mu-search-tika-backend

This is the Tika backend for mu-search. It is to be used in conjunction with that microservice. Please see the README of mu-search in order to use this component.

This image is based on Tika and includes the components which are commonly expected to be available for mu-search.

Tutorials

Add Tika to your stack

Add the Tika service to docker-compose.yml

services:
  tika:
    image: semtech/mu-search-tika-backend:1.0.0

Restart the stack using docker compose up -d. The tika service will be created.

How-to guides

How to customize the Tika config

Generate a default Tika config

mu script tika generate-config

Mount the generated config folder in /config

services:
  tika:
    image: semtech/mu-search-tika-backend:1.0.0
    volumes:
    - ./config/tika:/config

(Re)create the service using docker compose up -d tika.

How to increase the Java heap space

To increase the Java heap space, mount a custom config tika-config.xml and config the server's forkedJvmArgs

<properties>
  <server>
    <params>
      <forkedJvmArgs>
        <arg>-Xms4g</arg>
        <arg>-Xmx4g</arg>
       </forkedJvmArgs>
    </params>
  </server>
</properties>

How to disable the OCR parser

To disable the OCR parser, mount a custom config tika-config.xml containing the following content

<properties>
  <parsers>
    <parser class="org.apache.tika.parser.DefaultParser">
      <parser-exclude class="org.apache.tika.parser.ocr.TesseractOCRParser"/>
    </parser>
  </parsers>
</properties>

Reference

Tika configuration

Overwrite the default Tika config by mounting a folder containing tika-config.xml in /config as explained in 'How to customize the Tika config'.

All config options are documented in the official Tika documentation.

OCR support

As of v2 Tika has out-of-the-box support for performing automatic OCR on PDF documents. The official Tika Docker images provides support for:

• English
• French
• German
• Italian
• Spanish
• Japanese

This image additionally adds support for:

• Dutch