Conform to spec

[ian-r-rose]

1. Draws accuracy from a Rayleigh distribution with mode around 5 meters. May or may not be reasonable, but a fine place to start, I think.
2. Casts properties to an integer as the spec demands. I think the spec should probably relax on this front and allow floating point numbers, but this is what it says for now.

I have not updated the event_location field to be a Point, since the spec does not say that just yet.

Edit: whoops, I see that the spec does now specify point.

Fixes #3

[ian-r-rose]

Okay, Point is fixed.

[thekaveman]

I'd honestly like us to standardize around Feature and FeatureCollection in all of these cases, as hinted at in openmobilityfoundation/mobility-data-specification#74 and proposed in https://github.com/CityOfLosAngeles/mobility-data-specification#73. This would allow accuracy and other related properties to co-exist with the geo-data in an easily extensible manner. It also follows from our (buried) discussion on the JSON Schema work

[ian-r-rose]

Ah, I misunderstood. I can revert that last commit, or we can hang out here until openmobilityfoundation/mobility-data-specification#74 is resolved.

[thekaveman]

Yeah...that last comment got lost in there. I'm hoping we'll have that issue tied up in the next few days, but I'll leave it up to you. I'm good to merge the other changes if you want to revert, or I'm good to wait it out as things may still be changing...

[thekaveman]

@ian-r-rose can you revert that last commit, given the latest developments in MDS? Then I can get the other changes merged in.

[ian-r-rose]

Yep, reverted!