mirror of
				https://github.com/Legrandin/pycryptodome.git
				synced 2025-10-31 21:51:57 +00:00 
			
		
		
		
	
		
			
				
	
	
		
			395 lines
		
	
	
	
		
			11 KiB
		
	
	
	
		
			Python
		
	
	
	
	
	
			
		
		
	
	
			395 lines
		
	
	
	
		
			11 KiB
		
	
	
	
		
			Python
		
	
	
	
	
	
| # ===================================================================
 | |
| #
 | |
| # Copyright (c) 2014, Legrandin <helderijs@gmail.com>
 | |
| # All rights reserved.
 | |
| #
 | |
| # Redistribution and use in source and binary forms, with or without
 | |
| # modification, are permitted provided that the following conditions
 | |
| # are met:
 | |
| #
 | |
| # 1. Redistributions of source code must retain the above copyright
 | |
| #    notice, this list of conditions and the following disclaimer.
 | |
| # 2. Redistributions in binary form must reproduce the above copyright
 | |
| #    notice, this list of conditions and the following disclaimer in
 | |
| #    the documentation and/or other materials provided with the
 | |
| #    distribution.
 | |
| #
 | |
| # THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
 | |
| # "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
 | |
| # LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
 | |
| # FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
 | |
| # COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
 | |
| # INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
 | |
| # BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
 | |
| # LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
 | |
| # CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
 | |
| # LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
 | |
| # ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
 | |
| # POSSIBILITY OF SUCH DAMAGE.
 | |
| # ===================================================================
 | |
| 
 | |
| from ._IntegerBase import IntegerBase
 | |
| 
 | |
| from Crypto.Util.number import long_to_bytes, bytes_to_long
 | |
| 
 | |
| 
 | |
| class IntegerNative(IntegerBase):
 | |
|     """A class to model a natural integer (including zero)"""
 | |
| 
 | |
|     def __init__(self, value):
 | |
|         if isinstance(value, float):
 | |
|             raise ValueError("A floating point type is not a natural number")
 | |
|         try:
 | |
|             self._value = value._value
 | |
|         except AttributeError:
 | |
|             self._value = value
 | |
| 
 | |
|     # Conversions
 | |
|     def __int__(self):
 | |
|         return self._value
 | |
| 
 | |
|     def __str__(self):
 | |
|         return str(int(self))
 | |
| 
 | |
|     def __repr__(self):
 | |
|         return "Integer(%s)" % str(self)
 | |
| 
 | |
|     # Only Python 2.x
 | |
|     def __hex__(self):
 | |
|         return hex(self._value)
 | |
| 
 | |
|     # Only Python 3.x
 | |
|     def __index__(self):
 | |
|         return int(self._value)
 | |
| 
 | |
|     def to_bytes(self, block_size=0, byteorder='big'):
 | |
|         if self._value < 0:
 | |
|             raise ValueError("Conversion only valid for non-negative numbers")
 | |
|         result = long_to_bytes(self._value, block_size)
 | |
|         if len(result) > block_size > 0:
 | |
|             raise ValueError("Value too large to encode")
 | |
|         if byteorder == 'big':
 | |
|             pass
 | |
|         elif byteorder == 'little':
 | |
|             result = bytearray(result)
 | |
|             result.reverse()
 | |
|             result = bytes(result)
 | |
|         else:
 | |
|             raise ValueError("Incorrect byteorder")
 | |
|         return result
 | |
| 
 | |
|     @classmethod
 | |
|     def from_bytes(cls, byte_string, byteorder='big'):
 | |
|         if byteorder == 'big':
 | |
|             pass
 | |
|         elif byteorder == 'little':
 | |
|             byte_string = bytearray(byte_string)
 | |
|             byte_string.reverse()
 | |
|         else:
 | |
|             raise ValueError("Incorrect byteorder")
 | |
|         return cls(bytes_to_long(byte_string))
 | |
| 
 | |
|     # Relations
 | |
|     def __eq__(self, term):
 | |
|         if term is None:
 | |
|             return False
 | |
|         return self._value == int(term)
 | |
| 
 | |
|     def __ne__(self, term):
 | |
|         return not self.__eq__(term)
 | |
| 
 | |
|     def __lt__(self, term):
 | |
|         return self._value < int(term)
 | |
| 
 | |
|     def __le__(self, term):
 | |
|         return self.__lt__(term) or self.__eq__(term)
 | |
| 
 | |
|     def __gt__(self, term):
 | |
|         return not self.__le__(term)
 | |
| 
 | |
|     def __ge__(self, term):
 | |
|         return not self.__lt__(term)
 | |
| 
 | |
|     def __nonzero__(self):
 | |
|         return self._value != 0
 | |
|     __bool__ = __nonzero__
 | |
| 
 | |
|     def is_negative(self):
 | |
|         return self._value < 0
 | |
| 
 | |
|     # Arithmetic operations
 | |
|     def __add__(self, term):
 | |
|         try:
 | |
|             return self.__class__(self._value + int(term))
 | |
|         except (ValueError, AttributeError, TypeError):
 | |
|             return NotImplemented
 | |
| 
 | |
|     def __sub__(self, term):
 | |
|         try:
 | |
|             return self.__class__(self._value - int(term))
 | |
|         except (ValueError, AttributeError, TypeError):
 | |
|             return NotImplemented
 | |
| 
 | |
|     def __mul__(self, factor):
 | |
|         try:
 | |
|             return self.__class__(self._value * int(factor))
 | |
|         except (ValueError, AttributeError, TypeError):
 | |
|             return NotImplemented
 | |
| 
 | |
|     def __floordiv__(self, divisor):
 | |
|         return self.__class__(self._value // int(divisor))
 | |
| 
 | |
|     def __mod__(self, divisor):
 | |
|         divisor_value = int(divisor)
 | |
|         if divisor_value < 0:
 | |
|             raise ValueError("Modulus must be positive")
 | |
|         return self.__class__(self._value % divisor_value)
 | |
| 
 | |
|     def inplace_pow(self, exponent, modulus=None):
 | |
|         exp_value = int(exponent)
 | |
|         if exp_value < 0:
 | |
|             raise ValueError("Exponent must not be negative")
 | |
| 
 | |
|         if modulus is not None:
 | |
|             mod_value = int(modulus)
 | |
|             if mod_value < 0:
 | |
|                 raise ValueError("Modulus must be positive")
 | |
|             if mod_value == 0:
 | |
|                 raise ZeroDivisionError("Modulus cannot be zero")
 | |
|         else:
 | |
|             mod_value = None
 | |
|         self._value = pow(self._value, exp_value, mod_value)
 | |
|         return self
 | |
| 
 | |
|     def __pow__(self, exponent, modulus=None):
 | |
|         result = self.__class__(self)
 | |
|         return result.inplace_pow(exponent, modulus)
 | |
| 
 | |
|     def __abs__(self):
 | |
|         return abs(self._value)
 | |
| 
 | |
|     def sqrt(self, modulus=None):
 | |
| 
 | |
|         value = self._value
 | |
|         if modulus is None:
 | |
|             if value < 0:
 | |
|                 raise ValueError("Square root of negative value")
 | |
|             # http://stackoverflow.com/questions/15390807/integer-square-root-in-python
 | |
| 
 | |
|             x = value
 | |
|             y = (x + 1) // 2
 | |
|             while y < x:
 | |
|                 x = y
 | |
|                 y = (x + value // x) // 2
 | |
|             result = x
 | |
|         else:
 | |
|             if modulus <= 0:
 | |
|                 raise ValueError("Modulus must be positive")
 | |
|             result = self._tonelli_shanks(self % modulus, modulus)
 | |
| 
 | |
|         return self.__class__(result)
 | |
| 
 | |
|     def __iadd__(self, term):
 | |
|         self._value += int(term)
 | |
|         return self
 | |
| 
 | |
|     def __isub__(self, term):
 | |
|         self._value -= int(term)
 | |
|         return self
 | |
| 
 | |
|     def __imul__(self, term):
 | |
|         self._value *= int(term)
 | |
|         return self
 | |
| 
 | |
|     def __imod__(self, term):
 | |
|         modulus = int(term)
 | |
|         if modulus == 0:
 | |
|             raise ZeroDivisionError("Division by zero")
 | |
|         if modulus < 0:
 | |
|             raise ValueError("Modulus must be positive")
 | |
|         self._value %= modulus
 | |
|         return self
 | |
| 
 | |
|     # Boolean/bit operations
 | |
|     def __and__(self, term):
 | |
|         return self.__class__(self._value & int(term))
 | |
| 
 | |
|     def __or__(self, term):
 | |
|         return self.__class__(self._value | int(term))
 | |
| 
 | |
|     def __rshift__(self, pos):
 | |
|         try:
 | |
|             return self.__class__(self._value >> int(pos))
 | |
|         except OverflowError:
 | |
|             if self._value >= 0:
 | |
|                 return 0
 | |
|             else:
 | |
|                 return -1
 | |
| 
 | |
|     def __irshift__(self, pos):
 | |
|         try:
 | |
|             self._value >>= int(pos)
 | |
|         except OverflowError:
 | |
|             if self._value >= 0:
 | |
|                 return 0
 | |
|             else:
 | |
|                 return -1
 | |
|         return self
 | |
| 
 | |
|     def __lshift__(self, pos):
 | |
|         try:
 | |
|             return self.__class__(self._value << int(pos))
 | |
|         except OverflowError:
 | |
|             raise ValueError("Incorrect shift count")
 | |
| 
 | |
|     def __ilshift__(self, pos):
 | |
|         try:
 | |
|             self._value <<= int(pos)
 | |
|         except OverflowError:
 | |
|             raise ValueError("Incorrect shift count")
 | |
|         return self
 | |
| 
 | |
|     def get_bit(self, n):
 | |
|         if self._value < 0:
 | |
|             raise ValueError("no bit representation for negative values")
 | |
|         try:
 | |
|             try:
 | |
|                 result = (self._value >> n._value) & 1
 | |
|                 if n._value < 0:
 | |
|                     raise ValueError("negative bit count")
 | |
|             except AttributeError:
 | |
|                 result = (self._value >> n) & 1
 | |
|                 if n < 0:
 | |
|                     raise ValueError("negative bit count")
 | |
|         except OverflowError:
 | |
|             result = 0
 | |
|         return result
 | |
| 
 | |
|     # Extra
 | |
|     def is_odd(self):
 | |
|         return (self._value & 1) == 1
 | |
| 
 | |
|     def is_even(self):
 | |
|         return (self._value & 1) == 0
 | |
| 
 | |
|     def size_in_bits(self):
 | |
| 
 | |
|         if self._value < 0:
 | |
|             raise ValueError("Conversion only valid for non-negative numbers")
 | |
| 
 | |
|         if self._value == 0:
 | |
|             return 1
 | |
| 
 | |
|         bit_size = 0
 | |
|         tmp = self._value
 | |
|         while tmp:
 | |
|             tmp >>= 1
 | |
|             bit_size += 1
 | |
| 
 | |
|         return bit_size
 | |
| 
 | |
|     def size_in_bytes(self):
 | |
|         return (self.size_in_bits() - 1) // 8 + 1
 | |
| 
 | |
|     def is_perfect_square(self):
 | |
|         if self._value < 0:
 | |
|             return False
 | |
|         if self._value in (0, 1):
 | |
|             return True
 | |
| 
 | |
|         x = self._value // 2
 | |
|         square_x = x ** 2
 | |
| 
 | |
|         while square_x > self._value:
 | |
|             x = (square_x + self._value) // (2 * x)
 | |
|             square_x = x ** 2
 | |
| 
 | |
|         return self._value == x ** 2
 | |
| 
 | |
|     def fail_if_divisible_by(self, small_prime):
 | |
|         if (self._value % int(small_prime)) == 0:
 | |
|             raise ValueError("Value is composite")
 | |
| 
 | |
|     def multiply_accumulate(self, a, b):
 | |
|         self._value += int(a) * int(b)
 | |
|         return self
 | |
| 
 | |
|     def set(self, source):
 | |
|         self._value = int(source)
 | |
| 
 | |
|     def inplace_inverse(self, modulus):
 | |
|         modulus = int(modulus)
 | |
|         if modulus == 0:
 | |
|             raise ZeroDivisionError("Modulus cannot be zero")
 | |
|         if modulus < 0:
 | |
|             raise ValueError("Modulus cannot be negative")
 | |
|         r_p, r_n = self._value, modulus
 | |
|         s_p, s_n = 1, 0
 | |
|         while r_n > 0:
 | |
|             q = r_p // r_n
 | |
|             r_p, r_n = r_n, r_p - q * r_n
 | |
|             s_p, s_n = s_n, s_p - q * s_n
 | |
|         if r_p != 1:
 | |
|             raise ValueError("No inverse value can be computed" + str(r_p))
 | |
|         while s_p < 0:
 | |
|             s_p += modulus
 | |
|         self._value = s_p
 | |
|         return self
 | |
| 
 | |
|     def inverse(self, modulus):
 | |
|         result = self.__class__(self)
 | |
|         result.inplace_inverse(modulus)
 | |
|         return result
 | |
| 
 | |
|     def gcd(self, term):
 | |
|         r_p, r_n = abs(self._value), abs(int(term))
 | |
|         while r_n > 0:
 | |
|             q = r_p // r_n
 | |
|             r_p, r_n = r_n, r_p - q * r_n
 | |
|         return self.__class__(r_p)
 | |
| 
 | |
|     def lcm(self, term):
 | |
|         term = int(term)
 | |
|         if self._value == 0 or term == 0:
 | |
|             return self.__class__(0)
 | |
|         return self.__class__(abs((self._value * term) // self.gcd(term)._value))
 | |
| 
 | |
|     @staticmethod
 | |
|     def jacobi_symbol(a, n):
 | |
|         a = int(a)
 | |
|         n = int(n)
 | |
| 
 | |
|         if n <= 0:
 | |
|             raise ValueError("n must be a positive integer")
 | |
| 
 | |
|         if (n & 1) == 0:
 | |
|             raise ValueError("n must be odd for the Jacobi symbol")
 | |
| 
 | |
|         # Step 1
 | |
|         a = a % n
 | |
|         # Step 2
 | |
|         if a == 1 or n == 1:
 | |
|             return 1
 | |
|         # Step 3
 | |
|         if a == 0:
 | |
|             return 0
 | |
|         # Step 4
 | |
|         e = 0
 | |
|         a1 = a
 | |
|         while (a1 & 1) == 0:
 | |
|             a1 >>= 1
 | |
|             e += 1
 | |
|         # Step 5
 | |
|         if (e & 1) == 0:
 | |
|             s = 1
 | |
|         elif n % 8 in (1, 7):
 | |
|             s = 1
 | |
|         else:
 | |
|             s = -1
 | |
|         # Step 6
 | |
|         if n % 4 == 3 and a1 % 4 == 3:
 | |
|             s = -s
 | |
|         # Step 7
 | |
|         n1 = n % a1
 | |
|         # Step 8
 | |
|         return s * IntegerNative.jacobi_symbol(n1, a1)
 | 
