Port TESS user filtering to shape reducers

[CKrawczyk]

The Issue

At the moment if a user draws multiple shapes on a subject, those are allowed to cluster together in the reducer. Ideally this should not happen as a user should only draw one sahpe per object on a subject.

The fix

This has already been addressed and tested with the TESS column reducer. Each shape has a proxy for the user ID passed into the metric. The metric checks if the proxy values is the same and if so returns a distance of infinity otherwise it returns the typical euclidean distance.

How to apply the fix

• Adjust the shape process data function to calculate the proxy user ID. See the custom TESS reducer's process data function for how to do this.
• Adjust the shape metrics to account for this user ID proxy. The TESS reducer metric shows what this looks like in the case of the euclidean metric used for non-rotating shapes, but care needs to be taken when adjusting any of the angle metrics. This is the hard part
• In the shape reducers filter out blank rows.
• Around this line of code append the user ID proxy to the location data.
• On this line of code pass in the modifed location data with the custom metric
• Do the same for the other shape reducer
• Make a keyword switch to this this feature on or off
• Write tests for each shape with this featured turend on for both reducers (~24 new tests) This is the annoying part