[TR-553] Transition to a more generic JSON document reader for tweets Created: 19/Dec/18  Updated: 16/Feb/20  Resolved: 19/Dec/18

Status: Resolved
Project: Terrier Core
Component/s: .indexing
Affects Version/s: 5.0
Fix Version/s: 5.1

Type: Improvement Priority: Minor
Reporter: Richard McCreadie Assignee: Richard McCreadie
Resolution: Fixed  
Labels: None


 Description   
Twitter Indexing currently uses an old class called TwitterJSONDocument to convert JSON tweets into Terrier documents. This makes many assumptions about the format of tweets and is in general slower than desired.

Internally we replaced this class with a faster version based on Jackson (AIR sub-project), we should transition to this version instead.

 Comments   
Comment by Richard McCreadie [ 19/Dec/18 ]

Fix committed in d2d4e6fc. Passes unit tests

Comment by Richard McCreadie [ 19/Dec/18 ]

Moved to FlatJSONDocument. Pointed TwitterJSONCollection at this alternative document class. TwitterJSONDocument is left in the code base

Generated at Wed Apr 01 12:09:44 BST 2020 using JIRA 7.1.1#71004-sha1:d6b2c0d9b7051e9fb5e4eb8ce177ca56d91d7bd8.