Skip to content

krawler-meteofrance ​

A Krawler job designed to scrape data from the Météo-France public API

Overview ​

A Krawler based service to download diverse datasets from the public Météo-France API portal.

paquetobs ​

The paquetobs job files enable the scraping of real-time observations from all weather stations of the French network at hourly and/or sub-hourly intervals (every 6 minutes). These data are exposed by the public API Package Observations.

paquetobs-stations ​

This job scrapes the stations. The main collected properties are:

  • id, the id of the station
  • name, the name of the station
  • location, the location of the station

The station data are stored in compliance with the GeoJSON standard.

Configuration ​

VariableDescription
PAQUETOBS_TOKENThe token to use the paquetobs API
DEPARTMENTSThe list of departments used to filter the collected observations, e.g, "11,09,31"
STATIONSThe list of station ids used to filter the collected observations, e.g, "09099001,09301001"
DB_URLthe database url. By default: mongodb://localhost:27017/meteofrance.
DEBUGEnables debug output. Set it to krawler* to enable full output. By default it is undefined.

paquetobs-observations ​

This job scrapes the observations. The basic collected properties are:

  • id, the id of the station
  • name, the name of the station
  • temperature (°),
  • humidity (%),
  • wind direction (°),
  • wind speed (m/s),
  • precipitation (mm/h)

The observation data are stored in compliance with the GeoJSON standard.

Configuration ​

VariableDescription
PAQUETOBS_TOKENThe paquetobs API token.
FREQUENCYThe frequency of the observations to collect. It must be horaire or infrahoraire-6m. Default value is horaire.
LATENCYThe latency of the observations to collect. It must be horaire or infrahoraire-6m. Default value is horaire.
TTLThe retention period in seconds of the data. By default: 7 * 24 * 60 * 60 (~7 days)
DB_URLthe database url. By default: mongodb://localhost:27017/meteofrance.
DEBUGEnables debug output. Set it to krawler* to enable full output. By default it is undefined.

Forecast Pipeline (AROME / ARPEGE) ​

forecast-pipeline

These jobs download forecast data from the public Météo-France API:

  • AROME – High-resolution regional model
  • ARPEGE – Global model

Files are retrieved in GRIB2 format.

Available Jobs ​

JobTypeResolutionCoverage
arome-franceAROME0.025°France
arome-france-highAROME0.01°France
arpege-europeARPEGE0.1°Europe
arpege-worldARPEGE0.25°Global

Configuration ​

VariableDescription
DATA_SOURCESource of the data (meteofrance or data-gouv)
AROME_TOKENAPI token for AROME
ARPEGE_TOKENAPI token for ARPEGE
OUTPUT_DIRBase directory for downloaded files
WORKERS_LIMITMaximum number of concurrent downloads
OLDEST_RUN_INTERVAL_MSMax allowed run age
RUN_TIMESOverride model run times
PACKAGESOverride packages
FORECAST_TIMESOverride forecast steps

Deployment ​

We personally use Kargo to deploy the service.