Sample Code Conribution: Flume HBase Statistics Sink

Here is some sample source code I wrote at this month's hackathon - the HBaseIncrementerSink listens for Twitter feeds and for each bigram in such a feed, it updates a compound HBase key date/bigram, it also updates a summary count of events by date. The HBaseStatsServer provides a simple server that lets a webapp query the resulting HBase information, returning JSON data for the event counts and bigram counts for all time (it could be expanded to accept a time range query). This worked with a version of Flume that included Pat Hunt's inclusion of HBase suppport.

This could be a start on other more general/other Flume sink support for HBase.