fix: crash on tesseract during some ocr process

MTES-MCT · Oct 16, 2024 · ba20cd9 · ba20cd9
1 parent 51a96da
commit ba20cd9
Show file tree

Hide file tree

Showing 5 changed files with 18 additions and 21 deletions.
diff --git a/...rocess-file/src/main/java/fr/dossierfacile/process/file/amqp/AnalyzeDocumentReceiver.java b/...rocess-file/src/main/java/fr/dossierfacile/process/file/amqp/AnalyzeDocumentReceiver.java
@@ -33,10 +33,6 @@ public void startConsumer() {
 
     private void receiveDocument() {
         try {
-            if (!MemoryUtils.hasEnoughAvailableMemory()) {
-                log.warn("There is not currently enough memory to perform consumption ");
-                return;
-            }
             queueMessageService.consume(QueueName.QUEUE_DOCUMENT_ANALYSIS,
                     documentAnalysisDelay,
                     documentAnalysisTimeout,

diff --git a/...le-process-file/src/main/java/fr/dossierfacile/process/file/amqp/AnalyzeFileReceiver.java b/...le-process-file/src/main/java/fr/dossierfacile/process/file/amqp/AnalyzeFileReceiver.java
@@ -31,10 +31,6 @@ public void startConsumer() {
 
     private void receiveFile() {
         try {
-            if (!MemoryUtils.hasEnoughAvailableMemory()) {
-                log.warn("There is not currently enough memory to perform consumption ");
-                return;
-            }
             queueMessageService.consume(QueueName.QUEUE_FILE_ANALYSIS,
                     0,
                     fileAnalysisTimeout,

diff --git a/...ile-process-file/src/main/java/fr/dossierfacile/process/file/amqp/MinifyFileReceiver.java b/...ile-process-file/src/main/java/fr/dossierfacile/process/file/amqp/MinifyFileReceiver.java
@@ -33,10 +33,6 @@ public void startConsumer() {
 
     private void receiveFile() {
         try {
-            if (!MemoryUtils.hasEnoughAvailableMemory()) {
-                log.warn("There is not currently enough memory to perform consumption ");
-                return;
-            }
             queueMessageService.consume(QueueName.QUEUE_FILE_MINIFY,
                     0,
                     fileMinifyTimeout,

diff --git a/...ile/src/main/java/fr/dossierfacile/process/file/service/parsers/AbstractImagesParser.java b/...ile/src/main/java/fr/dossierfacile/process/file/service/parsers/AbstractImagesParser.java
@@ -3,6 +3,7 @@
 import fr.dossierfacile.common.entity.ocr.ParsedFile;
 import fr.dossierfacile.common.utils.FileUtility;
 import fr.dossierfacile.process.file.service.parsers.tools.PageExtractorModel;
+import fr.dossierfacile.process.file.util.MemoryUtils;
 import lombok.extern.slf4j.Slf4j;
 import net.sourceforge.tess4j.ITessAPI;
 import net.sourceforge.tess4j.Tesseract;
@@ -35,6 +36,7 @@ void init() {
             this.tesseract.setVariable("user_defined_dpi", "300");
         }
     }
+
     private BufferedImage[] getImages(File file) throws IOException {
         if ("pdf".equalsIgnoreCase(FilenameUtils.getExtension(file.getName()))) {
             BufferedImage[] images = FileUtility.convertPdfToImage(file);
@@ -76,7 +78,12 @@ protected T parse(BufferedImage... images) {
                         // rectangle exceeds image size
                         return null;
                     }
-                    extractedTexts.put(entry.getKey(), this.tesseract.doOCR(image, rect));
+                    MemoryUtils.logAvailableMemory(250);
+                    String text;
+                    synchronized (this) {
+                        text = this.tesseract.doOCR(image, rect);
+                    }
+                    extractedTexts.put(entry.getKey(), text);
                 }
                 T result = getResultFromExtraction(extractedTexts);
                 if (result != null) {
@@ -112,8 +119,11 @@ protected boolean modelMatches(PageExtractorModel model, BufferedImage image) th
                             if (image.getWidth() < (zone.rect().x + zone.rect().width)
                                     || image.getHeight() < (zone.rect().y + zone.rect().height))
                                 return false;
-
-                            String text = this.tesseract.doOCR(image, zone.rect());
+                            MemoryUtils.logAvailableMemory(250);
+                            String text;
+                            synchronized (this) {
+                                text = this.tesseract.doOCR(image,  zone.rect());
+                            }
                             log.debug("expected: " + zone.regexp() + " actual: " + text + "b=" + (text != null && text.trim().matches(zone.regexp())));
                             return text != null && text.trim().matches(zone.regexp());
                         } catch (Exception e) {

diff --git a/dossierfacile-process-file/src/main/java/fr/dossierfacile/process/file/util/MemoryUtils.java b/dossierfacile-process-file/src/main/java/fr/dossierfacile/process/file/util/MemoryUtils.java
@@ -4,13 +4,12 @@
 
 @Slf4j
 public class MemoryUtils {
-    // TODO currently we only log low availability
-    public static boolean hasEnoughAvailableMemory() {
+    public static int INT_1024X1024 = 1048576;
+
+    public static void logAvailableMemory(int thresholdInMB) {
         Runtime runtime = Runtime.getRuntime();
-        // Arbitrary choose 250MB as minimal requirement to perform
-        if (runtime.maxMemory() - runtime.totalMemory() + runtime.freeMemory() < 262144000) {
-            log.warn("Memory usage: (Total=" + runtime.totalMemory() / 1024 + " MB , max=" + runtime.maxMemory() / 1024 + " MB , free=" + runtime.freeMemory() / 1024 + " MB , avail=" + (runtime.maxMemory() - runtime.totalMemory() + runtime.freeMemory()));
+        if (runtime.maxMemory() - runtime.totalMemory() + runtime.freeMemory() < (long) thresholdInMB * INT_1024X1024) {
+            log.warn("Memory usage: (Total=" + runtime.totalMemory() / INT_1024X1024 + " MB , max=" + runtime.maxMemory() / INT_1024X1024 + " MB , free=" + runtime.freeMemory() / INT_1024X1024 + " MB , avail=" + ((runtime.maxMemory() - runtime.totalMemory() + runtime.freeMemory()) / INT_1024X1024));
         }
-        return true;
     }
 }