# -*- coding:utf-8 -*-
# 构造next数组函数
def gen_pnext(p):
i, k, m =
0, -
1, len(p)
pnext = [-
1]*m
# 初始数组元素全为-1
while i < m-
1:
# 生成下一个pnext元素值
if k == -
1
or p[i] == p[k]:
i, k = i+
1, k+
1
pnext[i] = k
# 设置pnext元素
else:
k = pnext[k]
# 退到更短相同前缀
return pnext
# KMP算法的改进
def gen_pnext_improve(p):
'''改进一点的版本，在下一跳字符和当前相同时，跳到下一跳的下一跳，跳的更远'''
i, k, m =
0, -
1, len(p)
pnext = [-
1] * m
while i < m-
1:
if k == -
1
or p[i] == p[k]:
i, k = i+
1, k+
1
if p[i] == p[k]:
pnext[i] = pnext[k]
else:
pnext[i] = k
else:
k = pnext[k]
return pnext
# 自己的神他喵麻烦构建next数组版本
def MINE_gen_pnext(p):
'''生成针对p中各个未知的i的下一跳位置表'''
pnext = [-
1,
0]
pnext=[]
count =
0
for j
in range(
0, len(p)+
1):
ps = p[
0:j]
if ps ==
'':
continue
for i
in range(len(ps),
1, -
1):
if ps[
0: i-
1] == ps[-i+
1:]:
pnext.append(len(ps[
0: i-
1]))
if count<len(pnext):
count +=
1
else:
pnext.append(
0)
count +=
1
for i
in range(len(pnext)-
1,
0, -
1):
pnext[i] = pnext[i-
1]
pnext[
0] = -
1
return pnext
# kmp匹配主函数
def matching_KMP(t, p):
'''KMP Main fuction, t for target, p for pattern'''
i, j =
0,
0
n, m = len(t), len(p)
pnext = gen_pnext(p)
while j < n
and i < m:
# i == m说明找到了匹配
if i == -
1
or t[j] == p[i]:
# 刚开始或者字符匹配时，考虑p中的下一字符
j, i = j+
1, i+
1
else:
# 字符不匹配，则考虑pnext对应的下一字符
i = pnext[i]
if i == m:
# 找到匹配，返回其下标
return j - i
return -
1
# 无匹配
print matching_KMP(
'ABCDEFG',
'EFG')