python - 如何使用 Python 将霍夫曼编码写入文件？

Question

我创建了一个 Python 脚本来使用 Huffman 算法压缩文本。假设我有以下字符串:

string = 'The quick brown fox jumps over the lazy dog'

运行我的算法会返回以下“位”:

result = '01111100111010101111010011111010000000011000111000010111110111110010100110010011010100101111100011110001000110101100111101000010101101110110111000111010101110010111111110011000101101000110111000'

通过将结果的位数与输入字符串进行比较，该算法似乎有效:

>>> print len(result), len(string) * 8
194 344

但现在问题来了:如何将其写入文件，同时仍然能够对其进行解码。您只能按字节写入文件，而不能按位写入。通过将“代码”写为字节，根本没有压缩!

我是计算机科学的新手，在线资源并不适合我。非常感谢所有帮助!

编辑:请注意，我的代码是这样的(如果是另一个输入字符串 'xxxxxxxyzz' ):

{'y': '00', 'x': '1', 'z': '10'}

我创建结果字符串的方法是按照输入字符串的顺序连接这些代码:

result = '1111111001010'

如何从这个结果返回到原始字符串？还是我完全错了？谢谢!

Answer 1

首先，您需要将输入字符串转换为字节:

def _to_Bytes(data):
  b = bytearray()
  for i in range(0, len(data), 8):
    b.append(int(data[i:i+8], 2))
  return bytes(b)

然后，打开一个文件以二进制模式写入:

result = '01111100111010101111010011111010000000011000111000010111110111110010100110010011010100101111100011110001000110101100111101000010101101110110111000111010101110010111111110011000101101000110111000'
with open('test.bin', 'wb') as f:
  f.write(_to_Bytes(result))

现在，将原始字符串写入文件，可以进行字节比较:

import os
with open('test_compare.txt', 'a') as f:
  f.write('The quick brown fox jumps over the lazy dog')

_o = os.path.getsize('test_compare.txt')
_c = os.path.getsize('test.bin')
print(f'Original file: {_o} bytes')
print(f'Compressed file: {_c} bytes')
print('Compressed file to about {}% of original'.format(round((((_o-_c)/_o)*100), 0)))

输出:

Original file: 43 bytes
Compressed file: 25 bytes
Compressed file to about 42.0% of original

要回到原来的状态，您可以编写一个函数来确定可能的字符顺序:

d = {'y': '00', 'x': '1', 'z': '10'}
result = '1111111001010'
from typing import Generator
def reverse_encoding(content:str, _lookup) -> Generator[str, None, None]:
  while content:
    _options = [i for i in _lookup if content.startswith(i) and (any(content[len(i):].startswith(b) for b in _lookup) or not content[len(i):])]
    if not _options:
      raise Exception("Decoding error")
    yield _lookup[_options[0]]
    content = content[len(_options[0]):]

print(''.join(reverse_encoding(result, {b:a for a, b in d.items()})))

输出:

'xxxxxxxyzz'