Tika App build for command line (with pdfbox jai-imageio sqlite-jdbc ocr)

Build(Tika 1.24.1)

• Install openjdk11
• Install Maven3
• Download tika full src
• Edit tika-parsers/pom.xml and remove the following scopes for jai-imageio & sqlite-jdbc

<scope>test</scope>
<scope>provided</scope>

• Edit tika-parsers/src/main/java/org/apache/tika/parser/ocr/TesseractOCRParser.java, comment everything in checkInitialization
• mvn install -Dossindex.fail=false -DskipTests=true
• jar output in tika-app/target/tika-app-*.jar
• On the server apt-get install tesseract-ocr tesseract-ocr-eng tesseract-ocr-fra ...(and all the other useful languages, simply put tesseract-ocr-all for all the available languages)

Exec

Default (no ocr nor extractInlineImages)

java -jar tika-app.jar --config=default-config.xml --help
java -jar tika-app.jar --config=default-config.xml --text /path/to/doc.pdf

OCR (ocr with extractInlineImages)

java -jar tika-app.jar --config=ocr-config.xml --help
java -jar tika-app.jar --config=ocr-config.xml --text /path/to/doc.pdf

https://cwiki.apache.org/confluence/display/tika/TikaOCR