src/genealgorithms.cpp

   1 /***************************************************************************
   2  *   Copyright (C) 2006 by Michael Andreen                                 *
   3  *   andreen@student.chalmers.se                                           *
   4  *                                                                         *
   5  *   This program is free software; you can redistribute it and/or modify  *
   6  *   it under the terms of the GNU General Public License as published by  *
   7  *   the Free Software Foundation; either version 2 of the License, or     *
   8  *   (at your option) any later version.                                   *
   9  *                                                                         *
  10  *   This program is distributed in the hope that it will be useful,       *
  11  *   but WITHOUT ANY WARRANTY; without even the implied warranty of        *
  12  *   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the         *
  13  *   GNU General Public License for more details.                          *
  14  *                                                                         *
  15  *   You should have received a copy of the GNU General Public License     *
  16  *   along with this program; if not, write to the                         *
  17  *   Free Software Foundation, Inc.,                                       *
  18  *   51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA          *
  19  ***************************************************************************/
  20
  21 #include "genealgorithms.h"
  22 #include "geneorder.h"
  23
  24 #include <algorithm>
  25 #include <set>
  26 #include <cstdlib>
  27 #include <iostream>
  28 using namespace std;
  29
  30 std::pair<int,int> longestSequences(const GeneOrder& go){
  31         vector<vector<int> > v = robinsonSchensted(go);
  32         return pair<int,int>(v[0].size(),v.size());
  33 }
  34
  35 std::vector<std::vector<int> > robinsonSchensted(const GeneOrder& go){
  36         vector<vector<int> > v;
  37         for (GeneOrder::iterator i = go.begin(); i != go.end(); ++i){
  38                 int n = abs(*i);
  39                 bool added = false;
  40                 for (vector<vector<int> >::iterator vs = v.begin();
  41                                 vs != v.end(); ++vs){
  42                         vector<int>::iterator bigger = upper_bound(vs->begin(),vs->end(),n);
  43                         if ( bigger == vs->end()){
  44                                 vs->push_back(n);
  45                                 added = true;
  46                                 break;
  47                         }else{
  48                                 swap(n,*bigger);
  49                         }
  50                 }
  51                 if (!added){
  52                         v.push_back(vector<int>());
  53                         v.back().push_back(n);
  54                 }
  55         }
  56         return v;
  57 }
  58
  59 struct FindP{
  60         size_t p;
  61         FindP(size_t p) : p(p) {}
  62         bool operator()(Interval i){
  63                 return (i.first == p || i.second == p);
  64         }
  65 };
  66
  67
  68 std::vector<Interval> findIntervalsAtPoints(const vector<Interval>& intervals){
  69         vector<Interval> points;
  70         points.push_back(Interval(intervals.size(),intervals.size())); //Dummy interval to match point and index
  71         for (size_t p = 1; p <= intervals.size(); ++p){
  72                 size_t f = 0;
  73                 size_t s = 0;
  74                 bool found = false;
  75                 size_t n = 0;
  76                 for (vector<Interval>::const_iterator i = intervals.begin(); i != intervals.end(); ++i, ++n){
  77                         if (i->first == p){
  78                                 if (!found){
  79                                         f = n;
  80                                         found = true;
  81                                 }else{
  82                                         s = n;
  83                                         break;
  84                                 }
  85                         }
  86                         if (i->second == p){
  87                                 if (!found){
  88                                         f = n;
  89                                         found = true;
  90                                 }else{
  91                                         s = n;
  92                                         break;
  93                                 }
  94                         }
  95                 }
  96                 points.push_back(Interval(f,s));
  97         }
  98         return points;
  99
 100 }
 101
 102 int countCycles(const GeneOrder& go){
 103         int cycles = 0;
 104         set<size_t> marked;
 105         vector<Interval> intervals = findIntervals(go);
 106         vector<Interval> points = findIntervalsAtPoints(intervals);
 107         for (size_t p = 1; p < go.size(); ++p){
 108                 if (marked.find(p) != marked.end())
 109                         continue;
 110                 Interval i = intervals[points[p].first];
 111                 while (marked.find(p) == marked.end()){
 112                         marked.insert(p);
 113                         if (i == intervals[points[p].first])
 114                                 i = intervals[points[p].second];
 115                         else
 116                                 i = intervals[points[p].first];
 117
 118                         if (p == i.first)
 119                                 p = i.second;
 120                         else
 121                                 p = i.first;
 122                 }
 123                 ++cycles;
 124         }
 125         return cycles;
 126 }
 127
 128 std::vector<Component> findComponents(const GeneOrder& go){
 129         return vector<Component>();
 130 }
 131
 132 /**
 133  * TODO: Think of a better than O(n^2) implementation
 134  * Possibly move it to GeneOrder too
 135  */
 136 std::vector<Interval> findIntervals(const GeneOrder& go){
 137         vector<Interval> intervals;
 138         for (size_t i = 0; i < go.size() - 1; ++i){
 139                 size_t f = 0;
 140                 size_t s = 0;
 141                 bool found = false;
 142                 size_t n = 0;
 143                 for (GeneOrder::iterator g = go.begin(); g != go.end(); ++g, ++n){
 144                         if (static_cast<size_t>(abs(*g)) == i){
 145                                 f = n;
 146                                 if (*g >= 0)
 147                                         ++f;
 148                                 if (found)
 149                                         break;
 150                                 found = true;
 151                         }
 152                         if(static_cast<size_t>(abs(*g)) == i+1){
 153                                 s = n;
 154                                 if (*g < 0)
 155                                         ++s;
 156                                 if (found)
 157                                         break;
 158                                 found = true;
 159                         }
 160                 }
 161                 intervals.push_back(Interval(f,s));
 162         }
 163         return intervals;
 164 }
 165