leptonica ASAN fixes

.gitignore

@@ -7,3 +7,4 @@ test_images
 bulk
 tests/bulk_test.cc
 CMakebkp.txt
+.vscode

src/conversion.cc

@@ -58,8 +58,8 @@ void createPDF(const std::string &input_path,
     delete renderer;
 }
 
-auto extractTextFromImageFileLeptonica(const std::string &file_path,
-                                       const std::string &lang) -> std::string {
+auto extractTextFromImageFileLeptonica(const std::string &file_path, const std::string &lang)
+    -> std::string {
     auto *api = new tesseract::TessBaseAPI();
     if (api->Init(nullptr, "eng") != 0) {
         fprintf(stderr, "Could not initialize tesseract.\n");
@@ -70,10 +70,12 @@ auto extractTextFromImageFileLeptonica(const std::string &file_path,
     // fully automatic - suitable for single columns of text
 
     api->SetPageSegMode(tesseract::PSM_AUTO);
-
     api->SetImage(image);
-    std::string outText(api->GetUTF8Text());
-    outText = api->GetUTF8Text();
+
+    // Get the text from the image.
+    char       *rawText = api->GetUTF8Text();
+    std::string outText(rawText);
+    delete[] rawText; // Free the memory allocated by GetUTF8Text
 
     api->End();
     delete api;
@@ -90,8 +92,12 @@ auto extractTextLSTM(const std::string &file_path, const std::string &lang) -> s
     Pix *image = pixRead(file_path.c_str());
 
     api->SetImage(image);
-    std::string outText(api->GetUTF8Text());
-    outText = api->GetUTF8Text();
+
+    // Get the text from the image.
+    char       *rawText = api->GetUTF8Text();
+    std::string outText(rawText);
+
+    delete[] rawText; // Free the memory allocated by GetUTF8Text
 
     api->End();
     delete api;

tests/ocr_test.cc

@@ -1,5 +1,6 @@
 #include <conversion.h>
 #include <curl/curl.h>
+#include <filesystem>
 #include <fs.h>
 #include <gtest/gtest.h>
 #include <gtest/internal/gtest-port.h>
@@ -49,14 +50,19 @@ TEST_F(ImageProcessingTests, EnvironmentTest) {
 TEST_F(ImageProcessingTests, ConvertSingleImageToTextFile) {
     imgstr::ImgProcessor imageTranslator;
 
-    imageTranslator.convertImageToTextFile(fpaths[0], tempDir);
+    std::string imagePath = fpaths[0];
 
-    auto filename = llvm::SmallString<256>(llvm::sys::path::filename(fpaths[0]));
+    imageTranslator.convertImageToTextFile(imagePath, tempDir);
+
+    llvm::SmallString<256> filename(llvm::sys::path::filename(imagePath));
 
     llvm::sys::path::replace_extension(filename, ".txt");
 
-    auto fname = tempDir + "/" + filename;
-    llvm::outs() << "Single Image Test File: " << fpaths[0] << ", LF: " << fname << '\n';
+    //    auto fname = tempDir + "/" + filename;
+
+    std::string fname = tempDir + "/" + std::string(filename.c_str());
+
+    llvm::outs() << "Single Image Test File: " << imagePath << ", LF: " << fname << '\n';
 
     bool fileExists = llvm::sys::fs::exists(tempDir + "/" + filename);