Merge pull request #2 from ANcpLua/chore/modernize-temp-paths

chore: modernize temp-path handling, harden ProcessRunner + Tesseract output

Author: ANcpLua

CreatePdf.NET.Tests/DocumentTests.cs

@@ -143,14 +143,20 @@ public async Task SaveAsync_WithFilename_CreatesFile()
     {
         var doc = Pdf.Create();
         doc.AddText("Test content");
-        var filename = Path.Combine(Path.GetTempPath(), $"{Guid.NewGuid():N}.pdf");
+        var tempDir = Directory.CreateTempSubdirectory("createpdf-doc-test-");
+        var filename = Path.Combine(tempDir.FullName, "document.pdf");
 
-        var path = await doc.SaveAsync(filename).ConfigureAwait(true);
-
-        path.Should().EndWith(".pdf");
-        File.Exists(path).Should().BeTrue();
+        try
+        {
+            var path = await doc.SaveAsync(filename).ConfigureAwait(true);
 
-        File.Delete(path);
+            path.Should().EndWith(".pdf");
+            File.Exists(path).Should().BeTrue();
+        }
+        finally
+        {
+            tempDir.Delete(recursive: true);
+        }
     }
 
     [Fact]

CreatePdf.NET.Tests/OcrServiceTests.cs

@@ -66,22 +66,23 @@ public async Task ProcessPdfStreamAsync_HandlesCleanupErrorsGracefully()
     }
 
     [Fact]
-    public void TryDeleteFile_ExistingFile_DeletesSuccessfully()
+    public void TryDeleteDirectory_ExistingDirectoryWithContents_DeletesRecursively()
     {
-        var tempFile = Path.GetTempFileName();
-        File.Exists(tempFile).Should().BeTrue();
+        var tempDir = Directory.CreateTempSubdirectory("createpdf-ocr-test-");
+        File.WriteAllText(Path.Combine(tempDir.FullName, "leftover.txt"), "x");
 
-        OcrService.TryDeleteFile(tempFile);
+        OcrService.TryDeleteDirectory(tempDir.FullName);
 
-        File.Exists(tempFile).Should().BeFalse();
+        Directory.Exists(tempDir.FullName).Should().BeFalse();
     }
 
     [Fact]
-    public void TryDeleteFile_NonExistentFile_DoesNotThrow()
+    public void TryDeleteDirectory_NonExistentDirectory_DoesNotThrow()
     {
-        var nonExistentFile = Path.Combine(Path.GetTempPath(), Guid.NewGuid().ToString());
+        var nonExistent = Path.Combine(Path.GetTempPath(), Guid.NewGuid().ToString("N"));
+
+        var act = () => OcrService.TryDeleteDirectory(nonExistent);
 
-        var act = () => OcrService.TryDeleteFile(nonExistentFile);
         act.Should().NotThrow();
     }
 }

CreatePdf.NET.Tests/RuntimeSystemEnvironmentTests.cs

@@ -17,14 +17,16 @@ public void Properties_MirrorUnderlyingEnvironment()
     [Fact]
     public void FileExists_DelegatesToFileSystem()
     {
-        var temp = Path.GetTempFileName();
+        var tempDir = Directory.CreateTempSubdirectory("createpdf-runtime-test-");
+        var temp = Path.Combine(tempDir.FullName, "probe.txt");
+        File.WriteAllText(temp, string.Empty);
         try
         {
             RuntimeSystemEnvironment.Instance.FileExists(temp).Should().BeTrue();
         }
         finally
         {
-            File.Delete(temp);
+            tempDir.Delete(recursive: true);
         }
 
         RuntimeSystemEnvironment.Instance.FileExists(temp).Should().BeFalse();

CreatePdf.NET/Internal/OcrService.cs

@@ -2,6 +2,8 @@ namespace CreatePdf.NET.Internal;
 
 internal sealed class OcrService
 {
+    private const string TempDirPrefix = "createpdf-ocr-";
+
     private readonly IOcrProvider _provider;
 
     public OcrService() : this(new TesseractOcrProvider())
@@ -16,49 +18,53 @@ internal OcrService(IOcrProvider provider)
     public async Task<string> ProcessPdfAsync(string pdfPath, OcrOptions options,
         CancellationToken cancellationToken = default)
     {
-        var tempDir = Path.GetTempPath();
-        var pdfName = Path.GetFileNameWithoutExtension(pdfPath);
-        var pngPath = Path.Combine(tempDir, $"{pdfName}_{Guid.NewGuid():N}.png");
-        var txtPath = Path.Combine(tempDir, $"{pdfName}_{Guid.NewGuid():N}.txt");
-
+        var workDir = Directory.CreateTempSubdirectory(TempDirPrefix);
         try
         {
-            await _provider.RasterizePdfToPngAsync(pdfPath, pngPath, options, cancellationToken).ConfigureAwait(false);
-            return await _provider.ExtractTextFromImageAsync(pngPath, txtPath, options, cancellationToken)
-                .ConfigureAwait(false);
+            return await OcrAsync(pdfPath, workDir.FullName, options, cancellationToken).ConfigureAwait(false);
         }
         finally
         {
-            TryDeleteFile(pngPath);
-            TryDeleteFile(txtPath);
+            TryDeleteDirectory(workDir.FullName);
         }
     }
 
     public async Task<string> ProcessPdfStreamAsync(Stream pdfStream, OcrOptions options,
         CancellationToken cancellationToken = default)
     {
-        var tempDir = Path.GetTempPath();
-        var pdfFileName = Path.ChangeExtension(Path.GetRandomFileName(), ".pdf");
-        var pdfPath = Path.Combine(tempDir, pdfFileName);
-
+        var workDir = Directory.CreateTempSubdirectory(TempDirPrefix);
         try
         {
-            await using (var fileStream = new FileStream(pdfPath, FileMode.CreateNew, FileAccess.Write, FileShare.None))
+            var pdfPath = Path.Combine(workDir.FullName, "input.pdf");
+            await using (var fileStream =
+                         new FileStream(pdfPath, FileMode.CreateNew, FileAccess.Write, FileShare.None))
             {
                 await pdfStream.CopyToAsync(fileStream, cancellationToken).ConfigureAwait(false);
             }
 
-            return await ProcessPdfAsync(pdfPath, options, cancellationToken).ConfigureAwait(false);
+            return await OcrAsync(pdfPath, workDir.FullName, options, cancellationToken).ConfigureAwait(false);
         }
         finally
         {
-            TryDeleteFile(pdfPath);
+            TryDeleteDirectory(workDir.FullName);
         }
     }
 
-    internal static void TryDeleteFile(string path)
+    private async Task<string> OcrAsync(string pdfPath, string workDir, OcrOptions options,
+        CancellationToken cancellationToken)
+    {
+        var pdfName = Path.GetFileNameWithoutExtension(pdfPath);
+        var pngPath = Path.Combine(workDir, $"{pdfName}.png");
+        var txtPath = Path.Combine(workDir, $"{pdfName}.txt");
+
+        await _provider.RasterizePdfToPngAsync(pdfPath, pngPath, options, cancellationToken).ConfigureAwait(false);
+        return await _provider.ExtractTextFromImageAsync(pngPath, txtPath, options, cancellationToken)
+            .ConfigureAwait(false);
+    }
+
+    internal static void TryDeleteDirectory(string path)
     {
-        if (File.Exists(path))
-            File.Delete(path);
+        if (Directory.Exists(path))
+            Directory.Delete(path, recursive: true);
     }
 }

CreatePdf.NET/Internal/ProcessRunner.cs

@@ -10,7 +10,9 @@ public async Task RunAsync(ProcessStartInfo startInfo, CancellationToken cancell
     {
         ArgumentNullException.ThrowIfNull(startInfo);
 
-        using var process = Process.Start(startInfo)!;
+        using var process = Process.Start(startInfo)
+            ?? throw new InvalidOperationException(
+                $"Process.Start returned null for '{startInfo.FileName}' — no new process was created.");
         await process.WaitForExitAsync(cancellationToken).ConfigureAwait(false);
     }
 }

CreatePdf.NET/Internal/TesseractOcrProvider.cs

@@ -50,7 +50,10 @@ await _processRunner.RunAsync(
     public async Task<string> ExtractTextFromImageAsync(string pngPath, string txtPath, OcrOptions options,
         CancellationToken cancellationToken = default)
     {
-        var outputBase = txtPath[..^4];
+        // Tesseract appends ".txt" itself, so strip the extension to give it the base path.
+        var outputBase = Path.Combine(
+            Path.GetDirectoryName(txtPath) ?? string.Empty,
+            Path.GetFileNameWithoutExtension(txtPath));
 
         await _processRunner.RunAsync(
                 CreateProcessInfo(