Refactor: Move things between Otava and Signal Processing

[henrikingo]

Otava is really the combination of two python projects, each with their own history and each in their own github repository. Otava was originally developed by Datastax engineers, then donated to Apache Foundation and renamed to Otava. The core math was developed by MongoDB engineers and released in a repository called signal processing algorithms.

There's code that would more naturally belong in the other repository, but today everything is located where it is, purely based on who originally wrote the code.

Pre-requisite: This ticket assumes that an ongoing discussion comes to a conclusion how we want to organize our code going forward. Either we merge the signal processing code into Otava (aka "vendoring"), or if we continue with 2 separate repositories, we at least have an open and ongoing communication and collaboration between the two.

The purpose of this ticket is to list pieces of code that could more naturally live on the other side of the divide, if it weren't for above historical reasons. Possibly when implementing this refactoring, it may make sense to split work into multiple sub tickets.

---

Move from Otava to Signal Processing Algorithms:
otava/analysis.py:

ExtendedSignificanceTester
TTestSignificanceTester

[henrikingo]

Update: we replaced signal_processing dependency with a full rewrite of the e-divisive algorithm. So the scope of this ticket is now to refactor the mentioned classes/files closer to the new implementation instead. See: #96