Xponents SDK v2.9 update 9

Xponents 2.9.9 Fourth of July

• Much improved stop filters for nearly all Solr-supported languages (using lucene stopword resources)
• Reviewed geocoding rules; Assigned default confidence to country matches, instead of 0.
• Solr6 staging
• SolrGazetteer: added "findPlace( name, parameters)" and the ability to find romanized place names for a given native language name.

XText 2.9.12

Keeps in line with Xponents releases as well as Tika releases. No functional changes here.

• Tika 1.15
• PDFBox 2.06
• POI 3.16
• Minor version updates for Solr 6.x staging, but not critical for XText
• Minor Apache Commons updates for deprecated libraries

Xponents 2.7.8

This Xponents release was focused on code clean up, documentation, stabilizing extractors for basic geoname tagging, keyword tagging and text extraction from binaries.

• Extraction: Initial working basic geotagger for experimentation (PlaceGeocoder, Alpha release). Introducing some early name and word filters into the tagger database (solr gazetteer core) as well as for runtime seemed important for performance and sanity. Such filters were stabilized in this release.
• Extraction: Improved API for interacting with SolrGazetteer - search by feature type, by location, etc.
• Extraction: solidified the solr/gazetteer and solr/taxcat cores for Solr 4.10 configuration
• XText: Tika 1.8 upgrade
• Code: various checkstyle fixes

Xponents 2.8 plans to have hopefully a Solr 5.x upgrade or support, and improved geotagging/geocoding rules in the Extraction "geo" module.

-Marc Ubaldino