Content
Installation of CSniper requires some prerequisites and some configuration steps which are outlined below.
Prerequisites
Before you start installing and configuring CSniper, you need to install following software:
- a Java application server (tested with Tomcat 6)
- a MySQL compatible RMDBS (tested with MySQL 5.5)
- an empty database (let’s call it csniper)
- IMS Corpus Workbench ((http://cwb.sourceforge.net/download.php)[http://cwb.sourceforge.net/download.php])
- (optionally) an installation of Tgrep2 (http://tedlab.mit.edu/~dr/Tgrep2/)
For sake of simplicity, this guide assumes that you are using Tomcat 6 and Mysql 5.5 on Ubuntu 12.
Deployment
Just copy the csniper.war
into your Tomcat working directory (e.g. /var/lib/tomcat6/webapps
) and restart Tomcat. Per default Tomcat settings this will unpack your csniper.war
. CSniper is not yet configured though, the next step is mandatory.
Configuration
As a first step, you need to create two directories:
- one configuration directory for CSniper, for this guide at
/srv/csniper
- one directory for your corpora, e.g.
/srv/csniper/corpora
To let Tomcat know where our configuration is located, set the environment variable
CATALINA_OPTS="-Dcsniper.home=/srv/csniper"
either in /etc/environment
or at the start of /usr/share/tomcat6/bin/startup.sh
. Copy the csniper.properties.example
to /srv/csniper
, and rename it to csniper.properties
. Then you need to specify the options:
-
First enter your database credentials:
database.dialect=org.hibernate.dialect.MySQL5Dialect database.driver=com.mysql.jdbc.Driver database.url=jdbc:mysql://localhost:3306/csniper (or whatever your database is called) database.username=username database.password=password
-
Then configure the path to your cqp executable, which comes with the installation of CWB:
engine.cqp.executable=path_to_your_cqp_executable_file engine.cqp.macrosLocation=classpath:/BNC_macros.txt
-
Optionally you can configure CSniper to use tgrep2, in which case you need to specify the executable here:
engine.tgrep.executable=path_to_your_tgrep2_executable_file
-
You also need to specify the path to the (converted) corpora you want to use:
corpus.service.info.file=corpus.properties corpus.service.path=/srv/csniper/corpora
CSniper uses opennlp components for tokenization and parsing. If you want to use other components (e.g. Stanford Parser), you can specfiy these options in segmenter.properties
and parser.properties
respectively (examples can be found in the wiki). These need to be placed into /srv/csniper as well. The components themselves need to reside in the same directory as the sql connector above.
After you have saved this configuration, you want to restart tomcat again.
CSniper is (almost) ready to be used now, and can be found at http://localhost:8080/csniper (default Tomcat configuration). But without data, CSniper doesn’t make a whole lot of sense.
Which means: it’s time to convert a corpus for CSniper.