Open-Meteo API Python Client

This ia an API client to get weather data from the Open-Meteo Weather API based on the Python library requests.

Instead of using JSON, the API client uses FlatBuffers to transfer data. Encoding data in FlatBuffers is more efficient for long time-series data. Data can be transferred to numpy or pandas using Zero-Copy to analyze large amount of data quickly. The schema definition files can be found on GitHub open-meteo/sdk.

This library is primarily designed for data-scientists to process weather data. In combination with the Open-Meteo Historical Weather API data from 1940 onwards can be analyzed quickly.

Basic Usage

The following example gets an hourly temperature and precipitation forecast for Berlin. Additionally, the current temperature is retrieved. It is recommended to only specify the required weather variables.

# pip install openmeteo-requests

import openmeteo_requests
from openmeteo_sdk.Variable import Variable

om = openmeteo_requests.Client()
params = {
    "latitude": 52.54,
    "longitude": 13.41,
    "hourly": ["temperature_2m", "precipitation"],
    "current": ["temperature_2m"]
}

responses = om.weather_api("https://api.open-meteo.com/v1/forecast", params=params)
response = responses[0]
print(f"Coordinates {response.Latitude()}°E {response.Longitude()}°N")
print(f"Elevation {response.Elevation()} m asl")
print(f"Timezone {response.Timezone()} {response.TimezoneAbbreviation()}")
print(f"Timezone difference to GMT+0 {response.UtcOffsetSeconds()} s")

# Current values
current = response.Current()
current_variables = list(map(lambda i: current.Variables(i), range(0, current.VariablesLength())))
current_temperature_2m = next(filter(lambda x: x.Variable() == Variable.temperature and x.Altitude() == 2, current_variables))
current_relative_humidity_2m = next(filter(lambda x: x.Variable() == Variable.relative_humidity and x.Altitude() == 2, current_variables))

print(f"Current time {current.Time()}")
print(f"Current temperature_2m {current_temperature_2m.Value()}")
print(f"Current relative_humidity_2m {current_relative_humidity_2m.Value()}")

Note 1: You can also supply a list of latitude and longitude coordinates to get data for multiple locations. The API will return a array of results, hence in this example, we only consider the first location with response = responses[0].

Note 2: Please note the function calls () for each attribute like Latitude(). Those function calls are necessary due to the FlatBuffers format to dynamically get data from an attribute without expensive parsing.

NumPy

If you are using NumPy you can easily get hourly or daily data as NumPy array of type float.

import numpy as np

hourly = response.Hourly()
hourly_time = range(hourly.Time(), hourly.TimeEnd(), hourly.Interval())
hourly_variables = list(map(lambda i: hourly.Variables(i), range(0, hourly.VariablesLength())))

hourly_temperature_2m = next(filter(lambda x: x.Variable() == Variable.temperature and x.Altitude() == 2, hourly_variables)).ValuesAsNumpy()
hourly_precipitation = next(filter(lambda x: x.Variable() == Variable.precipitation, hourly_variables)).ValuesAsNumpy()
hourly_wind_speed_10m = next(filter(lambda x: x.Variable() == Variable.wind_speed and x.Altitude() == 10, hourly_variables)).ValuesAsNumpy()

Pandas

For Pandas you can prepare a data-frame from hourly data like follows:

hourly_data = {"date": pd.date_range(
	start = pd.to_datetime(hourly.Time(), unit = "s"),
	end = pd.to_datetime(hourly.TimeEnd(), unit = "s"),
	freq = pd.Timedelta(seconds = hourly.Interval()),
	inclusive = "left"
)}
hourly_data["temperature_2m"] = hourly_temperature_2m
hourly_data["precipitation"] = hourly_precipitation
hourly_data["wind_speed_10m"] = hourly_wind_speed_10m

hourly_dataframe = pd.DataFrame(data = hourly_data)
print(hourly_dataframe)
#date  temperature_2m  precipitation
#0  2023-08-01 00:00:00       16.945999            1.7
#1  2023-08-01 01:00:00       16.996000            2.1
#2  2023-08-01 02:00:00       16.996000            1.0
#3  2023-08-01 03:00:00       16.846001            0.2

Caching Data

If you are working with large amounts of data, caching data can make it easier to develop. You can pass a cached session from the library requests-cache to the Open-Meteo API client.

The following example stores all data indefinitely (expire_after=-1) in a SQLite database called .cache.sqlite. For more options read the requests-cache documentation.

Additionally, retry-requests to automatically retry failed API calls in case there has been any unexpected network or server error.

# pip install openmeteo-requests
# pip install requests-cache retry-requests

import openmeteo_requests
import requests_cache
from retry_requests import retry

# Setup the Open-Meteo API client with a cache and retry mechanism
cache_session = requests_cache.CachedSession('.cache', expire_after=-1)
retry_session = retry(cache_session, retries=5, backoff_factor=0.2)
om = openmeteo_requests.Client(session=retry_session)

# Using the client object `om` will now cache all weather data

TODO

• Document multi location/timeinterval usage
• Document FlatBuffers data structure
• Document time start/end/interval
• Document timezones behavior
• Document pressure level and upper level
• Document endpoints for air quality, etc
• Consider dedicated pandas library to convert responses quickly

License

MIT