A Python interface for MET Norway's Locationforecast 2.0 service.
Project description
MET Norway Location Forecast
A Python interface for the MET Norway Locationforecast/2.0 service. This is a free weather data service provided by the Norwegian Meteorological Institute and Yr.
Contents
Features
- Get weather forecast data for anywhere in the world.
- Automatically take care of caching data.
- Helpful classes for managing forecast data.
- Convert between units of measurement.
Installation
Installing from PyPI:
pip install metno-locationforecast
Usage
Basics
Before using this package you should be aware of the terms of
service for using the MET Weather API.
The metno-locationforecast
package will not make requests unless current
data has expired and will send requests with the appropriate
If-Modified-Since
header if possible. Identification can be provided by
passing a User-Agent
string to the Forecast class, see more on this below.
After installing metno-locationforecast
the following commands can be run
in a python console. Start by importing the Place
and Forecast
classes, these are the main classes you will need to interact with.
>>> from metno-locationforecast import Place, Forecast
Create a Place
instance. Geographic coordinates are given by latitude,
longitude (in degrees) and altitude (in metres). The altitude parameter is
optional but recommended. Note that latitude and longitude are rounded to four
decimal places and altitude is rounded to the nearest integer, this is required
by the MET API. GeoNames is a helpful website for
finding the geographic coordinates of a place.
>>> new_york = Place("New York", 40.7, -74.0, 10)
Next create a Forecast
instance for the place. Here you need to supply the
type of forecast, options are "compact"
(a limited set of variables
suitable for most purposes) or "complete"
(an extensive set of weather
data). For more details on the differences check out the this
page. We also need to
supply a User-Agent
string, typically this will include the name and
version of your application as well as contact information (email address or
website) more details on what is expected
here. Do NOT use the string supplied
here as this does not apply to your site. Optionally, you can provide a
save_location
parameter, this is the folder where data will be stored. The
default save_location
is "./data/"
.
>>> ny_forecast = Forecast(new_york, "compact", "metno-locationforecast/1.0 https://github.com/Rory-Sullivan/metno-locationforecast")
Then run the update method. This will make a request to the MET API for data and
will save the data to the save location. If data already exists for the
forecast, this will only request new data if the data has expired and will make
the request using the appropriate If-Modified-Since
header. It returns a
string describing which process occurred, this will be one of
"Data-Not-Expired"
, "Data-Not-Modified"
or "Data-Modified"
.
Only in the case of "Data-Modified"
has any change to the data occurred.
>>> ny_forecast.update()
'Data-Modified'
>>> ny_forecast.update()
'Data-Not-Expired'
Finally we can print the forecast.
>>> print(ny_forecast)
Forecast for New York:
Forecast between 2020-07-21 14:00:00 and 2020-07-21 15:00:00:
air_pressure_at_sea_level: 1016.7hPa
air_temperature: 28.7celsius
cloud_area_fraction: 1.6%
...
Accessing Data
Printing forecasts to the terminal is great but most likely you want to use the forecast data in your own application. When the update method is run it parses the returned data which can then be accessed through attributes of the forecast instance.
The most notable of these is the data
attribute.
>>> type(ny_forecast.data)
<class 'dict'>
>>> ny_forecast.data.keys()
dict_keys(['last_modified', 'expires', 'updated_at', 'units', 'intervals'])
'last_modified'
, 'expires'
and 'updated_at'
are
datetime.datetime
objects for when the data was last modified, when it is
expected to expire and when the forecast was updated, respectively.
'units'
contains a dictionary mapping variable names to the units in which
they are provided by the API.
'intervals'
is where we find the actual weather data. It is a list of
intervals. Note that the MET API usually supplies multiple intervals for each
time point in the data set, the forecast parser takes the shortest supplied
interval for each time point.
>>> type(ny_forecast.data["intervals"])
<class 'list'>
>>> type(ny_forecast.data["intervals"][0])
<class 'metno-locationforecast.data_containers.Interval'>
>>> print(ny_forecast.data["intervals"][0])
Forecast between 2020-07-21 14:00:00 and 2020-07-21 15:00:00:
air_pressure_at_sea_level: 1016.7hPa
air_temperature: 28.7celsius
cloud_area_fraction: 1.6%
relative_humidity: 56.0%
wind_from_direction: 349.7degrees
wind_speed: 1.4m/s
precipitation_amount: 0.0mm
Each interval is a metno-locationforecast.data_containers.Interval
instance. This interval class has a 'variables'
attribute which is a
dictionary mapping variable names to
metno-locationforecast.data_containers.Variable
instances.
>>> first_interval = ny_forecast.data["intervals"][0]
>>> first_interval.start_time
datetime.datetime(2020, 7, 21, 14, 0)
>>> first_interval.end_time
datetime.datetime(2020, 7, 21, 15, 0)
>>> first_interval.duration
datetime.timedelta(0, 3600)
>>> first_interval.variables.keys()
dict_keys(['air_pressure_at_sea_level', 'air_temperature', 'cloud_area_fraction', 'relative_humidity', 'wind_from_direction', 'wind_speed', 'precipitation_amount'])
>>>
>>> rain = first_interval.variables["precipitation_amount"]
>>> print(rain)
precipitation_amount: 0.0mm
>>> rain.value
0.0
>>> rain.units
'mm'
For a full overview of the Interval
and Variable
classes see the
code.
Other attributes of the Forecast
class that could be useful are;
- response: This is the full
requests.Response
object received from the MET API. - json_string: A string containing all data in json format. This is what is saved.
- json: An object representation of the json_string.
The Forecast
class also has additional methods that may be of use.
- save: Save data to save location.
- load: Load data from saved file.
More Examples
For further usage examples see the examples folder.
Notes on Licensing
While the code in this package is covered by an MIT license and is free to use the weather data collected from the MET Weather API is covered by a separate license and has it's own terms of use.
Useful Links
- The Norwegian Meteorological Institute - https://www.met.no/en
- MET Weather API - https://api.met.no/
- MET Weather API Terms of Service - https://api.met.no/doc/TermsOfService
- Locationforecast/2.0 documentation - https://api.met.no/weatherapi/locationforecast/2.0
- Full list of variables and their names - https://api.met.no/doc/locationforecast/datamodel
- Yr - https://www.yr.no/en
- Yr Developer Portal - https://developer.yr.no/
- Yr Terms of Service (same as the MET API terms of service but perhaps more readable) - https://developer.yr.no/doc/TermsOfService/
- GeoNames - http://www.geonames.org/
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for metno-locationforecast-0.0.1b1.tar.gz
Algorithm | Hash digest | |
---|---|---|
SHA256 | 37d01dbd616e13c652197805fe6fa114ef52105b052c9e9c7902f0a0f015c884 |
|
MD5 | 730590fb838f49171bfc44ddbbb893c4 |
|
BLAKE2b-256 | 7b2ddc618a98b6a4f13a80720bfdbeb6f777cd17dc8c570a4f8a3d10b86e8679 |
Hashes for metno_locationforecast-0.0.1b1-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 076e66c270e02dd733c7cf05429ca62d761c28507169bbd3ccb563dda06aae5e |
|
MD5 | b6173a87397811842e33cc87a92bea02 |
|
BLAKE2b-256 | 695109c9861512fdcf5a057d6b7ff8bcc3c8201e2bd87bbd6658cf534f391859 |