API to request forecast track based on hour to landfall

[SorooshMani-NOAA]

Related to oceanmodeling/ondemand-storm-workflow#8

stormevents is useful for fetching storm data in a clean format, but still extracting the specific entries useful for a simulation falls on the shoulders of users. It'd be very useful to add some of these functionalities for extracting data to the package itself.

One such a functionality is to extract forecast track based on hours to landfall. As suggested by @WPringle, the same logic discussed in the linked issue above, can be directly brought to stormevents.

Feature request:
We need an API to retrieve a single OFCL forecast track for the storm with forecast start date at specified hours before the landfall. To account for the track history of the storm, for all times before the specified time, the track must be taken from hour-0 OFCL track starting at all prior cycles.

[WPringle]

@SorooshMani-NOAA I think inside StormEvents it may not be necessary to extract the hours to landfall, just need to extract the forecast for a given datetime. Working out how many hours before landfall could be done somewhere else perhaps. We would need to have like a shoreline database to workout the points inside that polygon.

[WPringle]

@SorooshMani-NOAA The other use case is getting the times before the forecast time. I suggest we use start_date to be the forecast time, and introduce another variable that gives the negative and positive forecast hours to extract. So can be like -72 hours to +120 hours.

[SorooshMani-NOAA]

@WPringle forecast_time is used in multiple places in the code, so it might not be very easy to change what it does without affecting a lot of other things. Can you please write an example use case and describe what you expect? Something like

my_extracted_data = track.get_track(etc., etc.)

[WPringle]

@SorooshMani-NOAA

    track = stormevents.nhc.VortexTrack(
       storm='al062018',
       start_date='2018091218', 
       forecast_hours=[-72,120],
       advisories='OFCL',
       file_deck='a',
    )

This should get the 2018091218 forecast and keep all forecast hours from 0 to 120, as well as fill in the previous 0-hr CARQs for three days prior to 2018081218 (forecast hours = [-72,-66,...,-6,0,6,...,120])