[OPENNLP-694] GeoEntityLinker gazatteerSearcher should sanitize lucene query strings

• Type: Improvement
• Status: Closed
• Priority: Minor
• Resolution: Fixed
• Affects Version/s: addons-1.6.0
• Fix Version/s: addons-1.6.0
• Component/s: Entity Linker
• Environment: java 7
• Assignee: Mark Giaconia
• Reporter: Mark Giaconia
• Created: 12/May/14 19:29
• Updated: 12/May/14 19:45
• Resolved: 12/May/14 19:45

Description

Currently the GeoEntityLinker takes Namefinder results and passes them directly to the GazetteerSearcher class. Noisy NER results can cause lucene syntax errors to be thrown and logged. A simple regex cleanup of the input would be helpful.