// src/screens/ai_quant_lab/QuantModulePanel_CFA_Display.cpp // // CFA Quant — result rendering. Hosts the per-command display branches of // display_cfa_result (13 commands) plus the file-local cfa_flatten_scalars // helper and the fmt_* utilities used only inside this TU. // // Part of the partial-class split of QuantModulePanel_CFA.cpp. // keep that file maintainable. #include "core/logging/Logger.h" #include "screens/ai_quant_lab/QuantModulePanel.h" #include "screens/ai_quant_lab/QuantModulePanel_Common.h" #include "screens/ai_quant_lab/QuantModulePanel_GsHelpers.h" #include "screens/ai_quant_lab/QuantModulePanel_Styles.h" #include "services/ai_quant_lab/AIQuantLabService.h" #include "services/file_manager/FileManagerService.h" #include "ui/theme/Theme.h" #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include namespace fincept::screens { using namespace fincept::services::quant; using namespace fincept::screens::quant_styles; using namespace fincept::screens::quant_common; using namespace fincept::screens::quant_gs_helpers; namespace { // Fallback flattener — only used for unknown commands. void cfa_flatten_scalars(const QJsonObject& obj, const QString& prefix, QStringList& keys, QHash& out) { for (auto it = obj.begin(); it != obj.end(); ++it) { const auto& v = it.value(); const QString key = prefix.isEmpty() ? it.key() : prefix + "." + it.key(); if (v.isObject()) { if (key.endsWith(".data") || key.endsWith(".fitted_values") || key.endsWith(".residuals") || key.endsWith(".model_summary") || key.endsWith(".predictions") || key.endsWith(".sample_means") || key.endsWith(".resampled_statistics") || key.endsWith(".cluster_labels") || key.endsWith(".principal_components") || key.endsWith(".cluster_centers")) continue; cfa_flatten_scalars(v.toObject(), key, keys, out); } else if (!v.isArray()) { keys.append(key); out.insert(key, v); } } } QString fmt_num_safe(const QJsonValue& v, int decimals = 4) { if (v.isNull() || v.isUndefined()) return QStringLiteral("—"); return QString::number(v.toDouble(), 'f', decimals); } [[maybe_unused]] QString fmt_pct(const QJsonValue& v, int decimals = 2) { if (v.isNull() || v.isUndefined()) return QStringLiteral("—"); return QString::number(v.toDouble() * 100.0, 'f', decimals) + "%"; } [[maybe_unused]] QString fmt_bool(const QJsonValue& v) { if (v.isNull() || v.isUndefined()) return QStringLiteral("—"); return v.toBool() ? QStringLiteral("YES") : QStringLiteral("NO"); } } // namespace void QuantModulePanel::display_cfa_result(const QString& command, const QJsonObject& payload) { clear_results(); if (!payload.value("success").toBool()) { display_error(payload.value("error").toString(tr("Unknown error"))); return; } const QJsonObject result = payload.value("result").toObject(); const QString method = result.value("method").toString(); const double calc_time = result.value("calculation_time").toDouble(); const QString analysis_type = payload.value("analysis_type").toString(command); const QString accent = module_.color.name(); // ── Section header ────────────────────────────────────────────────────── QString header_text = analysis_type.toUpper(); header_text.replace('_', ' '); if (!method.isEmpty()) header_text += QString(" | %1").arg(method.toUpper()); if (calc_time > 0.0) header_text += tr(" | %1 ms").arg(QString::number(calc_time * 1000.0, 'f', 1)); results_layout_->addWidget(gs_section_header(header_text, accent)); const QJsonValue value_raw = result.value("value"); const QJsonObject value_obj = value_raw.toObject(); const QJsonObject meta = result.value("metadata").toObject(); bool rendered = false; // ── 1. TREND ANALYSIS ─────────────────────────────────────────────────── if (command == "trend_analysis") { const double slope = value_obj.value("slope").toDouble(); const double intercept = value_obj.value("intercept").toDouble(); const double r2 = meta.value("r_squared").toDouble(); const double tstat = meta.value("slope_t_statistic").toDouble(); const double pval = meta.value("slope_p_value").toDouble(); const QString direction = meta.value("trend_direction").toString().toUpper(); const bool significant = meta.value("trend_significant").toBool(); auto dir_color = [&]() { if (direction == "UPWARD") return ui::colors::POSITIVE(); if (direction == "DOWNWARD") return ui::colors::NEGATIVE(); return ui::colors::TEXT_PRIMARY(); }; QList top = { gs_make_card(tr("DIRECTION"), direction.isEmpty() ? QStringLiteral("—") : direction, this, dir_color()), gs_make_card(tr("SLOPE"), gs_fmt_num(slope, 6), this, gs_pos_neg_color(slope)), gs_make_card(tr("INTERCEPT"), gs_fmt_num(intercept, 4), this), gs_make_card(tr("R²"), gs_fmt_num(r2, 4), this, r2 >= 0.7 ? ui::colors::POSITIVE() : r2 >= 0.3 ? ui::colors::WARNING() : ui::colors::NEGATIVE()), }; results_layout_->addWidget(gs_card_row(top, this)); QList stats = { gs_make_card(tr("t-STATISTIC"), gs_fmt_num(tstat, 3), this), gs_make_card(tr("p-VALUE"), gs_fmt_num(pval, 4), this, pval < 0.05 ? ui::colors::POSITIVE() : ui::colors::WARNING()), gs_make_card(tr("SIGNIFICANT"), significant ? tr("YES") : tr("NO"), this, significant ? ui::colors::POSITIVE() : ui::colors::WARNING()), gs_make_card(tr("OBSERVATIONS"), QString::number(result.value("parameters").toObject().value("data").toArray().size()), this), }; results_layout_->addWidget(gs_card_row(stats, this)); status_label_->setText(tr("Trend %1 slope=%2 R²=%3 p=%4") .arg(direction) .arg(slope, 0, 'g', 4) .arg(r2, 0, 'f', 3) .arg(pval, 0, 'f', 4)); rendered = true; } // ── 2. STATIONARITY TEST ──────────────────────────────────────────────── else if (command == "stationarity_test") { const QString test = result.value("parameters").toObject().value("test_type").toString().toUpper(); const double tstat = value_obj.value("test_statistic").toDouble(); const double pval = value_obj.value("p_value").toDouble(); const bool stationary = value_obj.value("is_stationary").toBool(); const QString h0 = value_obj.value("null_hypothesis").toString(); QList top = { gs_make_card(tr("TEST"), test.isEmpty() ? QStringLiteral("ADF") : test, this), gs_make_card(tr("TEST STAT"), gs_fmt_num(tstat, 4), this), gs_make_card(tr("p-VALUE"), gs_fmt_num(pval, 4), this, pval < 0.05 ? ui::colors::POSITIVE() : ui::colors::WARNING()), gs_make_card(tr("STATIONARY"), stationary ? tr("YES") : tr("NO"), this, stationary ? ui::colors::POSITIVE() : ui::colors::NEGATIVE()), }; results_layout_->addWidget(gs_card_row(top, this)); // Critical values row const QJsonObject crit = value_obj.value("critical_values").toObject(); if (!crit.isEmpty()) { QList crit_cards; for (auto it = crit.begin(); it != crit.end(); ++it) { crit_cards << gs_make_card(tr("CRIT %1").arg(it.key()), gs_fmt_num(it.value().toDouble(), 4), this); } results_layout_->addWidget(gs_card_row(crit_cards, this)); } if (!h0.isEmpty()) { auto* note = new QLabel(tr("H₀: %1").arg(h0)); note->setStyleSheet( QString("color:%1; font-size:10px; padding:6px 10px; background:%2; border-left:3px solid %3;") .arg(ui::colors::TEXT_SECONDARY(), ui::colors::BG_SURFACE(), accent)); note->setWordWrap(true); results_layout_->addWidget(note); } status_label_->setText(tr("%1: stat=%2 p=%3 → %4") .arg(test) .arg(tstat, 0, 'f', 3) .arg(pval, 0, 'f', 4) .arg(stationary ? tr("stationary") : tr("non-stationary"))); rendered = true; } // ── 3. ARIMA MODEL ────────────────────────────────────────────────────── else if (command == "arima_model") { const QJsonArray order = result.value("parameters").toObject().value("order").toArray(); const QString order_str = order.size() == 3 ? QString("(%1, %2, %3)").arg(order[0].toInt()).arg(order[1].toInt()).arg(order[2].toInt()) : QStringLiteral("—"); const double aic = meta.value("aic").toDouble(); const double bic = meta.value("bic").toDouble(); const double llf = meta.value("log_likelihood").toDouble(); const double lb_p = meta.value("ljung_box_p_value").toDouble(); const bool autocorr = meta.value("residuals_autocorrelated").toBool(); QList top = { gs_make_card(tr("ORDER (p,d,q)"), order_str, this), gs_make_card(tr("AIC"), gs_fmt_num(aic, 2), this), gs_make_card(tr("BIC"), gs_fmt_num(bic, 2), this), gs_make_card(tr("LOG-LIKELIHOOD"), gs_fmt_num(llf, 2), this), }; results_layout_->addWidget(gs_card_row(top, this)); QList diag = { gs_make_card(tr("LJUNG-BOX p"), gs_fmt_num(lb_p, 4), this, lb_p > 0.05 ? ui::colors::POSITIVE() : ui::colors::WARNING()), gs_make_card(tr("RESID. AUTOCORR."), autocorr ? tr("YES") : tr("NO"), this, autocorr ? ui::colors::WARNING() : ui::colors::POSITIVE()), gs_make_card(tr("FITTED POINTS"), QString::number(value_obj.value("fitted_values").toArray().size()), this), gs_make_card(tr("RESIDUALS"), QString::number(value_obj.value("residuals").toArray().size()), this), }; results_layout_->addWidget(gs_card_row(diag, this)); // Coefficient list (value.parameters can be an array OR a {const, ar.L1, ...} object) const QJsonValue coef_raw = value_obj.value("parameters"); QStringList coef_lines; if (coef_raw.isArray()) { const auto arr = coef_raw.toArray(); for (int i = 0; i < arr.size(); ++i) coef_lines << QString("β%1 = %2").arg(i).arg(arr[i].toDouble(), 0, 'f', 6); } else if (coef_raw.isObject()) { const auto o = coef_raw.toObject(); for (auto it = o.begin(); it != o.end(); ++it) coef_lines << QString("%1 = %2").arg(it.key()).arg(it.value().toDouble(), 0, 'f', 6); } if (!coef_lines.isEmpty()) { auto* coef_lbl = new QLabel(tr("Coefficients: %1").arg(coef_lines.join(" "))); coef_lbl->setWordWrap(true); coef_lbl->setStyleSheet(QString("color:%1; font-family:'Courier New'; font-size:10px;" "padding:6px 10px; background:%2; border-left:3px solid %3;") .arg(ui::colors::TEXT_PRIMARY(), ui::colors::BG_SURFACE(), accent)); results_layout_->addWidget(coef_lbl); } status_label_->setText( tr("ARIMA%1 AIC=%2 Ljung-Box p=%3").arg(order_str).arg(aic, 0, 'f', 2).arg(lb_p, 0, 'f', 3)); rendered = true; } // ── 4. FORECASTING ────────────────────────────────────────────────────── else if (command == "forecasting") { const QJsonObject params = result.value("parameters").toObject(); const QString fmethod = params.value("method").toString().toUpper(); const int horizon = meta.value("forecast_horizon").toInt(); const int train_n = meta.value("train_data_size").toInt(); const int test_n = meta.value("test_data_size").toInt(); const QJsonValue mae = meta.value("mae"); const QJsonValue rmse = meta.value("rmse"); const QJsonValue mape = meta.value("mape"); QList top = { gs_make_card(tr("METHOD"), fmethod.isEmpty() ? QStringLiteral("—") : fmethod, this), gs_make_card(tr("HORIZON"), QString::number(horizon), this), gs_make_card(tr("TRAIN"), tr("%1 obs").arg(train_n), this), gs_make_card(tr("TEST"), tr("%1 obs").arg(test_n), this), }; results_layout_->addWidget(gs_card_row(top, this)); QList err = { gs_make_card(tr("MAE"), fmt_num_safe(mae, 4), this, mae.isNull() ? ui::colors::TEXT_TERTIARY() : ui::colors::WARNING()), gs_make_card(tr("RMSE"), fmt_num_safe(rmse, 4), this, rmse.isNull() ? ui::colors::TEXT_TERTIARY() : ui::colors::WARNING()), gs_make_card(tr("MAPE"), mape.isNull() ? QStringLiteral("—") : QString::number(mape.toDouble(), 'f', 2) + "%", this), gs_make_card(tr("FCST POINTS"), QString::number(value_raw.toArray().size()), this), }; results_layout_->addWidget(gs_card_row(err, this)); // Forecast values table if (value_raw.isArray()) { const auto arr = value_raw.toArray(); auto* table = new QTableWidget(arr.size(), 2, this); table->setHorizontalHeaderLabels({tr("Step"), tr("Forecast")}); table->verticalHeader()->setVisible(false); table->setEditTriggers(QAbstractItemView::NoEditTriggers); table->horizontalHeader()->setSectionResizeMode(QHeaderView::Stretch); table->setStyleSheet(table_ss()); table->setMaximumHeight(qMin(arr.size() * 24 + 32, 280)); for (int i = 0; i < arr.size(); ++i) { table->setItem(i, 0, new QTableWidgetItem(QString("t+%1").arg(i + 1))); auto* it = new QTableWidgetItem(QString::number(arr[i].toDouble(), 'f', 6)); it->setTextAlignment(Qt::AlignRight | Qt::AlignVCenter); table->setItem(i, 1, it); table->setRowHeight(i, 24); } results_layout_->addWidget(table); } status_label_->setText( tr("%1 forecast (%2 steps) MAE=%3").arg(fmethod).arg(horizon).arg(fmt_num_safe(mae, 4))); rendered = true; } // ── 5. SUPERVISED LEARNING ────────────────────────────────────────────── else if (command == "supervised_learning") { const QJsonObject params = result.value("parameters").toObject(); const QString problem = params.value("problem_type").toString(); const bool is_reg = (problem == "regression"); const QString best = meta.value("best_algorithm").toString().toUpper(); const int train_n = meta.value("train_size").toInt(); const int test_n = meta.value("test_size").toInt(); const int n_feat = meta.value("n_features").toInt(); QList top = { gs_make_card(tr("PROBLEM"), problem.toUpper(), this), gs_make_card(tr("BEST ALGORITHM"), best.isEmpty() ? QStringLiteral("—") : best, this, ui::colors::POSITIVE()), gs_make_card(tr("FEATURES"), QString::number(n_feat), this), gs_make_card(tr("TRAIN / TEST"), QString("%1 / %2").arg(train_n).arg(test_n), this), }; results_layout_->addWidget(gs_card_row(top, this)); // Per-algorithm metrics table QStringList headers = is_reg ? QStringList{tr("Algorithm"), tr("MSE"), tr("RMSE"), tr("R²")} : QStringList{tr("Algorithm"), tr("Accuracy"), tr("Precision"), tr("Recall")}; QStringList score_keys = is_reg ? QStringList{"mse", "rmse", "r2_score"} : QStringList{"accuracy", "precision", "recall"}; QList> rows; for (auto it = value_obj.begin(); it != value_obj.end(); ++it) { if (it.value().isObject()) rows.append({it.key(), it.value().toObject()}); } if (!rows.isEmpty()) { auto* table = new QTableWidget(rows.size(), headers.size(), this); table->setHorizontalHeaderLabels(headers); table->verticalHeader()->setVisible(false); table->setEditTriggers(QAbstractItemView::NoEditTriggers); table->horizontalHeader()->setSectionResizeMode(QHeaderView::Stretch); table->setStyleSheet(table_ss()); table->setMaximumHeight(qMin(rows.size() * 26 + 32, 320)); for (int r = 0; r < rows.size(); ++r) { const auto& algo = rows[r].first; const auto& metrics = rows[r].second; auto* name_it = new QTableWidgetItem(algo.toUpper()); if (algo.toUpper() == best) name_it->setForeground(QColor(ui::colors::POSITIVE())); table->setItem(r, 0, name_it); if (metrics.contains("error")) { auto* err_it = new QTableWidgetItem(tr("ERR: %1").arg(metrics.value("error").toString())); err_it->setForeground(QColor(ui::colors::NEGATIVE())); table->setItem(r, 1, err_it); table->setSpan(r, 1, 1, headers.size() - 1); } else { for (int c = 0; c < score_keys.size(); ++c) { const double v = metrics.value(score_keys[c]).toDouble(); auto* it = new QTableWidgetItem(QString::number(v, 'f', 4)); it->setTextAlignment(Qt::AlignRight | Qt::AlignVCenter); // R² and accuracy: green if good if ((score_keys[c] == "r2_score" || score_keys[c] == "accuracy") && v > 0.5) it->setForeground(QColor(ui::colors::POSITIVE())); else if ((score_keys[c] == "r2_score" || score_keys[c] == "accuracy") && v < 0) it->setForeground(QColor(ui::colors::NEGATIVE())); table->setItem(r, c + 1, it); } } table->setRowHeight(r, 26); } results_layout_->addWidget(table); } status_label_->setText(tr("Supervised %1: best=%2 (%3 algos)").arg(problem).arg(best).arg(rows.size())); rendered = true; } // ── 6. UNSUPERVISED LEARNING ──────────────────────────────────────────── else if (command == "unsupervised_learning") { const int n_samples = meta.value("n_samples").toInt(); const int n_feat = meta.value("n_features").toInt(); QList top = { gs_make_card(tr("SAMPLES"), QString::number(n_samples), this), gs_make_card(tr("FEATURES"), QString::number(n_feat), this), gs_make_card(tr("METHODS"), QString::number(value_obj.size()), this), gs_make_card(tr("DATA SCALED"), meta.value("data_scaled").toBool() ? tr("YES") : tr("NO"), this), }; results_layout_->addWidget(gs_card_row(top, this)); // PCA section if (value_obj.contains("pca") && value_obj.value("pca").isObject()) { const QJsonObject pca = value_obj.value("pca").toObject(); results_layout_->addWidget(gs_section_header(tr("PCA"), accent)); const QJsonArray evr = pca.value("explained_variance_ratio").toArray(); const QJsonArray cum = pca.value("cumulative_variance").toArray(); const int n95 = pca.value("components_for_95_variance").toInt(); const double total = evr.isEmpty() ? 0.0 : cum[cum.size() - 1].toDouble(); QList pca_cards = { gs_make_card(tr("COMPONENTS"), QString::number(evr.size()), this), gs_make_card(tr("FOR 95% VAR"), QString::number(n95), this, ui::colors::POSITIVE()), gs_make_card( tr("TOP COMPONENT"), evr.isEmpty() ? QStringLiteral("—") : QString::number(evr[0].toDouble() * 100, 'f', 2) + "%", this), gs_make_card(tr("CUM. EXPLAINED"), QString::number(total * 100, 'f', 2) + "%", this), }; results_layout_->addWidget(gs_card_row(pca_cards, this)); if (!evr.isEmpty()) { const int rows = std::min(10, evr.size()); auto* table = new QTableWidget(rows, 3, this); table->setHorizontalHeaderLabels({tr("PC"), tr("Explained Var"), tr("Cumulative")}); table->verticalHeader()->setVisible(false); table->setEditTriggers(QAbstractItemView::NoEditTriggers); table->horizontalHeader()->setSectionResizeMode(QHeaderView::Stretch); table->setStyleSheet(table_ss()); table->setMaximumHeight(rows * 24 + 32); for (int i = 0; i < rows; ++i) { table->setItem(i, 0, new QTableWidgetItem(QString("PC%1").arg(i + 1))); auto* e = new QTableWidgetItem(QString::number(evr[i].toDouble() * 100, 'f', 3) + "%"); e->setTextAlignment(Qt::AlignRight | Qt::AlignVCenter); table->setItem(i, 1, e); auto* c = new QTableWidgetItem(QString::number(cum[i].toDouble() * 100, 'f', 3) + "%"); c->setTextAlignment(Qt::AlignRight | Qt::AlignVCenter); table->setItem(i, 2, c); table->setRowHeight(i, 24); } results_layout_->addWidget(table); } } // K-Means section if (value_obj.contains("kmeans") && value_obj.value("kmeans").isObject()) { const QJsonObject km = value_obj.value("kmeans").toObject(); results_layout_->addWidget(gs_section_header(tr("K-MEANS"), accent)); const int opt_k = km.value("optimal_k").toInt(); const QJsonArray sil = km.value("silhouette_by_k").toArray(); const QJsonArray ine = km.value("inertia_by_k").toArray(); double best_sil = 0.0; for (const auto& v : sil) { const auto pair = v.toArray(); if (pair.size() >= 2) best_sil = std::max(best_sil, pair[1].toDouble()); } const QJsonArray labels = km.value("cluster_labels").toArray(); QList km_cards = { gs_make_card(tr("OPTIMAL k"), QString::number(opt_k), this, ui::colors::POSITIVE()), gs_make_card(tr("BEST SILHOUETTE"), gs_fmt_num(best_sil, 4), this, best_sil >= 0.5 ? ui::colors::POSITIVE() : ui::colors::WARNING()), gs_make_card(tr("k RANGE"), QString::number(sil.size()), this), gs_make_card(tr("OBSERVATIONS"), QString::number(labels.size()), this), }; results_layout_->addWidget(gs_card_row(km_cards, this)); if (!sil.isEmpty()) { auto* table = new QTableWidget(sil.size(), 3, this); table->setHorizontalHeaderLabels({tr("k"), tr("Silhouette"), tr("Inertia")}); table->verticalHeader()->setVisible(false); table->setEditTriggers(QAbstractItemView::NoEditTriggers); table->horizontalHeader()->setSectionResizeMode(QHeaderView::Stretch); table->setStyleSheet(table_ss()); table->setMaximumHeight(qMin(sil.size() * 24 + 32, 280)); for (int i = 0; i < sil.size(); ++i) { const auto sp = sil[i].toArray(); const auto ip = i < ine.size() ? ine[i].toArray() : QJsonArray(); const int k = sp.size() > 0 ? sp[0].toInt() : 0; const double s = sp.size() > 1 ? sp[1].toDouble() : 0.0; const double in_v = ip.size() > 1 ? ip[1].toDouble() : 0.0; auto* k_it = new QTableWidgetItem(QString::number(k)); if (k == opt_k) k_it->setForeground(QColor(ui::colors::POSITIVE())); table->setItem(i, 0, k_it); auto* s_it = new QTableWidgetItem(QString::number(s, 'f', 4)); s_it->setTextAlignment(Qt::AlignRight | Qt::AlignVCenter); if (k == opt_k) s_it->setForeground(QColor(ui::colors::POSITIVE())); table->setItem(i, 1, s_it); auto* in_it = new QTableWidgetItem(QString::number(in_v, 'f', 2)); in_it->setTextAlignment(Qt::AlignRight | Qt::AlignVCenter); table->setItem(i, 2, in_it); table->setRowHeight(i, 24); } results_layout_->addWidget(table); } } // Hierarchical (simple) if (value_obj.contains("hierarchical") && value_obj.value("hierarchical").isObject()) { const QJsonObject h = value_obj.value("hierarchical").toObject(); QList h_cards = { gs_make_card(tr("HIERARCHICAL k"), QString::number(h.value("n_clusters").toInt()), this), gs_make_card(tr("LABELS"), QString::number(h.value("cluster_labels").toArray().size()), this), }; results_layout_->addWidget(gs_card_row(h_cards, this)); } status_label_->setText( tr("Unsupervised: %1 method(s) on %2×%3").arg(value_obj.size()).arg(n_samples).arg(n_feat)); rendered = true; } // ── 7. MODEL EVALUATION ───────────────────────────────────────────────── else if (command == "model_evaluation") { const QJsonObject params = result.value("parameters").toObject(); const QString model = params.value("model_type").toString().toUpper(); const QString problem = params.value("problem_type").toString().toUpper(); const int folds = params.value("cv_folds").toInt(); const double mean_cv = meta.value("mean_cv_score").toDouble(); const double std_cv = meta.value("std_cv_score").toDouble(); const double overfit = meta.value("overfitting_score").toDouble(); const bool is_overfit = meta.value("is_overfitting").toBool(); const double final_train = meta.value("final_train_score").toDouble(); const double final_val = meta.value("final_val_score").toDouble(); QList top = { gs_make_card(tr("MODEL"), model.isEmpty() ? QStringLiteral("—") : model, this), gs_make_card(tr("PROBLEM"), problem, this), gs_make_card(tr("CV FOLDS"), QString::number(folds), this), gs_make_card(tr("MEAN CV ± σ"), QString("%1 ± %2").arg(mean_cv, 0, 'f', 4).arg(std_cv, 0, 'f', 4), this, mean_cv > 0.5 ? ui::colors::POSITIVE() : mean_cv > 0 ? ui::colors::WARNING() : ui::colors::NEGATIVE()), }; results_layout_->addWidget(gs_card_row(top, this)); QList overfit_cards = { gs_make_card(tr("FINAL TRAIN"), gs_fmt_num(final_train, 4), this), gs_make_card(tr("FINAL VAL"), gs_fmt_num(final_val, 4), this, final_val > 0.5 ? ui::colors::POSITIVE() : ui::colors::WARNING()), gs_make_card(tr("OVERFIT GAP"), gs_fmt_num(overfit, 4), this, overfit > 0.1 ? ui::colors::NEGATIVE() : ui::colors::POSITIVE()), gs_make_card(tr("OVERFITTING"), is_overfit ? tr("YES") : tr("NO"), this, is_overfit ? ui::colors::NEGATIVE() : ui::colors::POSITIVE()), }; results_layout_->addWidget(gs_card_row(overfit_cards, this)); // CV scores + learning curves table const QJsonArray cv = value_obj.value("cv_scores").toArray(); const QJsonArray train_s = value_obj.value("train_scores").toArray(); const QJsonArray val_s = value_obj.value("val_scores").toArray(); if (!cv.isEmpty() || !train_s.isEmpty()) { auto* table = new QTableWidget(std::max(cv.size(), train_s.size()), 3, this); table->setHorizontalHeaderLabels({tr("Idx"), tr("CV / Train Score"), tr("Val Score")}); table->verticalHeader()->setVisible(false); table->setEditTriggers(QAbstractItemView::NoEditTriggers); table->horizontalHeader()->setSectionResizeMode(QHeaderView::Stretch); table->setStyleSheet(table_ss()); table->setMaximumHeight(qMin(table->rowCount() * 24 + 32, 280)); const int rows = table->rowCount(); for (int i = 0; i < rows; ++i) { table->setItem(i, 0, new QTableWidgetItem(QString::number(i))); if (i < cv.size()) { auto* it = new QTableWidgetItem(tr("%1 (cv)").arg(QString::number(cv[i].toDouble(), 'f', 4))); it->setTextAlignment(Qt::AlignRight | Qt::AlignVCenter); table->setItem(i, 1, it); } else if (i < train_s.size()) { auto* it = new QTableWidgetItem(tr("%1 (tr)").arg(QString::number(train_s[i].toDouble(), 'f', 4))); it->setTextAlignment(Qt::AlignRight | Qt::AlignVCenter); table->setItem(i, 1, it); } if (i < val_s.size()) { const double v = val_s[i].toDouble(); auto* it = new QTableWidgetItem(QString::number(v, 'f', 4)); it->setTextAlignment(Qt::AlignRight | Qt::AlignVCenter); if (v > 0.5) it->setForeground(QColor(ui::colors::POSITIVE())); else if (v < 0) it->setForeground(QColor(ui::colors::NEGATIVE())); table->setItem(i, 2, it); } table->setRowHeight(i, 24); } results_layout_->addWidget(table); } status_label_->setText(tr("%1: CV %2±%3 Overfit: %4") .arg(model) .arg(mean_cv, 0, 'f', 3) .arg(std_cv, 0, 'f', 3) .arg(is_overfit ? tr("YES") : tr("NO"))); rendered = true; } // ── 8. RESAMPLING METHODS ─────────────────────────────────────────────── else if (command == "resampling_methods") { const double orig = meta.value("original_statistic").toDouble(); const int data_n = meta.value("data_size").toInt(); const QString stat_name = meta.value("statistic_name").toString(); QList top = { gs_make_card(tr("STATISTIC"), stat_name.isEmpty() ? tr("mean") : stat_name, this), gs_make_card(tr("ORIGINAL VALUE"), gs_fmt_num(orig, 6), this, gs_pos_neg_color(orig)), gs_make_card(tr("DATA SIZE"), QString::number(data_n), this), gs_make_card(tr("METHODS"), QString::number(value_obj.size()), this), }; results_layout_->addWidget(gs_card_row(top, this)); // Bootstrap if (value_obj.contains("bootstrap") && value_obj.value("bootstrap").isObject()) { const QJsonObject b = value_obj.value("bootstrap").toObject(); results_layout_->addWidget(gs_section_header(tr("BOOTSTRAP"), accent)); const auto ci = b.value("confidence_interval").toArray(); const QString ci_str = ci.size() == 2 ? QString("[%1, %2]").arg(ci[0].toDouble(), 0, 'f', 4).arg(ci[1].toDouble(), 0, 'f', 4) : QStringLiteral("—"); QList b_cards = { gs_make_card(tr("MEAN"), gs_fmt_num(b.value("mean").toDouble(), 6), this), gs_make_card(tr("STD"), gs_fmt_num(b.value("std").toDouble(), 6), this), gs_make_card(tr("BIAS"), gs_fmt_num(b.value("bias").toDouble(), 6), this, gs_pos_neg_color(b.value("bias").toDouble())), gs_make_card(tr("95% CI"), ci_str, this, ui::colors::INFO()), }; results_layout_->addWidget(gs_card_row(b_cards, this)); } // Jackknife if (value_obj.contains("jackknife") && value_obj.value("jackknife").isObject()) { const QJsonObject j = value_obj.value("jackknife").toObject(); results_layout_->addWidget(gs_section_header(tr("JACKKNIFE"), accent)); QList j_cards = { gs_make_card(tr("BIAS"), gs_fmt_num(j.value("bias").toDouble(), 6), this, gs_pos_neg_color(j.value("bias").toDouble())), gs_make_card(tr("BIAS-CORRECTED"), gs_fmt_num(j.value("bias_corrected_estimate").toDouble(), 6), this), gs_make_card(tr("VARIANCE"), gs_fmt_num(j.value("variance_estimate").toDouble(), 6), this), gs_make_card(tr("STD ERROR"), gs_fmt_num(j.value("std_error").toDouble(), 6), this), }; results_layout_->addWidget(gs_card_row(j_cards, this)); } // Permutation if (value_obj.contains("permutation") && value_obj.value("permutation").isObject()) { const QJsonObject p = value_obj.value("permutation").toObject(); results_layout_->addWidget(gs_section_header(tr("PERMUTATION"), accent)); const double pv = p.value("p_value").toDouble(); const bool sig = p.value("significant").toBool(); QList p_cards = { gs_make_card(tr("p-VALUE"), gs_fmt_num(pv, 4), this, pv < 0.05 ? ui::colors::POSITIVE() : ui::colors::WARNING()), gs_make_card(tr("SIGNIFICANT"), sig ? tr("YES") : tr("NO"), this, sig ? ui::colors::POSITIVE() : ui::colors::WARNING()), gs_make_card(tr("RESAMPLES"), QString::number(p.value("resampled_statistics").toArray().size()), this), gs_make_card(tr("ORIGINAL"), gs_fmt_num(orig, 6), this), }; results_layout_->addWidget(gs_card_row(p_cards, this)); } status_label_->setText( tr("Resampling: original=%1 (%2 method(s))").arg(orig, 0, 'g', 4).arg(value_obj.size())); rendered = true; } // ── 9. SAMPLING TECHNIQUES ────────────────────────────────────────────── else if (command == "sampling_techniques") { const QJsonObject params = result.value("parameters").toObject(); const int pop_n = params.value("population_size").toInt(); const int sample_n = params.value("sample_size").toInt(); const double frac = meta.value("sampling_fraction").toDouble(); QList top = { gs_make_card(tr("POPULATION"), QString::number(pop_n), this), gs_make_card(tr("SAMPLE SIZE"), QString::number(sample_n), this), gs_make_card(tr("SAMPLING FRAC"), QString::number(frac * 100, 'f', 2) + "%", this), gs_make_card(tr("METHODS"), QString::number(value_obj.size()), this), }; results_layout_->addWidget(gs_card_row(top, this)); // Per-method comparison table — pick the first numeric column from each method's stats if (!value_obj.isEmpty()) { QList> methods; for (auto it = value_obj.begin(); it != value_obj.end(); ++it) if (it.value().isObject()) methods.append({it.key(), it.value().toObject()}); auto* table = new QTableWidget(methods.size(), 5, this); table->setHorizontalHeaderLabels( {tr("Method"), tr("Sample N"), tr("Sample Mean"), tr("Pop Mean"), tr("Bias")}); table->verticalHeader()->setVisible(false); table->setEditTriggers(QAbstractItemView::NoEditTriggers); table->horizontalHeader()->setSectionResizeMode(QHeaderView::Stretch); table->setStyleSheet(table_ss()); table->setMaximumHeight(qMin(methods.size() * 26 + 32, 240)); for (int r = 0; r < methods.size(); ++r) { const auto& name = methods[r].first; const auto& m = methods[r].second; table->setItem(r, 0, new QTableWidgetItem(name.toUpper())); if (m.contains("error")) { auto* e = new QTableWidgetItem(tr("ERR: %1").arg(m.value("error").toString())); e->setForeground(QColor(ui::colors::NEGATIVE())); table->setItem(r, 1, e); table->setSpan(r, 1, 1, 4); } else { table->setItem(r, 1, new QTableWidgetItem(QString::number(m.value("sample_size").toInt()))); auto first_val = [](const QJsonObject& o) -> double { if (o.isEmpty()) return 0.0; return o.begin().value().toDouble(); }; const double sm = first_val(m.value("sample_mean").toObject()); const double pm = first_val(m.value("population_mean").toObject()); const double bias = first_val(m.value("bias").toObject()); auto add_n = [&](int col, double v, bool pos_neg = false) { auto* it = new QTableWidgetItem(QString::number(v, 'f', 4)); it->setTextAlignment(Qt::AlignRight | Qt::AlignVCenter); if (pos_neg) it->setForeground(QColor(gs_pos_neg_color(v))); table->setItem(r, col, it); }; add_n(2, sm); add_n(3, pm); add_n(4, bias, true); } table->setRowHeight(r, 26); } results_layout_->addWidget(table); } status_label_->setText(tr("Sampling: %1 method(s) pop=%2 n=%3 (%4%)") .arg(value_obj.size()) .arg(pop_n) .arg(sample_n) .arg(frac * 100, 0, 'f', 1)); rendered = true; } // ── 10. CENTRAL LIMIT THEOREM ─────────────────────────────────────────── else if (command == "central_limit_theorem") { const QJsonObject params = result.value("parameters").toObject(); const QString dist = params.value("population_dist").toString().toUpper(); const double pmean = meta.value("population_mean").toDouble(); const double pstd = meta.value("population_std").toDouble(); const int pop_n = meta.value("population_size").toInt(); QList top = { gs_make_card(tr("DISTRIBUTION"), dist.isEmpty() ? QStringLiteral("—") : dist, this), gs_make_card(tr("POP MEAN"), gs_fmt_num(pmean, 4), this), gs_make_card(tr("POP STD"), gs_fmt_num(pstd, 4), this), gs_make_card(tr("POP SIZE"), QString::number(pop_n), this), }; results_layout_->addWidget(gs_card_row(top, this)); // Per-sample-size table QList> rows; for (auto it = value_obj.begin(); it != value_obj.end(); ++it) { if (!it.value().isObject()) continue; const QString k = it.key(); // "n_5", "n_30", etc const int n = k.startsWith("n_") ? k.mid(2).toInt() : 0; rows.append({n, it.value().toObject()}); } std::sort(rows.begin(), rows.end(), [](const auto& a, const auto& b) { return a.first < b.first; }); if (!rows.isEmpty()) { auto* table = new QTableWidget(rows.size(), 7, this); table->setHorizontalHeaderLabels({tr("n"), tr("Theor. Mean"), tr("Theor. SE"), tr("Empir. Mean"), tr("Empir. SE"), tr("Mean Bias"), tr("Normal?")}); table->verticalHeader()->setVisible(false); table->setEditTriggers(QAbstractItemView::NoEditTriggers); table->horizontalHeader()->setSectionResizeMode(QHeaderView::Stretch); table->setStyleSheet(table_ss()); table->setMaximumHeight(qMin(rows.size() * 26 + 32, 320)); for (int r = 0; r < rows.size(); ++r) { const auto& o = rows[r].second; table->setItem(r, 0, new QTableWidgetItem(QString::number(rows[r].first))); auto add = [&](int col, double v, int dec) { auto* it = new QTableWidgetItem(QString::number(v, 'f', dec)); it->setTextAlignment(Qt::AlignRight | Qt::AlignVCenter); table->setItem(r, col, it); }; add(1, o.value("theoretical_mean").toDouble(), 4); add(2, o.value("theoretical_std").toDouble(), 4); add(3, o.value("empirical_mean").toDouble(), 4); add(4, o.value("empirical_std").toDouble(), 4); const double mb = o.value("mean_bias").toDouble(); auto* mbi = new QTableWidgetItem(QString::number(mb, 'f', 4)); mbi->setTextAlignment(Qt::AlignRight | Qt::AlignVCenter); mbi->setForeground(QColor(gs_pos_neg_color(mb))); table->setItem(r, 5, mbi); const QJsonValue norm = o.value("appears_normal"); QString norm_str = norm.isNull() ? QStringLiteral("—") : (norm.toBool() ? tr("YES") : tr("NO")); auto* ni = new QTableWidgetItem(norm_str); ni->setTextAlignment(Qt::AlignCenter); if (!norm.isNull()) ni->setForeground(QColor(norm.toBool() ? ui::colors::POSITIVE() : ui::colors::WARNING())); table->setItem(r, 6, ni); table->setRowHeight(r, 26); } results_layout_->addWidget(table); } status_label_->setText(tr("CLT %1: μ=%2 σ=%3 (%4 sample sizes)") .arg(dist) .arg(pmean, 0, 'f', 3) .arg(pstd, 0, 'f', 3) .arg(rows.size())); rendered = true; } // ── 11. SAMPLING ERROR ANALYSIS ───────────────────────────────────────── else if (command == "sampling_error_analysis") { const int rec_n = meta.value("recommended_sample_size").toInt(); const double tgt_moe = meta.value("target_margin_of_error").toDouble(); const double zc = meta.value("z_critical").toDouble(); const double conf = meta.value("coverage_target").toDouble(); QList top = { gs_make_card(tr("RECOMMENDED n"), QString::number(rec_n), this, ui::colors::POSITIVE()), gs_make_card(tr("TARGET MoE"), gs_fmt_num(tgt_moe, 4), this), gs_make_card(tr("Z CRITICAL"), gs_fmt_num(zc, 4), this), gs_make_card(tr("CONFIDENCE"), QString::number(conf * 100, 'f', 1) + "%", this, ui::colors::INFO()), }; results_layout_->addWidget(gs_card_row(top, this)); QList> rows; for (auto it = value_obj.begin(); it != value_obj.end(); ++it) { if (!it.value().isObject()) continue; const auto o = it.value().toObject(); rows.append({o.value("sample_size").toInt(), o}); } std::sort(rows.begin(), rows.end(), [](const auto& a, const auto& b) { return a.first < b.first; }); if (!rows.isEmpty()) { auto* table = new QTableWidget(rows.size(), 6, this); table->setHorizontalHeaderLabels( {tr("n"), tr("Theor. SE"), tr("Empir. SE"), tr("MoE"), tr("Coverage"), tr("MSE")}); table->verticalHeader()->setVisible(false); table->setEditTriggers(QAbstractItemView::NoEditTriggers); table->horizontalHeader()->setSectionResizeMode(QHeaderView::Stretch); table->setStyleSheet(table_ss()); table->setMaximumHeight(qMin(rows.size() * 26 + 32, 320)); for (int r = 0; r < rows.size(); ++r) { const auto& o = rows[r].second; table->setItem(r, 0, new QTableWidgetItem(QString::number(rows[r].first))); auto add = [&](int col, double v, int dec) { auto* it = new QTableWidgetItem(QString::number(v, 'f', dec)); it->setTextAlignment(Qt::AlignRight | Qt::AlignVCenter); table->setItem(r, col, it); }; add(1, o.value("theoretical_se").toDouble(), 6); add(2, o.value("empirical_se").toDouble(), 6); add(3, o.value("theoretical_moe").toDouble(), 6); const double cov = o.value("coverage_probability").toDouble(); auto* ci = new QTableWidgetItem(QString::number(cov * 100, 'f', 2) + "%"); ci->setTextAlignment(Qt::AlignRight | Qt::AlignVCenter); ci->setForeground(QColor(cov >= conf - 0.02 ? ui::colors::POSITIVE() : ui::colors::WARNING())); table->setItem(r, 4, ci); add(5, o.value("mean_squared_error").toDouble(), 6); table->setRowHeight(r, 26); } results_layout_->addWidget(table); } status_label_->setText( tr("Sampling error: recommended n=%1 target MoE=%2").arg(rec_n).arg(tgt_moe, 0, 'f', 4)); rendered = true; } // ── 12. VALIDATE DATA ─────────────────────────────────────────────────── else if (command == "validate_data") { const double score = result.value("quality_score").toDouble(); const QString name = result.value("data_name").toString(); const auto issues = result.value("issues").toArray(); const auto warnings = result.value("warnings").toArray(); const auto recs = result.value("recommendations").toArray(); const auto stats_obj = result.value("statistics").toObject(); QList top = { gs_make_card(tr("QUALITY SCORE"), tr("%1 / 100").arg(QString::number(score, 'f', 1)), this, score >= 80 ? ui::colors::POSITIVE() : score >= 50 ? ui::colors::WARNING() : ui::colors::NEGATIVE()), gs_make_card(tr("DATA NAME"), name.isEmpty() ? tr("data") : name, this), gs_make_card(tr("ISSUES"), QString::number(issues.size()), this, issues.isEmpty() ? ui::colors::POSITIVE() : ui::colors::NEGATIVE()), gs_make_card(tr("WARNINGS"), QString::number(warnings.size()), this, warnings.isEmpty() ? ui::colors::POSITIVE() : ui::colors::WARNING()), }; results_layout_->addWidget(gs_card_row(top, this)); // Statistics table (if present) if (!stats_obj.isEmpty()) { QStringList sk; QHash sv; for (auto it = stats_obj.begin(); it != stats_obj.end(); ++it) { if (!it.value().isObject() && !it.value().isArray()) { sk << it.key(); sv.insert(it.key(), it.value()); } } if (!sk.isEmpty()) { results_layout_->addWidget(gs_section_header(tr("STATISTICS"), accent)); auto* table = new QTableWidget(sk.size(), 2, this); table->setHorizontalHeaderLabels({tr("Statistic"), tr("Value")}); table->verticalHeader()->setVisible(false); table->setEditTriggers(QAbstractItemView::NoEditTriggers); table->horizontalHeader()->setSectionResizeMode(QHeaderView::Stretch); table->setStyleSheet(table_ss()); table->setMaximumHeight(qMin(sk.size() * 24 + 32, 280)); for (int r = 0; r < sk.size(); ++r) { QString lbl = sk[r]; lbl.replace('_', ' '); table->setItem(r, 0, new QTableWidgetItem(lbl.toUpper())); auto* vi = new QTableWidgetItem(format_val(sv.value(sk[r]))); vi->setTextAlignment(Qt::AlignRight | Qt::AlignVCenter); table->setItem(r, 1, vi); table->setRowHeight(r, 24); } results_layout_->addWidget(table); } } // Issues / warnings / recommendations sections auto add_section = [this, accent](const QString& title, const QJsonArray& items, bool issue_shape, const QString& bar_color) { if (items.isEmpty()) return; auto* hdr = new QLabel(tr("%1 (%2)").arg(title).arg(items.size())); hdr->setStyleSheet(QString("color:%1; font-weight:700; font-size:10px; letter-spacing:1px;" "padding:6px 10px; background:%2; border-left:3px solid %3;") .arg(ui::colors::TEXT_PRIMARY(), ui::colors::BG_SURFACE(), bar_color)); results_layout_->addWidget(hdr); for (const auto& v : items) { QString text; QString fg = ui::colors::TEXT_SECONDARY(); if (issue_shape && v.isObject()) { const auto o = v.toObject(); const QString sev = o.value("severity").toString().toUpper(); if (sev == "CRITICAL" || sev == "HIGH") fg = ui::colors::NEGATIVE(); else if (sev == "MEDIUM") fg = ui::colors::WARNING(); text = tr("[%1] %2: %3") .arg(sev.isEmpty() ? QStringLiteral("—") : sev, o.value("type").toString(), o.value("description").toString(o.value("message").toString())); } else { text = QStringLiteral("• ") + v.toString(); } auto* item = new QLabel(text); item->setWordWrap(true); item->setStyleSheet(QString("color:%1; padding:2px 12px;").arg(fg)); results_layout_->addWidget(item); } }; add_section(tr("ISSUES"), issues, true, ui::colors::NEGATIVE()); add_section(tr("WARNINGS"), warnings, true, ui::colors::WARNING()); add_section(tr("RECOMMENDATIONS"), recs, false, ui::colors::INFO()); status_label_->setText(tr("Validate %1: score=%2 issues=%3 warnings=%4") .arg(name) .arg(score, 0, 'f', 1) .arg(issues.size()) .arg(warnings.size())); rendered = true; } // ── Fallback for unknown commands: flat metrics table ─────────────────── if (!rendered) { QStringList keys; QHash kv; if (value_obj.isEmpty() && !value_raw.isNull() && !value_raw.isObject() && !value_raw.isArray()) { keys.append("value"); kv.insert("value", value_raw); } else { cfa_flatten_scalars(value_obj, "value", keys, kv); } cfa_flatten_scalars(meta, "metadata", keys, kv); if (!keys.isEmpty()) { auto* table = new QTableWidget(keys.size(), 2, this); table->setHorizontalHeaderLabels({tr("Metric"), tr("Value")}); table->setEditTriggers(QAbstractItemView::NoEditTriggers); table->horizontalHeader()->setSectionResizeMode(QHeaderView::Stretch); table->verticalHeader()->setVisible(false); table->setMaximumHeight(qMin(keys.size() * 28 + 32, 480)); table->setStyleSheet(table_ss()); for (int r = 0; r < keys.size(); ++r) { QString label = keys[r]; label.replace('_', ' '); table->setItem(r, 0, new QTableWidgetItem(label.toUpper())); auto* vi = new QTableWidgetItem(format_val(kv.value(keys[r]))); vi->setTextAlignment(Qt::AlignRight | Qt::AlignVCenter); table->setItem(r, 1, vi); } results_layout_->addWidget(table); } } // ── Raw JSON viewer (collapsed by default — capped height) ────────────── auto* raw_hdr = new QLabel(tr("RAW JSON")); raw_hdr->setStyleSheet(QString("color:%1; font-size:9px; font-weight:700; letter-spacing:1px; margin-top:4px;") .arg(ui::colors::TEXT_TERTIARY())); results_layout_->addWidget(raw_hdr); auto* raw = new QTextEdit; raw->setReadOnly(true); raw->setMaximumHeight(180); raw->setPlainText(QJsonDocument(payload).toJson(QJsonDocument::Indented)); raw->setStyleSheet(output_ss()); results_layout_->addWidget(raw); // ── Export button ─────────────────────────────────────────────────────── auto* export_btn = new QPushButton(tr("EXPORT RESULTS")); export_btn->setCursor(Qt::PointingHandCursor); export_btn->setFixedHeight(28); export_btn->setStyleSheet(QString("QPushButton { background:transparent; color:%1; border:1px solid %2; " "font-size:%3px; font-family:%4; padding:0 12px; }" "QPushButton:hover { color:%1; background:rgba(255,255,255,0.05); }") .arg(module_.color.name(), ui::colors::BORDER_DIM()) .arg(ui::fonts::SMALL) .arg(ui::fonts::DATA_FAMILY)); const QString json_str = QJsonDocument(payload).toJson(QJsonDocument::Indented); connect(export_btn, &QPushButton::clicked, this, [this, json_str, command]() { QString safe_name = module_.label + "_" + command; safe_name.replace(QRegularExpression("[^a-zA-Z0-9_\\-]"), "_"); const QString stored_name = safe_name + "_" + QString::number(QDateTime::currentMSecsSinceEpoch()) + ".json"; const QString dest = services::FileManagerService::instance().storage_dir() + "/" + stored_name; QFile f(dest); if (f.open(QIODevice::WriteOnly | QIODevice::Text)) { f.write(json_str.toUtf8()); f.close(); services::FileManagerService::instance().register_file(stored_name, module_.label + "_" + command + ".json", QFileInfo(dest).size(), "application/json", "ai_quant_lab"); LOG_INFO("AIQuantLab", "Exported result: " + stored_name); } }); results_layout_->addWidget(export_btn); if (status_label_->text() == tr("Running...")) status_label_->setText(tr("Done")); } } // namespace fincept::screens