Changeset 14351 for branches/symbreg-factors-2650/HeuristicLab.Problems.DataAnalysis.Symbolic/3.4/Importer/InfixExpressionParser.cs

Timestamp:

10/23/16 09:44:29 (7 years ago)

Author:

gkronber

Message:

#2650: merged r14332:14350 from trunk to branch

Location:

branches/symbreg-factors-2650

Files:

• . (modified) (1 prop)
• HeuristicLab.Problems.DataAnalysis.Symbolic (modified) (1 prop)
• HeuristicLab.Problems.DataAnalysis.Symbolic/3.4/Importer/InfixExpressionParser.cs (modified) (11 diffs)

branches/symbreg-factors-2650/HeuristicLab.Problems.DataAnalysis.Symbolic/3.4/Importer/InfixExpressionParser.cs

@@ -26 +26 @@
 using System.Text;
 using HeuristicLab.Collections;
+using HeuristicLab.Common;
 using HeuristicLab.Encodings.SymbolicExpressionTreeEncoding;
 
@@ -48 +49 @@
   /// </summary>
   public sealed class InfixExpressionParser {
-    private enum TokenType { Operator, Identifier, Number, LeftPar, RightPar, Eq, End, NA };
+    private enum TokenType { Operator, Identifier, Number, LeftPar, RightPar, Comma, Eq, End, NA };
     private class Token {
       internal double doubleVal;
@@ -114 +115 @@
         { "MEAN", new Average()},
         { "IF", new IfThenElse()},
-        { ">", new GreaterThan()},
-        { "<", new LessThan()},
+        { "GT", new GreaterThan()},
+        { "LT", new LessThan()},
         { "AND", new And()},
         { "OR", new Or()},
@@ -121 +122 @@
         { "XOR", new Xor()},
         { "DIFF", new Derivative()},
+        { "LAG", new LaggedVariable() },
       };
 
@@ -150 +152 @@
         }
         if (char.IsDigit(str[pos])) {
-          // read number (=> read until white space or operator)
+          // read number (=> read until white space or operator or comma)
           var sb = new StringBuilder();
           sb.Append(str[pos]);
@@ -159 +161 @@
             && str[pos] != '*'
             && str[pos] != '/'
-            && str[pos] != ')') {
+            && str[pos] != ')'
+            && str[pos] != ',') {
             sb.Append(str[pos]);
             pos++;
@@ -226 +229 @@
           pos++;
           yield return new Token { TokenType = TokenType.Eq, strVal = "=" };
+        } else if (str[pos] == ',') {
+          pos++;
+          yield return new Token { TokenType = TokenType.Comma, strVal = "," };
         } else {
           throw new ArgumentException("Invalid character: " + str[pos]);
@@ -232 +238 @@
     }
 
+    // S       = Expr EOF
+    // Expr    = ['-' | '+'] Term { '+' Term | '-' Term }
+    // Term    = Fact { '*' Fact | '/' Fact }
+    // Fact    = '(' Expr ')' | funcId '(' ArgList ')' | varId | number
+    // ArgList = Expr { ',' Expr }
     private ISymbolicExpressionTreeNode ParseS(Queue<Token> tokens) {
       var expr = ParseExpr(tokens);
@@ -339 +350 @@
     }
 
-    // Fact = '(' Expr ')' | funcId '(' Expr ')' | varId [ = valId ] | number
+    // Fact = '(' Expr ')' | 'LAG' '(' varId ',' ['+' | '-' ] number ')' | funcId '(' Expr ')' | varId [ = valId ] | number
     private ISymbolicExpressionTreeNode ParseFact(Queue<Token> tokens) {
       var next = tokens.Peek();
@@ -359 +370 @@
           if (lPar.TokenType != TokenType.LeftPar)
             throw new ArgumentException("expected (");
-          var expr = ParseExpr(tokens);
+
+          // handle 'lag' specifically
+          if (funcNode.Symbol is LaggedVariable) {
+            var varId = tokens.Dequeue();
+            if (varId.TokenType != TokenType.Identifier) throw new ArgumentException("Identifier expected. Format for lagged variables: \"lag(x, -1)\"");
+            var comma = tokens.Dequeue();
+            if (comma.TokenType != TokenType.Comma) throw new ArgumentException("',' expected, Format for lagged variables: \"lag(x, -1)\"");
+            double sign = 1.0;
+            if (tokens.Peek().strVal == "+" || tokens.Peek().strVal == "-") {
+              // read sign
+              var signTok = tokens.Dequeue();
+              if (signTok.strVal == "-") sign = -1.0;
+            }
+            var lagToken = tokens.Dequeue();
+            if (lagToken.TokenType != TokenType.Number) throw new ArgumentException("Number expected, Format for lagged variables: \"lag(x, -1)\"");
+            if (!lagToken.doubleVal.IsAlmost(Math.Round(lagToken.doubleVal)))
+              throw new ArgumentException("Time lags must be integer values");
+            var laggedVarNode = funcNode as LaggedVariableTreeNode;
+            laggedVarNode.VariableName = varId.strVal;
+            laggedVarNode.Lag = (int)Math.Round(sign * lagToken.doubleVal);
+            laggedVarNode.Weight = 1.0;
+          } else {
+            // functions
+            var args = ParseArgList(tokens);
+            // check number of arguments
+            if (funcNode.Symbol.MinimumArity > args.Length || funcNode.Symbol.MaximumArity < args.Length) {
+              throw new ArgumentException(string.Format("Symbol {0} requires between {1} and  {2} arguments.", funcId,
+                funcNode.Symbol.MinimumArity, funcNode.Symbol.MaximumArity));
+            }
+            foreach (var arg in args) funcNode.AddSubtree(arg);
+          }
+
           var rPar = tokens.Dequeue();
           if (rPar.TokenType != TokenType.RightPar)
             throw new ArgumentException("expected )");
 
-          funcNode.AddSubtree(expr);
           return funcNode;
         } else {
@@ -396 +437 @@
     }
 
+    // ArgList = Expr { ',' Expr }
+    private ISymbolicExpressionTreeNode[] ParseArgList(Queue<Token> tokens) {
+      var exprList = new List<ISymbolicExpressionTreeNode>();
+      exprList.Add(ParseExpr(tokens));
+      while (tokens.Peek().TokenType != TokenType.RightPar) {
+        var comma = tokens.Dequeue();
+        if (comma.TokenType != TokenType.Comma) throw new ArgumentException("expected ',' ");
+        exprList.Add(ParseExpr(tokens));
+      }
+      return exprList.ToArray();
+    }
   }
 }