AgGarage

An open-source python package with multiple python tools.

Installation

pip install aggarage

Usage

from aggarage import *

@blitz # Example of a decorator from the package, it optimizes heavy calculations
def heavy_calc(x):
    total = 0
    for i in range(x):
        total += i ** 2
    return total

print(heavy_calc(10_000_000))

Timed Variables

AgGarage allows you to store variables over time and retrieve past values.

from aggarage import create_timed_variable, get_value_of_timed_variable, clear_all_timed_variables

create_timed_variable("score", 100)
create_timed_variable("score", 200)

print(get_value_of_timed_variable("score"))       # latest value
print(get_value_of_timed_variable("score", "-1m")) # value 1 minute ago

clear_all_timed_variables()  # removes all timed variable data

Utilities

AgGarage also provides common utility functions:

fib(n) – Fibonacci number factorial(n) / factorial_recursive(n) – Factorial is_prime(num) – Prime check gcd(a, b) / lcm(a, b) – Greatest/common divisors reverse_string(s) / is_palindrome(s) – String utilities sum_of_squares(n) – Sum of squares printf(text, color="red") – Colored terminal output

Notes

Timed variables are stored with second-level precision and automatically prune entries older than 30 minutes. The @blitz decorator uses JIT compilation if available or threads as a fallback to speed up computations.

lightning

The @lightning decorator uses aggressive parallelization and multiprocessing to optimize performance for CPU-bound tasks. It is much faster than @blitz for heavy computations but has more overhead, so it is best suited for very intensive tasks. Use @lightning when you need maximum speedup for complex calculations that can be parallelized.

from aggarage import *

@lightning
def _heavy_computation_lightning(n):
        total = 0
        for i in range(n):
            total += sum(j * j for j in range(1000))
        return total

@blitz
def _heavy_computation_blitz(n):
        total = 0
        for i in range(n):
            total += sum(j * j for j in range(1000))
        return total

if __name__ == "__main__":
    import time

    n = 1000

    start = time.time()
    result_lightning = _heavy_computation_lightning(n)
    end = time.time()
    print(f"Lightning result: {result_lightning}, Time taken: {end - start:.4f} seconds")

    start = time.time()
    result_blitz = _heavy_computation_blitz(n)
    end = time.time()
    print(f"Blitz result: {result_blitz}, Time taken: {end - start:.4f} seconds")

Methods Comparison

Result: 332,833,500,000 (all methods produce identical results)

Key Observations:

• Lightning has higher initial overhead but benefits significantly from caching
• Blitz provides good performance with minimal overhead
• Both decorators implement effective caching mechanisms