TreeTagger のバックアップ(No.4) - 投野由紀夫授業用ホームページ

バックアップ一覧
差分を表示
現在との差分を表示
ソースを表示
TreeTagger へ行く。
- 1 (2012-05-24 (木) 17:57:51)
- 2 (2012-05-24 (木) 18:06:58)
- 3 (2012-06-05 (火) 23:23:34)
- 4 (2012-06-26 (火) 00:19:58)
- 5 (2013-01-14 (月) 23:50:52)
- 6 (2020-03-05 (木) 14:39:40)

Download

2. Download the English parameter file from here

3. Uncompress 1. and 2. and read INSTALL.txt (below)

Installation

1. Install a Perl interpreter (if you have not already installed one).

  You can download a Perl interpreter for Windows for free at
  http://www.activestate.com/activeperl/

2. Move the TreeTagger directory to the root directory of drive C:.

3. Download the PC parameter files for the languages you need, decompress

  them (e.g. using Winzip or 7zip) and move them to the subdirectory lib.
  Rename the parameter files to <language>.par
  Example: Rename french-par-linux-3.1.bin to french.par

4. Add the path C:\TreeTagger\bin to the PATH environment variable.

5. Open a shell and type the command

  set PATH=C:\TreeTagger\bin;%PATH%

6. Change to the directory C:\TreeTagger

7. Now you can test the tagger, e.g. by analyzing this file with the command

  tag-english INSTALL.txt

Examples

c:\TreeTagger>set PATH=C:\TreeTagger\bin;%PATH%

c:\TreeTagger>tag-english INSTALL.txt > INSTALL.tag.txt
       reading parameters ...
       tagging ...
       finished.

Chunker (for Mac-Intel)

english-chunker-par-linux-3.2.bin.gz --> should be downloaded

unzip the parameter file and move it to lib directory:

gzip -cd english-chunker-par-linux-3.2.bin.gz > lib/english-chunker.par

test:

$ echo 'I was reading a book.' | cmd/tagger-chunker-english > hello.txt

$ cat hello.txt
<NC>
I	PP	I 
</NC>
<VC>
was	VBD	be
reading	VBG	read
</VC>
<NC>
a	DT	a
book	NN	book
</NC>
.	SENT	.

Notes

For Linux and Mac, Penn TreeBank? tagset is default. If you want VV for lexical verbs, VB for be-verbs, and VH for have, you should edit the file 'tree-tagger-english' as follows:

#last line
perl -pe 's/\tV[BDHV]/\VB/;s/IN\/that/\tIN/;'

--> perl -pe 's/IN\/that/\tIN/;'