void initializeModels() {
// Create a country model.
unsigned countryModelRows = countries_.size() + 1;
const unsigned countryModelColumns = 1;
countryModel_ =
new Wt::WStandardItemModel(countryModelRows, countryModelColumns,
this);
// The initial text shown in the country combo box should be an empty
// string.
int row = 0;
countryModel_->setData(row, 0, std::string(" "), Wt::DisplayRole);
countryModel_->setData(row, 0, std::string(), Wt::UserRole);
// For each country, update the model based on the key (corresponding
// to the country code):
// - set the country name for the display role,
// - set the city names for the user role.
row = 1;
for (CountryMap::const_iterator i = countries_.begin();
i != countries_.end(); ++i, ++row) {
countryModel_->setData(row, 0, i->second, Wt::DisplayRole);
countryModel_->setData(row, 0, i->first, Wt::UserRole);
}
// Create a city model.
cityModel_ = new Wt::WStandardItemModel(this);
updateCityModel(std::string());
}
/*this will convert data internal form to Wt::WStandardItemModel ( table format for wt widget YahooStockHistory
**/
Wt::WAbstractItemModel * YahooStockHistory::provideModelObject( Wt::WContainerWidget *parent)
{
Wt::WStandardItemModel *model = new Wt::WStandardItemModel(366, 5, parent);
model->setHeaderData(0, std::string("Date"));
model->setHeaderData(1, std::string("Price"));
model->setHeaderData(2, std::string("PricePredictionmoins1"));
model->setHeaderData(3, std::string("PricePredictionmoins2"));
model->setHeaderData(4, std::string("PricePredictionmoins3"));
updateModelWithFakePrices(model);
return model;
}
Wt::WStandardItemModel *csvToModel(const std::string& csvFile,
Wt::WObject *parent,
bool firstLineIsHeaders)
{
std::ifstream f(csvFile.c_str());
if (f) {
Wt::WStandardItemModel *result = new Wt::WStandardItemModel(0, 0, parent);
result->setItemPrototype(new NumericItem());
readFromCsv(f, result, -1, firstLineIsHeaders);
return result;
} else
return 0;
}
/////////1/////////2/////////3/////////4/////////5/////////6/////////7/////////8/////////9/////////A
void Populations::ShowPopulation()
{
try
{
const Wt::WString &site_name = areas_->currentText();
Wt::WApplication::instance()->log("notice") << "Populations::ShowPopulation() for " << site_name;
delete chart_;
chart_ = 0;
delete footertext_;
footertext_ = 0;
delete algo_legend_;
algo_legend_ = 0;
pqxx::connection conn(db_conn_str_);
pqxx::transaction<> trans(conn, "web population");
std::stringstream sstr;
sstr << "SELECT * FROM trained_solutions INNER JOIN pred_sites ON "
<< "trained_solutions.pred_site_id = pred_sites.pred_site_id WHERE "
<< "site_name='" << site_name.narrow() << "' "
<< "ORDER BY train_time ASC";
pqxx::result res = trans.exec(sstr.str());
if(!res.size())
{
Wt::WApplication::instance()->log("warn") << "critical error during session initialization : ";
footertext_ = new Wt::WText("no population found", impl_);
return;
}
boost::regex regx("\\w+\\(\\w+");
set<string> confclasses;
typedef map<size_t, vector<pair<string, double> > > SolutionsT; // train_time, algorithm, validation_error
SolutionsT solutions;
double max_train_time = 0.0;
static const double train_time_treshold = 60.0;
for(size_t i=0; i<res.size(); ++i)
{
size_t train_sol_id;
res[i]["train_sol_id"].to(train_sol_id);
size_t generation;
res[i]["generation"].to(generation);
string config;
res[i]["configuration"].to(config);
boost::smatch regxmatch;
if(boost::regex_search(config, regxmatch, regx))
{
string confclass = regxmatch[0] + ")";
double train_time;
res[i]["train_time"].to(train_time);
double validation_error;
res[i]["validation_error"].to(validation_error);
if(criteria_->currentText() == "Generation")
confclass = (boost::format("%03d") % generation).str();
confclasses.insert(confclass);
if(train_time <= train_time_treshold)
solutions[train_time * 100].push_back(make_pair(confclass, validation_error));
max_train_time = std::max(max_train_time, train_time);
}
}
Wt::WStandardItemModel *model = new Wt::WStandardItemModel(res.size(), 1 + confclasses.size());
model->setHeaderData(0, Wt::Horizontal, any(string("train_time")));
for(set<string>::iterator it = confclasses.begin(); it != confclasses.end(); ++it)
model->setHeaderData(1 + std::distance(confclasses.begin(), it), Wt::Horizontal, any(*it));
for(SolutionsT::iterator it = solutions.begin(); it != solutions.end(); ++it)
{
const size_t modrow = std::distance(solutions.begin(), it);
model->setData(modrow, 0, any(it->first / 100.0));
for(vector<pair<string, double> >::iterator itv = it->second.begin(); itv != it->second.end(); ++itv)
{
const size_t modcol = 1 + std::distance(confclasses.begin(), confclasses.find(itv->first));
model->setData(modrow, modcol, any(itv->second));
}
}
chart_ = new Wt::Chart::WCartesianChart(impl_);
chart_->setType(Wt::Chart::ScatterPlot);
chart_->resize(800, 500);
chart_->setTitle("Population performance");
chart_->setPlotAreaPadding(200, Wt::Right);
chart_->setPlotAreaPadding(30, Wt::Bottom);
chart_->setModel(model);
chart_->setXSeriesColumn(0);
chart_->setLegendEnabled(true);
chart_->axis(Wt::Chart::XAxis).setScale(Wt::Chart::LinearScale);
// chart_->axis(Wt::Chart::XAxis).setMaximum(60.0);
chart_->axis(Wt::Chart::XAxis).setLabelFormat("%.2f");
chart_->axis(Wt::Chart::XAxis).setTitle("training_time");
chart_->axis(Wt::Chart::YAxis).setTitle("error");
for(set<string>::iterator it = confclasses.begin(); it != confclasses.end(); ++it)
{
const size_t npos = std::distance(confclasses.begin(), it);
Wt::Chart::WDataSeries data1(Wt::Chart::WDataSeries(1 + npos, Wt::Chart::PointSeries, Wt::Chart::Y1Axis));
data1.setLegendEnabled(true);
if(criteria_->currentText() == "Generation")
{
// make older generations fade
const float brightness = 1.0 - (static_cast<float>(npos) / confclasses.size()); // 0.0 to 1.0
data1.setBrush(Wt::WBrush(Wt::WColor(255, brightness * 255, brightness * 255)));
}
chart_->addSeries(data1);
}
sstr.str("");
sstr << "Longest training time : " << max_train_time;
footertext_ = new Wt::WText(sstr.str(), impl_);
Wt::WApplication::instance()->log("notice") << sstr.str();
if(criteria_->currentText() == "Algorithm")
{
Wt::WTable *algo_legend_ = new Wt::WTable(impl_);
algo_legend_->setStyleClass("algo_desc");
new Wt::WText("<b>Algorithm descriptions:</b>", Wt::XHTMLText, algo_legend_->elementAt(0, 0));
new Wt::WText("<b>DLIB_KRLS:</b>", Wt::XHTMLText, algo_legend_->elementAt(1, 0));
new Wt::WText("<a href='http://dclib.sourceforge.net/ml.html#krls'>kernel recursive least squares</a>"
" -> <a href='http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.2.9587&rep=rep1&type=pdf'>detailed paper</a>", Wt::XHTMLText, algo_legend_->elementAt(1, 1));
new Wt::WText("<b>DLIB_RVM:</b>", Wt::XHTMLText, algo_legend_->elementAt(2, 0));
new Wt::WText("<a href='http://dclib.sourceforge.net/ml.html#rvm_regression_trainer'>relevance vector machine</a>", Wt::XHTMLText, algo_legend_->elementAt(2, 1));
new Wt::WText("<b>Kernel descriptions:</b>", Wt::XHTMLText, algo_legend_->elementAt(3, 0));
new Wt::WText("<b>RBF:</b>", Wt::XHTMLText, algo_legend_->elementAt(4, 0));
new Wt::WText("<a href='http://dclib.sourceforge.net/ml.html#radial_basis_kernel'>radial basis function</a>", Wt::XHTMLText, algo_legend_->elementAt(4, 1));
new Wt::WText("<b>SIG:</b>", Wt::XHTMLText, algo_legend_->elementAt(5, 0));
new Wt::WText("<a href='http://dclib.sourceforge.net/ml.html#sigmoid_kernel'>sigmoid</a>", Wt::XHTMLText, algo_legend_->elementAt(5, 1));
new Wt::WText("<b>POLY:</b>", Wt::XHTMLText, algo_legend_->elementAt(6, 0));
new Wt::WText("<a href='http://dclib.sourceforge.net/ml.html#polynomial_kernel'>polynomial</a>", Wt::XHTMLText, algo_legend_->elementAt(6, 1));
}
}
catch(std::exception& ex)
{
Wt::WApplication::instance()->log("error") << "Exception in Populations::ShowPopulation() : " << ex.what();
}
}
