EDIT

/**

* Copyright (C) 2017 EDIT

* European Distributed Institute of Taxonomy

* http://www.e-taxonomy.eu

*

* The contents of this file are subject to the Mozilla Public License Version 1.1

* See LICENSE.TXT at the top of this package for the full license terms.

*/

package eu.etaxonomy.cdm.app.common.tasks;

import org.apache.log4j.Logger;

import eu.etaxonomy.cdm.api.application.CdmApplicationController;

import eu.etaxonomy.cdm.app.common.CdmDestinations;

import eu.etaxonomy.cdm.database.DbSchemaValidation;

import eu.etaxonomy.cdm.database.ICdmDataSource;

import eu.etaxonomy.cdm.model.agent.Person;

import eu.etaxonomy.cdm.model.agent.Team;

import eu.etaxonomy.cdm.model.reference.Reference;

/**

 * Deduplicates certain classes.

 *

 * CAUTION: Deduplicating teams does not yet work correctly. Team members are often duplicated.

 *

 * @author a.mueller

 * @date 24.11.2017

 *

 */

public class Deduplicator {

    @SuppressWarnings("unused")

    private static final Logger logger = Logger.getLogger(Deduplicator.class);

    //database validation status (create, update, validate ...)

//  static DbSchemaValidation hbm2dll = DbSchemaValidation.VALIDATE;

//  static final ICdmDataSource cdmDestination = CdmDestinations.cdm_production_cichorieae();

//  static final ICdmDataSource cdmDestination = CdmDestinations.localH2();

    static final ICdmDataSource cdmDestination = CdmDestinations.cdm_production_caryophyllales_spp();

//  static final ICdmDataSource cdmDestination = CdmDestinations.cdm_local_caryo_spp();

   static boolean doPersons = false;

   static boolean doTeams = true;

   static boolean doReferences = false;

    private void doInvoke(ICdmDataSource destination){

        try {

            CdmApplicationController app = CdmApplicationController.NewInstance(destination, DbSchemaValidation.VALIDATE);

            if (doPersons){

                app.getAgentService().deduplicate(Person.class, null, null);

            }

            if (doTeams){

                app.getAgentService().deduplicate(Team.class, null, null);

            }

            if (doReferences){

                app.getReferenceService().deduplicate(Reference.class, null, null);

            }

        } catch (Exception e) {

            e.printStackTrace();

        }

    }

    /**

     * @param args

     */

    public static void main(String[] args) {

        ICdmDataSource destination = CdmDestinations.chooseDestination(args) != null ? CdmDestinations.chooseDestination(args) : cdmDestination;

        System.out.println("Start updating caches for "+ destination.getDatabase() + "...");

        Deduplicator me = new Deduplicator();

        me.doInvoke(destination);

        System.exit(0);

    }

}

	static final ICdmDataSource cdmDestination = CdmDestinations.cdm_redlist_animalia_production_final();

//    static final ICdmDataSource cdmDestination = CdmDestinations.cdm_test_redlist_mammalia();

//		        //band1

//				"rldb_print_v4_0_1_0_Brutvoegel.xml",BfnXmlImportReferences.uuidBand1_brutvoegel.toString(), //Brutvögel

//				"rldb_print_v4_0_1_0_Reptilien_1.xml",uuidBand1_kriechtiere.toString(),

//                "rldb_print_v4_0_1_0_Amphibien.xml",uuidBand1_lurche.toString(),  //Kriechtiere

//                "RoteListe_v4_0_6_0_BFN_Saeuger_korr.xml",uuidBand1_saeugetiere.toString(), //Säugetiere

//                "rldb_print_v4_0_1_0_Fische.xml",uuidBand1_suessfische.toString(),

//                //

//		        //band2

//		        "rldb_print_v4_0_1_0_artenarmeWeichtiergruppen_121127_verantw_syn.xml",uuidBand2_bodenlebendenWirbellosenMeerestiere.toString(),

//		        "rldb_print_v4_0_1_0_Asselspinnen_120907_verantw_syn.xml",uuidBand2_bodenlebendenWirbellosenMeerestiere.toString(),

//		        "rldb_print_v4_0_1_0_Flohkrebse_121128_verantw_syn.xml",uuidBand2_bodenlebendenWirbellosenMeerestiere.toString(),

//		        "rldb_print_v4_0_1_0_Igelwuermer_120907_verantw.xml",uuidBand2_bodenlebendenWirbellosenMeerestiere.toString(),

//		        "rldb_print_v4_0_1_0_Kumazeen_120709_verantw_syn.xml",uuidBand2_bodenlebendenWirbellosenMeerestiere.toString(),

//		        "rldb_print_v4_0_1_0_Asseln_121128_verantw_syn.xml",uuidBand2_bodenlebendenWirbellosenMeerestiere.toString(),

//		        "rldb_print_v4_0_1_0_Moostierchen_121128_verantw_syn.xml",uuidBand2_bodenlebendenWirbellosenMeerestiere.toString(),

//		        "rldb_print_v4_0_1_0_Muscheln_121128_verantw_syn.xml",uuidBand2_bodenlebendenWirbellosenMeerestiere.toString(),

//		        "rldb_print_v4_0_1_0_Schnecken_130206_verantw_syn.xml",uuidBand2_bodenlebendenWirbellosenMeerestiere.toString(),

//		        "rldb_print_v4_0_1_0_Meeresfische_syn.xml",uuidBand2_meeresfischeUndNeunaugen.toString(),

//		        "rldb_print_v4_0_1_0_Nesseltiere_130104_verantw_syn.xml",uuidBand2_bodenlebendenWirbellosenMeerestiere.toString(),

//		        "rldb_print_v4_0_1_0_Schaedellose_120907_verantw_syn.xml",uuidBand2_bodenlebendenWirbellosenMeerestiere.toString(), //Schädellose

//		        "rldb_print_v4_0_1_0_Schwaemme_121127_verantw_syn.xml",uuidBand2_bodenlebendenWirbellosenMeerestiere.toString(), //Schwämme

//		        "rldb_print_v4_0_1_0_Seepocken_121128_verantw_syn.xml",uuidBand2_bodenlebendenWirbellosenMeerestiere.toString(),

//		        "rldb_print_v4_0_1_0_Seescheiden_121128_verantw_syn.xml",uuidBand2_bodenlebendenWirbellosenMeerestiere.toString(),

//		        "rldb_print_v4_0_1_0_Stachelhaeuter_121128_verantw_syn.xml",uuidBand2_bodenlebendenWirbellosenMeerestiere.toString(), //Stachelhäuter

//		        "rldb_print_v4_0_1_0_Vielborster_130206_verantw_syn.xml",uuidBand2_bodenlebendenWirbellosenMeerestiere.toString(),

//		        "rldb_print_v4_0_1_0_Wenigborster_121128_verantw_syn.xml",uuidBand2_bodenlebendenWirbellosenMeerestiere.toString(),

//		        "rldb_print_v4_0_1_0_Zehnfusskrebse_130104_verantw_syn.xml", uuidBand2_bodenlebendenWirbellosenMeerestiere.toString(),

//		        //

//		        //band3

//			    "rldb_print_v4_0_1_0_Ameisen_110609_rev120113_syn.xml", uuidBand3_ameisen.toString(), //Hymenoptera: Formicidae

//			    "rldb_print_v4_0_1_0_Bienen_PWKorr_HG_120413_DF_120612_syn.xml",uuidBand3_bienen.toString(),  //Hymnenoptera: Apidae

//			    "rldb_print_v4_0_1_0_Binnenmollusken_0alle_120413_DF_syn.xml",uuidBand3_binnenmollusken.toString(),

//			    //eulenfalter - korrekt?

//			    "rldb_print_v4_0_1_0_Eulen_Korruebern_23-05-2012_KorrV_syn.xml",uuidBand3_eulenfalter.toString(), //Eulen & Korrübern

//			    "rldb_print_v4_0_1_0_Thysanoptera_120413_DF_korrV_Verantw.xml",uuidBand3_fransenfluegler.toString(),

//			    "rldb_print_v4_0_1_0_Heuschrecken_syn.xml",uuidBand3_heuschrecken.toString(),  //Saltatoria

//			    "rldb_print_v4_0_1_0_Ohrwuermer_DF_syn.xml",uuidBand3_ohrwuermer.toString(),   //Dermaptera

//			    "rldb_print_v4_0_1_0_Pflanzenwespen_280711_Autor_110815_HG2_120413_DF_syn.xml",uuidBand3_pflanzenwespen.toString(), //Hymenoptera: Symphata

//			    "rldb_print_v4_0_1_0_Asilidae_GMH_Wolff_110314_HGxls_120413_DF_korrV_Verantw_syn.xml",uuidBand3_raubfliegen.toString(),

//			    "rldb_print_v4_0_1_0_Blattoptera_140413_DF_syn.xml",uuidBand3_schaben.toString(),

//			    "rldb_print_v4_0_1_0_Schwebfliegen_111103_KorrAS_120413_DF_syn.xml",uuidBand3_schwebfliegen.toString(), //Diptera: Syrphidae

//			    //spanner - korrekt?

//				"rldb_print_v4_0_1_0_Eulenspinner_Spanner_13-06-2012_KorrV_syn.xml",uuidBand3_spanner.toString(), //Eulenspinner & Spanner

//			    "rldb_print_v4_0_1_0_Spinner_Oktober2011_eingearbKorr_120124_Korruebern_MB_02-05-2012_KorrV_syn.xml",uuidBand3_spinner.toString(),

//			    "rldb_print_v4_0_1_0_Tagfalter_06-06-2012_KorrV_syn.xml",uuidBand3_tagfalter.toString(),

//			    "rldb_print_v4_0_1_0_Empidoidea_120413_DF.xml",uuidBand3_tanzfliegen.toString(),  //Empidoidea

//			    //wespen - fehlen ????? => siehe auch titel des Referenz Word Files

//			    "rldb_print_v4_0_1_0_Pyraloidea_Februar_2012_Korruebern_MB_24-04-2012_syn.xml",uuidBand3_zuenslerfalter.toString()

//		        "RoteListe_v4_0_6_0_BFN_Saeuger_korr.xml",BfnXmlImportReferences.uuidBand1_saeugetiere.toString(),

                "rldb_print_v4_0_1_0_Amphibien.xml",BfnXmlImportReferences.uuidBand1_lurche.toString(),

                "rldb_print_v4_0_1_0_Reptilien_1.xml",BfnXmlImportReferences.uuidBand1_kriechtiere.toString(),

                "rldb_print_v4_0_1_0_Heuschrecken_syn.xml", BfnXmlImportReferences.uuidBand3_heuschrecken.toString()

    private static UUID pteridophytaUuid = UUID.fromString("cd2a7c42-4e2b-4a42-8044-18c8e3ccb76d");

            rootNode = getTaxonNodeService().find(pteridophytaUuid);

//        UUID altFamUuid1;

//            altFamUuid1 = state.getTransformer().getFeatureUuid("Alt.Fam.");

        //TextData  //not used anymore

//        Feature feature1 = getFeature(state, altFamUuid1, "Families in other Floras (Text)", "Families in other Floras (Text)", "Other floras", null);

//        feature1.addRepresentation(Representation.NewInstance("Familias en otras Floras", "Familias en otras Floras", null, Language.SPANISH_CASTILIAN()));

////        TextData textData = TextData.NewInstance(feature1, famStr, Language.DEFAULT(), null);

//        TextData textData = TextData.NewInstance(feature1, null, Language.DEFAULT(), null);

//        textData.putText(Language.SPANISH_CASTILIAN(), "Familias en otras Floras");

//        textData.addSource(OriginalSourceType.PrimaryTaxonomicSource, null,null, famRef, null, famTaxon.getName(),null);

//        desc.addElement(textData);

        Feature feature2 = getFeature(state, altFamUuid2, "Families in other Floras", "Families in other Floras", "Other floras", null);

        //feature should exist already

//        feature2.setSupportsTaxonInteraction(true);

//        feature2.addRepresentation(Representation.NewInstance("Familias en otras Floras", "Familias en otras Floras", null, Language.SPANISH_CASTILIAN()));

        //Concept Relation  //not used anymore

//        famTaxon.addTaxonRelation(taxon, TaxonRelationshipType.INCLUDES(), taxon.getSec(), null);

/**

* Copyright (C) 2017 EDIT

* European Distributed Institute of Taxonomy

* http://www.e-taxonomy.eu

*

* The contents of this file are subject to the Mozilla Public License Version 1.1

* See LICENSE.TXT at the top of this package for the full license terms.

*/

package eu.etaxonomy.cdm.io.greece;

import java.io.File;

import java.io.IOException;

import java.net.URI;

import java.util.ArrayList;

import org.apache.http.HttpException;

import org.apache.log4j.Logger;

import org.apache.sanselan.ImageReadException;

import org.apache.sanselan.Sanselan;

import org.apache.sanselan.common.IImageMetadata;

import org.apache.sanselan.common.ImageMetadata.Item;

import org.joda.time.DateTime;

import org.joda.time.DateTimeZone;

import org.joda.time.format.DateTimeFormat;

import org.joda.time.format.DateTimeFormatter;

import eu.etaxonomy.cdm.common.media.ImageInfo;

import eu.etaxonomy.cdm.io.common.utils.ImportDeduplicationHelper;

import eu.etaxonomy.cdm.model.agent.AgentBase;

import eu.etaxonomy.cdm.model.description.TaxonDescription;

import eu.etaxonomy.cdm.model.description.TextData;

import eu.etaxonomy.cdm.model.media.ImageFile;

import eu.etaxonomy.cdm.model.media.Media;

import eu.etaxonomy.cdm.model.taxon.Taxon;

/**

 * @author a.mueller

 * @date 13.05.2017

 *

 */

public class ImageImportTest {

    @SuppressWarnings("unused")

    private static final Logger logger = Logger.getLogger(FloraHellenicaImageImport.class);

    private static final String BASE_URL = "https://media.e-taxonomy.eu/flora-greece/";

    private static final String IMAGE_FOLDER = "////BGBM-PESIHPC/Greece/thumbs/";

    private ImportDeduplicationHelper deduplicationHelper = null;

    protected void doInvoke() {

        for (int plate = 1; plate < 22 ; plate++){

            System.out.println("Plate: " + plate);

            String fill = plate < 10 ? "0" : "";

            String plateStr = "Plate_" + fill + plate + "/";

            String fullFolderUrl = BASE_URL + plateStr;

            String fullThumbUrl = BASE_URL + "thumbs/" + plateStr;

            String folderStr = IMAGE_FOLDER + plateStr;

            File file = new File(folderStr);

            String[] list = file.list();

            System.out.println(DateTimeZone.getAvailableIDs());

            for (String fileStr : list){

                String[] taxonNameAndArtist = getTaxonName(fileStr);

                String taxonNameStr = taxonNameAndArtist[0];

                String artistStr = taxonNameAndArtist[1];

                if(false){

                    continue;

                }

                Taxon taxon = getAcceptedTaxon(taxonNameStr);

                TaxonDescription imageGallery = taxon.getImageGallery(true);

                TextData textData = TextData.NewInstance();

                URI uri = URI.create(fullFolderUrl + fileStr);

                //image metadata

                File imageFile = new File("");

                IImageMetadata metadata;

                try {

                    metadata = Sanselan.getMetadata(uri.toURL().openStream(), null);

                    ArrayList<?> items = metadata.getItems();

                    for (Object object : items){

                        Item item = (Item) object;

                        System.out.println(item.getKeyword() +  ":    " + item.getText());

                        String value = removeQuots(item.getText());

                        if("Image Description".equalsIgnoreCase(item.getKeyword())){

//                          media.putDescription(Language.DEFAULT(), item.getText());

                        }else if ("date time original".equalsIgnoreCase(item.getKeyword())){

                            DateTimeFormatter f = DateTimeFormat.forPattern("yyyy:MM:dd HH:mm:ss");

                            DateTime created = f.withZone(DateTimeZone.forID("Europe/Athens")).parseDateTime(value);

                            System.out.println(created);

                        }

                    }

                } catch (ImageReadException | IOException e1) {

                    // TODO Auto-generated catch block

                    e1.printStackTrace();

                }

                ImageInfo imageMetaData;

                try {

                    imageMetaData = ImageInfo.NewInstance(uri, 0);

                    String mimeType = imageMetaData.getMimeType();

                    String suffix = null;

                    int height = imageMetaData.getHeight();

                    int width = imageMetaData.getWidth();

                    Integer size = null;

                    DateTime mediaCreated = null;

                    AgentBase<?> artist = null;

                    Media media = ImageFile.NewMediaInstance(mediaCreated, artist, uri, mimeType, suffix, size, height, width);

                    textData.addMedia(media);

                    imageGallery.addElement(textData);

                } catch (IOException | HttpException e) {

                    // TODO Auto-generated catch block

                    e.printStackTrace();

                }

            }

        }

    }

    /**

     * @param text

     * @return

     */

    private String removeQuots(String text) {

        if (text.startsWith("'") && text.endsWith("'")){

            return text.substring(1, text.length() -1);

        }else{

            return text;

        }

    }

    /**

     * @param taxonNameStr

     * @return

     */

    private Taxon getAcceptedTaxon(String taxonNameStr) {

        return Taxon.NewInstance(null, null);

    }

    /**

     * @param fileStr

     * @return

     */

    private String[] getTaxonName(String fileStr) {

        String[] result = new String[2];

        fileStr = fileStr.split("\\.")[0];

        fileStr = fileStr.replaceAll("[0-9]", "");

        String[] x = fileStr.split("_");

        if (x.length == 2){

            result[1] = x[1];

        }

        fileStr = splitCamelCase(x[0]);

        String[] split = fileStr.split(" ");

        String name = split[0] + " " + split[1].toLowerCase() +

                (split.length > 2 ? " subsp. " + split[2].toLowerCase() : "");

        result[0] = name;

        System.out.println(result[0] + (result[1] != null ?  "   Artist: " + result[1]: ""));

        return result;

    }

    //from http://stackoverflow.com/questions/2559759/how-do-i-convert-camelcase-into-human-readable-names-in-java

    static String splitCamelCase(String s) {

        return s.replaceAll(

           String.format("%s",

//              "(?<=[A-Z])(?=[A-Z][a-z])",

              "(?<=[^A-Z])(?=[A-Z])"

//              "(?<=[A-Za-z])(?=[^A-Za-z])"

           ),

           " "

        );

     }

    public static void main(String[] str){

        ImageImportTest test = new ImageImportTest();

        test.doInvoke();

    }

}