eu.itesla_project.offline.forecast_errors.ForecastErrorsAnalysisParameters.java Source code

Introduction

Here is the source code for eu.itesla_project.offline.forecast_errors.ForecastErrorsAnalysisParameters.java

Source

/**
 * Copyright (c) 2016, All partners of the iTesla project (http://www.itesla-project.eu/consortium)
 * This Source Code Form is subject to the terms of the Mozilla Public
 * License, v. 2.0. If a copy of the MPL was not distributed with this
 * file, You can obtain one at http://mozilla.org/MPL/2.0/.
 */
package eu.itesla_project.offline.forecast_errors;

import org.joda.time.DateTime;
import org.joda.time.Interval;

import eu.itesla_project.commons.config.ModuleConfig;
import eu.itesla_project.commons.config.PlatformConfig;
import eu.itesla_project.iidm.network.Country;
import eu.itesla_project.cases.CaseType;

import java.io.Serializable;
import java.util.EnumSet;
import java.util.Set;

/**
 *
 * @author Quinary <itesla@quinary.com>
 */
public class ForecastErrorsAnalysisParameters implements Serializable {

    /*
     * example of fea-parameters.properties file.
     *
        
     #this points to the network to be analyzed (ref. CaseRepository impl)
     baseCaseDate=2013-01-15T18:45:00+01:00
        
    # interval of data to be analysed
    histoInterval=2013-03-01T00:00:00+01:00/2013-12-31T00:00:00+01:00
        
    # id of the forecast errors analysis
     # it should encode meaningful information for identifying it: e.g. network id and some analysis parameters
     # it should not contain "strange" chars and spaces
     feAnalysisId = France_flagPQ0_ir06
        
     # fraction of explained variance for PCA
     ir=0.9
        
    # flagPQ for all injections: if 1, P and Q vars are separately sampled, if 0, Q is sampled with the same initial pf starting from P vars.
    flagPQ=0
        
     # method for missing data imputation: if 1, method = proposed method by RSE; if 2, method = gaussian conditional sampling; if 3, method = gaussian mixture imputation; if 4, interpolation based method
     method=3
        
     # number of clusters for PCA
    nClusters=3
        
    # quantile of the distribution of historical data related to Q vars, to set realistic limits of Q samples in case of using a constant power factor to produce Q samples starting from P samples
    percentileHistorical=0.05
        
    # 1 = fictitious gaussians simulate forecast errors, 0 = historical data with copula estimation are considered
    modalityGaussian=0
        
    # % 0 = outliers are included as valid samples, 1 = outliers are excluded
    outliers=1
        
    # 1 = activated conditional sampling, 0 = not active
    conditionalSampling=0
        
    # number of samples to generate -offline-
    nSamples=100
        
        
    */

    public static final Set<Country> DEFAULT_COUNTRIES = EnumSet.of(Country.FR);
    public static final CaseType DEFAULT_CASE_TYPE = CaseType.FO;

    private final DateTime baseCaseDate;
    private final Interval histoInterval;
    private final String feAnalysisId;
    private final double ir;
    private final Integer flagPQ;
    private final Integer method;
    private final Integer nClusters;
    private final double percentileHistorical;
    private final Integer modalityGaussian;
    private final Integer outliers;
    private final Integer conditionalSampling;
    private final Integer nSamples;
    private final Set<Country> countries;
    private final CaseType caseType;

    public ForecastErrorsAnalysisParameters(DateTime baseCaseDate, Interval histoInterval, String feAnalysisId,
            double ir, Integer flagPQ, Integer method, Integer nClusters, double percentileHistorical,
            Integer modalityGaussian, Integer outliers, Integer conditionalSampling, Integer nSamples,
            Set<Country> countries, CaseType caseType) {
        this.baseCaseDate = baseCaseDate;
        this.histoInterval = histoInterval;
        this.feAnalysisId = feAnalysisId;
        this.ir = ir;
        this.flagPQ = flagPQ;
        this.method = method;
        this.nClusters = nClusters;
        this.percentileHistorical = percentileHistorical;
        this.modalityGaussian = modalityGaussian;
        this.outliers = outliers;
        this.conditionalSampling = conditionalSampling;
        this.nSamples = nSamples;
        this.countries = countries;
        this.caseType = caseType;
    }

    public static ForecastErrorsAnalysisParameters load() {
        ModuleConfig config = PlatformConfig.defaultConfig().getModuleConfig("fea-parameters");

        DateTime baseCaseDate = DateTime.parse(config.getStringProperty("baseCaseDate"));
        Interval histoInterval = Interval.parse(config.getStringProperty("histoInterval"));
        String feAnalysisId = config.getStringProperty("feAnalysisId");
        double ir = config.getDoubleProperty("ir");
        Integer flagPQ = config.getIntProperty("flagPQ");
        Integer method = config.getIntProperty("method");
        Integer nClusters = config.getIntProperty("nClusters");
        double percentileHistorical = config.getDoubleProperty("percentileHistorical");
        Integer modalityGaussian = config.getOptionalIntProperty("modalityGaussian");
        Integer outliers = config.getOptionalIntProperty("outliers");
        Integer conditionalSampling = config.getOptionalIntProperty("conditionalSampling");
        Integer nSamples = config.getIntProperty("nSamples");
        Set<Country> countries = config.getEnumSetProperty("countries", Country.class, DEFAULT_COUNTRIES);
        CaseType caseType = config.getEnumProperty("caseType", CaseType.class, DEFAULT_CASE_TYPE);

        return new ForecastErrorsAnalysisParameters(baseCaseDate, histoInterval, feAnalysisId, ir, flagPQ, method,
                nClusters, percentileHistorical, modalityGaussian, outliers, conditionalSampling, nSamples,
                countries, caseType);
    }

    public DateTime getBaseCaseDate() {
        return baseCaseDate;
    }

    public Interval getHistoInterval() {
        return histoInterval;
    }

    public String getFeAnalysisId() {
        return feAnalysisId;
    }

    public double getIr() {
        return ir;
    }

    public Integer getFlagPQ() {
        return flagPQ;
    }

    public Integer getMethod() {
        return method;
    }

    public Integer getConditionalSampling() {
        return conditionalSampling;
    }

    public Integer getModalityGaussian() {
        return modalityGaussian;
    }

    public Integer getnClusters() {
        return nClusters;
    }

    public Integer getOutliers() {
        return outliers;
    }

    public double getPercentileHistorical() {
        return percentileHistorical;
    }

    public Integer getnSamples() {
        return nSamples;
    }

    public Set<Country> getCountries() {
        return countries;
    }

    public CaseType getCaseType() {
        return caseType;
    }

    @Override
    public String toString() {
        return "ForecastErrorsAnalsysParameters [baseCaseDate=" + baseCaseDate + ", histoInterval=" + histoInterval
                + ", feAnalysisId=" + feAnalysisId + ", ir=" + ir + ", flagPQ=" + flagPQ + ", method=" + method
                + ", nClusters=" + nClusters + ", percentileHistorical=" + percentileHistorical
                + ", modalityGaussian=" + modalityGaussian + ", outliers=" + outliers + ", conditionalSampling="
                + conditionalSampling + ", nSamples=" + nSamples + ", countries=" + countries + ", caseType="
                + caseType + "]";
    }

}