[TIKA-2906] Modularize tika-eval's language stats from the application

• Type: Task
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 1.23
• Assignee: Tim Allison
• Reporter: Tim Allison
• Created: 15/Jul/19 13:04
• Updated: 17/Aug/19 00:07
• Resolved: 13/Aug/19 17:37

Description

Tika-eval's language stats are tightly coupled to the application and the initial workflow of running against a directory of extracts and reporting info to an H2 db.

It would be helpful for large-scale data processing pipelines to modularize some of tika-eval's stats so that they can be applied to, e.g. a full Solr/ES cluster. We won't build the actual connectors to Solr/ES/other on this ticket, but we will make it easier for integrators to build their own.

This is slated for 1.23/2.0...not 1.22.