[decimal] Re-write Decimal constructor from floating-point value from_float()

[forfudan]

This PR re-writes the Decimal constructor from floating-point value by adding a new static method from_float(). It also adds tests and benches.

• Purpose: Converts a Float64 (double-precision floating-point) value into a Decimal object, enabling precise decimal arithmetic.
• IEEE 754: Leverages IEEE 754 standard for floating-point representation, using bit manipulation for efficient exponent extraction.
• Special Value Handling: Explicitly checks for and handles special floating-point values like zero, denormalized numbers, positive/negative infinity, and NaN (Not a Number).
• Decimal Exponent Approximation: Approximates the decimal exponent from the binary exponent using the mathematical relationship log10(2^exp) = exp * log10(2).
• Coefficient Extraction: Extracts the integer coefficient representing the significant digits of the Float64.
• Range Limiting: Ensures that the calculated scale remains within the valid range (0 to Decimal.MAX_SCALE) to prevent errors.