1 import re
2
3
4 strs = "if sum >= 1000 then x : x - 1;#"+" "
5
6 types = {'begin':1,
7 'if':2,
8 'then':3,
9 'while':4,
10 'do':5,
11 'end':6,
12 'l(l|d)*':10,
13 'dd*':11,
14 '+':13,
15 '-':14,
16 '*':15,
17 '/':16,
18 ':':17,
19 ':=':18,
20 '<':20,
21 '<=':21,
22 '<>':22,
23 '>':23,
24 '>=':24,
25 '=':25,
26 ';':26,
27 '(':27,
28 ')':28,
29 '#':0
30 }
31
32 if __name__ == '__main__':
33 # strs = input('请输入程序代码：')+" " #补位
34
35 index = 0
36 while index < len(strs):
37 keyIndex = 0
38 for key in types.keys():
39 if index+len(key) < len(strs):
40 if strs[index:index+len(key)] == key and not re.match('^[=a-zA-Z0-9_-]$', strs[index+len(key)]):
41 if not(strs[index] == '=' and re.match('^[<>]$', strs[index-1])):
42 ss = strs[index:index+len(key)]
43 print((ss, types.get(ss)))
44 elif re.match('^[a-zA-Z0-9_]+', strs[index:]):
45 ss = re.match('^([a-zA-Z0-9_]+)', strs[index:]).group()
46 if not types.get(ss):
47 if re.match('[a-zA-Z]+', ss):
48 print((ss, '标识符'))
49 elif re.match('\d+', ss):
50 print((ss, '数字'))
51 else:
52 print((ss, '其他'))
53 index += len(ss)
54 keyIndex+=1
55 index+=1
