Update boto version.

Add host argument to s3_connect.  Update unit test patches to match.
Host argument is now required by boto.  Latest version of boto provides a
hook to read from a .boto file, and there is a way for remote-task
to set such a file up and then read from it.  However, it's not
clear how to get the .boto file to the task instances in a multi-instance
cluster so that the file would be read by boto when reducers open
their own streams to output to an S3 file.

use video duration from events

EDUCATOR-1411

Added parameter import_credentials to HistogramFromSqoopToMySQLWorkflowBase

Allow empty inserts for ModuleEngagementSummaryMetricRangesMysqlTask

for sites with periods of low activity.

Inspect traceback for handling end of output error.

Add missing backslash for null ascii.

Brian/also load json events

Translation of MySQL's LONGBLOB to Vertica's LONG VARBINARY.  

It coexists with regular event output, and is controlled by
an optional parameter.  By default it runs with
event_record_type equal to 'EventRecord', but can
be overridden by running with --event-record-type 'JsonEventRecord'.

Includes bug fix to timestamp handling:  Add validation screening dates < 1900.

Also includes support for event loading to BigQuery, by adding
support for partitioning to bigquery_load.

* Use records for warehouse loading where defined.
* Check bigquery availability in load code.

Add support for loading to S3 by interval or by date.  PerDate loading
checks whether the data already exists, which is good for incremental runs.
Bulk loading just runs over an interval, and assumes the data isn't already
present on S3.  This is better for processing many days more efficiently.

To address issue with loading into BigQuery, null characters in column values
are encoded as the string '\0'.

Check Sqoop job completion before running.

Make sure enrollment output is encoded

Including a best effort translation of MySQL's LONGBLOB to Vertica's LONG VARBINARY.  Data loss is possible here.

Fix enrollment validation to use unicode version of course_id.

Add non-ASCII testing for course-id and related values to
ensure that encoding is properly done elsewhere.

Also check for the marker file when checking for completeness.

Incremental video.

Added tasks to load warehouse data into BigQuery.

and tidy up the confusion between credentials and import_credentials.

Use merchant close date to compute interval.

Upgrade setuptools, and other supporting packages.

Cleaned up and added logic to figure out the latest completion date.

Default to empty function for hadoop counter callback.

Speed up finance reporting workflow.

Event Exporter Improved Encryption Logging & Error Handling 

Improving logging and adding a counter when there is a failure during the event-exporter encryption process using gpg

Workaround for travis boto issue with GCE. 

Changed active users computation to a per week basis.