(sakuli/sakuli-ocr#7) Added tesseract 4.1.1 port for ubuntu 18.04

.test/goss.yaml

@@ -60,11 +60,7 @@ package:
     installed: true
   dnsutils:
     installed: true
-  tesseract-ocr:
-    installed: true
-  tesseract-ocr-deu:
-    installed: true
-  tesseract-ocr-eng:
+  tesseract:
     installed: true
 port:
   tcp:5901:
@@ -92,3 +88,14 @@ command:
     exec: "echo ${LOG_MODE}"
     stdout:
       - "/^ci$/"
+  tesseract-languages:
+    exit-status: 0
+    exec: "tesseract --list-langs"
+    stdout:
+      - "deu"
+      - "eng"
+  tesseract-version:
+    exit-status: 0
+    exec: "tesseract -v"
+    stdout:
+      - "tesseract 4.1.1"

Dockerfile.sakuli-base

@@ -1,6 +1,6 @@
 # This Dockerfile is used to build an headles vnc image based on Ubuntu
 # Used to run CI jobs for nut.js
-ARG BASE_IMAGE_VERSION=20.04
+ARG BASE_IMAGE_VERSION=18.04
 FROM ubuntu:$BASE_IMAGE_VERSION
 
 ARG NODE_VERSION
@@ -27,7 +27,8 @@ ENV REFRESHED_AT=${BUILD_DATE} \
     VNC_RESOLUTION=1280x1024 \
     VNC_PW=vncpassword \
     VNC_VIEW_ONLY=false \
-    IMG=taconsol/sakuli-base
+    IMG=taconsol/sakuli-base \
+    TESSDATA_PREFIX=/usr/local/share/tessdata
 
 EXPOSE $VNC_PORT $NO_VNC_PORT
 
@@ -37,6 +38,9 @@ WORKDIR $HOME
 COPY ./src/common/install/ $INST_SCRIPTS/
 RUN find $INST_SCRIPTS -name '*.sh' -exec chmod a+x {} +
 
+### Copy tesseract
+COPY ./src/tesseract /tmp/tesseract
+
 ### Install required packages and add configuration
 RUN apt-get update && \
     apt-get install -y \
@@ -60,6 +64,7 @@ RUN apt-get update && \
                     firefox-geckodriver \
                     openbox \
                     obconf \
+                    obmenu \
                     xterm \
                     x11-xkb-utils \
                     libnss-wrapper \
@@ -68,9 +73,12 @@ RUN apt-get update && \
                     iputils-ping \
                     iputils-tracepath \
                     dnsutils \
-                    tesseract-ocr \
-                    tesseract-ocr-deu \
-                    tesseract-ocr-eng && \
+                    libleptonica-dev \
+                    libpng-dev \
+                    libjpeg8-dev \
+                    libtiff5-dev \
+                    zlib1g-dev \
+                    libtool && \
     apt-get clean -y && \
     apt-get autoremove -y && \
     rm -rf /var/lib/apt/lists/* /var/cache/apt/* && \
@@ -86,7 +94,11 @@ RUN apt-get update && \
     chmod +x -v $NO_VNC_HOME/utils/*.sh && \
     ln -s $NO_VNC_HOME/vnc_lite.html $NO_VNC_HOME/index.html && \
     ln -s /usr/bin/chromium-browser /usr/bin/google-chrome && \
-    echo "CHROMIUM_FLAGS='--no-sandbox --user-data-dir'" > $HOME/.chromium-browser.init
+    echo "CHROMIUM_FLAGS='--no-sandbox --user-data-dir'" > $HOME/.chromium-browser.init && \
+    dpkg -i /tmp/tesseract/tesseract_4.1.1-1_amd64.deb && \
+    ldconfig && \
+    cp /tmp/tesseract/*traineddata /usr/local/share/tessdata && \
+    rm -rf /tmp/tesseract
 
 COPY ./src/common/wm/ $HOME/
 COPY ./src/common/config/openbox /etc/xdg/openbox