v0.4.7: Bugfixes and papercuts.

This release fixes an issue with reading the hyphenation patterns for certain languages (like french, #27) and correctly handles filenames with spaces when resuming work. Additionally, some help and logging messages have been improved and the documentation has been updated.

The wrapper is now able to dump the annotated text in the UIMA XMI format in addition to the tabular DOF format (option -xmi).

Attention: Starting with this version, the wrapper is looking for the config folder with the default configuration and support files next to the jar file instead of in the current directory (#34). So, unzip the zip file and leave everything where it is, no need to move the folder around. As before, use the -config option to specify your own modified configuration file.

v0.4.6

Bugfix release:

Since version 0.4.5, writing files on windows failed (#24).

Additionally, this release contains some minor cleanups of on-screen messages and logging.

v0.4.5

• Based on DKPro Core 1.8.0 (stable)
• Improved error/message logging.
• Improved error handling, e.g., for missing config files
• Release bundles config folder again

v0.4.4

Bugfix Release:

Since v0.4.2, some sentences that span multiple paragraphs were annotated with an incorrect start offset, causing very long sentences that could let the dependency parser run outof memory.

v0.4.3

• Added hyphenation algorithm (for more than 70 languages)
• Added Coreference Resolution (for English only)

v0.4.2

• Critical bugfix. In the previous version, some sentences were skipped if they span over more than one paragraph, i.e. over more than two line breaks.

v0.4.1

Changes in version v0.4.1:

• Name of the jar changed to ddw-0.4.1.jar
• Integrated a resume function (can be enabled by the command line parameter -resume)
• Add language support for 21 languages
• Bugfixes

v0.4.0

Release v0.4.0:

• Changed name of jar to wrapper-0.4.0.jar
• Implemented a XMLReader, run the pipeline with the argument -reader xml
• Catched out of memory exception, pipeline keeps running if a file causes a out of memory exception
• Added documentation on how to read in directories and how to use patterns.

v0.3.6

• Bugfix: A tab was added at the end of each line, which caused an error in Pandas -> the values for the columns were displaced by 1 column.

v0.3.5

• Bugfix: Path detection for config files should work on every system
• Bugfix: SRL for English
• Parser and SRL no longer exectued in the default variant
• Light version removed (is now enabled by default)
• Added comments to the default.properties
• Updated the process how arguments passed to pipeline components. treetagger-config must be updated.