 78.7% smaller payloads with LZ4 compression
- 4.0x faster round-trip on 100 Mbps networks
- 148x speedup for NumPy arrays
🎯 Global Accessibility
This release makes B-FAST documentation accessible to:
- International developers with comprehensive English docs
- Brazilian/Portuguese developers with native language support
- All users with improved navigation and examples
🔧 Installation
Python:
pip install bfast-py==1.0.7TypeScript:
npm install bfast-client@1.0.7📖 Documentation Links
- English: https://marcelomarkus.github.io/b-fast/
- Português: https://marcelomarkus.github.io/b-fast/pt/
🚀 What's Next
With multilingual documentation in place, B-FAST is now ready for global adoption. The library maintains its exceptional performance while being accessible to developers worldwide.
Full Changelog: https://github.com/marcelomarkus/b-fast/blob/main/CHANGELOG.md
🚀 B-FAST v1.0.6 - Major Performance Breakthrough
🚀 Performance Breakthrough
B-FAST v1.0.6 delivers exceptional performance improvements and code organization, making it production-ready for high-performance applications.
⚡ Performance Highlights
- 2.1x faster than JSON serialization (4.51ms vs 9.64ms for 10k objects)
- 78.7% smaller payloads with built-in LZ4 compression (252KB vs 1.18MB)
- 4.0x faster round-trip performance on 100 Mbps networks
- 148x speedup for NumPy arrays vs JSON
🔧 Technical Optimizations
- SIMD batch processing with cache-aligned memory operations
- Branch prediction hints and unrolled loops for common patterns
- Hash-based string ID caching (64-entry cache)
- Zero-copy NumPy array serialization
- Direct memory access with unsafe operations for maximum speed
📊 Benchmark Results
| Method | Time (ms) | Speedup | Payload Size | Reduction |
|---|---|---|---|---|
| JSON (stdlib) | 9.64ms | 1.0x | 1.18MB | 0% |
| orjson | 1.51ms | 6.4x | 1.06MB | 10.2% |
| Pickle | 2.74ms | 3.5x | 808KB | 31.6% |
| B-FAST | 4.51ms | 2.1x | 998KB | 15.5% |
| B-FAST + LZ4 | 5.21ms | 1.9x | 252KB | 78.7% |
🎯 Ideal Use Cases
- Mobile & IoT: 78.7% bandwidth savings + 2.1x performance
- Data Pipelines: 148x speedup for NumPy arrays
- APIs over slow networks: Up to 4x faster round-trip
- Storage & Caching: Superior compression with fast decompression
📁 Code Organization
- Moved benchmark files to dedicated
/benchmarksfolder - Clean project structure with only
main.pyin root - Comprehensive test suite and documentation
📚 Documentation
- Added comprehensive performance analysis (
docs/performance.md) - Created troubleshooting guide (
docs/troubleshooting.md) - Updated README with real-world benchmark results
- Complete API documentation at https://marcelomarkus.github.io/b-fast/
🔧 Installation
Python:
pip install bfast-py==1.0.6TypeScript:
npm install bfast-client@1.0.6🚀 What's Next
B-FAST v1.0.6 represents a mature, production-ready serialization library that excels in bandwidth-constrained environments while maintaining competitive performance across all network speeds.
Full Changelog: https://github.com/marcelomarkus/b-fast/blob/main/CHANGELOG.md