Migrate to LocalDate, Fix some bugs

[dickermoshe]

This PR will replace ZonedDateTime with LocalDate.
The benefit of this change is that the time zone is only in a single location. It is impossible for the timezone in the geolocation class to differ from the timezone in the ZonedDateTime. It also generally makes the entire API more user-friendly. In the past, I've been confused about what time of day the ZonedDateTime should use.

Additionally, this PR fixes a small error in the getSunriseOffsetByDegrees function, Where SolarEvent.SUNSET was being passed to getInstantFromTime.

This PR fixes a bug in getInstantFromTime, which simplifies the calculation by offloading the math to the Java Duration class. There was also an issue where the Instant being returned was incorrectly converting to the local timezone.

[KosherJava]

Thank you for the extensive set of changes. It is interesting that we still need a ZonedDateTime in a few instances (not surprising), but it is not in the publicly facing API.

[dickermoshe]

Thanks, I will now create a set of tests which will compare the results of the old and new api.

Keep in mind that it is impossible to do extensive testing comparing the two becuase:

1. The new java.time classes behave slightly differently than java.util
2. The old API had some other bugs which were fixed here Migrate to ICU #249

I'll see what I can do

[KosherJava]

Thank you @dickermoshe . While edge cases were fixed by ICU, the main focus is on day-to-day minor changes in old vs. the new, and I appreciate your testing. Try current year for a number of cities with all zmanim compared.

[dickermoshe]

Testing java against java is something that I do not know how to do.
The maven/gradle setup is way to confusing to me