source: cpp/frams/_demos/simil_test.cpp @ 1282

Last change on this file since 1282 was 1262, checked in by Maciej Komosinski, 18 months ago

Added missing #include for *nix compilers

File size: 7.3 KB
Line 
1// This file is a part of Framsticks SDK.  http://www.framsticks.com/
2// Copyright (C) 1999-2020  Maciej Komosinski and Szymon Ulatowski.
3// See LICENSE.txt for details.
4
5
6#include <vector>
7#include <string>
8#include "common/loggers/loggertostdout.h"
9#include "frams/_demos/genotypeloader.h"
10#include "frams/genetics/preconfigured.h"
11#include "common/virtfile/stdiofile.h"
12#include "frams/model/similarity/measure-distribution.h"
13#include "frams/model/similarity/measure-greedy.h"
14#include "frams/model/similarity/measure-hungarian.h"
15#include <stdexcept> //std::invalid_argument
16
17using namespace std;
18
19int add_double_param(std::vector<string> *args, int pos, std::vector<double> *params, std::vector<string> *params_names)
20{
21    for (unsigned int i = 0; i < params_names->size(); i++)
22    {
23        try
24        {
25            params->push_back(std::stod(args->at(pos)));
26            pos++;
27        }
28        catch (const std::invalid_argument&)
29        {
30            printf("%s should be a number\n", params_names->at(i).c_str());
31            return -1;
32        }
33        catch (const std::out_of_range&)
34        {
35            printf("%s should be inside double range\n", params_names->at(i).c_str());
36            return -1;
37        }
38    }
39    return 0;
40}
41
42/** Computes a matrix of distances between all genotypes in the specified
43        .gen file, using the matching and measure weights as specified in the
44        command line. */
45int main(int argc, char *argv[])
46{
47    typedef double *pDouble;
48    LoggerToStdout messages_to_stdout(LoggerBase::Enable);
49    SimilMeasureBase *simil_measure = nullptr;
50    if (argc < 5)
51    {
52        printf("Too few parameters!\n");
53        printf("Command line: [-names] <genotypesFile> <measure (greedy/hungarian)> <w_dP> <w_dDEG> <w_dNEU> <w_dGEO> <fixZaxis?>\n\n");
54        printf("Command line: [-names] <genotypesFile> <measure (distribution)> <desc> <simil> <dens> <bins> <samp_num>\n\n");
55        printf("Parameters:\n");
56        printf("  <genotypesFile> name of a file with genotypes\n");
57        printf("  <measure> similarity measure name (greedy/hungarian/distribution)\n");
58        printf("\n");
59        printf("Parameters of greedy and hungarian measures:\n");
60        printf("  <w_dP> weight of the difference in the number of parts\n");
61        printf("  <w_dDEG> weight of the difference in degrees of matched parts\n");
62        printf("  <w_dNEU> weight of the difference in neurons of matched parts\n");
63        printf("  <w_dGEO> weight of the distance of matched parts\n");
64        printf("  <fixZaxis?> should the 'z' (vertical) coordinate be fixed during the alignment? (0 or 1)\n\n");
65        printf("Parameters of distribution measure:\n");
66        printf("  <dens> sampling density\n");
67        printf("  <bins> number of histogram bins\n");
68        printf("  <samp_num> number of samples taken\n\n");
69
70        printf("Switches:\n");
71        printf("  -names specifies that the number and names of genotypes are to be printed to output\n");
72        printf("  before the distance matrix; by default the number and names are not printed\n\n");
73
74        printf("Outputs a symmetric distance matrix in the format:\n");
75        printf("  <row_1> (columns in a row are separated by TABs)\n");
76        printf("  ...\n");
77        printf("  <row_n>\n");
78
79        return -1;
80    }
81
82    std::vector<string> args;
83    for (int i = 1; i < argc; i++)
84        args.push_back(std::string(argv[i]));
85
86    bool print_names = false;
87
88    int pos = 1;
89    if (args.at(0).compare("-names")==0)
90    {
91        print_names = true;
92        pos = 2;
93    }
94
95    string measure_name = args.at(pos);
96    pos++;
97    std::vector<double> params;
98
99    if (measure_name.compare("greedy")==0 || measure_name.compare("hungarian")==0)
100    {
101        std::vector<string> params_names{ "<w_dP>", "<w_dDEG>", "<w_dNEU>", "<w_dGEO>",  "<fixZaxis?>" };
102
103        if (add_double_param(&args, pos, &params, &params_names) == -1)
104            return -1;
105
106        if (measure_name.compare("greedy")==0)
107            simil_measure = new SimilMeasureGreedy();
108        else
109            simil_measure = new SimilMeasureHungarian();
110    }
111
112
113    else if (measure_name.compare("distribution")==0)
114    {
115        std::vector<string> params_names{ "<dens>", "<bins>", "<samp_num>" };
116
117        if (add_double_param(&args, pos, &params, &params_names)==-1)
118            return -1;
119
120        simil_measure = new SimilMeasureDistribution();
121    }
122
123    else
124    {
125        printf("Measure type should be greedy (flexible criteria order and optimal matching), hungarian (vertex degree order and greedy matching) or distribution!\n");
126        return -1;
127    }
128
129    simil_measure->setParams(params);
130
131    // read the input file
132    // prepare loading of genotypes from a .gen file
133    // create some basic genotype converters
134    PreconfiguredGenetics genetics;
135    StdioFileSystem_autoselect stdiofilesys;
136
137    // prepare output parameters from .gen file
138    vector<Geno *> pvGenos;
139    vector<char *> pvNames;
140
141    long count = 0, totalsize = 0;
142    GenotypeMiniLoader loader(args.at(0).c_str());
143    GenotypeMini *loaded;
144    while (loaded = loader.loadNextGenotype())
145    {
146        // while a valid genotype was loaded
147        count++;
148        totalsize += loaded->genotype.length();
149        // create a Geno object based on the MiniGenotype
150        Geno *pNextGenotype = new Geno(loaded->genotype);
151        if ((pNextGenotype != NULL) && (pNextGenotype->isValid()))
152        {
153            pvGenos.push_back(pNextGenotype);
154            char *szNewName = new char[loaded->name.length() + 1];
155            strcpy(szNewName, loaded->name.c_str());
156            pvNames.push_back(szNewName);
157        }
158        else
159        {
160            printf("Genotype %2li is not valid\n", count);
161            if (pNextGenotype != NULL) delete pNextGenotype;
162        }
163    }
164    if (loader.getStatus() == GenotypeMiniLoader::OnError)
165    {
166        printf("Error: %s", loader.getError().c_str());
167    }
168
169    double dSimilarity = 0.0;
170    double **aaSimil = NULL; // array of similarities
171
172    // create the empty array of similarities
173    aaSimil = new pDouble[pvGenos.size()];
174    for (unsigned int k = 0; k < pvGenos.size(); k++)
175    {
176        aaSimil[k] = new double[pvGenos.size()];
177        for (unsigned int l = 0; l < pvGenos.size(); l++)
178            aaSimil[k][l] = 0.0;
179    }
180
181
182
183    // compute and remember similarities
184    for (unsigned int i = 0; i < pvGenos.size(); i++)
185    {
186        for (unsigned int j = 0; j < pvGenos.size(); j++)
187        {
188            dSimilarity = simil_measure->evaluateDistance(pvGenos.operator[](i), pvGenos.operator[](j));
189            aaSimil[i][j] = dSimilarity;
190        }
191    }
192
193    if (print_names)
194    {
195        // if "-names" switch was given, print the number of genotypes and their names
196        printf("%li\n", pvGenos.size());
197        for (unsigned int iGen = 0; iGen < pvNames.size(); iGen++)
198        {
199            printf("%s\n", pvNames.at(iGen));
200        }
201    }
202
203    // print out the matrix of similarities
204    for (unsigned int i = 0; i < pvGenos.size(); i++)
205    {
206        for (unsigned int j = 0; j < pvGenos.size(); j++)
207        {
208            printf("%.2lf\t", aaSimil[i][j]);
209        }
210        printf("\n");
211    }
212
213    // delete vectors and arrays
214    for (unsigned int i = 0; i < pvGenos.size(); i++)
215    {
216        delete pvGenos.operator[](i);
217        delete[] pvNames.operator[](i);
218        delete[] aaSimil[i];
219    }
220
221    delete[] aaSimil;
222    delete simil_measure;
223
224    return 0;
225}
Note: See TracBrowser for help on using the repository browser.