SyncFloat

class thread_factory.concurrency.sync_types.sync_float.SyncFloat(*args, **kwargs)

Bases: ISync

🔐 Core Purpose

SyncFloat provides synchronized access to floating-point operations, ensuring that multiple threads can safely read and modify the value without race conditions.

🔧 Features

• Full support for arithmetic operations (+, -, *, /, //, %, **, etc.)

• Comparison operators (==, <, >, etc.)

• Type conversion (int(), float(), str())

• Manual and atomic value access via .get() and .set()

• Safe thread-aware binary operations with locking on both operands

• Cross-compatible operations with other Sync types (e.g., SyncInt, SyncBool)

• Reverse operation support (__radd__, __rsub__, etc.)

• CPython compatibility with as_integer_ratio, hex, and fromhex

🧠 Deadlock Prevention

For binary operations involving multiple Sync types, locks are always acquired in order of object ID to avoid deadlocks.

This makes operations like SyncFloat(3.14) + SyncFloat(2.0) safe across threads.

🔁 Sync Type Interop

The _unwrap_other() method allows flexible interaction between SyncFloat and:

• Other SyncFloat, SyncInt, or any object with a get() method

• Primitive types like float, int, str (auto-converted via float())

✅ Recommended Usage:

a = SyncFloat(3.0) b = SyncInt(2) c = a + b # Thread-safe: unwrapped and locked d = SyncFloat(2.5) e = SyncFloat.safe_pow(a, d, 2) # Apply ternary pow safely (custom)

as_integer_ratio()

Return a pair of integers whose ratio is exactly equal to the float.

Returns:

The numerator and denominator.

Return type:

tuple[int, int]

Raises:

• OverflowError – If the float is infinity.

• ValueError – If the float is NaN.

conjugate()

Return the complex conjugate of the float (always self for real numbers).

Returns:

The same value.

Return type:

float

decrement(value: Real | ISync = 1.0)

Atomically subtract value from the internal float and return the new total.

static fromhex(s)

Parameters:

s (str) – A hexadecimal float string (e.g., ‘0x1.91eb851eb851fp+1’).

Returns:

A new instance initialized with the parsed float.

Return type:

SyncFloat

get() → float

Get the current float value in a thread-safe way.

Returns:

The current value.

Return type:

float

hex()

Return a hexadecimal string representation of the float.

Returns:

Hexadecimal float string (e.g., ‘0x1.999999999999ap-4’).

Return type:

str

property imag: float

Since SyncFloat represents real numbers, this always returns 0.0.

Returns:

0.0

Return type:

float

increment(value: Real | ISync = 1.0)

Atomically add value to the internal float and return the new total.

is_integer()

Check if the float is equivalent to an integer.

Returns:

True if value is an integer.

Return type:

bool

property real: float

This mirrors float behavior. For SyncFloat, it’s always equal to self.get().

Returns:

The real part of the float.

Return type:

float

set(new_value: Real | str)

Set a new float value in a thread-safe way.

Parameters:

new_value (Real | str) – A numeric or string value convertible to float.