0 to 1 data normalization

[brandomr]

To accomplish this task we should make an "interim" or "dev" release of Mixmasta based on the ongoing refactor and then update mixmasta_processors.py to perform the 0 to 1 normalization based on the code found in transition-data-normalization.py.

We'll need to update the Flow.js to include the data_paths_normalized (locations on S3/filesystem) and should update the IndicatorSchema.py to optionally handle data_paths_normalized as well.

Note: transition-data-normalization.py contains updated (much faster) functions for performing the 0 to 1 normalization!