Write unit tests

[vasudevanv]

Testing will include 3 different components

1. Testing for data
2. Testing for the classifier
3. Testing for the pipeline

Possible test frameworks to use: pytest, unittest, nose

[vasudevanv]

We need to have a good dummy test dataset that needs to be small enough to run tests quickly but large enough to cover corner cases.

[dhimmel]

It might make sense to create a notebook in machine-learning that generates these test cases. I'm willing to do this, but if anyone else wants to claim the task, go ahead.

[vasudevanv]

That is a good idea. If you could generate the test data, I can take a shot at writing the tests. I have been playing around with pytest, and it's pretty neat. I think we should use it.

[jessept]

Pytest sounds like a good choice. I read in this article that nose (one of the modules we were discussing before) does not seem to get along very well with the logging module.

[jessept]

We are going to split out the unit testing into 2 pieces.
@vasudevanv will work on unit tests for data.py and figshare.py
@jessept will work on unit tests for analysis.py and utils.py

@vasudevanv is using pytest so we will continue on that front.