Fix observation events times with local time dates

[rhannequin]

Before, times from Astronoby::Events::ObservationEvents could produce wrong results when the location was far away from the Greenwich meridian.

The reason was a bug when converting times (hour+minute+second) into Time objects. The date was correctly added by the time was in UTC. When converting back the UTC time to local time, the date could change.

Now, the date is added depending on the time with concern of the user's UTC offset.

The offset must be provided hen instantiating the Astronoby::Observer object using a new attribute: #utc_offset.

utc_offset = "-05:00"
time = Time.new(2015, 2, 5, 0, 0, 0, utc_offset)
observer = Astronoby::Observer.new(
  latitude: Astronoby::Angle.from_degrees(38),
  longitude: Astronoby::Angle.from_degrees(-78),
  utc_offset: utc_offset
)
sun = Astronoby::Sun.new(time: time)
observation_events = sun.observation_events(observer: observer)

observation_events.rising_time.getlocal(utc_offset)
 # => 2015-02-05 07:12:59 -0500

observation_events.transit_time.getlocal(utc_offset)
 # => 2015-02-05 12:25:59 -0500

observation_events.setting_time.getlocal(utc_offset)
 # => 2015-02-05 17:39:27 -0500

Fixes #97

[rhannequin]

Here are some tests to compare results over several days, from different sources, for different locations. The point is to notice a significant difference that would make believe that the fix is actually a bug taking the data for the next or previous day.

• 2024-11-23 41.87,-87.62 UTC-06:00

  • Astronoby n-1: 06:49 - 16:24
  • Astronoby n: 06:50 - 16:24
  • Astronoby n+1: 06:51 - 16:23
  • Astronoby n [fixed]: 06:50 - 16:24
  • IMCCE n-1: 06:50 - 16:23
  • IMCCE n: 06:52 - 16:22
  • IMCCE n+1: 06:53 - 16:22
  • OpenWeather: 06:50 - 16:24

• 2024-09-11 41.87,-87.62 UTC-06:00

  • Astronoby n-1: 05:26 - 18:09 (previous day)
  • Astronoby n: 05:27 - 18:08 (previous day)
  • Astronoby n+1: 05:28 - 18:06 (previous day)
  • Astronoby n [fixed]: 05:27 - 18:08
  • IMCCE n-1: 05:27 - 18:06
  • IMCCE: 05:28 - 18:05
  • IMCCE n+1: 05:29 - 18:03

• 2000-01-02 -33.8688,151.2093 UTC+11:00

  • Astronoby n-1: 05:48 (next day) - 20:09
  • Astronoby n: 05:48 (next day) - 20:10
  • Astronoby n+1: 05:49 (next day) - 20:10
  • Astronoby n [fixed]: 05:48 - 20:10
  • OpenWeather: 05:49 - 20:09

The test don't show a significant error. This makes me think the fix is correct.

[trevorturk]

🥇