BIDI_ML_INFO_EXTRACTION/BiddingKG/dl/fingerprint/makeMD5.py

import math

def F(x,y,z):
    return (x&y) | (~x&z)

def G(x,y,z):
    return (x&z) | (y&~z)

def H(x,y,z):
    return x ^ y ^z

def I(x,y,z):
    return y ^ (x | ~z)

def get_s16(val):
    if val < 0x80000000:
        return val
    else:
        return (val - 0x100000000)

def XX(a,b,c,d,k,i,s,_func):
    a = a+ (_func(b,c,d)&0xFFFFFFFF)+k+get_s16(i)
    a = ((a&0xFFFFFFFF)<<s) | ((a&0xFFFFFFFF)>>(32-s))
    a += b
    return a

def text2hex(_str):
    list_hex = []
    for _hex in map(hex,_str.encode()):
        _i = int(_hex[2:],16)
        while(True):
            a = _i&0xFF
            list_hex.append(hex(a)[2:].rjust(2,"0"))
            _i = _i >>8
            if _i==0:
                break
    return "".join(list_hex)

def paddingBits(_str):
    padding_length = (512-(len(_str)*4+64)%512)//4
    _str += "8"+"0"*(padding_length-1)
    return _str

def paddingLength(_str,length):
    bit_length = length&0xFFFFFFFFFFFFFFFF
    padding_str = ""
    for i in range(8):
        padding_str += hex(bit_length&0xFF)[2:].rjust(2,"0")
        bit_length = bit_length >> 8
    _str += padding_str
    return _str

def byte2unsign(b):
    return b & 0x7F + 128 if b<0 else b

def hex_lowFirst(_str):
    _result = 0
    for _i in range(len(_str)//2):
        _result = _result | (int(_str[_i*2:(_i+1)*2],16)<<(8*_i))
    return _result

def trans(A,B,C,D,_str):
    list_s = [[7,12,17,22],[5,9,14,20],[4,11,16,23],[6,10,15,21]]
    list_k = [int(math.floor(abs(math.sin(i+1))*(2**32))) for i in range(64)]
    list_begin_step = [[0,1],[1,5],[5,3],[0,7]]
    list_hex = [hex_lowFirst(_str[i*8:(i+1)*8]) for i in range(16)]
    # list_hex.reverse()
    list_ABCD = [A,B,C,D]
    _A,_B,_C,_D = list_ABCD
    list_Func = [F,G,H,I]
    #四轮计算
    for i in range(4):
        _s = list_s[i]
        _k = list_k[i*16:(i+1)*16]
        _step_begin,_step = list_begin_step[i]
        _s_begin = -1
        _func = list_Func[i]
        ABCD_begin = 1
        for j in range(16):
            _s_begin = (_s_begin+1)%4
            ABCD_begin = (ABCD_begin-1)%4
            _begin = (_step_begin+j*_step)%16
            list_ABCD[ABCD_begin] = XX(list_ABCD[ABCD_begin],list_ABCD[(ABCD_begin+1)%4],list_ABCD[(ABCD_begin+2)%4],list_ABCD[(ABCD_begin+3)%4],list_hex[_begin],_k[j],_s[_s_begin],_func)
    list_ABCD[0] += _A
    list_ABCD[1] += _B
    list_ABCD[2] += _C
    list_ABCD[3] += _D
    return list_ABCD


def getMD5(_str):
    A = 0x67452301
    B = 0xefcdab89
    C = 0x98badcfe
    D = 0x10325476

    hex_str = text2hex(_str)

    _length = len(hex_str)*4
    hex_str = paddingBits(hex_str)
    hex_str = paddingLength(hex_str,_length)
    for i in range(len(hex_str)//128):
        _str = hex_str[i*128:(i+1)*128]
        A,B,C,D = trans(A,B,C,D,_str)
    mdf_str = ""
    for _i in [A,B,C,D]:
        list_j = []
        list_j.append(hex(_i&0xFF)[2:].rjust(2,'0'))
        list_j.append(hex((_i>>8)&0xFF)[2:].rjust(2,'0'))
        list_j.append(hex((_i>>16)&0xFF)[2:].rjust(2,'0'))
        list_j.append(hex((_i>>24)&0xFF)[2:].rjust(2,'0'))
        # _j = hex(_i)[2:].rjust(8,'0')
        # for h in range(len(_j)//2):
        #     list_j.append(_j[h*2:(h+1)*2])
        # list_j.reverse()
        mdf_str += "".join(list_j)
    return mdf_str


if __name__=="__main__":
    import codecs
    sourceHtml = codecs.open("C:\\Users\\User\\Desktop\\2.html","r",encoding="utf8").read()
    # sourceHtml = "abcddafafffffffffffffffffffffffff你"
    print(getMD5(sourceHtml))