atamer / transform Goto Github PK

This is the source code of highly effective transforming process
There is two concurrent operations in the program ,
first one is TransformService, and second one is FileWriterService

TransformServices works on a maxCPU - 1 threads. These threads making transformation (Filtering , Renaming vs..)
Benchmarking in different systems may result different outcome , I have just optimized on my machine

FileWriterService works on single thread and it writes the tranformed data to output file

I assume that data file can be very big , but conf files are small enough to hold in memory.

IO operations(Read/Write) works on single thread because the main bottleneck in these operations is Disk IO

I have used Executor Services for multithreading .
I have used BlockingQueue for data passing between threads
I have used CountDownLatch in order to check termination

TECH-STACK:
        Java 8 , Maven , log4j , intellij


HOW-TO-WORK :
        Program takes files locations from argument parameters.
        Run unit tests and create data , conf files, then you can give these arguments as program parameter.
        sample arguments: "src/test/resources/bigfile.txt src/test/resources/conf1.txt src/test/resources/conf2.txt src/test/resources/output.txt"


TO-DO: need to improve unit test coverage for each method
       need to improve solution in very large configuration files.