Skip to content

lojban/jbokorp

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

41 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

korpora zei sisku

Hosted at:

[https://korp.alexburka.com](https://korp.alexburka.com)

To add new corpus to "ralju korpora", add them to the directory corpus_import/ralju/original, in a format that parallels the other entries in that directory, and send me a pull request.

Installation

This is the repository for the "corpus pipeline" of korpora zei sisku.

Try running the install.sh script. It should install the pipeline and put makefiles in the right directories. If it does not work, visit this page: http://spraakbanken.gu.se/swe/forskning/infrastruktur/korp/distribution/corpuspipeline. The additional dependencies you will need are GNU Make, Python 2.7 and Corpus Workbench.

You also need the modified version of jbofi'e, which you can get at https://github.com/danr/jbofihe. The jbofihe and cmafihe executables should be in your $PATH.

Python dependencies:

* BeautifulSoup 4
* unidecode
* dateutil

You will need to set the CWB_DATADIR and CORPUS_REGISTRY environment variables, and also you PYTHONPATH to the corpus_import/annotate/python directory. The SB_MODELS variable also needs to be set, but it can be set to anything since we won't be using those models anyway.

Example:

export PYTHONPATH=$HOME/code/jbokorp/corpus_import/annotate/python:$PYTHONPATH
export SB_MODELS=$HOME/code/jbokorp/corpus_import/annotate/models
export CWB_DATADIR=$HOME/corpora/data
export CORPUS_REGISTRY=$HOME/corpora/registry

Test

Go to the ralju directory, and issue

make export

After a while, you should have an export directory, with segmented and annotated texts. Check that it has been reasonably well done.

To install on a remote server, do:

make installcorpus installtimespans installinfo

You will need to change the remote, its directories and the mysql_db in Makefile.common.

Releases

No releases published

Packages

No packages published

Languages

  • Python 50.2%
  • HTML 23.2%
  • JavaScript 20.0%
  • Makefile 6.1%
  • Shell 0.5%