This PR adds the capability of getting time series data also for coverages which are related to the main one - this is useful in the context of showing the forecasts of several models for a given climatic variable.
The proposed implementation adds a self-referential many-to-many relationship to the coverageconfiguration table, and then exposes it in the admin, thus allowing a coverage configuration to be related to others. This is then used in the time-series API endpoint to retrieve data for these related coverages.
For example, a request like this:
GET http://webapp:5001/api/v2/coverages/time-series/tas_absolute_value_ensemble-rcp26-DJF?
coords=POINT(11.5469 44.9524)&
datetime=../..&
include_coverage_data=true&
coverage_data_smoothing=NO_SMOOTHING&
coverage_data_smoothing=MOVING_AVERAGE_11_YEARS&
include_coverage_related_data=true
Would cause the API to return:
a time series with the values for tas_absolute_value_ensemble-rcp26-DJF unsmoothed
a time series with the values for tas_absolute_value_ensemble-rcp26-DJF smoothed with a moving average
and for each related coverage:
a time series for the related coverage unsmoothed
a time series for the related coverage smoothed with a moving average
The below image illustrates a possible outcome, where we show only the unsmoothed values of the main coverage and the smoothed values of all other time series:
In this example we are plotting the TAS absolute values of the 5-model ensemble for the RCP26 scenario in the DJF season (i.e. winter) as the main time series and also the individual values of four of those models (EC-EARTH-CCLM4-8-17, EC-EARTH-RACMO22E, EC-EARTH_RCA4 and HadGEM2-ES_RACMO22E)
This PR adds the capability of getting time series data also for coverages which are related to the main one - this is useful in the context of showing the forecasts of several models for a given climatic variable.
The proposed implementation adds a self-referential many-to-many relationship to the
coverageconfigurationtable, and then exposes it in the admin, thus allowing a coverage configuration to be related to others. This is then used in the time-series API endpoint to retrieve data for these related coverages.For example, a request like this:
Would cause the API to return:
tas_absolute_value_ensemble-rcp26-DJFunsmoothedtas_absolute_value_ensemble-rcp26-DJFsmoothed with a moving averageThe below image illustrates a possible outcome, where we show only the unsmoothed values of the main coverage and the smoothed values of all other time series:
In this example we are plotting the TAS absolute values of the 5-model ensemble for the
RCP26scenario in theDJFseason (i.e. winter) as the main time series and also the individual values of four of those models (EC-EARTH-CCLM4-8-17,EC-EARTH-RACMO22E,EC-EARTH_RCA4andHadGEM2-ES_RACMO22E)