SignatureCollection.h

   1 // Affixes and signatures in signature/suffix-based model of morphology
   2 // Copyright © 2009 The University of Chicago
   3 #ifndef SIGNATURECOLLECTION_H
   4 #define SIGNATURECOLLECTION_H
   5
   6 // See the CMiniLexicon class in MiniLexicon.h for an overview of
   7 // suffix/signature-based discovery of morphology.
   8
   9 #include "CollectionTemplate.h"
  10 #include "generaldefinitions.h"
  11 #include "AffixLocation.h"
  12 #include "Allomorphy.h"
  13
  14 template<class K, class V> class QMap;
  15 class QString;
  16
  17 class CSignatureCollection: public TCollection<class CSignature> {
  18         class CSuffixCollection* MySuffixes;
  19         class CPrefixCollection* MyPrefixes;
  20         /// prefix or suffix
  21         enum eAffixLocation m_SignatureType;
  22         double m_DLofPointersToMyMembers;
  23         double m_SumOfDLofPointersInternalToEachMember;
  24 public:
  25         // construction/destruction.
  26
  27         CSignatureCollection();
  28         CSignatureCollection(enum eAffixLocation prefix_or_suffix);
  29         CSignatureCollection(CMiniLexicon* mini, CSuffixCollection* suffixes,
  30                 enum eAffixLocation prefix_or_suffix);
  31         CSignatureCollection(CMiniLexicon* mini, CPrefixCollection* prefixes,
  32                 enum eAffixLocation prefix_or_suffix);
  33         /// deprecated
  34         CSignatureCollection(CMiniLexicon* mini);
  35         ~CSignatureCollection();
  36
  37         // disable copy
  38 private:
  39         CSignatureCollection(const CSignatureCollection& x);
  40         CSignatureCollection& operator=(const CSignatureCollection& x);
  41 public:
  42         // assign.
  43
  44         void Empty();
  45         void RemoveAll();
  46
  47         // insert.
  48
  49         CSignature* operator<<(CSignature* sig);
  50         CSignature* operator<<(CParse* affixes);
  51         void AddPointer(CSignature* sig);
  52         CSignature* AddToCollection(CParse& affixes);
  53         CSignature* AddToCollection(CStringSurrogate& dot_delimited);
  54
  55         // find.
  56
  57         CSignature* operator^=(CParse& affixes);
  58         CSignature* operator^=(CParse* affixes);
  59         /// str should be a .-delimited list of affixes
  60         /// result is 0 if no such signature exists
  61         CSignature* operator^=(CStringSurrogate& str);
  62         /// *this ^= CStringSurrogate(str)
  63         CSignature* operator^=(QString str);
  64
  65         // remove.
  66
  67         /// doesn't delete sig
  68         bool Remove(CSignature* sig);
  69         /// deletes sig
  70         bool RemoveMember(CSignature* sig);
  71         /// deletes *this ^= dot_delimited
  72         bool RemoveMember(CStringSurrogate& dot_delimited);
  73         bool RemoveMember(CStringSurrogate& dot_delimited, bool delete_it);
  74         /// remove each item from m_DeletionArray
  75         void DeleteMarkedMembers();
  76
  77         // union of contained signatures.
  78
  79         void SetMyPrefixes(CPrefixCollection* prefixes);
  80         CPrefixCollection* GetMyPrefixes() { return MyPrefixes; }
  81         void SetMySuffixes(CSuffixCollection* suffixes);
  82         CSuffixCollection* GetMySuffixes() { return MySuffixes; }
  83
  84         // affix location (prefix or suffix).
  85
  86         eAffixLocation GetSignatureType() { return m_SignatureType; }
  87
  88         // Compare pairs of sigs, and identify pairs where one is exactly a suffix of the other
  89
  90         void CompareSignaturePairsForTotalOverlap();
  91         void RecutLongerSigToMatchTheShorter(CSignatureAlignment* pSigAlignment);
  92
  93         // allomorphy.
  94
  95         void FindAllomorphy();
  96         void PutAffixesOfRegularSignaturesIntoNewSuffixes(
  97                 CSuffixCollection* NewSuffixes,
  98                 int MinimumNumberOfStems);
  99         void PutAffixesOfRegularSignaturesIntoNewPrefixes(
 100                 CPrefixCollection* NewPrefixes);
 101
 102         // output to GUI.
 103
 104         void BorrowedSigsDisplay(class Q3ListView* widget,
 105                 QMap<QString, QString>* filter);
 106         void ListDisplay(class Q3ListView* widget,
 107                 QMap<QString, QString>* filter = 0);
 108
 109         // serialization and deserialization (file I/O).
 110
 111         void OutputXfst (QString);
 112         void OutputSignatures(QString filename);
 113         void ReadSignatureFile(QString filename,
 114                 enum eAffixLocation prefix_or_suffix);
 115         void ReadSignatureFileBis(QString filename);
 116
 117         /// scan all sigs, and if the stems of pSig do not begin
 118         /// with DiffLetter, then Diff
 119         void TryToRemove(const QString Suffix, const QString DiffLetter);
 120         double GetMDL();
 121         void CalculateFrequencies();
 122         /// For each signature, determines if a signature
 123         /// with more robustness is contained in it; if so, takes the
 124         /// larger robustness as its own. This measures our certain of
 125         /// the STEMS as such for the signature.
 126         void CheckRobustness();
 127         void FindDisplayOrdering();
 128         void CleanUp();
 129         /// sum for each signature the number of stems * number of affixes
 130         int GetTotalNumberOfWords();
 131         int TheseTwoSuffixesShareHowManyStems(
 132                 class CSuffix* suf1, class CSuffix* suf2);
 133         double ComputeDLofInternalPointersOfEachMember(enum eMDL_STYLE style);
 134         double ComputeLengthOfPointersToEachOfMyMembers(enum eMDL_STYLE style);
 135         void GetIndividualCountsForEachStem();
 136 };
 137
 138 #endif // SIGNATURECOLLECTION_H