removes exjs dependency

[mejackreed]

Part of #41

Though, I'm not sure how much this does in reducing actual bundle size.

[edsilv]

Hey, please see the failing getRangeByPath test I added.
I think it's related to this: https://github.com/IIIF-Commons/manifesto/pull/42/files#diff-01f13f6d1a0f04758601d8128e56b2efL1429

[stephenwf]

Could reduce it down using something like:

const reducer = (acc, next) => {
  acc.push(next);
  const nextRanges = next.getRanges();
  if (nextRanges.length) {
     return nextRanges.reduce(reducer, acc);
  }
  return acc;
}

and use it like:

const flatRanges = topRange.getRanges.reduce(reducer, [topRange]);

If it needs to be unique, could make it a Set or a Map mapped by the id field. (if this is the failing test that you mentioned)

[edsilv]

Yep, we need a flat array of unique ranges. Using Set sounds good.

[mejackreed]

Thanks for the comments here. I've used your approach @stephenwf and pushed that up in
4576dc8.

[stephenwf]

A couple of console logs made its way in 👀

[mejackreed]

Thanks for catching that.

[mejackreed]

Thanks for the great review (and the additional tests added). Is there a way you would like me to resolve the dist conflicts?

Also I noticed that the uniqueness needed in the added test wasn't caught when just using an Array. @edsilv is there a manifest/scenario where I could write a failing test for that?

[edsilv]

I've altered the biocrats.js test to check the number of returned ranges. The root range was being included when it didn't have an id. Root ranges without ids are "placeholders" for generating trees I believe.

I've also removed the dist directory, but this doesn't seem to have fixed the conflicts?

[mejackreed]

I've force pushed a rebase that should handle the dist thrashing. I'm going to submit a separate PR that should help with the release process for this package, so that built versions do not need to be submitted in PR's or committed.

[edsilv]

Cool. Well, the UV seems happy now. Think it's mergeable?