|
55
|
1 using Implab;
|
|
|
2 using System;
|
|
|
3 using System.Collections.Generic;
|
|
|
4 using System.Linq;
|
|
|
5 using System.Text;
|
|
|
6 using System.Threading.Tasks;
|
|
|
7
|
|
|
8 namespace Implab.Parsing {
|
|
|
9 /// <summary>
|
|
|
10 /// Базовый абстрактный класс. Грамматика, позволяет формулировать выражения над алфавитом типа <c>char</c>.
|
|
|
11 /// </summary>
|
|
|
12 /// <typeparam name="TGrammar"></typeparam>
|
|
|
13 public abstract class Grammar<TGrammar> where TGrammar: Grammar<TGrammar>, new() {
|
|
|
14 Alphabet m_alphabet = new Alphabet();
|
|
|
15 static TGrammar _instance;
|
|
|
16
|
|
|
17 public static TGrammar Instance{
|
|
|
18 get {
|
|
|
19 if (_instance == null)
|
|
|
20 _instance = new TGrammar();
|
|
|
21 return _instance;
|
|
|
22 }
|
|
|
23 }
|
|
|
24
|
|
|
25 public SymbolToken UnclassifiedToken() {
|
|
|
26 return new SymbolToken(Alphabet.UNCLASSIFIED);
|
|
|
27 }
|
|
|
28
|
|
|
29 public void DefineAlphabet(IEnumerable<char> alphabet) {
|
|
|
30 Safe.ArgumentNotNull(alphabet, "alphabet");
|
|
|
31
|
|
|
32 foreach (var ch in alphabet)
|
|
|
33 m_alphabet.DefineSymbol(ch);
|
|
|
34 }
|
|
|
35 public Token SymbolRangeToken(char start, char end) {
|
|
|
36 return SymbolToken(Enumerable.Range(start, end - start + 1).Select(x => (char)x));
|
|
|
37 }
|
|
|
38
|
|
|
39 public Token SymbolToken(char symbol) {
|
|
|
40 return Token.New(TranslateOrAdd(symbol));
|
|
|
41 }
|
|
|
42
|
|
|
43 public Token SymbolToken(IEnumerable<char> symbols) {
|
|
|
44 Safe.ArgumentNotNull(symbols, "symbols");
|
|
|
45
|
|
|
46 return Token.New(TranslateOrAdd(symbols).ToArray());
|
|
|
47 }
|
|
|
48
|
|
|
49 public Token SymbolSetToken(params char[] set) {
|
|
|
50 return SymbolToken(set);
|
|
|
51 }
|
|
|
52
|
|
|
53 int TranslateOrAdd(char ch) {
|
|
|
54 var t = m_alphabet.Translate(ch);
|
|
|
55 if (t == Alphabet.UNCLASSIFIED)
|
|
|
56 t = m_alphabet.DefineSymbol(ch);
|
|
|
57 return t;
|
|
|
58 }
|
|
|
59
|
|
|
60 IEnumerable<int> TranslateOrAdd(IEnumerable<char> symbols) {
|
|
|
61 return symbols.Distinct().Select(TranslateOrAdd);
|
|
|
62 }
|
|
|
63
|
|
|
64 int TranslateOrDie(char ch) {
|
|
|
65 var t = m_alphabet.Translate(ch);
|
|
|
66 if (t == Alphabet.UNCLASSIFIED)
|
|
|
67 throw new ApplicationException(String.Format("Symbol '{0}' is UNCLASSIFIED", ch));
|
|
|
68 return t;
|
|
|
69 }
|
|
|
70
|
|
|
71 IEnumerable<int> TranslateOrDie(IEnumerable<char> symbols) {
|
|
|
72 return symbols.Distinct().Select(TranslateOrDie);
|
|
|
73 }
|
|
|
74
|
|
|
75 public Token SymbolTokenExcept(IEnumerable<char> symbols) {
|
|
|
76 Safe.ArgumentNotNull(symbols, "symbols");
|
|
|
77
|
|
|
78 return Token.New( Enumerable.Range(0, m_alphabet.Count).Except(TranslateOrDie(symbols)).ToArray());
|
|
|
79 }
|
|
|
80
|
|
|
81 protected CDFADefinition BuildDFA(Token lang) {
|
|
|
82 Safe.ArgumentNotNull(lang, "lang");
|
|
|
83
|
|
|
84 var dfa = new CDFADefinition(m_alphabet);
|
|
|
85
|
|
|
86 var builder = new DFABuilder();
|
|
|
87
|
|
|
88 lang.Accept( builder );
|
|
|
89
|
|
|
90 builder.BuildDFA(dfa);
|
|
|
91 if (dfa.InitialStateIsFinal)
|
|
|
92 throw new ApplicationException("The specified language contains empty token");
|
|
|
93
|
|
|
94 return dfa.Optimize();
|
|
|
95 }
|
|
|
96
|
|
|
97
|
|
|
98
|
|
|
99 //protected abstract TGrammar CreateInstance();
|
|
|
100 }
|
|
|
101
|
|
|
102
|
|
|
103 }
|
