# Building Flow endpoint $ cd flow/ ; mvn clean install; $ cd launchers/fullflow ; mvn clean install; # Starting Stanbol $ cd flow/fullflow/target ; java -jar org.apache.stanbol.launchers.fullflow-0.9.0-incubating.jar; # Testing Flow : ## Pool flow feature By default a file endpoint will fetch all files in /tmp/chaininput folder, process it with the langId engine and then write the xml-rdf result in /tmp/chainoutput folder. ## On demand flow feature The flow endpoint offer sub-ressource that allow you to directly call some predefined particular flow. Template for this request is : $ curl -X POST -H "Accept: text/turtle" -H "Content-type: text/plain" --data "Here comes a little test with Paris as content and also Berlin but why not detect city as Boston and some well know people like Bob Marley." http://localhost:8080/flow/{{flowName}} Where {{flowName}} can have the values : - default : default chain : all registred engines sorted by weight - tika : only call tika engine, no visible result (as the engine don't create enhancements) - tika2 : tika then langID then NER then save the result in /tmp/chainoutput folder - tika-default : tika then default flow Flow/Route definition for default can be find here : https://svn.apache.org/repos/asf/incubator/stanbol/branches/cameltrial/flow/weightedgraphflow/src/main/java/org/apache/stanbol/flow/weightedgraphflow/WeightedChain.java Others Flow/routes definition resides here : https://svn.apache.org/repos/asf/incubator/stanbol/branches/cameltrial/flow/filepoolgraphflow/src/main/java/org/apache/stanbol/flow/filepoolgraphflow/FileRoute.java