Extended the existing HadoopGreenstoneIngest with proper Reduce phase - this also required the creation of several new classes to handle the partitioning and grouping of Map phase output, and a class to wrap access to the txt2(t)db processes.