High-performance N-dimensional arrays for PHP, powered by Rust via FFI

• Blazing Fast - Zero-copy views, Rust backend, minimal FFI overhead
• NumPy Compatible - Familiar API for Python developers
• Memory Safe - Automatic memory management via Rust
• N-Dimensional - Support for arbitrary dimensions
• Type Safe - Multiple data types with automatic inference
• Scientific Computing - Complete toolkit for ML and data science

composer require phpmlkit/ndarray

Requirements: PHP 8.1+ with FFI extension enabled

use PhpMlKit\NDArray\NDArray; // Create arrays $a = NDArray::array([[1, 2], [3, 4]]); $b = NDArray::ones([2, 2]); // Mathematical operations $c = $a->add($b); $d = $a->matmul($b); $e = $a->sum(); // Slicing with zero-copy views $first_row = $a[0]; // View: no data copied! $sub_matrix = $a->slice(['0:1', '0:1']); // Linear algebra $dot = $a->dot($b); $trace = $a->trace();

Work with large datasets efficiently:

$data = NDArray::random([10000, 1000]); // 80 MB array $batch = $data->slice(['0:1000']); // View: 0 bytes copied! $mean = $batch->mean(); // Process on view

Coming from Python? You'll feel at home (with PHP syntax):

• Full Documentation
• Installation Guide
• Quick Start
• NumPy Migration
• API Reference

Views share memory with the parent array (zero-copy):

$arr = NDArray::array([[1, 2], [3, 4]]); $row = $arr[0]; // View $row[0] = 999; // Modifies $arr!

Copies are independent:

$copy = $arr->copy(); $copy->set([0, 0], 999); // Doesn't modify $arr

Operations create copies:

$result = $arr->multiply(2); // Copy - original unchanged

PHP does NOT support operator overloading. Use method calls:

// ❌ This doesn't work in PHP $c = $a + $b; $c = $a * 2; // ✅ Use method calls instead $c = $a->add($b); $c = $a->multiply(2);

Multi-dimensional indexing uses strings:

// ❌ Invalid PHP syntax $value = $matrix[0, 0]; // ✅ Use string syntax $value = $matrix['0,0']; // or $value = $matrix->get(0, 0);

Slicing uses string syntax or method calls:

// ❌ Invalid PHP syntax $slice = $arr[0:5]; // ✅ Use string syntax $slice = $arr['0:5']; // or $slice = $arr->slice(['0:5']);

• Integers: Int8, Int16, Int32, Int64
• Unsigned: UInt8, UInt16, UInt32, UInt64
• Floats: Float32, Float64
• Boolean: Bool

• ✅ Array creation (zeros, ones, random, arange, linspace)
• ✅ Indexing and slicing (multi-dimensional, negative indices, steps)
• ✅ Views and copies (zero-copy slicing, transpose)
• ✅ Arithmetic operations (add, subtract, multiply, divide, mod, power)
• ✅ Mathematical functions (abs, sqrt, exp, log, trig, etc.)
• ✅ Reductions (sum, mean, std, min, max, argmin, argmax)
• ✅ Linear algebra (dot, matmul, trace, diagonal)
• ✅ Shape manipulation (reshape, transpose, squeeze, expandDims)
• ✅ Comparisons and boolean operations
• ✅ Sorting and searching

# Clone repository git clone https://github.com/phpmlkit/ndarray.git cd ndarray # Build the rust library ./scripts/build.sh debug # Install dependencies composer install # Run tests composer test # Run static analysis composer lint # Format code composer cs:fix

# Install Node dependencies npm install # Start docs dev server npm run docs:dev # Build docs npm run docs:build

Contributions are welcome! Please see CONTRIBUTING.md for guidelines.

MIT License - see LICENSE file for details.

Created by CodeWithKyrian

Powered by Rust ndarray

⭐ Star this repo if you find it helpful!