Closes #200: use tesseract command line

sismics · Mar 23, 2018 · abde9b7 · abde9b7
1 parent be1c2a7
commit abde9b7
Show file tree

Hide file tree

Showing 14 changed files with 67 additions and 1,368 deletions.
diff --git a/.travis.yml b/.travis.yml
@@ -17,8 +17,6 @@ after_success:
   - docker push $REPO
 env:
   global:
-  - TESSDATA_PREFIX=/usr/share/tesseract-ocr
-  - LC_NUMERIC=C
   - secure: LRGpjWORb0qy6VuypZjTAfA8uRHlFUMTwb77cenS9PPRBxuSnctC531asS9Xg3DqC5nsRxBBprgfCKotn5S8nBSD1ceHh84NASyzLSBft3xSMbg7f/2i7MQ+pGVwLncusBU6E/drnMFwZBleo+9M8Tf96axY5zuUp90MUTpSgt0=
   - secure: bCDDR6+I7PmSkuTYZv1HF/z98ANX/SFEESUCqxVmV5Gs0zFC0vQXaPJQ2xaJNRop1HZBFMZLeMMPleb0iOs985smpvK2F6Rbop9Tu+Vyo0uKqv9tbZ7F8Nfgnv9suHKZlL84FNeUQZJX6vsFIYPEJ/r7K5P/M0PdUy++fEwxEhU=
   - secure: ewXnzbkgCIHpDWtaWGMa1OYZJ/ki99zcIl4jcDPIC0eB3njX/WgfcC6i0Ke9mLqDqwXarWJ6helm22sNh+xtQiz6isfBtBX+novfRt9AANrBe3koCMUemMDy7oh5VflBaFNP0DVb8LSCnwf6dx6ZB5E9EB8knvk40quc/cXpGjY=

diff --git a/Dockerfile b/Dockerfile
@@ -4,9 +4,6 @@ MAINTAINER b.gamard@sismics.com
 RUN apt-get update && apt-get -y -q install ffmpeg mediainfo tesseract-ocr tesseract-ocr-fra tesseract-ocr-ita tesseract-ocr-kor tesseract-ocr-rus tesseract-ocr-ukr tesseract-ocr-spa tesseract-ocr-ara tesseract-ocr-hin tesseract-ocr-deu tesseract-ocr-pol tesseract-ocr-jpn tesseract-ocr-por tesseract-ocr-tha tesseract-ocr-jpn tesseract-ocr-chi-sim tesseract-ocr-chi-tra && \
     apt-get clean && rm -rf /var/lib/apt/lists/*
 
-ENV TESSDATA_PREFIX /usr/share/tesseract-ocr/4.00/
-ENV LC_NUMERIC C
-
 # Remove the embedded javax.mail jar from Jetty
 RUN rm -f /opt/jetty/lib/jndi/javax.mail.glassfish-*.jar
 

diff --git a/docs-core/pom.xml b/docs-core/pom.xml
@@ -138,46 +138,41 @@
       <artifactId>bcprov-jdk15on</artifactId>
     </dependency>
 
-		<dependency>
-			<groupId>fr.opensagres.xdocreport</groupId>
-			<artifactId>org.odftoolkit.odfdom.converter.pdf</artifactId>
-		</dependency>
-
-		<dependency>
-			<groupId>fr.opensagres.xdocreport</groupId>
-			<artifactId>org.apache.poi.xwpf.converter.pdf</artifactId>
-		</dependency>
-
     <dependency>
-      <groupId>net.java.dev.jna</groupId>
-      <artifactId>jna</artifactId>
+      <groupId>fr.opensagres.xdocreport</groupId>
+      <artifactId>org.odftoolkit.odfdom.converter.pdf</artifactId>
     </dependency>
 
-    <!-- ImageIO plugins -->
     <dependency>
-      <groupId>com.levigo.jbig2</groupId>
-      <artifactId>levigo-jbig2-imageio</artifactId>
-    </dependency> 
+      <groupId>fr.opensagres.xdocreport</groupId>
+      <artifactId>org.apache.poi.xwpf.converter.pdf</artifactId>
+    </dependency>
 
+    <!-- ImageIO plugins -->
     <dependency>
       <groupId>com.twelvemonkeys.imageio</groupId>
       <artifactId>imageio-jpeg</artifactId>
     </dependency>
-    
+
     <dependency>
-      <groupId>com.github.jai-imageio</groupId>
-      <artifactId>jai-imageio-core</artifactId>
+      <groupId>com.twelvemonkeys.imageio</groupId>
+      <artifactId>imageio-tiff</artifactId>
     </dependency>
 
     <dependency>
       <groupId>com.github.jai-imageio</groupId>
       <artifactId>jai-imageio-jpeg2000</artifactId>
     </dependency>
 
+    <dependency>
+      <groupId>com.levigo.jbig2</groupId>
+      <artifactId>levigo-jbig2-imageio</artifactId>
+    </dependency>
+
+    <!-- Only for connecting to PostgreSQL database -->
     <dependency>
       <groupId>org.postgresql</groupId>
       <artifactId>postgresql</artifactId>
-      <version>42.2.2.jre7</version>
     </dependency>
 
     <!-- Test dependencies -->

diff --git a/docs-core/src/main/java/com/sismics/docs/core/util/ActionUtil.java b/docs-core/src/main/java/com/sismics/docs/core/util/ActionUtil.java
@@ -5,6 +5,7 @@
 import com.sismics.docs.core.util.action.Action;
 import com.sismics.docs.core.util.action.AddTagAction;
 import com.sismics.docs.core.util.action.RemoveTagAction;
+import org.slf4j.Logger;
 import org.slf4j.LoggerFactory;
 
 import javax.json.JsonObject;
@@ -18,7 +19,7 @@ public class ActionUtil {
     /**
      * Logger.
      */
-    private static final org.slf4j.Logger log = LoggerFactory.getLogger(LuceneUtil.class);
+    private static final Logger log = LoggerFactory.getLogger(ActionUtil.class);
 
     /**
      * Find the action associated to an action type.

diff --git a/docs-core/src/main/java/com/sismics/docs/core/util/FileUtil.java b/docs-core/src/main/java/com/sismics/docs/core/util/FileUtil.java
@@ -1,44 +1,40 @@
 package com.sismics.docs.core.util;
 
+import com.google.common.base.Charsets;
 import com.google.common.base.Strings;
+import com.google.common.collect.Lists;
+import com.google.common.io.CharStreams;
 import com.sismics.docs.core.constant.Constants;
 import com.sismics.docs.core.dao.jpa.FileDao;
 import com.sismics.docs.core.dao.jpa.UserDao;
 import com.sismics.docs.core.event.DocumentUpdatedAsyncEvent;
 import com.sismics.docs.core.event.FileCreatedAsyncEvent;
 import com.sismics.docs.core.model.jpa.File;
 import com.sismics.docs.core.model.jpa.User;
-import com.sismics.tess4j.Tesseract;
 import com.sismics.util.ImageDeskew;
 import com.sismics.util.Scalr;
 import com.sismics.util.context.ThreadLocalContext;
+import com.sismics.util.io.InputStreamReaderThread;
 import com.sismics.util.mime.MimeTypeUtil;
 import org.apache.commons.lang.StringUtils;
-import org.slf4j.Logger;
-import org.slf4j.LoggerFactory;
 
 import javax.crypto.Cipher;
 import javax.crypto.CipherInputStream;
+import javax.imageio.ImageIO;
 import java.awt.image.BufferedImage;
 import java.io.IOException;
 import java.io.InputStream;
+import java.io.InputStreamReader;
 import java.nio.file.Files;
 import java.nio.file.Path;
-import java.util.Collections;
-import java.util.HashSet;
-import java.util.Set;
+import java.util.*;
 
 /**
  * File entity utilities.
  * 
  * @author bgamard
  */
 public class FileUtil {
-    /**
-     * Logger.
-     */
-    private static final Logger log = LoggerFactory.getLogger(FileUtil.class);
-
     /**
      * File ID of files currently being processed.
      */
@@ -50,28 +46,30 @@ public class FileUtil {
      * @param language Language to OCR
      * @param image Buffered image
      * @return Content extracted
+     * @throws Exception e
      */
-    public static String ocrFile(String language, BufferedImage image) {
+    public static String ocrFile(String language, BufferedImage image) throws Exception {
         // Upscale, grayscale and deskew the image
-        String content = null;
         BufferedImage resizedImage = Scalr.resize(image, Scalr.Method.AUTOMATIC, Scalr.Mode.AUTOMATIC, 3500, Scalr.OP_ANTIALIAS, Scalr.OP_GRAYSCALE);
         image.flush();
         ImageDeskew imageDeskew = new ImageDeskew(resizedImage);
         BufferedImage deskewedImage = Scalr.rotate(resizedImage, - imageDeskew.getSkewAngle(), Scalr.OP_ANTIALIAS, Scalr.OP_GRAYSCALE);
         resizedImage.flush();
-        image = deskewedImage;
+        Path tmpFile = ThreadLocalContext.get().createTemporaryFile();
+        ImageIO.write(deskewedImage, "tiff", tmpFile.toFile());
 
-        // OCR the file
-        try {
-            Tesseract instance = Tesseract.getInstance();
-            log.info("Starting OCR with TESSDATA_PREFIX=" + System.getenv("TESSDATA_PREFIX") + ";LC_NUMERIC=" + System.getenv("LC_NUMERIC"));
-            instance.setLanguage(language);
-            content = instance.doOCR(image);
-        } catch (Throwable e) {
-            log.error("Error while OCR-izing the image", e);
-        }
+        List<String> result = Lists.newLinkedList(Arrays.asList("tesseract", tmpFile.toAbsolutePath().toString(), "stdout", "-l", language));
+        ProcessBuilder pb = new ProcessBuilder(result);
+        Process process = pb.start();
 
-        return content;
+        // Consume the process error stream
+        final String commandName = pb.command().get(0);
+        new InputStreamReaderThread(process.getErrorStream(), commandName).start();
+
+        // Consume the data as text
+        try (InputStream is = process.getInputStream()) {
+            return CharStreams.toString(new InputStreamReader(is, Charsets.UTF_8));
+        }
     }
 
     /**

diff --git a/docs-core/src/main/java/com/sismics/docs/core/util/format/ImageFormatHandler.java b/docs-core/src/main/java/com/sismics/docs/core/util/format/ImageFormatHandler.java
@@ -17,7 +17,6 @@
 
 import javax.imageio.ImageIO;
 import java.awt.image.BufferedImage;
-import java.io.IOException;
 import java.io.InputStream;
 import java.nio.file.Files;
 import java.nio.file.Path;
@@ -45,19 +44,16 @@ public boolean accept(String mimeType) {
     }
 
     @Override
-    public BufferedImage generateThumbnail(Path file) throws IOException {
+    public BufferedImage generateThumbnail(Path file) throws Exception {
         try (InputStream inputStream = Files.newInputStream(file)) {
             return ImageIO.read(inputStream);
         }
     }
 
     @Override
-    public String extractContent(String language, Path file) {
+    public String extractContent(String language, Path file) throws Exception {
         try (InputStream inputStream = Files.newInputStream(file)) {
             return FileUtil.ocrFile(language, ImageIO.read(inputStream));
-        } catch (IOException e) {
-            log.error("Error reading the image", e);
-            return null;
         }
     }
 

diff --git a/docs-core/src/main/java/com/sismics/docs/core/util/format/PdfFormatHandler.java b/docs-core/src/main/java/com/sismics/docs/core/util/format/PdfFormatHandler.java
@@ -58,6 +58,7 @@ public String extractContent(String language, Path file) {
                  PDDocument pdfDocument = PDDocument.load(inputStream)) {
                 PDFRenderer renderer = new PDFRenderer(pdfDocument);
                 for (int pageIndex = 0; pageIndex < pdfDocument.getNumberOfPages(); pageIndex++) {
+                    log.info("OCR page " + (pageIndex + 1) + "/" + pdfDocument.getNumberOfPages() + " of PDF file containing only images");
                     sb.append(" ");
                     sb.append(FileUtil.ocrFile(language, renderer.renderImage(pageIndex)));
                 }