NOTE: THIS PAGE IS IN PROGRESS.  PLEASE CHECK BACK FOR MORE DETAILS.

For now, see: https://downloads.apache.org/tika/2.0.0/CHANGES-2.0.0.txt

Metadata

Removed duplicate/triplicate keys

Background: In early 1.x, we had basic metadata keys that were created somewhat ad hoc.  We then added normalized metadata keys based on standards such as Dublin Core, or we at least tried to add namespaces to the metadata keys for specific file formats.  To maintain backwards compatibility, we kept the old keys and added new keys.  This led to quite a bit of metadata bloat, where we'd have the same information two or three times.  In Tika 2.x, we slimmed down the metadata keys and relied only on, say Dublin Core if it exists.

Changed Metadata Keys

• Metadata.RESOURCE_NAME_KEY has been renamed TikaCoreProperties.RESOURCE_NAME_KEY.
• TikaCoreProperties.KEYWORDS has been renamed Office.KEYWORDS.
• Meta X-Parsed-By has changed to X-TIKA:Parsed-By
• X-TIKA:EXCEPTION:runtime has been changed to X-TIKA:EXCEPTION:container_exception

tika-parsers – specific parser changes

tika-parsers module

When using tika-parsers in your project, you need to change the dependencies from

<dependency>
  <groupId>org.apache.tika</groupId>
  <artifactId>tika-parsers</artifactId>
  <version>1.27</version>
</dependency>

to

<dependency>
  <groupId>org.apache.tika</groupId>
  <artifactId>tika-parsers-standard-package</artifactId>
  <version>2.0.0</version>
</dependency>
<dependency>
  <groupId>org.apache.tika</groupId>
  <artifactId>tika-parsers-scientific-module</artifactId>
  <version>2.0.0</version>
</dependency>
<dependency>
  <groupId>org.apache.tika</groupId>
  <artifactId>tika-parsers-sqlite3-module</artifactId>
  <version>2.0.0</version>
</dependency>

Also, there's a small transitive dependency issue with jcl-over-slf4j between tika-parsers-standard-package 2.0.0 and tika-parser-scientific-module:2.0.0. So if you are using maven enforcer plugin, you will need to fix it by adding this:

<!-- Fix tika-parsers-standard-package 2.0.0 vs tika-parser-scientific-module:2.0.0 transitive dependency -->
<dependency>
    <groupId>org.slf4j</groupId>
    <artifactId>jcl-over-slf4j</artifactId>
    <version>1.7.31</version>
</dependency>

If you are checking for CVEs (recommended), the tika-parser-scientific-module:2.0.0 comes with a transitive dependency on quartz 2.2.0 which should be fixed like this:

  <dependency>
    <groupId>edu.ucar</groupId>
    <artifactId>netcdf4</artifactId>
    <version>${netcdf-java.version}</version>
    <exclusions>
      ...
      <exclusion>
        <groupId>org.quartz-scheduler</groupId>
        <artifactId>quartz</artifactId>
      </exclusion>
    </exclusions>
  </dependency>
  <dependency>
    <groupId>org.quartz-scheduler</groupId>
    <artifactId>quartz</artifactId>
    <version>2.3.2</version>
  </dependency>

When using lang detection, you need to change now use:

<dependency>
  <groupId>org.apache.tika</groupId>
  <artifactId>tika-langdetect-optimaize</artifactId>
  <version>2.0.0</version>
</dependency>

Also note that org.apache.tika.langdetect.OptimaizeLangDetector.getDefaultLanguageDetector has moved to org.apache.tika.langdetect.optimaize.OptimaizeLangDetector.getDefaultLanguageDetector.

For OCR, you can not use anymore TesseractOCRConfig.setTesseractPath(String) and TesseractOCRConfig.setTessdataPath(String) methods. They moved to the TesseractOCRParser class.

tika-parsers-module optional dependencies

zstd

The zstd dependency includes native libs and is not packaged with the tika-parsers-module.  If you'd like to parse zstd files, include:

    <dependency>
      <groupId>com.github.luben</groupId>
      <artifactId>zstd-jni</artifactId>
      <version>1.5.0-4</version>
    </dependency>

TIFF and JPEG2000

If you plan to write TIFFs with Tika (rendering of PDF pages for OCR), and if the BSD-3 with nuclear disclaimer license is acceptable to you, include:

<dependency>
  <groupId>com.github.jai-imageio</groupId>
  <artifactId>jai-imageio-core</artifactId>
  <version>1.4.0</version>
</dependency>

If you plan on processing JPEG2000 images (most common use case would be rendering PDF pages for OCR), and if the BSD-3 with nuclear disclaimer license is acceptable to you,  include:

<dependency>
  <groupId>com.github.jai-imageio</groupId>
  <artifactId>jai-imageio-jpeg2000</artifactId>
  <version>1.4.0</version>
</dependency>

Note! In 2.x, Tika will not warn you if a PDF page that you're trying to render has a JPEG2000 in it.  PDFBox will log a warning.

tika-app

tika-server

General

• enableFileUrl has been removed in favor of two separate fetchers, one for files and one for URLs (see tika-pipes#FetchersInClassicServerEndpoints).
  
  • FileSystemFetcher (which is packaged with tika-core) for files
  • HttpFetcher (requires an external jar from https://mvnrepository.com/artifact/org.apache.tika/tika-fetcher-http) for URLs.

Configuration

tika-pipes

See the tika-pipes page.

tika-eval

tika-langid

In the 1.x branch, the default (hardwired) language identification component was the wrapper around Optimaize.  If you used the following in 1.x:

<dependency>
  <groupId>org.apache.tika</groupId>
  <artifactId>tika-langdetect</artifactId>
  <version>1.27</version>
</dependency>

In 2.x, change this to:

<dependency>
  <groupId>org.apache.tika</groupId>
  <artifactId>tika-langdetect-optimaize</artifactId>
  <version>2.0.x</version>
</dependency>

The original  language id component that was built by Tika devs and that used to be in tika-core is now in the tika-langdetect-tika module.