CQL Search Installation (The part of the text about context.xml and web.xml files can be used to build NetBeans projects as well.) ------------------------------------ Contents: - Requirements - Installation and Configuration of CQLSearch -- placing war-file -- The fragment web.txt to be present in Tomcat's web.xml file - Testing URI's file. - For the developer: information about Federated-Content-Search subproject of CLARIN and over the external library that supplies interfaces and parent classes = Requirements = * cqlsearch2-XXX package * Tomcat >=5, with the Trova search engine deployed = Installation and Configuration of CQLSearch = ========================================================================================= == PLACING WAR-file == You must deploy the cqlsearch2's .war file to the Tomcat instance where Trova search engine is deployed. E.g. on lux16 it is tomcat-webuser. ============================================================================================= == web.xml === Find the following in this tar-ball: -- DeploymentSettings/web.txt Check if you have listed there parameters in TomCat's web.xml and re-set them properly if necessary. Th file consist of two parts: the first part contains the parameters shared by Trova, Annex, cqlsearch. The second part consists of the server-dependent parameters specific only for CQLSearch. There are comments near these parameters on how to set the values. Read these comments. ============================================================================================= = Testing URI's = You need to run a couple of url-requests to see if the things work after installation. Take the url's from the corresponding file DeploymentSettings/TestURIs.txt. The base URI depends on a server, so read the first sentence where the base URI's for all three srevers lux16, lux17 and corpus1 are listed. If you are a deploying sysadmin, you do not have to analyse and report the search or scan results, you just have to check if there are no exceptions, errors and warnings and report about them if necessary. ============================================================================================= = For the developer: information over the external library (supplies interfaces and parent classes) = CQL search is a part of the Federated (aggregated) search project with the idea that via one common interface a user can make a SRU/CQL request which is dispatched across all the institutions participating in the project. See http://weblicht.sfs.uni-tuebingen.de/Aggregator/. As a result the user obtains hits from all the participants, including MPI. Each participant implements its own search engine, and responsible to make it compatible with the common interface. That's why all the participants are using the same library of interfaces and parent classes, created in Mannheim by Oliver Schoenefeld. The commented sources are available (for reading only!!!) at https://svn.clarin.eu/ (e.g. SRUServer).