How phpy 2.2 Enables Seamless PHP‑Python Operator Overloading for Scientific Computing

phpy enables PHP‑Python interoperability, allowing PHP code to import and manipulate Python libraries such as numpy. Prior to version 2.2, operations on NumPy arrays required explicit method calls (e.g., __mul__), which made the code verbose.

Numpy operation examples (pre‑2.2)

Array multiplication

$np = PyCore::import("numpy");
$array = $np->array([1, 2, 3]);
$a2 = $array->__mul__(5);
PyCore::print($a2); // [ 5 10 15 ]

Matrix multiplication

$np = PyCore::import("numpy");
$array = $np->array([[1, 2, 3], [4, 5, 6]]);
$a2 = $array->__mul__(5);
PyCore::print($a2); // [[ 5 10 15]
//  [20 25 30]]

Operator overloading in phpy 2.2

phpy 2.2 introduces custom PHP bytecode handlers that map native PHP operators to Python’s special methods, enabling direct use of PHP operators on PyObject instances.

Arithmetic operators

+

→ __add__ (addition) - → __sub__ (subtraction) * → __mul__ (multiplication) / → __floordiv__ (floor division; true division is not the default) % → __mod__ (modulo) ** → __pow__ (exponentiation)

Bitwise operators

<<

→ __lshift__ (left shift) >> → __rshift__ (right shift) & → __and__ (bitwise AND) | → __or__ (bitwise OR) ^ → __xor__ (bitwise XOR) ~ → __invert__ (bitwise NOT)

Comparison operators

==

→ __eq__ (equality) != → __ne__ (inequality) < → __lt__ (less‑than) <= → __le__ (less‑than‑or‑equal)

Assignment (in‑place) operators

+=

→

__iadd__

-=

→

__isub__

*=

→

__imul__

/=

→

__ifloordiv__

%=

→

__imod__

**=

→

__ipow__

<<=

→

__ilshift__

>>=

→

__irshift__

&=

→

__iand__

|=

→

__ior__

^=

→

__ixor__

Important notes

In‑place operators correspond to Python’s __i* special methods.

Python strings cannot be concatenated with += or .= through this mechanism.

Division uses floor division ( __floordiv__) by default; use truediv explicitly if needed.

New syntax examples (phpy 2.2)

Arithmetic

$np = PyCore::import('numpy');
$arr = $np->array([3, 4]);
$arr2 = $arr * 3;      // [9, 12]
$arr3 = $arr2 + 5;     // [14, 17]
$arr4 = $arr3 - 5;     // [9, 12]
$arr5 = $arr4 / 3;     // [3, 4]
$arr6 = $arr5 ** 2;    // [9, 16]
$arr7 = $arr6 % 5;     // [4, 1]
$arr8 = -$arr;        // [-3, -4]

Bitwise

$np = PyCore::import('numpy');
$arr = $np->array([7, 12]);
$arr2 = $arr & 3;   // [3, 0]
$arr3 = $arr | 3;   // [7, 15]
$arr4 = $arr ^ 3;   // [4, 15]
$arr5 = $arr << 2;  // [28, 48]
$arr6 = $arr >> 2;  // [1, 3]
$arr7 = ~$arr;      // [-8, -13]

Assignment (in‑place)

$np = PyCore::import('numpy');
$arr = $np->array([7, 12]);
$arr += 5;   // [12, 17]
$arr -= 5;   // [7, 12]
$arr *= 5;   // [35, 60]
$arr /= 5;   // [7, 12]
$arr **= 3;  // [343, 1728]
$arr %= 17;  // [3, 11]
$arr &= 3;   // [3, 3]
$arr2 = $np->array([7, 12]);
$arr2 <<= 4; // [112, 192]
$arr2 >>= 3; // [14, 24]
$arr2 ^= 3;  // [13, 27]
$arr2 |= 9;  // [13, 27]

Conclusion

The operator overloading introduced in phpy 2.2 allows PHP developers to write concise, readable code for scientific calculations by using native PHP operators on Python objects. This brings PHP’s computational expressiveness close to that of Python, especially when leveraging libraries such as numpy.