Support matching on variables that are collections

[felipefzdz]

This PR addresses #123 issue.

I was TDDing this feature outside-in and I've figured out that production code was already able to handle collections for multi prime matching. I'd like to keep the tests though as they'll serve as documentation and to avoid regressions.

I've updated the documentation with an example about using collections when matching variables.

Let me know if you're happy with the PR and I'll squash the commits.

Meanwhile reading the code I've found a couple of TODOs to make the code more robust and maintainable, e.g.

• validation error if variable types length != outcome variable matcher length
• validate PrimePreparedMulti
• deal with a prime that doesn't exist
• maybe say the last outcome is the default or make a default mandatory?
• generalise all the prepared stores, very little difference

I'll address those in a future PR, if that's ok.

[chbatey]

Somehow had turned off my notifications for scassandra :( sorry @felipefzdz looking now

[felipefzdz]

No worries :)

[chbatey]

Sorry been out out of the (c|s)assandra world lately. Back on this week so will review this.

[tolbertam]

D'oh, just realized my recent PR created some conflicts here, however I think that PR added support matching of collections, but i'll have to test it out. I'll use the test cases here to see if that was the case.