Implement Matlab bindings

[BenBrock]

Implement Matlab Bindings for Binsparse

This PR adds Matlab bindings, supporting (hopefully) Matlab and Octave. These add the functions binsparse_read and binsparse_write, which read and write, respectively, sparse matrices to or from Binsparse files.

This PR returns matrices as a "native Matlab struct," which contains the component arrays that compose the sparse matrix. Is does not convert to or from a Matlab sparse matrix. (We probably want that, but it should be done at the Matlab layer.)

Below are some AI-generated notes on what's changed:

New Features:

Core MEX Functions:

• binsparse_read.c - Read Binsparse matrix files into MATLAB structs
• binsparse_write.c - Write MATLAB structs to Binsparse format files
• Complete type support - All Binsparse types including complex numbers (BSP_COMPLEX_FLOAT32/64)
• Full API coverage - Optional groups, JSON metadata, compression levels

Build Infrastructure:

• build_matlab_bindings.m - Main MATLAB build script with auto-detection
• build_octave_bindings.m - Main Octave build script using mkoctfile
• compile_octave.sh - Command-line shell script for Octave compilation
• Individual compilation scripts for both MATLAB and Octave

Utilities & Testing:

• bsp_matrix_create.m - Helper for creating matrix structs
• bsp_matrix_info.m - Matrix information display utility
• test_binsparse_read.m / test_binsparse_write.m - Comprehensive test suites
• README.md - Complete documentation with usage examples

API:

% Reading
matrix = binsparse_read('file.bsp.h5');
matrix = binsparse_read('file.bsp.h5', 'group_name');

% Writing  
binsparse_write('output.bsp.h5', matrix);
binsparse_write('output.bsp.h5', matrix, 'group', '{"meta": "data"}', 6);

Key Technical Features:

• Efficient type conversion - Direct memcpy for compatible types, proper complex number handling
• Memory management - Safe allocation/cleanup with comprehensive error handling
• Cross-platform - Works on Windows, macOS, Linux with both MATLAB and Octave
• Round-trip compatibility - Perfect read → write → read workflows

Configuration Updates:

• Updated .gitignore for MATLAB/Octave artifacts
• Modified .pre-commit-config.yaml to exclude MATLAB files from clang-format
• Enhanced CMakeLists.txt for bindings support

This provides a complete, production-ready interface for MATLAB/Octave users to work with Binsparse matrices, enabling seamless integration into scientific computing workflows.