Update stream dm job - batchInterval is not hardcoded anymore

[nhnminh]

This pull request has one change in streamDMJob.scala:

Change:

• batchInterval is set by parameter from command line (arguments of spark program)
• This will process the list of arguments as follow: ++ Check the first argument (args(0)) if it is an integer. If yes, check if it is >0 and < maxInt. If yes, set it as the batchInterval. Then drop the first element of args ++ Pass the newArgs (args dropping the first element) to process in Task.
• batchInterval is now set by Milliseconds (before it was Seconds).

• Default value of batchInterval is 1000 ms.

  Reason: User can set batchInterval from arguments. Because batchInterval needs to be tuned a lot to help SparkStreaming program have a good performance, it should not be hardcoded.

Test: To verify this, you could run the following test:

1. Test with batchInterval:

./spark.sh "200 EvaluatePrequential -l (trees.HoeffdingTree -l 0 -t 0.05 -g 200 -o) -s (FileReader -f ../data/electNormNew.arff -k 4000 -d 10)" 1> result.res 2> log.log

This sets batchInterval to be 200 ms

2. Test with negative batchInterval:

./spark.sh "-300 EvaluatePrequential -l (trees.HoeffdingTree -l 0 -t 0.05 -g 200 -o) -s (FileReader -f ../data/electNormNew.arff -k 4000 -d 10)" 1> result.res 2> log.log

This sets batchInterval to be default value = 1000 ms

3. Test with no user-defined value: ./spark.sh "EvaluatePrequential -l (trees.HoeffdingTree -l 0 -t 0.05 -g 200 -o) -s (FileReader -f ../data/electNormNew.arff -k 4000 -d 10)" 1> result.res 2> log.log

This sets batchInterval to be default value = 1000 ms