Dockerization + Fix for Tesseract 3.03

Dockerfile

@@ -0,0 +1,10 @@
+FROM sismics/debian-java7-jetty9
+MAINTAINER benjamin.gam@gmail.com
+
+RUN apt-get -y -q install tesseract-ocr tesseract-ocr-fra
+
+ENV TESSDATA_PREFIX /usr/share/tesseract-ocr
+ENV LC_NUMERIC C
+
+ADD docs-web/target/docs-web-*.war /opt/jetty/webapps/docs.war
+ADD docs.xml /opt/jetty/webapps/docs.xml

build.sh

@@ -0,0 +1,2 @@
+#!/bin/sh
+docker build -t sismics/docs .

docs-core/src/main/java/com/sismics/tess4j/Tesseract.java

@@ -53,7 +53,7 @@ public class Tesseract {
     private static Tesseract instance;
     private final static Rectangle EMPTY_RECTANGLE = new Rectangle();
     private String language = "eng";
-    private String datapath = "tessdata";
+    private String datapath = null;
     private int psm = TessAPI.TessPageSegMode.PSM_AUTO;
     private boolean hocr;
     private int pageNum;

docs-docker-data/Dockerfile

@@ -0,0 +1,5 @@
+FROM busybox
+VOLUME ["/var/docs"]
+ADD data /var/docs/
+RUN chown -R 1001:1001 /var/docs
+CMD ["/bin/sh"]

docs-docker-data/build.sh

@@ -0,0 +1,3 @@
+#!/bin/sh
+
+docker build -t sismics/docs_data .

docs-docker-data/run.sh

@@ -0,0 +1,3 @@
+#!/bin/sh
+docker rm -f sismics_docs_data
+docker run --name sismics_docs_data sismics/docs_data

docs.xml

@@ -0,0 +1,10 @@
+<Configure class="org.eclipse.jetty.webapp.WebAppContext">
+  <Set name="contextPath">/</Set>
+  <Set name="war"><SystemProperty name="jetty.data" default="."/>/webapps/docs.war</Set>
+
+  <Call class="java.lang.System" name="setProperty">
+    <Arg>docs.home</Arg>
+    <Arg>/var/docs</Arg>
+  </Call>
+
+</Configure>

run-service.sh

@@ -0,0 +1,7 @@
+#!/bin/bash
+docker rm -f sismics_docs
+docker run \
+    -d --name=sismics_docs --restart=always \
+    --volumes-from=sismics_docs_data \
+    -e 'VIRTUAL_HOST_SECURE=docs.sismics.com' -e 'VIRTUAL_PORT=80' \
+    sismics/docs:latest