#349: Use js-yaml to parse json, yaml, and yml config files

[fillerwriter]

Addresses #349

Summary of changes: Refactored buildPatternData to use js-yaml library to parse all the config files. Added a unit test to ensure we're picking up the appropriate files.

[bmuenzenmeyer]

hello @fillerwriter this is certainly a welcome surprise! thanks for this initial bit of work - I am excited to round out PL Node support with yaml.

[bmuenzenmeyer]

I will review this in earnest, but can already see that this may cause problems in the dev-3.0 branch, where we made a bunch of var to const or let changes already. If you'd be so kind as to alter statements back to var, this will make for an easier merge later.

[fillerwriter]

@bmuenzenmeyer No problem. I've swapped out the consts that fall outside the additional work I've done.

[bmuenzenmeyer]

Thanks for making those changes. Will pay dividends down the road.

Individual patterns can also have data associated with them, which we implement within pattern_assembler.js here: https://github.com/pattern-lab/patternlab-node/blob/master/core/lib/pattern_assembler.js#L319-L358

For this to be truly feature complete against #349 and the specification, I'd ask that we support pattern-level yaml too.

[fillerwriter]

@bmuenzenmeyer Apologies for letting this sit for a bit. I ended up refactoring the logic for loading the config files into helper functions, and swapped out the implementation across the board. I'd like to tighten up the implementation before merging in. If you have any input, I'm interested in hearing it. :-)

[bmuenzenmeyer]

this was initially extremely confusing to me because I thought configFileLoader was referring to patternlab-config.json when it actuality it was dealing with data.

I like the direction of further encapsulating this data, somewhat similar to how we have a dedicated markdown_parser now.

would you be willing to change this to, say, perhaps data_loader?

[fillerwriter]

Sweet. I like data_loader as a name, and can definitely switch. The other thing I'm wanting to do is adjust the footprint of the methods so that fs can be an optional dependency, and ensuring we have at least minimal testing setup for the new code.

I'm hoping to get to this today/tomorrow.

[fillerwriter]

@bmuenzenmeyer Ready for review. Thanks for your patience.

[bmuenzenmeyer]

Great work! Sorry this took so long! It will be in the next release