Files
Platform/src/main/Experiment.cpp

302 lines
15 KiB
C++

#include "common/Datasets.h"
#include "reports/ReportConsole.h"
#include "common/Paths.h"
#include "Models.h"
#include "Scores.h"
#include "RocAuc.h"
#include "Experiment.h"
namespace platform {
using json = nlohmann::ordered_json;
void Experiment::saveResult()
{
result.save();
std::cout << "Result saved in " << Paths::results() << result.getFilename() << std::endl;
}
void Experiment::report(bool classification_report)
{
ReportConsole report(result.getJson());
report.show();
if (classification_report) {
std::cout << report.showClassificationReport(Colors::BLUE());
}
}
void Experiment::show()
{
std::cout << result.getJson().dump(4) << std::endl;
}
void Experiment::saveGraph()
{
std::cout << "Saving graphs..." << std::endl;
auto data = result.getJson();
for (const auto& item : data["results"]) {
auto graphs = item["graph"];
int i = 0;
for (const auto& graph : graphs) {
i++;
auto fileName = Paths::graphs() + result.getFilename() + "_graph_" + item["dataset"].get<std::string>() + "_" + std::to_string(i) + ".dot";
auto file = std::ofstream(fileName);
file << graph.get<std::string>();
file.close();
std::cout << "Graph saved in " << fileName << std::endl;
}
}
}
void Experiment::go(std::vector<std::string> filesToProcess, bool quiet, bool no_train_score, bool generate_fold_files, bool graph)
{
for (auto fileName : filesToProcess) {
if (fileName.size() > max_name)
max_name = fileName.size();
}
std::cout << Colors::MAGENTA() << "*** Starting experiment: " << result.getTitle() << " ***" << Colors::RESET() << std::endl << std::endl;
auto clf = Models::instance()->create(result.getModel());
auto version = clf->getVersion();
std::cout << Colors::BLUE() << " Using " << result.getModel() << " ver. " << version << std::endl << std::endl;
if (!quiet) {
std::cout << Colors::GREEN() << " Status Meaning" << std::endl;
std::cout << " ------ --------------------------------" << Colors::RESET() << std::endl;
std::cout << " ( " << Colors::GREEN() << "a" << Colors::RESET() << " ) Fitting model with train dataset" << std::endl;
std::cout << " ( " << Colors::GREEN() << "b" << Colors::RESET() << " ) Scoring train dataset" << std::endl;
std::cout << " ( " << Colors::GREEN() << "c" << Colors::RESET() << " ) Scoring test dataset" << std::endl << std::endl;
std::cout << Colors::YELLOW() << "Note: fold number in this color means fitting had issues such as not using all features in BoostAODE classifier" << std::endl << std::endl;
std::cout << Colors::GREEN() << left << " # " << setw(max_name) << "Dataset" << " #Samp #Feat Seed Status" << std::endl;
std::cout << " --- " << string(max_name, '-') << " ----- ----- ---- " << string(4 + 3 * nfolds, '-') << Colors::RESET() << std::endl;
}
int num = 0;
for (auto fileName : filesToProcess) {
if (!quiet)
std::cout << " " << setw(3) << right << num++ << " " << setw(max_name) << left << fileName << right << flush;
cross_validation(fileName, quiet, no_train_score, generate_fold_files, graph);
if (!quiet)
std::cout << std::endl;
}
if (!quiet)
std::cout << std::endl;
}
std::string getColor(bayesnet::status_t status)
{
switch (status) {
case bayesnet::NORMAL:
return Colors::GREEN();
case bayesnet::WARNING:
return Colors::YELLOW();
case bayesnet::ERROR:
return Colors::RED();
default:
return Colors::RESET();
}
}
void showProgress(int fold, const std::string& color, const std::string& phase)
{
std::string prefix = phase == "-" ? "" : "\b\b\b\b";
std::cout << prefix << color << fold << Colors::RESET() << "(" << color << phase << Colors::RESET() << ")" << flush;
}
void generate_files(const std::string& fileName, bool discretize, bool stratified, int seed, int nfold, torch::Tensor X_train, torch::Tensor y_train, torch::Tensor X_test, torch::Tensor y_test, std::vector<int>& train, std::vector<int>& test)
{
std::string file_name = Paths::experiment_file(fileName, discretize, stratified, seed, nfold);
auto file = std::ofstream(file_name);
json output;
output["seed"] = seed;
output["nfold"] = nfold;
output["X_train"] = json::array();
auto n = X_train.size(1);
for (int i = 0; i < X_train.size(0); i++) {
if (X_train.dtype() == torch::kFloat32) {
auto xvf_ptr = X_train.index({ i }).data_ptr<float>();
auto feature = std::vector<float>(xvf_ptr, xvf_ptr + n);
output["X_train"].push_back(feature);
} else {
auto feature = std::vector<int>(X_train.index({ i }).data_ptr<int>(), X_train.index({ i }).data_ptr<int>() + n);
output["X_train"].push_back(feature);
}
}
output["y_train"] = std::vector<int>(y_train.data_ptr<int>(), y_train.data_ptr<int>() + n);
output["X_test"] = json::array();
n = X_test.size(1);
for (int i = 0; i < X_test.size(0); i++) {
if (X_train.dtype() == torch::kFloat32) {
auto xvf_ptr = X_test.index({ i }).data_ptr<float>();
auto feature = std::vector<float>(xvf_ptr, xvf_ptr + n);
output["X_test"].push_back(feature);
} else {
auto feature = std::vector<int>(X_test.index({ i }).data_ptr<int>(), X_test.index({ i }).data_ptr<int>() + n);
output["X_test"].push_back(feature);
}
}
output["y_test"] = std::vector<int>(y_test.data_ptr<int>(), y_test.data_ptr<int>() + n);
output["train"] = train;
output["test"] = test;
file << output.dump(4);
file.close();
}
void Experiment::cross_validation(const std::string& fileName, bool quiet, bool no_train_score, bool generate_fold_files, bool graph)
{
//
// Load dataset and prepare data
//
auto datasets = Datasets(discretized, Paths::datasets(), discretization_algo);
auto& dataset = datasets.getDataset(fileName);
dataset.load();
auto [X, y] = dataset.getTensors(); // Only need y for folding
auto features = dataset.getFeatures();
auto n_features = dataset.getNFeatures();
auto n_samples = dataset.getNSamples();
auto className = dataset.getClassName();
auto labels = dataset.getLabels();
int num_classes = dataset.getNClasses();
if (!quiet) {
std::cout << " " << setw(5) << n_samples << " " << setw(5) << n_features << flush;
}
//
// Prepare Result
//
auto partial_result = PartialResult();
partial_result.setSamples(n_samples).setFeatures(n_features).setClasses(num_classes);
partial_result.setHyperparameters(hyperparameters.get(fileName));
//
// Initialize results std::vectors
//
int nResults = nfolds * static_cast<int>(randomSeeds.size());
auto accuracy_test = torch::zeros({ nResults }, torch::kFloat64);
auto accuracy_train = torch::zeros({ nResults }, torch::kFloat64);
auto auc_test = torch::zeros({ nResults }, torch::kFloat64);
auto auc_train = torch::zeros({ nResults }, torch::kFloat64);
auto train_time = torch::zeros({ nResults }, torch::kFloat64);
auto test_time = torch::zeros({ nResults }, torch::kFloat64);
auto nodes = torch::zeros({ nResults }, torch::kFloat64);
auto edges = torch::zeros({ nResults }, torch::kFloat64);
auto num_states = torch::zeros({ nResults }, torch::kFloat64);
json confusion_matrices = json::array();
json confusion_matrices_train = json::array();
std::vector<std::string> notes;
std::vector<std::string> graphs;
Timer train_timer, test_timer;
int item = 0;
bool first_seed = true;
//
// Loop over random seeds
//
for (auto seed : randomSeeds) {
if (!quiet) {
string prefix = " ";
if (!first_seed) {
prefix = "\n" + string(18 + max_name, ' ');
}
std::cout << prefix << setw(4) << right << seed << " " << flush;
first_seed = false;
}
folding::Fold* fold;
if (stratified)
fold = new folding::StratifiedKFold(nfolds, y, seed);
else
fold = new folding::KFold(nfolds, n_samples, seed);
//
// Loop over folds
//
for (int nfold = 0; nfold < nfolds; nfold++) {
auto clf = Models::instance()->create(result.getModel());
if (!quiet)
showProgress(nfold + 1, getColor(clf->getStatus()), "-");
setModelVersion(clf->getVersion());
auto valid = clf->getValidHyperparameters();
hyperparameters.check(valid, fileName);
clf->setHyperparameters(hyperparameters.get(fileName));
//
// Split train - test dataset
//
train_timer.start();
auto [train, test] = fold->getFold(nfold);
auto [X_train, X_test, y_train, y_test] = dataset.getTrainTestTensors(train, test);
auto states = dataset.getStates(); // Get the states of the features Once they are discretized
if (generate_fold_files)
generate_files(fileName, discretized, stratified, seed, nfold, X_train, y_train, X_test, y_test, train, test);
if (!quiet)
showProgress(nfold + 1, getColor(clf->getStatus()), "a");
//
// Train model
//
clf->fit(X_train, y_train, features, className, states, smooth_type);
if (!quiet)
showProgress(nfold + 1, getColor(clf->getStatus()), "b");
auto clf_notes = clf->getNotes();
std::transform(clf_notes.begin(), clf_notes.end(), std::back_inserter(notes), [nfold](const std::string& note)
{ return "Fold " + std::to_string(nfold) + ": " + note; });
nodes[item] = clf->getNumberOfNodes();
edges[item] = clf->getNumberOfEdges();
num_states[item] = clf->getNumberOfStates();
train_time[item] = train_timer.getDuration();
double accuracy_train_value = 0.0;
//
// Score train
//
double auc_train_value = 0;
if (!no_train_score) {
auto roc_auc = RocAuc();
auto y_proba_train = clf->predict_proba(X_train);
Scores scores(y_train, y_proba_train, num_classes, labels);
accuracy_train_value = scores.accuracy();
auc_train_value = roc_auc.compute(y_proba_train, y_train);
confusion_matrices_train.push_back(scores.get_confusion_matrix_json(true));
}
//
// Test model
//
if (!quiet)
showProgress(nfold + 1, getColor(clf->getStatus()), "c");
test_timer.start();
// auto y_predict = clf->predict(X_test);
auto y_proba_test = clf->predict_proba(X_test);
Scores scores(y_test, y_proba_test, num_classes, labels);
auto accuracy_test_value = scores.accuracy();
auto roc_auc = RocAuc();
double auc_test_value = roc_auc.compute(y_proba_test, y_test);
test_time[item] = test_timer.getDuration();
auc_train[item] = auc_train_value;
auc_test[item] = auc_test_value;
accuracy_train[item] = accuracy_train_value;
accuracy_test[item] = accuracy_test_value;
confusion_matrices.push_back(scores.get_confusion_matrix_json(true));
if (!quiet)
std::cout << "\b\b\b, " << flush;
//
// Store results and times in std::vector
//
partial_result.addAucTrain(auc_train_value);
partial_result.addAucTest(auc_test_value);
partial_result.addScoreTrain(accuracy_train_value);
partial_result.addScoreTest(accuracy_test_value);
partial_result.addTimeTrain(train_time[item].item<double>());
partial_result.addTimeTest(test_time[item].item<double>());
item++;
if (graph) {
std::string result = "";
for (const auto& line : clf->graph()) {
result += line + "\n";
}
graphs.push_back(result);
}
}
if (!quiet)
std::cout << "end. " << flush;
delete fold;
}
//
// Store result totals in Result
//
partial_result.setGraph(graphs);
partial_result.setScoreTest(torch::mean(accuracy_test).item<double>()).setScoreTrain(torch::mean(accuracy_train).item<double>());
partial_result.setScoreTestStd(torch::std(accuracy_test).item<double>()).setScoreTrainStd(torch::std(accuracy_train).item<double>());
partial_result.setAucTest(torch::mean(auc_test).item<double>()).setAucTrain(torch::mean(auc_train).item<double>());
partial_result.setAucTestStd(torch::std(auc_test).item<double>()).setAucTrainStd(torch::std(auc_train).item<double>());
partial_result.setTrainTime(torch::mean(train_time).item<double>()).setTestTime(torch::mean(test_time).item<double>());
partial_result.setTestTimeStd(torch::std(test_time).item<double>()).setTrainTimeStd(torch::std(train_time).item<double>());
partial_result.setNodes(torch::mean(nodes).item<double>()).setLeaves(torch::mean(edges).item<double>()).setDepth(torch::mean(num_states).item<double>());
partial_result.setDataset(fileName).setNotes(notes);
partial_result.setConfusionMatrices(confusion_matrices);
if (!no_train_score)
partial_result.setConfusionMatricesTrain(confusion_matrices_train);
addResult(partial_result);
}
}