MeteoFrance_PublicAPI: A Python Wrapper for the MétéoFrance API !

python wrapper of the portail-api.meteofrance.fr datasets

It especially allows to manage easily the API key.

Documentation: antoinetavant.github.io/meteofranceapi (WIP)

Source Code: github.com/antoinetavant/meteofranceapi

Disclaimer

The tool is not officially provided by MétéoFrance.

It is build by trials-and-errors and revers-engineering by an amateur.

Most of the functionalities should work, but it may not be the best way to achieve it.

Installation

Using pip

pip install meteofrance-publicapi

Usage

To use the ressources, you need an account at MeteoFrance. It is free (as in free-beer).

How to get the API key

First register on portail-api.meteofrance.fr

Once your account is activated, several options are possible to use meteofrance-publicapi :

• an API key for the API requested
• a token for the API requested (limited to max 1 hour)
• an Application ID to the classes which will manage the API token.

The Application ID is the recommended use, as it allows to requests dynamically new tokens.

However, you need to keep it a secret !

Obtaining the Application ID

From the website portail-api.meteofrance.fr

• click on the Bonjour User Name button (upper right corner)
• click Mes API
• click "Générer Token" for any API
• scroll down to the curl field
• The Application Key is the last field of the curl commend : curl -k -X POST [... ... ...] "Authorization: Basic <YourAPPLICATION_ID>

Examples

Accessing observation data

Observation data are measured by a station. The list of the stations can be accessed by

import meteofrance_publicapi as mpa
observation_client = mpa.Observations(application_id=<YourApplicationID>)
observation_client.list_stations()

Accessing Arome Forecasts

See the notebook Arome Forecast for examples of accessing the forecast of the AROME model.

TODO

• Add local cache capabilities, for instance with joblib
• Add persistent storage, locally or with a could provider