[TIKA-1797] Pick best location by analysing context in GeoTopic parser

• Type: Sub-task
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Component/s: parser

Environment

ALL

• Assignee: Unassigned
• Reporter: Madhav Sharan
• Created: 17/Nov/15 21:41
• Updated: 25/Nov/15 23:11

Description

lucene-geo-gazetteer is enhanced to return multiple results for a string.

GeoTopic parser should start using new fields country code and admin codes to determine best possible locations within context of a document.

Proposal:
1. Get top 3 results for each location string
2. Group all locations on country code, admin code1, admin code2
3. Give priority to most occuring country, admin code1, admin code 2