[ https://issues.apache.org/jira/browse/ATLAS-57?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14637896#comment-14637896 ]
Venkatesh Seetharam commented on ATLAS-57: ------------------------------------------ Not sure how this is related to Atlas, closing this as invalid. > Allow Kuromoji to load a custom dictionary from the file system instead of > the classpath > ---------------------------------------------------------------------------------------- > > Key: ATLAS-57 > URL: https://issues.apache.org/jira/browse/ATLAS-57 > Project: Atlas > Issue Type: Improvement > Reporter: Hayden Muhl > > The JapaneseTokenizer currently loads its dictionary directly from the > classpath. We create a custom dictionary with additional words, and in order > to get the JapaneseTokenizer to pick up our custom dictionary, we have to > repackage the WAR file (we currently use 4.10) to include our custom > dictionary. > Proposed changes: > * Publish the Kuromoji tools as a JAR to Maven to make compiling a custom > dictionary easier. > * Move loading of the dictionary out to the JapaneseTokenizerFactory, and > pass the dictionary into the JapaneseTokenizer constructor. > * Add an argument to the JapaneseTokenizerFactory to load a dictionary off > the file system. If this argument is not passed, then load the default > dictionary from the classpath. -- This message was sent by Atlassian JIRA (v6.3.4#6332)