318
|
1 #!/usr/bin/python
|
|
2
|
|
3 """ Parser generator utility """
|
|
4
|
|
5 import argparse
|
|
6 import re
|
|
7 import sys
|
|
8 import datetime
|
|
9 from pyyacc import Grammar, print_grammar
|
|
10
|
|
11
|
|
12 class XaccLexer:
|
|
13 def __init__(self):
|
|
14 pass
|
|
15
|
|
16 def feed(self, txt):
|
|
17 # Create a regular expression for the lexing part:
|
|
18 tok_spec = [
|
|
19 ('ID', r'[A-Za-z][A-Za-z\d_]*'),
|
|
20 ('STRING', r"'[^']*'"),
|
|
21 ('BRACEDCODE', r"\{[^\}]*\}"),
|
|
22 ('OTHER', r'[:;\|]'),
|
|
23 ('SKIP', r'[ ]')
|
|
24 ]
|
|
25 tok_re = '|'.join('(?P<%s>%s)' % pair for pair in tok_spec)
|
|
26 gettok = re.compile(tok_re).match
|
|
27
|
|
28 lines = txt.split('\n')
|
|
29
|
|
30 def tokenize_line(line):
|
|
31 """ Generator that splits up a line into tokens """
|
|
32 mo = gettok(line)
|
|
33 pos = 0
|
|
34 while mo:
|
|
35 typ = mo.lastgroup
|
|
36 val = mo.group(typ)
|
|
37 if typ == 'ID':
|
|
38 yield (typ, val)
|
|
39 elif typ == 'STRING':
|
|
40 typ = 'ID'
|
|
41 yield (typ, val[1:-1])
|
|
42 elif typ == 'OTHER':
|
|
43 typ = val
|
|
44 yield (typ, val)
|
|
45 elif typ == 'BRACEDCODE':
|
|
46 yield (typ, val)
|
|
47 elif typ == 'SKIP':
|
|
48 pass
|
|
49 else:
|
|
50 raise NotImplementedError(str(typ))
|
|
51 pos = mo.end()
|
|
52 mo = gettok(line, pos)
|
|
53 if len(line) != pos:
|
|
54 raise ParseError('Lex fault at {}'.format(line))
|
|
55
|
|
56 def tokenize():
|
|
57 section = 0
|
|
58 for line in lines:
|
|
59 line = line.strip()
|
|
60 if not line:
|
|
61 continue # Skip empty lines
|
|
62 if line == '%%':
|
|
63 section += 1
|
|
64 yield('%%', '%%')
|
|
65 continue
|
|
66 if section == 0:
|
|
67 if line.startswith('%tokens'):
|
|
68 yield('%tokens', '%tokens')
|
|
69 yield from tokenize_line(line[7:])
|
|
70 else:
|
|
71 yield ('HEADER', line)
|
|
72 elif section == 1:
|
|
73 yield from tokenize_line(line)
|
|
74 yield ('eof', 'eof')
|
|
75 self.tokens = tokenize()
|
|
76 self.token = self.tokens.__next__()
|
|
77
|
|
78 def next_token(self):
|
|
79 t = self.token
|
|
80 if t[0] != 'eof':
|
|
81 self.token = self.tokens.__next__()
|
|
82 #print(t)
|
|
83 return t
|
|
84
|
|
85
|
|
86 class ParseError(Exception):
|
|
87 pass
|
|
88
|
|
89
|
|
90 class XaccParser:
|
|
91 """ Implements a recursive descent parser to parse grammar rules.
|
|
92 We could have made an generated parser, but that would yield a chicken
|
|
93 egg issue.
|
|
94 """
|
|
95 def __init__(self, lexer):
|
|
96 self.lexer = lexer
|
|
97
|
|
98 @property
|
|
99 def Peak(self):
|
|
100 """ Sneak peak to the next token in line """
|
|
101 return self.lexer.token[0]
|
|
102
|
|
103 def next_token(self):
|
|
104 """ Take the next token """
|
|
105 return self.lexer.next_token()
|
|
106
|
|
107 def consume(self, typ):
|
|
108 """ Eat next token of type typ or raise an exception """
|
|
109 if self.Peak == typ:
|
|
110 return self.next_token()
|
|
111 else:
|
|
112 raise ParseError('Expected {}, but got {}'.format(typ, self.Peak))
|
|
113
|
|
114 def has_consumed(self, typ):
|
|
115 """ Consume typ if possible and return true if so """
|
|
116 if self.Peak == typ:
|
|
117 self.consume(typ)
|
|
118 return True
|
|
119 return False
|
|
120
|
|
121 def parse_grammar(self):
|
|
122 """ Entry parse function into recursive descent parser """
|
|
123 # parse header
|
|
124 headers = []
|
|
125 terminals = []
|
|
126 while self.Peak in ['HEADER', '%tokens']:
|
|
127 if self.Peak == '%tokens':
|
|
128 self.consume('%tokens')
|
|
129 while self.Peak == 'ID':
|
|
130 terminals.append(self.consume('ID')[1])
|
|
131 else:
|
|
132 headers.append(self.consume('HEADER')[1])
|
|
133 self.consume('%%')
|
|
134 self.grammar = Grammar(terminals)
|
|
135 while self.Peak != 'eof':
|
|
136 self.parse_rule()
|
|
137 return self.grammar
|
|
138
|
|
139 def parse_symbol(self):
|
|
140 return self.consume('ID')[1]
|
|
141
|
|
142 def parse_rhs(self):
|
|
143 symbols = []
|
|
144 while self.Peak not in [';', 'BRACEDCODE', '|']:
|
|
145 symbols.append(self.parse_symbol())
|
|
146 if self.Peak == 'BRACEDCODE':
|
|
147 action = self.consume('BRACEDCODE')[1]
|
|
148 action = action[1:-1].strip()
|
|
149 else:
|
|
150 action = None
|
|
151 return symbols, action
|
|
152
|
|
153 def parse_rule(self):
|
|
154 p = self.parse_symbol()
|
|
155 self.consume(':')
|
|
156 symbols, action = self.parse_rhs()
|
|
157 self.grammar.add_production(p, symbols, action)
|
|
158 while self.has_consumed('|'):
|
|
159 symbols, action = self.parse_rhs()
|
|
160 self.grammar.add_production(p, symbols, action)
|
|
161 self.consume(';')
|
|
162
|
|
163
|
|
164 class XaccGenerator:
|
|
165 """ Generator that writes generated parser to file """
|
|
166 def __init__(self):
|
|
167 pass
|
|
168
|
|
169 def generate(self, grammar, output_file):
|
|
170 print_grammar(grammar)
|
|
171 self.grammar = grammar
|
|
172 self.action_table, self.goto_table = grammar.doGenerate()
|
|
173 self.generate_python_script(output_file)
|
|
174
|
|
175 def generate_python_script(self, f):
|
|
176 """ Generate python script with the parser table """
|
|
177 print('#!/usr/bin/python', file=f)
|
|
178 stamp = datetime.datetime.now().ctime()
|
|
179 print('""" Automatically generated by xacc on {} """'.format(stamp), file=f)
|
|
180 print('from pyyacc import LRParser, Reduce, Shift, Accept, Production, Grammar', file=f)
|
|
181 print('from ppci import Token', file=f)
|
|
182 print(file=f)
|
|
183 print('class Parser(LRParser):', file=f)
|
|
184 print(' def __init__(self):', file=f)
|
|
185 # Generate rules:
|
|
186 print(' self.start_symbol = "{}"'.format(self.grammar.start_symbol), file=f)
|
|
187 print(' self.grammar = Grammar({})'.format(self.grammar.terminals), file=f)
|
|
188 for rule_number, rule in enumerate(self.grammar.productions):
|
|
189 rule.f_name = 'action_{}_{}'.format(rule.name, rule_number)
|
|
190 print(' self.grammar.add_production("{}", {}, self.{})'.format(rule.name, rule.symbols, rule.f_name), file=f)
|
|
191 # Fill action table:
|
|
192 print(' self.action_table = {}', file=f)
|
|
193 for state in self.action_table:
|
|
194 action = self.action_table[state]
|
|
195 print(' self.action_table[{}] = {}'.format(state, action), file=f)
|
|
196 print('', file=f)
|
|
197
|
|
198 # Fill goto table:
|
|
199 print(' self.goto_table = {}', file=f)
|
|
200 for gt in self.goto_table:
|
|
201 to = self.goto_table[gt]
|
|
202 print(' self.goto_table[{}] = {}'.format(gt, to), file=f)
|
|
203 print('', file=f)
|
|
204
|
|
205 # Generate a function for each action:
|
|
206 for rule in self.grammar.productions:
|
|
207 M = len(rule.symbols)
|
|
208 args = ', '.join('arg{}'.format(n + 1) for n in range(M))
|
|
209 print(' def {}(self, {}):'.format(rule.f_name, args), file=f)
|
|
210 if rule.f == None:
|
|
211 semantics = 'pass'
|
|
212 else:
|
|
213 semantics = str(rule.f)
|
|
214 if semantics.strip() == '':
|
|
215 semantics = 'pass'
|
|
216 for n in range(M):
|
|
217 semantics = semantics.replace('${}'.format(n + 1), 'arg{}'.format(n + 1))
|
|
218 print(' {}'.format(semantics), file=f)
|
|
219 print('', file=f)
|
|
220
|
|
221
|
|
222 def main():
|
|
223 # Parse arguments:
|
|
224 parser = argparse.ArgumentParser(description='xacc compiler compiler')
|
|
225 parser.add_argument('source', type=argparse.FileType('r'), \
|
|
226 help='the parser specification')
|
|
227 parser.add_argument('-o', '--output', type=argparse.FileType('w'), \
|
|
228 default=sys.stdout)
|
|
229 args = parser.parse_args()
|
|
230 src = args.source.read()
|
|
231 args.source.close()
|
|
232
|
|
233 # Construction of generator parts:
|
|
234 lexer = XaccLexer()
|
|
235 parser = XaccParser(lexer)
|
|
236 generator = XaccGenerator()
|
|
237
|
|
238 # Sequence source through the generator parts:
|
|
239 lexer.feed(src)
|
|
240 grammar = parser.parse_grammar()
|
|
241 generator.generate(grammar, args.output)
|
|
242 args.output.close()
|
|
243
|
|
244
|
|
245 if __name__ == '__main__':
|
|
246 main()
|