leodido/treetagger repository overview

⁠treetagger.docker

This repository⁠ contains docker images to build and ship ready to use TreeTagger⁠ instances.

You will not have to manually install TreeTagger in your system again.

Detailed info here⁠.

⁠What it is

A tool for annotating text with part-of-speech - i.e., POS tagging - and lemma information.

⁠Supported languages

17 languages are supported: bulgarian, dutch, english, estonian, finnish, french, galician, german, italian, latin, portuguese, polish, russian, slovak, spanish, swahili, mongolian (only parameter file provided, no scripts).

Some of them have also alternative parameter files.

⁠Tagging

⁠Italian

Suppose you want to (tokenize and) tag an Italian text.

The script to use is tree-tagger-italian.

It expects UTF8 encoded input files as arguments. If no files have been specified, input from stdin is expected.

echo 'Proviamo semplicemente a eseguire un test di prova.' | \
          docker run --rm -i leodido/treetagger tree-tagger-italian

Outputs:

Proviamo	     VER:pres	   provare
semplicemente	ADV	        semplicemente
a	            PRE	        a
eseguire	     VER:infi	   eseguire
un	           DET:indef	  un
test	         NOM	        test
di	           PRE	        di
prova	        NOM	        prova
.	            SENT	       .```


#### Portuguese

Now, try with some Portuguese.

```bash
echo 'Qual é o seu nome?' | \
         docker run --rm -i leodido/treetagger tree-tagger-portuguese

Results:

Qual	PT0	 qual
é	   VM      ser
o	   DA0	 o
seu	 DP3	 seu
nome	NCMS	nome
?	   Fit	 ?

And so on for other supported languages.

⁠Chunking

Suppose you want to tokenize, tag and annotate a German text with nominal and verbal chunks.

echo 'Das ist ein Test.' | \
        docker run -i leodido/treetagger tagger-chunker-german

Which outputs:

<NC>
Das	    PDS	    die
</NC>
<VC>
ist	    VAFIN	  sein
</VC>
<NC>
ein	    ART	    eine
Test	   NN	     Test
</NC>
.	      $.	     .

⁠Build

This image is tested, built and pushed using CircleCI.

See the repository⁠ for further information about TreeTagger, about manual building, testing, and so on.

⁠Credits

• Helmut Schmid, University of Stuttgart, Germany - TreeTagger⁠.

Last update: 28/05/2015