Add correlation function

[asizemore]

Resolves #33 Requires https://github.com/VEuPathDB/veupathUtils/pull/20

Adds a core correlation function that accepts data tables. Then there are other functions that accept abundance data and metadata that just do formatting really and pass the data to the core corr computation (say that 5 times fast!). Added spearman and pearson methods because we got both for free.

This PR also adds a SampleMetadata class

To Dos

• [x] make core corr function
• [x] make SampleMetadata class
• [x] make corr fn that accepts samplemetadata and abundance data
• [x] change stats names
• [x] test stats names
• [x] document
• [x] add checks to data.table x data.table fn

I started making the toJSON changes, but realized that will require work in veupathutils. I think it'd be better if i made that a separate PR. This one is getting large anyways.

[asizemore]

a note to my future self: we're going to have a lot of different ways to use the correlation function. Some with metadata, sometimes with two assays, sometimes with one, sometmes with random genomics data, or just metadata. My current idea for how this should grow is to have one basic correlation function that takes exactly two data frames (or matrices whatevs) and does whatever correlation method. Then we have lots of correlation methods that do different validations and data wrangling based on the inputs, then eventually make two data.frames and send them to the basic correlation function. That way we're not crazy repeating code and can appropriately handle what to do with lots o' different inputs

[asizemore]

Update before i switch my brain to diff abund

• added support for dates
• added support for real variable metadata
• not totally convinced on what to do about the var1 var2 thing. The elements of these columns are variables so maybe the column name should be correlate1Name correlate2Name? There's also a discussion around about naming them based on taxa or assay or something that's coming in. That has to do also with breaking up the plugin. Worth some more thinking before deciding.
• should shore up how handling the variables input within the function. Should it be part of the signature? Is NULL okay for the default or should it be an empty VariableMetadataList? Probably the latter...

[asizemore]

updated based on the feedback. Removed variables entirely, which will help us get things moving. In the future we need to revisit the conversation so that we can optimally handle dates. I expect i'll be back here in a month anyways so we can do that then after we learn a bit more about how it all works together.