API Calls

Function forecast handles all request parameters and returns a Forecast object.

>>> from darksky import forecast
>>> boston = forecast(key, 42.3601, -71.0589)
>>>

The first 3 positional arguments are identical to the 3 required parameters for API call. The optional query parameters need to be provided as keyword arguments.

Using time argument will get you a time machine call. Using timeout argument will set default request timeout .

>>> BOSTON = key, 42.3601, -71.0589
>>> from datetime import datetime as dt
>>> t = dt(2013, 5, 6, 12).isoformat()
>>> boston = forecast(*BOSTON, time=t)
>>> boston.time
1367866800

Data Points and Data Blocks

The values as well as DataPoint and DataBlock objects are accessed using instance attributes or dictionary keys. You can access current values directly, without going through currently data point.

>>> boston['currently']['temperature']
60.72
>>> boston.temperature
60.72

Data blocks are indexable and iterable by their data values.

>>> len(boston.hourly)
24
>>>
>>> boston.hourly[1].temperature
59.49
>>>
>>> # list temperatures for next 10 hours
... [hour.temperature for hour in boston.hourly[:10]]
[60.83, 59.49, 58.93, 57.95, 56.01, 53.95, 51.21, 49.21, 47.95, 46.31]

Nonexistent attributes will raise AttributeError and dictionary keys KeyError the way you’d expect.

Raw data

To get the raw data dictionary, you can either access it through instance attributes or navigate to it through dictionary keys, the same way you would navigate the actual dictionary.

>>> boston.hourly[2]
{'ozone': 290.06, 'temperature': 58.93, 'pressure': 1017.8, 'windBearing': 274, 'dewPoint': 52.58, 'cloudCover': 0.29, 'apparentTemperature': 58.93, 'windSpeed': 7.96, 'summary': 'Partly Cloudy', 'icon': 'partly-cloudy-night', 'humidity': 0.79, 'precipProbability': 0, 'precipIntensity': 0, 'visibility': 8.67, 'time': 1476410400}
>>>
>>> boston['hourly']['data'][2]
{'ozone': 290.06, 'temperature': 58.93, 'pressure': 1017.8, 'windBearing': 274, 'dewPoint': 52.58, 'cloudCover': 0.29, 'apparentTemperature': 58.93, 'windSpeed': 7.96, 'summary': 'Partly Cloudy', 'icon': 'partly-cloudy-night', 'humidity': 0.79, 'precipProbability': 0, 'precipIntensity': 0, 'visibility': 8.67, 'time': 1476410400}

Flags and Alerts

All dashes - in attribute names of Flags objects are replaced by underscores _. This doesn’t affect the dictionary keys.

>>> # instead of 'boston.flags.isd-stations'
... boston.flags.isd_stations
['383340-99999', '383390-99999', '383410-99999', '384620-99999', '384710-99999']
>>>
>>> boston.flags['isd-stations']
['383340-99999', '383390-99999', '383410-99999', '384620-99999', '384710-99999']

Even though Alerts are represented by a list, the data accessibility through instance attributes is preserved for alerts in the list.

>>> boston.alerts[0].title
'Freeze Watch for Norfolk, MA'

Updating data

Use refresh() method to update data of a Forecast object. The refresh() method takes optional queries (including time, making it a Time machine object) as keyword arguments. Calling refresh() without any arguments will set all queries to default values. Use timeout argument to set the request timeout.

>>> boston.refresh()
>>> (boston.time, boston.temperature, len(boston.hourly))
(1476403500, 60.72, 49)
>>>
>>> boston.refresh(units='si', extend='hourly')
>>> (boston.time, boston.temperature, len(boston.hourly))
(1476404205, 15.81, 169)
>>>
>>> boston.refresh(units='us')
>>> (boston.time, boston.temperature, len(boston.hourly))
(1476404489, 60.57, 49)

For Developers

Response headers are stored in a dictionary under response_headers attribute.

>>> boston.response_headers['X-response-Time']
'146.035ms'