Add column to CSV export toggle normalized grades

[pwilkins]

Another episode in the Remote Gradebook saga. When we last visited (#147) our hero, Remote Gradebook, she was mourning the loss of her companion, Legacy Dashboard (edx #10890), deprecated by the villainous Open EdX (edx #10942).
Add another column to the CSV file that contains grades to toggle whether the CSV file exported to the remote gradebook contains normalized grades (0-1, the default and current behavior) or raw scores (0-max-points, the new feature).

[pwilkins]

Currently edX's remote gradebook sends this CSV data to MIT's enterprise gradebook[1].

+-----------------+----------------+
| External email  | Ex 01          |
+-----------------+----------------+
| johndoe@mit.edu | 0.0            |
+-----------------+----------------+
| janedoe@mit.edu | 0.272727272727 |
+-----------------+----------------+

In edX #10890, we proposed adding columns to the CSV file. There is a more detailed explanation of the proposal there. Here we propose to add yet another column to indicate which grading scheme the course staff wants to export to the enterprise grade system.

To add support for toggling the grading scheme, we propose adding another column titled normalize?. This column, like the max points column, actually describes the assignment, not the student, since these columns contain the same value for all students receiving a grade for the assignment. In other words, the values in the max points and normalize? columns should be the same for all rows.

+-----------------+-------+------------+------------+
| External email  | Ex 01 | max points | normalize? |
+-----------------+-------+------------+------------+
| johndoe@mit.edu | 5.0   | 5          | 0          |
+-----------------+-------+------------+------------+
| janedoe@mit.edu | 3.0   | 5          | 0          |
+-----------------+-------+------------+------------+

The normalize? column will contain either 1 for True (grades 0-1), or 0 for False (grades 0-max points).

@pdpinch

[pdpinch]

There were performance issues with #221 and courses like https://github.mit.edu/mitx/content-mit-8S05r-2016_Spring and https://github.mit.edu/mitx/content-mit-802r_2-2016_Spring

8.S05 may be a good test case because it would time out on exporting grades with just 35 students.

8.02 has 700 students in production. It would timeout even if grades had been pre-computed.

[pdpinch]

One place to start would be to compare performance for 8.S05 with #221 and with a similar PR against Dogwood.