Changeset 2263 for trunk/tools/CedmaImporter

Timestamp:

08/08/09 11:17:28 (15 years ago)

Author:

gkronber

Message:

Refactoring: removed duplicate methods. #719

Location:

trunk/tools/CedmaImporter

Files:

• Importer.cs (modified) (4 diffs)
• SymbolicExpressionImporter.cs (modified) (1 diff)

trunk/tools/CedmaImporter/Importer.cs

@@ -37 +37 @@
       using (StreamReader reader = File.OpenText(fileName)) {
         ReadResultsAndInputVariables(reader);
+        reader.ReadLine();
         ImportAllModels(dirName, reader, database);
       }
@@ -67 +68 @@
         string targetVariableName = modelData[TARGETVARIABLE_COLUMN].Trim();
         string algoName = modelData[ALGORITHM_COLUMN].Trim();
-        HeuristicLab.Core.IItem modelItem = ParseModel(dirName, modelData[FILENAME_COLUMN].Trim(), algoName);
-        HeuristicLab.Modeling.Database.SQLServerCompact.Variable targetVariable = new HeuristicLab.Modeling.Database.SQLServerCompact.Variable(targetVariableName);
-        Algorithm algorithm = new Algorithm(algoName);
-        Model model = new Model(targetVariable, algorithm);
-        model.TrainingSamplesStart = problem.TrainingSamplesStart;
-        model.TrainingSamplesEnd = problem.TrainingSamplesEnd;
-        model.ValidationSamplesStart = problem.ValidationSamplesStart;
-        model.ValidationSamplesEnd = problem.ValidationSamplesEnd;
-        model.TestSamplesStart = problem.TestSamplesStart;
-        model.TestSamplesEnd = problem.TestSamplesEnd;
+        try {
+          HeuristicLab.Core.IItem modelItem = ParseModel(dirName, modelData[FILENAME_COLUMN].Trim(), algoName);
+          HeuristicLab.Modeling.Database.SQLServerCompact.Variable targetVariable = new HeuristicLab.Modeling.Database.SQLServerCompact.Variable(targetVariableName);
+          Algorithm algorithm = new Algorithm(algoName);
+          Model model = new Model(targetVariable, algorithm);
+          model.TrainingSamplesStart = problem.TrainingSamplesStart;
+          model.TrainingSamplesEnd = problem.TrainingSamplesEnd;
+          model.ValidationSamplesStart = problem.ValidationSamplesStart;
+          model.ValidationSamplesEnd = problem.ValidationSamplesEnd;
+          model.TestSamplesStart = problem.TestSamplesStart;
+          model.TestSamplesEnd = problem.TestSamplesEnd;
 
-        IEnumerable<ModelResult> qualityModelResults = GetModelResults(model, modelData);
-        IEnumerable<InputVariableResult> inputVariableResults = GetInputVariableResults(model, modelData);
+          IEnumerable<ModelResult> qualityModelResults = GetModelResults(model, modelData);
+          IEnumerable<InputVariableResult> inputVariableResults = GetInputVariableResults(model, modelData);
 
-        // TODO
-        //database.Persist(model);
-        //foreach (ModelResult modelResult in qualityModelResults)
-        //  database.Persist(modelResult);
-        //foreach (InputVariableResult inputVariableResult in inputVariableResults)
-        //  database.Persist(inputVariableResult);
-
+          // TODO
+          //database.Persist(model);
+          //foreach (ModelResult modelResult in qualityModelResults)
+          //  database.Persist(modelResult);
+          //foreach (InputVariableResult inputVariableResult in inputVariableResults)
+          //  database.Persist(inputVariableResult);
+        }
+        catch (Exception ex) {
+        }
       }
     }
@@ -104 +108 @@
     }
 
-    Dictionary<string, IFunction> knownFunctions = new Dictionary<string, IFunction>() {
-    {"+", new Addition()},
-    {"and", new And()},
-    {"mean", new Average()},
-    {"cos", new Cosinus()},
-    {"/", new Division()},
-    {"equ", new Equal()},
-    {"exp", new Exponential()},
-    {">", new GreaterThan()},
-    {"if", new IfThenElse()},
-    {"<", new LessThan()},
-    {"log", new Logarithm()},
-    {"*", new Multiplication()},
-    {"not", new Not()},
-    {"or", new Or()},
-    {"expt", new Power()},
-    {"sign", new Signum()},
-    {"sin",new Sinus()},
-    {"sqrt", new Sqrt()},
-    {"-", new Subtraction()},
-    {"tan", new Tangens()},
-    {"xor", new Xor()}
-    };
-    Constant constant = new Constant();
-    HeuristicLab.GP.StructureIdentification.Variable variable = new HeuristicLab.GP.StructureIdentification.Variable();
-    Differential differential = new Differential();
-
     private HeuristicLab.Core.IItem ParseModel(string dirName, string modelFileName, string algoName) {
+      foreach (char c in Path.GetInvalidFileNameChars()) {
+        modelFileName = modelFileName.Replace(c, '_');
+      }
       if (algoName == "SupportVectorRegression") {
         HeuristicLab.Data.SVMModel model = new HeuristicLab.Data.SVMModel();
@@ -138 +118 @@
         return model;
       } else {
+        SymbolicExpressionImporter sexpImporter = new SymbolicExpressionImporter();
         GeneticProgrammingModel model = new GeneticProgrammingModel();
-        IEnumerable<Token> tokens = GetTokenStream(File.OpenText(Path.Combine(dirName, modelFileName) + ".gp.txt"));
-        model.FunctionTree = ParseSexp(new Queue<Token>(tokens));
+        model.FunctionTree = sexpImporter.Import(File.OpenText(Path.Combine(dirName, modelFileName) + ".gp.txt"));
         return model;
-      }
-    }
-
-    private IEnumerable<Token> GetTokenStream(StreamReader reader) {
-      return from line in GetLineStream(reader)
-             let strTokens = line.Split(new string[] { " ", "\t", Environment.NewLine }, StringSplitOptions.RemoveEmptyEntries).AsEnumerable()
-             from strToken in strTokens
-             let t = Token.Parse(strToken)
-             where t != null
-             select t;
-    }
-
-    private IEnumerable<string> GetLineStream(StreamReader reader) {
-      while (!reader.EndOfStream) yield return reader.ReadLine().Replace("(", " ( ").Replace(")", " ) ");
-      yield break;
-    }
-
-    private HeuristicLab.GP.Interfaces.IFunctionTree ParseSexp(Queue<Token> tokens) {
-      Expect(Token.LPAR, tokens);
-
-      if (tokens.Peek().Symbol == TokenSymbol.SYMB) {
-        if (tokens.Peek().StringValue.Equals("variable")) {
-          return ParseVariable(tokens);
-        } else if (tokens.Peek().StringValue.Equals("differential")) {
-          return ParseDifferential(tokens);
-        } else {
-          Token curToken = tokens.Dequeue();
-          IFunctionTree tree = CreateTree(curToken);
-          while (!tokens.Peek().Equals(Token.RPAR)) {
-            tree.AddSubTree(ParseSexp(tokens));
-          }
-          Expect(Token.RPAR, tokens);
-          return tree;
-        }
-      } else if (tokens.Peek().Symbol == TokenSymbol.NUMBER) {
-        ConstantFunctionTree t = (ConstantFunctionTree)constant.GetTreeNode();
-        t.Value = tokens.Dequeue().DoubleValue;
-        return t;
-      } else {
-        throw new FormatException("Expected function or constant symbol");
-      }
-    }
-
-    private IFunctionTree ParseDifferential(Queue<Token> tokens) {
-      Debug.Assert(tokens.Dequeue().StringValue == "differential");
-      VariableFunctionTree t = (VariableFunctionTree)differential.GetTreeNode();
-      t.Weight = tokens.Dequeue().DoubleValue;
-      t.VariableName = tokens.Dequeue().StringValue;
-      t.SampleOffset = (int)tokens.Dequeue().DoubleValue;
-      Expect(Token.RPAR, tokens);
-      return t;
-    }
-
-    private IFunctionTree ParseVariable(Queue<Token> tokens) {
-      Debug.Assert(tokens.Dequeue().StringValue == "variable");
-      VariableFunctionTree t = (VariableFunctionTree)variable.GetTreeNode();
-      t.Weight = tokens.Dequeue().DoubleValue;
-      t.VariableName = tokens.Dequeue().StringValue;
-      t.SampleOffset = (int)tokens.Dequeue().DoubleValue;
-      Expect(Token.RPAR, tokens);
-      return t;
-    }
-
-    private IFunctionTree CreateTree(Token token) {
-      if (token.Symbol != TokenSymbol.SYMB) throw new FormatException("Expected function symbol, but got: " + token.StringValue);
-      return knownFunctions[token.StringValue].GetTreeNode();
-    }
-
-    private void Expect(Token token, Queue<Token> tokens) {
-      Token cur = tokens.Dequeue();
-      if (!token.Equals(cur)) throw new FormatException("Expected: " + token.StringValue + ", but got found: " + cur.StringValue);
-    }
-
-    private enum TokenSymbol { LPAR, RPAR, SYMB, NUMBER };
-    private class Token {
-      public static readonly Token LPAR = Token.Parse("(");
-      public static readonly Token RPAR = Token.Parse(")");
-
-      public TokenSymbol Symbol { get; set; }
-      public string StringValue { get; set; }
-      public double DoubleValue { get; set; }
-      public Token() { }
-
-      public override bool Equals(object obj) {
-        Token other = (obj as Token);
-        if (other == null) return false;
-        if (other.Symbol != Symbol) return false;
-        return other.StringValue == this.StringValue;
-      }
-
-      public static Token Parse(string strToken) {
-        strToken = strToken.Trim();
-        Token t = new Token();
-        t.StringValue = strToken.Trim();
-        double temp;
-        if (strToken == "") {
-          t = null;
-        } else if (strToken == "(") {
-          t.Symbol = TokenSymbol.LPAR;
-        } else if (strToken == ")") {
-          t.Symbol = TokenSymbol.RPAR;
-        } else if (double.TryParse(strToken, out temp)) {
-          t.Symbol = TokenSymbol.NUMBER;
-          t.DoubleValue = double.Parse(strToken);
-        } else {
-          t.Symbol = TokenSymbol.SYMB;
-        }
-        return t;
       }
     }

trunk/tools/CedmaImporter/SymbolicExpressionImporter.cs

@@ -43 +43 @@
 
     public SymbolicExpressionImporter() {
+    }
+
+    internal IFunctionTree Import(StreamReader streamReader) {
+      return ParseSexp(new Queue<Token>(GetTokenStream(streamReader)));
     }