---

RetParamManager.hpp

Go to the documentation of this file.

/*==========================================================================
 * Copyright (c) 2001 Carnegie Mellon University.  All Rights Reserved.
 *
 * Use of the Lemur Toolkit for Language Modeling and Information Retrieval
 * is subject to the terms of the software license set forth in the LICENSE
 * file included with this software, and also available at
 * http://www.cs.cmu.edu/~lemur/license.html
 *
 *==========================================================================
*/

#ifndef _RETRIEVALPARAMETER_HPP
#define _RETRIEVALPARAMETER_HPP

#include "Param.hpp"

#include "TFIDFRetMethod.hpp"
#include "OkapiRetMethod.hpp"
#include "SimpleKLRetMethod.hpp"
#include "CORIRetMethod.hpp"
#include "CosSimRetMethod.hpp"
#include "InQueryRetMethod.hpp"

namespace RetrievalParameter {



  static String databaseIndex;
  static String retModel;
  static String textQuerySet;
  static String resultFile;
  static bool TRECresultFileFormat;
  static int fbDocCount;
  static int resultCount;
  static bool cacheDocReps;
  static bool useWorkingSet;
  static String workSetFile;


  static string getLower(char *parm, char *def) {
    string tmpString = ParamGetString(parm, def);
    // make it all lowercase
    for (int i = 0; i < tmpString.length(); i++)
      tmpString[i] = tolower(tmpString[i]);
    return tmpString;
  }

  static void get() {
    databaseIndex = ParamGetString("index","");
    retModel = getLower("retModel","kl");
    // backwards compatibility.
    if (retModel == "0") retModel = "tfidf";
    if (retModel == "1") retModel = "okapi";
    if (retModel == "2") retModel = "kl";
    if (retModel == "3") retModel = "inquery";
    if (retModel == "4") retModel = "cori_cs";
    if (retModel == "5") retModel = "cos";

    string tmp = getLower("cacheDocReps", "true");
    cacheDocReps = (tmp == "true" || tmp == "1");

    tmp = getLower("useWorkingSet", "false"); 
    useWorkingSet = (tmp == "true" || tmp == "1");
    // working set file name
    workSetFile = ParamGetString("workingSetFile",""); 
    
    textQuerySet = ParamGetString("textQuery","");
    resultFile = ParamGetString("resultFile","");

    tmp = getLower("resultFileFormat","trec");
    TRECresultFileFormat = (tmp == "trec" || tmp == "1");

    // default being no feedback
    fbDocCount = ParamGetInt("feedbackDocCount",0); 
    resultCount = ParamGetInt("resultCount", 1000); 
    
  }
};


namespace TFIDFParameter {


  static WeightParam docTFPrm;
  static WeightParam qryTFPrm;
  static FeedbackParam fbPrm;
  
  static void get()
  {
    string tfmethod = RetrievalParameter::getLower("doc.tfMethod", "bm25");
    if (tfmethod == "rawtf") docTFPrm.tf = RAWTF;
    else if (tfmethod == "logf") docTFPrm.tf = LOGTF;    
    else if (tfmethod == "bm25") docTFPrm.tf = BM25;

    docTFPrm.bm25K1 = ParamGetDouble("doc.bm25K1",defaultDocK1);
    docTFPrm.bm25B = ParamGetDouble("doc.bm25B",defaultDocB);

    tfmethod = RetrievalParameter::getLower("query.tfMethod", "bm25");
    if (tfmethod == "rawtf") qryTFPrm.tf = RAWTF;
    else if (tfmethod == "logf") qryTFPrm.tf = LOGTF;    
    else if (tfmethod == "bm25") qryTFPrm.tf = BM25;

    qryTFPrm.bm25K1 = ParamGetDouble("query.bm25K1",defaultQryK1);
    qryTFPrm.bm25B = defaultQryB;
    
    fbPrm.howManyTerms = ParamGetInt("feedbackTermCount",defaultHowManyTerms);
    fbPrm.posCoeff = ParamGetDouble("feedbackPosCoeff", defaultPosCoeff); 
  }
};

namespace OkapiParameter {


  static TFParam tfPrm;
  static FeedbackParam fbPrm;


  static void get()
  {
    tfPrm.k1 = ParamGetDouble("BM25K1",defaultK1);
    tfPrm.b =  ParamGetDouble("BM25B",defaultB);
    tfPrm.k3 = ParamGetDouble("BM25K3", defaultK3);
    fbPrm.expQTF = ParamGetDouble("BM25QTF", defaultExpQTF);
    fbPrm.howManyTerms = ParamGetInt("feedbackTermCount",defaultHowManyTerms);
    
  }
};

namespace SimpleKLParameter {

  static SimpleKLParameter::DocSmoothParam docPrm;
  static SimpleKLParameter::QueryModelParam qryPrm;
  static String smoothSupportFile;
    
  static void get()
  {
    smoothSupportFile = ParamGetString("smoothSupportFile", "");

    string tmpString = RetrievalParameter::getLower("adjustedScoreMethod", 
                                                    "negativekld");
    if (tmpString == "querylikelihood" || tmpString == "ql") {
      qryPrm.adjScoreMethod = SimpleKLParameter::QUERYLIKELIHOOD;
    } else if (tmpString == "crossentropy" ||tmpString == "ce") {
      qryPrm.adjScoreMethod = SimpleKLParameter::CROSSENTROPY;
    } else if (tmpString == "negativekld" || tmpString == "-d") {
      qryPrm.adjScoreMethod = SimpleKLParameter::NEGATIVEKLD;
    } else {
      cerr << "Unknown scoreMethod " << tmpString << ". Using NEGATIVEKLD" 
           << endl;
      qryPrm.adjScoreMethod = SimpleKLParameter::NEGATIVEKLD;
    }

    tmpString = RetrievalParameter::getLower("smoothMethod", 
                                             "dirichletprior");
    if (tmpString == "jelinikmercer" || tmpString == "jm" || tmpString == "0")
      docPrm.smthMethod = SimpleKLParameter::JELINEKMERCER;
    else if (tmpString == "dirichletprior" || tmpString == "dir" || 
             tmpString == "1")
      docPrm.smthMethod = SimpleKLParameter::DIRICHLETPRIOR;
    else if (tmpString == "absolutediscount" || tmpString == "ad" || 
             tmpString == "2")
      docPrm.smthMethod = SimpleKLParameter::ABSOLUTEDISCOUNT;
    else if (tmpString == "twostage" || tmpString == "2s" || tmpString == "3")
      docPrm.smthMethod = SimpleKLParameter::TWOSTAGE;
    else {
      cerr << "Unknown smoothMethod " << tmpString << ". Using DIRICHLET" 
           << endl;
      docPrm.smthMethod = SimpleKLParameter::defaultSmoothMethod;
    }
    

    tmpString = RetrievalParameter::getLower("smoothStrategy", "interpolate");
    if (tmpString == "interpolate" || tmpString == "int" || tmpString == "0")
      docPrm.smthStrategy= SimpleKLParameter::INTERPOLATE;
    else if (tmpString == "backoff" || tmpString == "bo" || tmpString == "1")
      docPrm.smthStrategy= SimpleKLParameter::BACKOFF;
    else {
      cerr << "Unknown smoothStrategy " << tmpString << ". Using INTERPOLATE" 
           << endl;
      docPrm.smthStrategy= SimpleKLParameter::defaultSmoothStrategy;
    }
    

    docPrm.ADDelta = ParamGetDouble("discountDelta",defaultADDelta);
    docPrm.JMLambda = ParamGetDouble("JelinekMercerLambda",defaultJMLambda);
    docPrm.DirPrior = ParamGetDouble("DirichletPrior",defaultDirPrior);
    
    tmpString = RetrievalParameter::getLower("queryUpdateMethod", "mixture");

    if (tmpString == "mixture" || tmpString == "mix" || tmpString == "0")
      qryPrm.fbMethod = SimpleKLParameter::MIXTURE;
    else if (tmpString == "divmin" || tmpString == "div" || tmpString == "1")
      qryPrm.fbMethod = SimpleKLParameter::DIVMIN;
    else if (tmpString == "markovchain" || tmpString == "mc" || 
             tmpString == "2")
      qryPrm.fbMethod = SimpleKLParameter::MARKOVCHAIN;
    else if (tmpString == "relevancemodel1" || tmpString == "rm1" || 
             tmpString == "3")
      qryPrm.fbMethod = SimpleKLParameter::RM1;
    else if (tmpString == "relevancemodel2" || tmpString == "rm2" || 
             tmpString == "4")
      qryPrm.fbMethod = SimpleKLParameter::RM1;
    else {
      cerr << "Unknown queryUpdateMethod " << tmpString 
           << ". Using MIXTURE" 
           << endl;
      qryPrm.fbMethod = SimpleKLParameter::MIXTURE;
    }
    

    qryPrm.fbCoeff = ParamGetDouble("feedbackCoefficient", defaultFBCoeff);
    qryPrm.fbPrTh = ParamGetDouble("feedbackProbThresh", defaultFBPrTh);
    qryPrm.fbPrSumTh = ParamGetDouble("feedbackProbSumThresh",
                                      defaultFBPrSumTh);
    qryPrm.fbTermCount = ParamGetInt("feedbackTermCount", defaultFBTermCount);
    qryPrm.fbMixtureNoise = ParamGetDouble("feedbackMixtureNoise",
                                           defaultFBMixNoise);
    qryPrm.emIterations = ParamGetInt("emIterations", defaultEMIterations);
                                              
  }
};

namespace CORIParameter {
  static String collectionCounts;
  static double cstffactor;
  static double cstfbaseline;
  static double doctffactor;
  static double doctfbaseline;
  static void get() {
    collectionCounts = ParamGetString("collCounts", "USE_INDEX_COUNTS");
    cstffactor = ParamGetDouble("CSCTF_factor", 150);
    cstfbaseline = ParamGetDouble("CSCTF_baseline", 50);
    doctffactor = ParamGetDouble("DOCCTF_factor", 1.5);
    doctfbaseline = ParamGetDouble("DOCCTF_baseline", 0.5);
  }
};

namespace CosSimParameter {


  static FeedbackParam fbPrm;
  static String L2NormFile;
  
  static void get()
  {
    fbPrm.howManyTerms = ParamGetInt("feedbackTermCount",defaultHowManyTerms);
    fbPrm.posCoeff = ParamGetDouble("feedbackPosCoeff", defaultPosCoeff); 
    L2NormFile = ParamGetString("L2File", defaultL2File);
  }
};

namespace InQueryParameter {
  static double fbCoeff = 0.5;
  static int fbTermCount = 50;
  static double defaultBelief = 0.4;
  static bool cacheIDF = false;
  static void get()
  {
    defaultBelief = ParamGetDouble("defaultBelief", defaultBelief);
    fbCoeff = ParamGetDouble("feedbackPosCoeff", fbCoeff);
    fbTermCount = ParamGetInt("feedbackTermCount", fbTermCount);
    string tmpString = RetrievalParameter::getLower("cacheIDF", "true");
    cacheIDF = (tmpString == "true" || tmpString == "1");
  }
};

#endif

---