Is a modular framework for publishing (distributed and heterogeneous) language resources.
It consists of self-contained "loosely-coupled" services, with (hopfully) clearly defined tasks and interfaces, the basic protocol (the linuga franca) being the FCS/SRU-protocol [ http://clarin.eu/fcs ], "FCS" standing for federated content search.
We recognize following basic components to the framework:
- Aggregator
- a service that is able to dispatch a query to multiple target repositories, and merge the result back. Currently we have a basic php-implementation fcs/aggreagtor/switch.php which however is not yet able to query multiple target repositories at once.
- Wrapper
services that translate between existing systems and the protocol. Currently we have (in an early dirty version):
- php - implementations accessing MySQL-db for Dictionaries
- xquery implementation for eXist-based content repository
- perl implementation mapping to ddc-api (corpus search engine) allowing to access our (public) corpora (not checked in yet)
- Workspace
- main user-interface allowing to select targets, issues queries, see the results and invoke specialized viewer for the resources. [ index.html ]
- ResourceViewer
- separate ui-components that are able to display/visualize specific resource types. (Currently we have only the basic XSL-transformed HTML-views of the results returned by the services)
- Simple answer
- This is easy if you just want to see what corpus_shell does if you point your browser at it and perhaps change some of it's behaviour to your needs by adapting the JavaScript, CSS and perhaps HTML parts: There is nothing to do but open index.html right from where you downloaded it. All your settings stay local in your browser and you query the publicly available projects at ICLTT's server.
- Advanced answer
- If you want to expose your own data to the world then it gets way more complicated. You need a server capable of interpreting PHP 5.3 or better and perhaps an exsit-db or mysql/mariadb database for storing and retrieving your data. Detailed instructions on how to do such a setup (may) follow (soon).