| commit | 162357616c97a56c8e2dc7c8eeb4eb116def0035 | [log] [tgz] |
|---|---|---|
| author | margaretha <margaretha@ids-mannheim.de> | Tue Mar 07 18:51:18 2017 +0100 |
| committer | margaretha <margaretha@ids-mannheim.de> | Tue Mar 07 18:56:36 2017 +0100 |
| tree | 33cc393b327fa886e52fd088a1105642a1205b90 | |
| parent | e07c421c512f3dacbcdb2546d67026022a18ec06 [diff] |
Changed get resources from virtual collection to corpus. Change-Id: I7c2448305220a544328c207c28995484d525dbcb
KorapSRU is the CLARIN Federated Content Search (FCS) endpoint for KorAP. It implements FCS specifications and connects the CLARIN FCS client Aggregator and KorAP. Thus, public resources in KorAP is accessible from Aggregator through KorapSRU.
CLARIN defines FCS specifications to allow distributed search across multiple heterogenous search engines in a uniform way. FCS specifications are built on the SRU/CQL protocol for communications between its client and endpoint. FCS 1.0 specification supports SRU (Search Retrieve via URL) 1.2 and FCS 2.0 specification SRU 2.0.
[KorapSRU 1.0.1 release] (https://github.com/KorAP/KorapSRU/releases/tag/release-1.0.1) implements FCS 1.0 specification and supports basic search using simple CQL (Contextual Query Language) for term query, phrase query and boolean query. FCS 2.0 specification is implemented in the newest version of KorapSRU, but it has not been released yet. It supports extended search (e.g. annotation search) that can be formulated using FCS Query Language (FCSQL) developed based on Corpus Query Processor(CQP).
Usually CQL and FCSQL queries are translated into the native language of a search engine in an FCS endpoint. Since KorAP supports multiple query languages and has its own query translator Koral, the translation is implemented in Koral, not in KorapSRU. Therefore, KorAP users will also be able to use CQL and FCSQL.
SRU explain request
gives general information about KorapSRU and some default search settings, for instance the number of records it retrieves per page (see http://clarin.ids-mannheim.de/korapsru?operation=explain). To obtain more information such as supported annotation layers needed for requesting an extended search,
x-fcs-endpoint-description=true
must be added as an extra request parameter (see [http://clarin.ids-mannheim.de/korapsru?operation=explain&x-fcs-endpoint-description=true] (http://clarin.ids-mannheim.de/korapsru?operation=explain&x-fcs-endpoint-description=true)).
SRU search retrieve request
contains a CQL or FCSQL query, for example [http://clarin.ids-mannheim.de/korapsru?operation=searchRetrieve&query=das%20Buch&version=1.2] (http://clarin.ids-mannheim.de/korapsru?operation=searchRetrieve&query=das%20Buch&version=1.2). KorapSRU forwards the CQL or FCSQL query in an SRU search retrieve request URL to Kustvakt, the API provider of KorAP managing the communications among all KorAP components. Moreover, KorapSRU transforms the query results from Kustvakt into an SRU response.
Java 7 (JDK 1.7 with JCE or OpenJDK 7)
Tomcat 7
Kustvakt (currently unavailable for public)
Configure the service URI in the src/main/resource/kustvakt.conf file to a Kustvakt server URI, for example:
korapsru.client.service.uri=http://localhost:8089/api/v0.1/
KorapSRU is built based on the FCSSimpleEndpoint library provided by CLARIN. KorapSRU 1.0.2-SNAPSHOT uses FCSSimpleEndpoint version 1.3.0 available from CLARIN Nexus repository. To allow Maven to download the library using JDK 1.7, an additional Java Cryptography Extension (JCE) Unlimited Strength Jurisdiction Policy Files 7 is needed.
To install a war file of KorapSRU, go to the root directory of the project and run
mvn install -Dhttps.protocols=TLS1.2
in a terminal.