Adds option for loading program CSVs right from s3

[higs4281]

Our data pipeline stores incoming school program data in s3; this allows the load_programs script to load these files directly fro s3 rather than requiring the moving and handling of files.

Additions

• --s3 option for the load_programs script and management command
• related tests

  Testing

• In standalone, try running load_programs with an s3 URL:

./manage.py load_programs --s3 true \
"http://files.consumerfinance.gov.s3.amazonaws.com/pb/paying_for_college/csv/CFPBDATAFILE713%20(2).CSV"

If you run ./manage.py purge programs first, it should load 1,522 programs.
If you run it without purging, it should report updating 1,522 programs.

Review

• @amymok
• [x] Project documentation has been updated (including the "Unreleased" section of the CHANGELOG)

[coveralls]

Coverage remained the same at 100.0% when pulling 1b0205e74387df6d537672696f844184f6df5699 on load-s3 into 503ad3882338e8f1120b9fa2b3aa88b4d3b165fe on master.

[amymok]

Looks good.