syNumpy Public Release v1.9.7

syNumpy 1.9.7

syNumpy 1.9.7 is the first public open-source release of the library from Symisc Systems. It delivers a modern C++17 implementation for reading and writing NumPy .npy files, built for straightforward integration into native codebases.

Highlights

• New clean C++17 implementation
• Standalone .npy reader and writer
• Single public header: synumpy.hpp
• Single implementation file: synumpy.cpp
• Core parser built around syNumpy::loadNpyBuffer()
• BSD 3-Clause licensed
• Production use inside FACEIO and PixLab internal systems

Included API Surface

• syNumpy::NpyArray
• syNumpy::loadNpyBuffer()
• syNumpy::loadNpy()
• syNumpy::saveNpyRaw()
• Typed syNumpy::saveNpy() overloads
• Append mode support on axis 0 for compatible arrays

What’s New

• Robust parsing and validation of NumPy .npy headers
• In-memory loading for embedded and buffer-driven workflows
• File-based loading and saving helpers
• Typed save helpers for common native C++ numeric types
• Explicit runtime error reporting via syNumpy::Error
• Minimal example program in example.cpp
• Ready-to-use CMakeLists.txt and Makefile

Integration

The recommended integration path is simple:

• Copy synumpy.hpp and synumpy.cpp into your project
• Compile with a C++17-compatible toolchain
• Start reading or writing .npy arrays immediately

No external runtime dependency stack is required for the core library.

Production Usage

syNumpy is used internally by:

• FACEIO for facial feature extraction workflows
• PixLab and Symisc Systems production pipelines
• PixLab document-processing and visual-search workflows
• DOCSCAN-related identity document scanning systems

Notes

• This release focuses on NumPy .npy support
• .npz archives are not part of this release
• Unsupported or malformed inputs fail explicitly rather than silently degrading behavior

Licensing

Released under the BSD 3-Clause License.

Project Links

• Homepage: https://pixlab.io/numpy-cpp-library
• Source code: https://github.com/symisc/sy-numpy-cpp
• FACEIO: https://faceio.net/
• PixLab: https://pixlab.io/