diff Implab/Formats/JSON/JSONGrammar.cs @ 163:419aa51b04fd ref20160224

JSON moved to Formats namespace Working in RegularDFA
author cin
date Wed, 24 Feb 2016 20:12:52 +0300
parents
children e227e78d72e4
line wrap: on
line diff
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/Implab/Formats/JSON/JSONGrammar.cs	Wed Feb 24 20:12:52 2016 +0300
@@ -0,0 +1,99 @@
+using System.Linq;
+using Implab.Automaton.RegularExpressions;
+
+namespace Implab.Formats.JSON {
+    class JSONGrammar : Grammar<JSONGrammar> {
+        public enum TokenType {
+            None,
+            BeginObject,
+            EndObject,
+            BeginArray,
+            EndArray,
+            String,
+            Number,
+            Literal,
+            NameSeparator,
+            ValueSeparator,
+
+            StringBound,
+            EscapedChar,
+            UnescapedChar,
+            EscapedUnicode,
+
+            Minus,
+            Plus,
+            Sign,
+            Integer,
+            Dot,
+            Exp
+        }
+
+        readonly CDFADefinition m_jsonDFA;
+        readonly CDFADefinition m_stringDFA;
+
+        public JSONGrammar() {
+            DefineAlphabet(Enumerable.Range(0, 0x20).Select(x => (char)x));
+            var hexDigit = SymbolRangeToken('a','f').Or(SymbolRangeToken('A','F')).Or(SymbolRangeToken('0','9'));
+            var digit9 = SymbolRangeToken('1', '9');
+            var zero = SymbolToken('0');
+            var digit = zero.Or(digit9);
+            var dot = SymbolToken('.');
+            var minus = SymbolToken('-');
+            var sign = SymbolSetToken('-', '+');
+            var expSign = SymbolSetToken('e', 'E');
+            var letters = SymbolRangeToken('a', 'z');
+            var integer = zero.Or(digit9.Cat(digit.EClosure()));
+            var frac = dot.Cat(digit.Closure());
+            var exp = expSign.Cat(sign.Optional()).Cat(digit.Closure());
+            var quote = SymbolToken('"');
+            var backSlash = SymbolToken('\\');
+            var specialEscapeChars = SymbolSetToken('\\', '"', '/', 'b', 'f', 't', 'n', 'r');
+            var unicodeEspace = SymbolToken('u').Cat(hexDigit.Repeat(4));
+            var whitespace = SymbolSetToken('\n', '\r', '\t', ' ').EClosure();
+            var beginObject = whitespace.Cat(SymbolToken('{')).Cat(whitespace);
+            var endObject = whitespace.Cat(SymbolToken('}')).Cat(whitespace);
+            var beginArray = whitespace.Cat(SymbolToken('[')).Cat(whitespace);
+            var endArray = whitespace.Cat(SymbolToken(']')).Cat(whitespace);
+            var nameSep = whitespace.Cat(SymbolToken(':')).Cat(whitespace);
+            var valueSep = whitespace.Cat(SymbolToken(',')).Cat(whitespace);
+            
+            var number = minus.Optional().Cat(integer).Cat(frac.Optional()).Cat(exp.Optional());
+            var literal = letters.Closure();
+            var unescaped = SymbolTokenExcept(Enumerable.Range(0, 0x20).Union(new int[] { '\\', '"' }).Select(x => (char)x));
+
+            var jsonExpression =
+                number.Tag(TokenType.Number)
+                .Or(literal.Tag(TokenType.Literal))
+                .Or(quote.Tag(TokenType.StringBound))
+                .Or(beginObject.Tag(TokenType.BeginObject))
+                .Or(endObject.Tag(TokenType.EndObject))
+                .Or(beginArray.Tag(TokenType.BeginArray))
+                .Or(endArray.Tag(TokenType.EndArray))
+                .Or(nameSep.Tag(TokenType.NameSeparator))
+                .Or(valueSep.Tag(TokenType.ValueSeparator));
+
+
+            var jsonStringExpression =
+                quote.Tag(TokenType.StringBound)
+                .Or(backSlash.Cat(specialEscapeChars).Tag(TokenType.EscapedChar))
+                .Or(backSlash.Cat(unicodeEspace).Tag(TokenType.EscapedUnicode))
+                .Or(unescaped.Closure().Tag(TokenType.UnescapedChar));
+                    
+
+            m_jsonDFA = BuildDFA(jsonExpression);
+            m_stringDFA = BuildDFA(jsonStringExpression);
+        }
+
+        public CDFADefinition JsonDFA {
+            get {
+                return m_jsonDFA;
+            }
+        }
+
+        public CDFADefinition JsonStringDFA {
+            get {
+                return m_stringDFA;
+            }
+        }
+    }
+}