org.apache.ctakes.ytex.kernel.dao

Interface ConceptDao

All Known Implementing Classes:

ConceptDaoImpl

public interface ConceptDao

create/retrieve concept graphs. store concept graph on file system as they can get big (>10MB). This is not a problem for sql server/oracle, but may require increasing the max_packet_size on mysql.

Author:

vijay

Method Summary

Modifier and Type	Method and Description
void	createConceptGraph(String dir, String name, String query, boolean checkCycle, Set<String> forbiddenConcepts) create the concept graph with specified name using specified query.
ConceptGraph	getConceptGraph(String name) retrieve an existing concept graph.

Method Detail

getConceptGraph

ConceptGraph getConceptGraph(String name)

retrieve an existing concept graph. This attempts to find the concept graph as follows:

• classpath: attempt to load org/apache/ctakes/ytex/conceptGraph/[name].gz
• relative to org.apache.ctakes.ytex.conceptGraphDir: if that fails then relative to the directory defined by the system property/ytex property org.apache.ctakes.ytex.conceptGraphDir
• relative to ytex.properties: if the org...conceptGraphDir property is not defined, then look for [directory of ytex.properties]/conceptGraph/[name].gz

Parameters:

name - name of concept graph. Will retrieve from classpath/file system. @see #createConceptGraph

Returns:

createConceptGraph

void createConceptGraph(String dir,
                        String name,
                        String query,
                        boolean checkCycle,
                        Set<String> forbiddenConcepts)
                 throws IOException

create the concept graph with specified name using specified query. Will create a file [name].gz under dir.

Parameters:

dir - directory of concept graph (optional). If null will use directory specified by the system property/ytex property org.apache.ctakes.ytex.conceptGraphDir. If the property is not defined, use [directory of ytex.properties]/conceptGraph

name - name of concept graph

query - returns 2 string columns, 1st column is the child concept, 2nd column is the parent concept.

checkCycle - if true will check for cycles and remove them (default true). If this is set to false, then only pagerank can be used on this graph.

forbiddenConcepts - set of concepts whose edges will not be added to the concept graph (optional). By default, this includes umls concepts like C1274012 (Ambiguous concept).

Throws:

IOException