基于Qt框架，使用C++调用本地Python.exe执行深度学习脚本，生成热图和预测结果（QT源码）

#include "../include/python_deep_learning.h" #include <QCryptographicHash> #include <QFileDialog> #include <QProcess> #include "../include/mainwindow.h" #include "ui_python_deep_learning.h" using namespace enum_class_python_deep_learning; Python_deep_learning::Python_deep_learning(QWidget *parent, int model, QFont *fontt) : QMainWindow(parent), ui(new Ui::Python_deep_learning), deep_learning_model(model), font(*fontt) { ui->setupUi(this); if (model == Model_categories::Example_1) { this->setWindowTitle("Example_1"); } else if (model == Model_categories::Example_2) { this->setWindowTitle("Example_2"); } } Python_deep_learning::~Python_deep_learning() { delete ui; } void Python_deep_learning::on_btn_switch_model_clicked() { if (deep_learning_model == Model_categories::Example_1) { this->setWindowTitle("Example_2"); deep_learning_model = Model_categories::Example_2; } else if (deep_learning_model == Model_categories::Example_2) { this->setWindowTitle("Example_1"); deep_learning_model = Model_categories::Example_1; } ui->upload_image->clear(); ui->heatmap->clear(); ui->predict_label->clear(); ui->upload_image->setText("上传图片"); } void Python_deep_learning::on_btn_select_image_clicked() { QString directory = QFileDialog::getOpenFileName(this, "选择图片", "./", "Image (*.jpg *.jpeg *.png)"); if (!directory.isEmpty()) { // 获取QLabel的当前大小 QSize labelSize = ui->upload_image->size(); // 使用scaled()方法调整图像的大小 QPixmap pixmap(directory); // 确保图像按比例缩放来适应 QLabel 的尺寸 pixmap = pixmap.scaled(labelSize, Qt::KeepAspectRatio, Qt::SmoothTransformation); // 将缩放后的图像设置到 QLabel ui->upload_image->setPixmap(pixmap); imagePath = directory; ui->heatmap->clear(); ui->predict_label->clear(); } } // 计算文件的哈希值 QString Python_deep_learning::calculateFileHash(const QString &filePath) { QFile file(filePath); if (!file.open(QIODevice::ReadOnly)) { return QString(); } QCryptographicHash hash(QCryptographicHash::Sha256); // 使用 SHA-256 哈希算法 if (hash.addData(&file)) { return hash.result().toHex(); } return QString(); } // 比较 tempPath 和 resourcePath 对应的文件是否相同 bool Python_deep_learning::areFilesIdentical(const QString &tempPath, const QString &resourcePath) { // 获取 tempPath 文件的哈希值 QString tempFileHash = calculateFileHash(tempPath); if (tempFileHash.isEmpty()) { return false; // 文件无法读取或哈希计算失败 } // 获取 resourcePath 文件的哈希值 QFile resourceFile(resourcePath); if (!resourceFile.open(QIODevice::ReadOnly)) { return false; // 资源文件无法读取 } // 获取资源文件的哈希值 QByteArray resourceData = resourceFile.readAll(); QCryptographicHash hash(QCryptographicHash::Sha256); hash.addData(resourceData); QString resourceFileHash = hash.result().toHex(); // 比较两个哈希值 return tempFileHash == resourceFileHash; } // 提取文件复制函数 bool Python_deep_learning::copyFileIfNotExists(const QString &srcDir, const QString &srcName, const QString &dstPath) { if (srcName.isEmpty()) { return false; } QString srcPath = srcDir + srcName; if (QFile::exists(dstPath)) { bool result = areFilesIdentical(srcPath, dstPath); if (!result) { int counts = 1; QString tempPath = "Backup_" + QString::number(counts) + dstPath; while (QFile::exists(tempPath)) { ++counts; tempPath = "Backup_" + QString::number(counts) + dstPath; } QFile::copy(dstPath, tempPath); } else { return true; } } QFile file(srcPath); if (file.open(QIODevice::ReadOnly)) { QFile localFile(dstPath); if (localFile.open(QIODevice::WriteOnly)) { localFile.write(file.readAll()); qDebug() << srcPath << "copied to" << dstPath; return true; } } return false; } // 提取Python脚本执行函数 QString Python_deep_learning::executePythonScript(const QStringList &arguments, const QString &workingDir) { QProcess process; process.setProgram("D:/python.exe"); process.setArguments(arguments); process.setWorkingDirectory(workingDir); process.start(); if (!process.waitForFinished()) { qDebug() << "Python script execution failed!"; return QString(); } return process.readAllStandardOutput().trimmed(); } void Python_deep_learning::on_btn_predict_clicked() { if (ui->upload_image->pixmap().isNull()) { return; } // 设置模型目录和资源文件 QString targetDir_name; QString modelWeightName, modelName, dropblockName, datasetName, modulesName, heatmapScriptName, predictScriptName, preGCNScriptName, dataName; QString modelWeightPath, modelPath, dropblockPath, datasetPath, modulesPath, heatmapScriptPath, predictScriptPath, preGCNScriptPath, dataPath; if (deep_learning_model == Model_categories::Example_1) { modelWeightName = ".pth"; modelName = ".py"; dropblockName = ".py"; modulesName = ".py"; heatmapScriptName = ".py"; predictScriptName = ".py"; preGCNScriptName = ".py"; targetDir_name = "./data/Example_1/"; } else if (deep_learning_model == Model_categories::Example_2) { modelWeightName = ".pth"; modelName = ".py"; datasetName = ".py"; modulesName = ".py"; heatmapScriptName = ".py"; predictScriptName = ".py"; preGCNScriptName = ".py"; dataName = ".csv"; targetDir_name = "./data/Example_2/"; } modelWeightPath = targetDir_name + modelWeightName; modelPath = targetDir_name + modelName; dropblockPath = targetDir_name + dropblockName; datasetPath = targetDir_name + datasetName; modulesPath = targetDir_name + modulesName; heatmapScriptPath = targetDir_name + heatmapScriptName; predictScriptPath = targetDir_name + predictScriptName; preGCNScriptPath = targetDir_name + preGCNScriptName; dataPath = targetDir_name + dataName; // 创建目标目录 QDir dir; if (!dir.exists(targetDir_name)) { dir.mkpath(targetDir_name); } // 复制资源文件 copyFileIfNotExists(":/", modelWeightName, modelWeightPath); copyFileIfNotExists(":/resources/", modelName, modelPath); copyFileIfNotExists(":/resources/", dropblockName, dropblockPath); copyFileIfNotExists(":/resources/", datasetName, datasetPath); copyFileIfNotExists(":/resources/", modulesName, modulesPath); copyFileIfNotExists(":/resources/", heatmapScriptName, heatmapScriptPath); copyFileIfNotExists(":/resources/", predictScriptName, predictScriptPath); copyFileIfNotExists(":/resources/", preGCNScriptName, preGCNScriptPath); copyFileIfNotExists(":/", dataName, dataPath); // 执行 heatmap 脚本 QString output = executePythonScript({heatmapScriptName, "--r_heatmap", imagePath}, targetDir_name); if (output.isEmpty()) return; // 显示 heatmap 图像 QPixmap heatmapPixmap(targetDir_name + output); QSize labelSize = ui->heatmap->size(); heatmapPixmap = heatmapPixmap.scaled(labelSize, Qt::KeepAspectRatio, Qt::SmoothTransformation); ui->heatmap->setPixmap(heatmapPixmap); // 执行 predict 脚本 output = executePythonScript({predictScriptName, "--r", imagePath}, targetDir_name); if (output.isEmpty()) return; // 显示预测结果图像 QPixmap predictPixmap(t