Upgrade to Spatial4j 0.4.1 and JTS 1.13

[dsmiley]

This fixes #5279

The main think irking me is that ElasticSearch uses a global hard-coded instance of JtsSpatialContext.GEO as the one and only SpatialContext as a static final in ShapeBuilder.SPATIAL_CONTEXT. That wasn't what I had in mind when I (or was it Ryan or Chris, I forget) devised SpatialContext concept. If you look at JtsSpatialContextFactory (or don't even use JTS, look at SpatialContextFactory) you'll see a bunch of options that trigger various behavior. The most important one is "geo" (aka geodetic or geodesic, synonyms) which is a boolean that chooses between a latitude-longitude spherical world model or a flat plane (Euclidean geometry). It's not quite clear to me at this time how ElasticSearch users that want a flat world model and who pre-project their data are using this.

[dsmiley]

Notify @chilling

Question: Should commentary/feedback go on the pull-request or does it go on the issue ( #5279 ) referenced by the pull-request? For that matter is there even a point to creating an issue if I know I'm going to create a pull-request any way?

[chilling]

@dsmiley cool stuff, Thanks. I'm currently looking at it. Feedback and comments on the actual PR goes here. If we need to discuss the Issue itself, it should be there. So I guess further discussion is on the PR.

[dsmiley]

For the record, my GitHub ID is @dsmiley, not "david". I wonder how much GH mistaken identity email he gets for all the David's out there.

[chilling]

oh, just saw saw it.thx

[david]

@dsmiley about once a month or so. It could definitely be worse. :)

[dsmiley]

@jpountz I don't see evidence that this was in fact merged into ElasticSearch.

[jpountz]

I didn't mean to close this pull request, it looks to me that Github automatically closed it because I deleted the elasticsearch:upgrade/lucene_47 branch, that this pull request is derived from. Maybe you can try to reopen it by branching master?

Sorry for the inconvenience!