From 016089aac16f76f213c7b24e17c9c080bbf58ace Mon Sep 17 00:00:00 2001
From: Christian Raymond <christian.raymond@irisa.fr>
Date: Fri, 10 Jun 2022 13:29:43 +0200
Subject: [PATCH] better alignment with almost similar objects
---
README.md | 2 ++
SequenceAligner.hpp | 34 +++++++++++++++++++++++++++++++++-
example.cpp | 3 ++-
3 files changed, 37 insertions(+), 2 deletions(-)
diff --git a/README.md b/README.md
index 1b82d63..0eaa119 100644
--- a/README.md
+++ b/README.md
@@ -96,6 +96,7 @@ The C++ version is template and can align any Objects with the folwing propertie
3) a `size()` method that return the number of character used to print the object
4) an `operator==()` to compare objects
5) the `operator<<` in order to print the object in `size()` character
+6) the `operator-` to privilegiate substitution over insertion/deletion when different object share similarity: return the percentage of similarity
### Exemple
@@ -109,6 +110,7 @@ class Integer
int size() const {return std::to_string(_i).length();} //size() should provide the nbsymbols to write object on screen
bool empty() const {return _i==std::numeric_limits<int>::max();}//empty() should say if the object is not empty (not constructed by the constructor without arg)
bool operator==(const Integer& g) const {return _i==g._i;} //must provide a comparator
+ int operator-(const Integer& g) const {return std::min(_i,g._i)*100/std::max(_i,g._i);} //to privilegiate substitution error instead of insertion/deletion when operator== say false but objects share similarity (return PERCENTAGE of similarity so return always 0; to ignore this function)
friend std::ostream& operator<<(std::ostream& o,const Integer& i) {return o<<i._i;}
};
```
diff --git a/SequenceAligner.hpp b/SequenceAligner.hpp
index 507c0dd..44e824f 100644
--- a/SequenceAligner.hpp
+++ b/SequenceAligner.hpp
@@ -12,6 +12,8 @@
#include "asciitablestream.hpp"
+
+
template<class T=std::string>
class SequenceAligner
{
@@ -316,7 +318,9 @@ class SequenceAligner
else
{
subOp = SUB;
- cs = cost[i-1][j-1] + substitutionPenalty;
+ mint discountCost=abs(reference[i-1]-hypothesis[j-1]);
+ if(discountCost>0) discountCost = substitutionPenalty*discountCost/100;
+ cs = cost[i-1][j-1] + substitutionPenalty - discountCost ;
}
cmint ci = cost[i][j-1] + insertionPenalty;
cmint cd = cost[i-1][j] + deletionPenalty;
@@ -401,3 +405,31 @@ class SequenceAligner
}
};
+//utility for string
+
+class Char
+{
+ char _c;
+ public:
+ Char(): _c(0) {} //Default constructor should mark the object as empty in some way
+ Char(const char c): _c(c){}
+ int size() const {return 1;} //size() should provide the nbsymbols to write object on screen
+ bool empty() const {return _c==0;}//empty() should say if the object is not empty (not constructed by the constructor without arg)
+ bool operator==(const Char& g) const {return _c==g._c;} //must provide a comparator
+ int operator-(const Char& g) const {return 0;} //to privilegiate substitution error instead of insertion/deletion when operator== say false but objects share similarity (return 0; to ignore this function)
+ friend std::ostream& operator<<(std::ostream& o,const Char& i)
+ {
+ if(i._c==0) return o<<"";
+ return o<<i._c;
+ }
+};
+
+
+int operator-(const std::string& one,const std::string& two)
+{
+ SequenceAligner<Char> wordAligner(false);
+ std::vector<Char> un{one.begin(),one.end()};
+ std::vector<Char> deux{two.begin(),two.end()};
+ const auto& res=wordAligner.align(un ,deux);
+ return res.f1();
+}
diff --git a/example.cpp b/example.cpp
index 2ce041d..f92aa63 100644
--- a/example.cpp
+++ b/example.cpp
@@ -15,6 +15,7 @@ class Integer
int size() const {return std::to_string(_i).length();} //size() should provide the nbsymbols to write object on screen
bool empty() const {return _i==std::numeric_limits<int>::max();}//empty() should say if the object is not empty (not constructed by the constructor without arg)
bool operator==(const Integer& g) const {return _i==g._i;} //must provide a comparator
+ int operator-(const Integer& g) const {return std::min(_i,g._i)*100/std::max(_i,g._i);} //to privilegiate substitution error instead of insertion/deletion when operator== say false but objects share similarity (return 0; to ignore this function), otherwise return the percentage of similarity
friend std::ostream& operator<<(std::ostream& o,const Integer& i)
{
if(i._i==std::numeric_limits<int>::max()) return o<<"";
@@ -71,7 +72,7 @@ int main(int argc,char*argv[0])
std::cout<<"\n\n"<<ss<<std::endl;
- std::vector<Integer> num1={1,2,3};
+ std::vector<Integer> num1={400,2,3,300,308,306};
std::vector<Integer> num2={2,3,400};
SequenceAligner<Integer> numEval;
--
GitLab