test stuff

build.gradle

@@ -19,6 +19,10 @@ dependencies {
 	
 	implementation 'org.apache.logging.log4j:log4j-core:2.20.0'
 	
+	// https://mvnrepository.com/artifact/org.apache.pdfbox/jbig2-imageio
+	implementation group: 'org.apache.pdfbox', name: 'jbig2-imageio', version: '3.0.4'
+	
+		
 	//general PDF
     implementation 'org.apache.pdfbox:pdfbox:2.0.27'
         

src/main/java/stirling/software/SPDF/controller/CompressController.java

@@ -69,7 +69,7 @@ public class CompressController {
         command.add(tempInputFile.toString());
         command.add(tempOutputFile.toString());
 
-        int returnCode = ProcessExecutor.runCommandWithOutputHandling(command);
+        int returnCode = ProcessExecutor.getInstance(ProcessExecutor.Processes.OCR_MY_PDF).runCommandWithOutputHandling(command);
         
         // Read the optimized PDF file
         byte[] pdfBytes = Files.readAllBytes(tempOutputFile);

src/main/java/stirling/software/SPDF/controller/OCRController.java

@@ -28,6 +28,7 @@ import org.springframework.web.servlet.ModelAndView;
 
 import stirling.software.SPDF.utils.ProcessExecutor;
 //import com.spire.pdf.*;
+import java.util.concurrent.Semaphore;
 @Controller
 public class OCRController {
 
@@ -41,11 +42,14 @@ public class OCRController {
 		return modelAndView;
 	}
 
+	private final Semaphore semaphore = new Semaphore(2);
+	
 	@PostMapping("/ocr-pdf")
 	public ResponseEntity<byte[]> processPdfWithOCR(@RequestParam("fileInput") MultipartFile inputFile,
 			@RequestParam("languages") List<String> selectedLanguages,
 			@RequestParam(name = "sidecar", required = false) Boolean sidecar) throws IOException, InterruptedException {
 
+
 		//--output-type pdfa
 		if (selectedLanguages == null || selectedLanguages.size() < 1) {
 			throw new IOException("Please select at least one language.");
@@ -60,18 +64,26 @@ public class OCRController {
 
 		// Run OCR Command
 	    String languageOption = String.join("+", selectedLanguages);
-	    List<String> command = new ArrayList<>(Arrays.asList("ocrmypdf","--verbose", "2", "--language", languageOption,
+	    
-	            tempInputFile.toString(), tempOutputFile.toString()));
+	    List<String> command = new ArrayList<>(Arrays.asList("ocrmypdf","--verbose", "2"));
+	    		
+	    
 	    String sidecarFile = tempOutputFile.toString().replace(".pdf", ".txt");
 	    if (sidecar != null && sidecar) {
 	        command.add("--sidecar");
 	        command.add(sidecarFile);
 	    }
-	    int returnCode = ProcessExecutor.runCommandWithOutputHandling(command);
+	    
-
+	    command.addAll(Arrays.asList("--language", languageOption,
+	            tempInputFile.toString(), tempOutputFile.toString()));
+	    
+	    //Run CLI command
+	    int returnCode = ProcessExecutor.getInstance(ProcessExecutor.Processes.OCR_MY_PDF).runCommandWithOutputHandling(command);
+        
 		// Read the OCR processed PDF file
 		byte[] pdfBytes = Files.readAllBytes(tempOutputFile);
-
+		
+	    
 		// Clean up the temporary files
 		Files.delete(tempInputFile);
 		// Return the OCR processed PDF as a response

src/main/java/stirling/software/SPDF/controller/converters/ConvertOfficeController.java

@@ -53,7 +53,7 @@ public byte[] convertToPdf(MultipartFile inputFile) throws IOException, Interrup
             "-o",
             tempOutputFile.toString(),
             tempInputFile.toString()));
-    int returnCode = ProcessExecutor.runCommandWithOutputHandling(command);
+    int returnCode = ProcessExecutor.getInstance(ProcessExecutor.Processes.LIBRE_OFFICE).runCommandWithOutputHandling(command);
 
     // Read the converted PDF file
     byte[] pdfBytes = Files.readAllBytes(tempOutputFile);

src/main/java/stirling/software/SPDF/utils/ProcessExecutor.java

@@ -6,60 +6,94 @@ import java.io.InputStreamReader;
 import java.nio.charset.StandardCharsets;
 import java.util.ArrayList;
 import java.util.List;
+import java.util.Map;
+import java.util.HashMap;
+import java.util.concurrent.Semaphore;
 public class ProcessExecutor {
-	public static int runCommandWithOutputHandling(List<String> command) throws IOException, InterruptedException {
+	
-	    ProcessBuilder processBuilder = new ProcessBuilder(command);
+	public enum Processes {
-	    Process process = processBuilder.start();
+        LIBRE_OFFICE,
+        OCR_MY_PDF
+    }
 
-	    // Read the error stream and standard output stream concurrently
+    private static final Map<Processes, ProcessExecutor> instances = new HashMap<>();
-	    List<String> errorLines = new ArrayList<>();
-	    List<String> outputLines = new ArrayList<>();
 
-	    Thread errorReaderThread = new Thread(() -> {
+    private final Semaphore semaphore;
-	        try (BufferedReader errorReader = new BufferedReader(new InputStreamReader(process.getErrorStream(), StandardCharsets.UTF_8))) {
-	            String line;
-	            while ((line = errorReader.readLine()) != null) {
-	                errorLines.add(line);
-	            }
-	        } catch (IOException e) {
-	            e.printStackTrace();
-	        }
-	    });
 
-	    Thread outputReaderThread = new Thread(() -> {
+    private ProcessExecutor(int semaphoreLimit) {
-	        try (BufferedReader outputReader = new BufferedReader(new InputStreamReader(process.getInputStream(), StandardCharsets.UTF_8))) {
+        this.semaphore = new Semaphore(semaphoreLimit);
-	            String line;
+    }
-	            while ((line = outputReader.readLine()) != null) {
-	                outputLines.add(line);
-	            }
-	        } catch (IOException e) {
-	            e.printStackTrace();
-	        }
-	    });
-
-	    errorReaderThread.start();
-	    outputReaderThread.start();
-
-	    // Wait for the conversion process to complete
-	    int exitCode = process.waitFor();
-
-	    // Wait for the reader threads to finish
-	    errorReaderThread.join();
-	    outputReaderThread.join();
-
-	    if (outputLines.size() > 0) {
-	        String outputMessage = String.join("\n", outputLines);
-	        System.out.println("Command output:\n" + outputMessage);
-	    }
-
-	    if (errorLines.size() > 0) {
-	        String errorMessage = String.join("\n", errorLines);
-	        System.out.println("Command error output:\n" + errorMessage);
-	        if (exitCode != 0) {
-	            throw new IOException("Command process failed with exit code " + exitCode + ". Error message: " + errorMessage);
-	        }
-	    }
 
+    public static ProcessExecutor getInstance(Processes processType) {
+        return instances.computeIfAbsent(processType, key -> {
+            int semaphoreLimit = switch (key) {
+                case LIBRE_OFFICE -> 2; 
+                case OCR_MY_PDF -> 2; 
+            };
+            return new ProcessExecutor(semaphoreLimit);
+        });
+    }
+    
+	public int runCommandWithOutputHandling(List<String> command) throws IOException, InterruptedException {
+		int exitCode = 1;
+		semaphore.acquire();
+		try {
+			
+		
+		    ProcessBuilder processBuilder = new ProcessBuilder(command);
+		    Process process = processBuilder.start();
+	
+		    // Read the error stream and standard output stream concurrently
+		    List<String> errorLines = new ArrayList<>();
+		    List<String> outputLines = new ArrayList<>();
+	
+		    Thread errorReaderThread = new Thread(() -> {
+		        try (BufferedReader errorReader = new BufferedReader(new InputStreamReader(process.getErrorStream(), StandardCharsets.UTF_8))) {
+		            String line;
+		            while ((line = errorReader.readLine()) != null) {
+		                errorLines.add(line);
+		            }
+		        } catch (IOException e) {
+		            e.printStackTrace();
+		        }
+		    });
+	
+		    Thread outputReaderThread = new Thread(() -> {
+		        try (BufferedReader outputReader = new BufferedReader(new InputStreamReader(process.getInputStream(), StandardCharsets.UTF_8))) {
+		            String line;
+		            while ((line = outputReader.readLine()) != null) {
+		                outputLines.add(line);
+		            }
+		        } catch (IOException e) {
+		            e.printStackTrace();
+		        }
+		    });
+	
+		    errorReaderThread.start();
+		    outputReaderThread.start();
+	
+		    // Wait for the conversion process to complete
+		    exitCode = process.waitFor();
+	
+		    // Wait for the reader threads to finish
+		    errorReaderThread.join();
+		    outputReaderThread.join();
+	
+		    if (outputLines.size() > 0) {
+		        String outputMessage = String.join("\n", outputLines);
+		        System.out.println("Command output:\n" + outputMessage);
+		    }
+	
+		    if (errorLines.size() > 0) {
+		        String errorMessage = String.join("\n", errorLines);
+		        System.out.println("Command error output:\n" + errorMessage);
+		        if (exitCode != 0) {
+		            throw new IOException("Command process failed with exit code " + exitCode + ". Error message: " + errorMessage);
+		        }
+		    }
+		} finally {
+			semaphore.release();
+		}
 	    return exitCode;
 	}
    		

src/main/resources/application.properties

@@ -1,12 +1,12 @@
-spring.http.multipart.max-file-size=1GB
+spring.http.multipart.max-file-size=2GB
-spring.http.multipart.max-request-size=1GB
+spring.http.multipart.max-request-size=2GB
 
 multipart.enabled=true
-multipart.max-file-size=1000MB
+multipart.max-file-size=2000MB
-multipart.max-request-size=1000MB
+multipart.max-request-size=2000MB
 
-spring.servlet.multipart.max-file-size=1000MB
+spring.servlet.multipart.max-file-size=2000MB
-spring.servlet.multipart.max-request-size=1000MB
+spring.servlet.multipart.max-request-size=2000MB
 
 server.forward-headers-strategy=NATIVE
 

src/main/resources/templates/ocr-pdf.html

@@ -26,10 +26,30 @@
 							        </div>
 							    </div>
 							</div>
-						<!--  	<div class="form-group">
+						  	<div class="form-group">
 						        <input type="checkbox" class="form-check-input" name="sidecar" id="sidecar" />
 						        <label class="form-check-label" for="sidecar" th:text="#{ocr.selectText.2}"></label>
-						    </div> -->
+						    </div>
+						    <div class="form-group">
+						        <input type="checkbox" class="form-check-input" name="deskew" id="deskew" />
+						        <label class="form-check-label" for="deskew" th:text="#{ocr.selectText.2}"></label>
+						    </div>
+						    <div class="form-group">
+						        <input type="checkbox" class="form-check-input" name="clean" id="clean" />
+						        <label class="form-check-label" for="clean" th:text="#{ocr.selectText.2}"></label>
+						    </div>
+						    <div class="form-group">
+						        <input type="checkbox" class="form-check-input" name="clean-final" id="clean-final" />
+						        <label class="form-check-label" for="clean-final" th:text="#{ocr.selectText.2}"></label>
+						    </div>
+						    <div class="form-group">
+                                <label th:text="#{pdfToImage.selectText}"></label> 
+                                <select class="form-control" name="ocrType">
+                                    <option value="skip-text">Ignores pages that have interacive text on them, only OCRs pages that are images</option>
+                                    <option value="force-ocr">Force OCR, will OCR Every page removing all original text</option>
+                                    <option value="Normal">Normal (Will error if contains text)</option>
+                                </select>
+                            </div>
 				            <button type="submit" class="btn btn-primary" th:text="#{ocr.submit}"></button>
 				        </form>
 				        <p th:text="#{ocr.credit}"></p>