Python wrapper for the Dark Sky API
Clone or download
Latest commit 1a9e5a7 Jan 14, 2018
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
darksky fix attr issue Jan 14, 2018
.gitignore update readme Jan 13, 2018
LICENSE init v0.2.2 Sep 6, 2016
README.rst update readme Jan 13, 2018
setup.py version up Jan 14, 2018

README.rst

darkskylib

This library for the Dark Sky API provides access to detailed weather information from around the globe.

Quick start

Before you start using this library, you need to get your API key here.

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'

Example script

from darksky import forecast
from datetime import date, timedelta

BOSTON = 42.3601, 71.0589

weekday = date.today()
with forecast('API_KEY', *BOSTON) as boston:
    print(boston.daily.summary, end='\n---\n')
    for day in boston.daily:
        day = dict(day = date.strftime(weekday, '%a'),
                   sum = day.summary,
                   tempMin = day.temperatureMin,
                   tempMax = day.temperatureMax
                   )
        print('{day}: {sum} Temp range: {tempMin} - {tempMax}'.format(**day))
        weekday += timedelta(days=1)

Output:

Light rain on Friday and Saturday, with temperatures bottoming out at 48°F on Tuesday.
---
Sun: Partly cloudy in the morning. Temp range: 44.86 - 57.26°F
Mon: Mostly cloudy in the morning. Temp range: 44.26 - 55.28°F
Tue: Clear throughout the day. Temp range: 36.85 - 47.9°F
Wed: Partly cloudy starting in the afternoon, continuing until evening. Temp range: 33.23 - 47.93°F
Thu: Light rain overnight. Temp range: 35.75 - 49.71°F
Fri: Light rain in the morning and afternoon. Temp range: 45.47 - 57.11°F
Sat: Drizzle in the morning. Temp range: 43.3 - 62.08°F
Sun: Clear throughout the day. Temp range: 39.81 - 60.84°F

License

The code is available under terms of MIT License