float language_model::perplexity_per_word(const sentence& tokens) const
{
if (tokens.size() == 0)
throw language_model_exception{
"perplexity_per_word() called on empty sentence"};
return perplexity(tokens) / tokens.size();
}
/* This function1 searches for a string in a sentence and return true/false*/
bool search_string(string str, sentence sen, int& j) {
j=0;
while (j!=sen.size()) {
if (!sen[j].compare(str)) {
break;
}
j++;
}
if (j==sen.size()) return false; else return true;
}
void do_add(vector<fd>& x, sentence line) {
fd temp;
int length=line.size();
temp.property=line[length-1];
temp.filename=line[length-2];
vector <int> index;
int minlength=10000000;
for(int i=0;i!=length-3;i++)
index.push_back(-1);
for(int j=1;j!=length-2;j++)
for(int i=0;i!=x.size();i++)
if ((!x[i].property.compare(line[j])) && (!x[i].filename.compare(line[length-2])))
index[j-1]=i;
for(int i=0;i!=length-3;i++) {
if (index[i]>-1)
if (minlength>x[index[i]].value.size())
minlength=x[index[i]].value.size();
}
for (int j=0;j!=minlength;j++) {
float value=0;
for (int i=0;i!=length-2;i++)
if (index[i]>0)
value+=x[index[i]].value[j];
temp.value.push_back(value);
}
x.push_back(temp);
}
vector<fd> convert_to_float(sentence s, map<string, vector<int> > x, sentence sf) {
vector<fd> datatemp;
int style,set;
if (!s[0].compare("s")) style=1; else style=2;
for(int i=1;i!=s.size();i++) {
fd temp;
switch (style) {
case 1: { // This is series case
set=0;
for(int j=0;j!=f.size(); j++) {
if (x[s[i]][j]) {
if (!set) {
temp.filename="Allfiles";
temp.property=s[i];
set=1;
}
for(int tt=1;tt!=f[j].size();tt++) {
float t=atof(&f[j][tt][x[s[i]][j]][0]);
temp.value.push_back(t);
}
}
// datatemp.push_back(temp);
}
datatemp.push_back(temp);
break;
}
case 2: { // This is parallel case.
for(int j=0;j!=f.size(); j++) {
if (x[s[i]][j]) {
temp.filename=sf[j];
temp.property=s[i];
temp.value.clear();
for(int tt=1;tt!=f[j].size();tt++) {
float t= 0.0;
if ((f[j][tt][x[s[i]][j]].compare("nan")))
t = atof(&f[j][tt][x[s[i]][j]][0]);
temp.value.push_back(t);
}
datatemp.push_back(temp);
}
else if (!s[i].compare("Step")) {
temp.filename=sf[j];
temp.property=s[i];
temp.value.clear();
for(int tt=1;tt!=f[j].size();tt++) {
float t=atof(&f[j][tt][x[s[i]][j]][0]);
temp.value.push_back(t);
}
datatemp.push_back(temp);
}
}
break;
}
}
}
return datatemp;
}
/*This function replaces certain input keywords to their appropriate words found in lammps file*/
void transform_userinput(sentence& temp) {
if (temp.size()) {
for(int i=1;i!=temp.size();i++) {
if (!temp[i].compare("t")) temp[i].replace(0,temp[i].length(),"Temp");
if (!temp[i].compare("p")) temp[i].replace(0,temp[i].length(),"Press");
if (!temp[i].compare("v")) temp[i].replace(0,temp[i].length(),"Volume");
if (!temp[i].compare("vd")) temp[i].replace(0,temp[i].length(),"E_vdwl");
if (!temp[i].compare("te")) temp[i].replace(0,temp[i].length(),"TotEng");
if (!temp[i].compare("ke")) temp[i].replace(0,temp[i].length(),"KinEng");
if (!temp[i].compare("pe")) temp[i].replace(0,temp[i].length(),"PotEng");
if (!temp[i].compare("be")) temp[i].replace(0,temp[i].length(),"E_bond");
if (!temp[i].compare("ae")) temp[i].replace(0,temp[i].length(),"E_angle");
if (!temp[i].compare("de")) temp[i].replace(0,temp[i].length(),"E_dihed");
if (!temp[i].compare("ce")) temp[i].replace(0,temp[i].length(),"E_coul");
if (!temp[i].compare("le")) temp[i].replace(0,temp[i].length(),"E_long");
}
}
}
/*This function prints out the parameter associated index in successfully read file */
void print_file_index_output(sentence s, map<string, vector<int> > x, sentence sf) {
size_t maxlen=0;
for(int i=0;i!=sf.size();i++)
maxlen=max(maxlen,sf[i].length());
cout << print_space(maxlen+11);
for(sentence::iterator aa=++s.begin(); aa!=s.end(); aa++)
cout << *aa << "\t";
cout << endl;
for(int j=0;j!=f.size();j++) {
cout << "File-> " << sf[j] << print_space(maxlen-sf[j].length()+1) <<": ";
for(sentence::iterator aa=++s.begin(); aa!=s.end(); aa++) {
cout << x[*aa][j] << "\t";
}
cout << endl;
}
}
float language_model::perplexity(const sentence& tokens) const
{
if (tokens.size() == 0)
throw language_model_exception{"perplexity() called on empty sentence"};
return std::pow(10.0f, -(log_prob(tokens) / tokens.size()));
}
void plot_xmgrace(sentence s, vector <fd> data) {
ofstream out("temp");
// Some string definitions
string s1 = "@ s";
string s2 = " comment ";
string s3 = " legend ";
int plotiterator=0; int set; int style;
int half_data=data.size()/2;
// Printing out X and Y axis properties
out << "@ xaxis label char size 1.500000\n";
out << "@ xaxis ticklabel char size 1.250000\n";
if (s[1].compare("cross"))
out << "@ xaxis label " << "\"time\""<< endl;
else
out << "@ xaxis label " << "\"" << s[2] << "\""<< endl;
out << "@ yaxis label char size 1.500000\n";
out << "@ yaxis ticklabel char size 1.250000\n";
// Printing out legend properties
out << "@ legend box linestyle 0\n";
out << "@ legend 0.25, 0.8\n";
// Selecting style for series or parallel plotting.
if (!s[0].compare("s")) style=1; else style=2;
if (s[1].compare("cross")) {
if (s.size()==2) {
out << "@ yaxis label " << "\"";
for(int i=1;i!=s.size();i++)
out << s[i] << " ";
out << "\""<< endl;
}
}
else
out << "@ yaxis label " << "\"" << s[3] << "\""<< endl;
int length=data.size();
if (!s[1].compare("cross")) length/=2;
// Storing data in file "temp" as asked (series or parallel)
for(int i=0;i!=length;i++) {
switch (style) {
case 1: { // Series Case
if (s[1].compare("cross")) { // This is executed if not cross
out << s1 << plotiterator << s2 << "\""<< data[i].property << "\"" << endl;
out << s1 << plotiterator << s3 << "\""<< data[i].property << "\"" << endl;
}
else
{
out << s1 << plotiterator << s2 << "\""<< data[i+half_data].property << " vs " << data[i].property << "\"" << endl;
out << s1 << plotiterator << s3 << "\""<< data[i+half_data].property << " vs " << data[i].property << "\"" << endl;
}
out << s1 << plotiterator << " symbol 1" << endl;
out << s1 << plotiterator << " symbol size 0.50000" << endl;
out << s1 << plotiterator << " symbol color 1" << endl;
out << s1 << plotiterator << " symbol pattern 1" << endl;
out << s1 << plotiterator << " symbol fill color "<< plotiterator+1 << endl;
out << s1 << plotiterator << " symbol fill pattern 1" << endl;
out << s1 << plotiterator << " symbol linewidth 1.0" << endl;
out << s1 << plotiterator << " symbol linestyle 1" << endl;
out << s1 << plotiterator << " symbol char 65" << endl;
out << s1 << plotiterator << " symbol char font 0" << endl;
plotiterator++;
for(int j=0;j!=data[i].value.size(); j++)
if (s[1].compare("cross"))
out << data[i].value[j] << endl; // This is executed if not cross
else
out << data[i].value[j] << " " << data[i+half_data].value[j]<< endl; // This is executed if cross
out << "&" << endl;
break;
}
case 2: { // Parallel Case
if (s[1].compare("cross")) {
out << s1 << plotiterator << s2 << "\""<< data[i].property << ": " << data[i].filename << "\"" << endl;
if (s.size()<=2)
out << s1 << plotiterator << s3 << "\"" << data[i].filename << "\"" << endl;
else
out << s1 << plotiterator << s3 << "\""<< data[i].property << ": " << data[i].filename << "\"" << endl;
}
else {
out << s1 << plotiterator << s2 << "\""<< data[i+half_data].property << " vs " << data[i].property << ": " << data[i].filename << "\"" << endl;
out << s1 << plotiterator << s3 << "\""<< data[i+half_data].property << " vs " << data[i].property << ": " << data[i].filename << "\"" << endl;
}
out << s1 << plotiterator << " symbol 0" << endl;
out << s1 << plotiterator << " symbol size 0.50000" << endl;
out << s1 << plotiterator << " symbol color 1" << endl;
out << s1 << plotiterator << " symbol pattern 1" << endl;
out << s1 << plotiterator << " symbol fill color "<< plotiterator+1 << endl;
out << s1 << plotiterator << " symbol fill pattern 1" << endl;
out << s1 << plotiterator << " symbol linewidth 1.0" << endl;
out << s1 << plotiterator << " symbol linestyle 1" << endl;
out << s1 << plotiterator << " symbol char 65" << endl;
out << s1 << plotiterator << " symbol char font 0" << endl;
plotiterator++;
for(int j=0;j!=data[i].value.size(); j++)
if (s[1].compare("cross")) {
out << data[i].value[j] << endl; // This is executed if not cross
}
else
out << data[i].value[j] << " " << data[i+half_data].value[j]<< endl; // This is executed if cross
out << "&" << endl;
break;
}
}
}
int rv=system("xmgrace temp &");
}
