Merge pull request #102 from sakuli/feature/(sakuli/sakuli-ocr#7)/add…

…-ocr-to-docker

Feature/(sakuli/sakuli ocr#7)/add ocr to docker

.test/e2e/e2e-ocr/ocr/ocr.js

@@ -0,0 +1,16 @@
+(async () => {
+  const testCase = new TestCase("OCR");
+  const url = "https://sakuli.io/e2e-pages/sandbox/#";
+  try {
+    await testCase.startStep("Navigate to website");
+    await _navigateTo(url);
+    await _wait(2000);
+
+    await testCase.startStep("find 'element in one place'");
+    await _getRegionByText("element in one place").highlight(1);
+  } catch (e) {
+    await testCase.handleException(e);
+  } finally {
+    await testCase.saveResult();
+  }
+})();

.test/e2e/e2e-ocr/package.json

@@ -0,0 +1,17 @@
+{
+  "name": "e2e-ocr",
+  "version": "2.3.0",
+  "description": "An OCR test suite",
+  "private": true,
+  "contributors": [
+    "Tim Keiner <tim.keiner@consol.de>",
+    "Simon Hofmann <simon.hofmann@consol.de>",
+    "Sven Hettwer <sven.hettwer@consol.de>",
+    "Cliff Le Van <cliff.le.van@consol.de>"
+  ],
+  "homepage": "https://github.com/sakuli/sakuli#readme",
+  "scripts": {
+    "test": "sakuli run .",
+    "test:debug": "node --inspect node_modules/@sakuli/cli/dist/index run ."
+  }
+}

.test/e2e/e2e-ocr/testsuite.properties

@@ -0,0 +1,4 @@
+testsuite.id=ocr
+testsuite.browser=chrome
+#log.level=debug
+selenium.chrome.arguments=--incognito --disable-dev-shm-usage --no-sandbox

.test/e2e/e2e-ocr/testsuite.suite

@@ -0,0 +1 @@
+ocr/ocr.js https://sakuli.io

.test/goss.yaml

@@ -60,6 +60,8 @@ package:
     installed: true
   dnsutils:
     installed: true
+  tesseract:
+    installed: true
 port:
   tcp:5901:
     listening: true
@@ -86,3 +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"

.test/test.sh

@@ -103,3 +103,11 @@ docker run \
     --shm-size=2G \
     taconsol/sakuli:${1:-latest}
 
+# OCR run
+docker run \
+    --rm \
+    -e SAKULI_LICENSE_KEY=${SAKULI_LICENSE_KEY} \
+    -e SAKULI_TEST_SUITE=/sakuli-project/e2e-ocr \
+    -v $(pwd)/e2e:/sakuli-project \
+    --shm-size=2G \
+    taconsol/sakuli:${1:-latest}

CHANGELOG.md

@@ -12,6 +12,7 @@ All notable changes to this project will be documented in this file.
 - Maintenance: Use node 14 in container  [(#82)](https://github.com/sakuli/sakuli-docker/issues/82)
 - Feature: Enable package installation [(#48)](https://github.com/sakuli/sakuli-docker/issues/48)
 - Enhancement: Link log files instead of copying them [(#98)](https://github.com/sakuli/sakuli-docker/issues/98)
+- Feature: Add ocr module to docker [(sakuli/sakuli-ocr#7)](https://github.com/sakuli/sakuli-ocr/issues/7)
 
 ## v2.4.0-2
 

Dockerfile.sakuli-base

@@ -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 \
@@ -68,7 +72,13 @@ RUN apt-get update && \
                     curl \
                     iputils-ping \
                     iputils-tracepath \
-                    dnsutils && \
+                    dnsutils \
+                    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/* && \
@@ -84,7 +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

run.sh

@@ -1,9 +1,7 @@
-#!/usr/bin/env bash
-
 docker run \
     --rm \
     -it \
     -e SAKULI_LICENSE_KEY=${SAKULI_LICENSE_KEY} \
-    -p 5902:5901 \
-    -p 6902:6901 \
+    -p 5901:5901 \
+    -p 6901:6901 \
     taconsol/sakuli

src/sakuli/sakuli.sh

@@ -24,6 +24,9 @@ do
     npm i -g -E @sakuli/forwarder-$fwd@$1
 done
 
+echo "Installing Sakuli OCR v$1"
+npm i -g -E @sakuli/ocr@$1
+
 echo "Installing TypeScript v3.8.3"
 npm i -g -E typescript@3.8.3