D-Net

package eu.dnetlib.data.mapreduce.hbase.bulktag;

import com.googlecode.protobuf.format.JsonFormat;

import eu.dnetlib.data.bulktag.CommunityConfiguration;

import eu.dnetlib.data.bulktag.CommunityConfigurationFactory;

import eu.dnetlib.data.proto.FieldTypeProtos;

import eu.dnetlib.data.proto.OafProtos;

import eu.dnetlib.data.proto.ResultProtos;

import org.apache.commons.io.IOUtils;

import org.apache.commons.lang3.StringUtils;

import org.apache.hadoop.mapreduce.Counter;

import org.apache.hadoop.mapreduce.Mapper;

import org.dom4j.DocumentException;

import org.junit.Before;

import org.junit.Test;

import org.junit.runner.RunWith;

import org.mockito.Mock;

import org.mockito.junit.MockitoJUnitRunner;

import java.io.IOException;

import java.util.HashSet;

import java.util.List;

import java.util.Set;

import java.util.stream.Collectors;

import static org.junit.Assert.*;

import static org.mockito.Mockito.*;

/**

 * Created by miriam on 02/08/2018.

 */

@RunWith(MockitoJUnitRunner.class)

public class ResultTaggerTest {

    private String xml;

    private CommunityConfiguration cc;

    @Mock

    private Mapper.Context context;

    @Mock

    private Counter counter;

    private ResultTagger resultTagger = new ResultTagger();

    @Before

    public void setUp() throws IOException, DocumentException {

        xml = IOUtils.toString(getClass().getResourceAsStream("community_configuration.xml"));

        when(context.getCounter(anyString(), anyString())).thenReturn(counter);

        cc = CommunityConfigurationFactory.newInstance(xml);

        assertEquals(cc.size(),4);

        cc.getCommunityList().forEach(c -> assertTrue(StringUtils.isNoneBlank(c.getId())));

    }

    @Test

    public void testResultTaggerSubject() throws IOException {

        OafProtos.Oaf oaf = getOaf("oaf_subject.json");

        assertTrue(oaf.getEntity().getResult().getMetadata().getContextList().isEmpty());

        oaf = resultTagger.enrichContext(oaf, cc, context);

        List<ResultProtos.Result.Context> contextList = oaf.getEntity().getResult().getMetadata().getContextList();

        assertNotNull(contextList);

        assertEquals(contextList.size(),1);

        assertEquals(contextList.get(0).getId(),"ni");

    }

    @Test

    public void testResultTaggerSubject2() throws IOException {

        OafProtos.Oaf oaf = getOaf("oaf_subject_2.jason");

        assertTrue(oaf.getEntity().getResult().getMetadata().getContextList().isEmpty());

        oaf = resultTagger.enrichContext(oaf, cc, context);

        List<ResultProtos.Result.Context> contextList = oaf.getEntity().getResult().getMetadata().getContextList();

        assertNotNull(contextList);

        assertEquals(contextList.size(),2);

        Set<String> ids = new HashSet<>();

        ids.addAll(contextList.stream().map(c->c.getId()).collect(Collectors.toList()));

        assertTrue(ids.contains("mes"));

        assertTrue(ids.contains("ni"));

        System.out.println(oaf.toString());

    }

    @Test

    public void testResultTaggerExistingContext() throws IOException {

        OafProtos.Oaf oaf = getOaf("oaf_existingcontext.json");

        assertFalse(oaf.getEntity().getResult().getMetadata().getContextList().isEmpty());

        oaf = resultTagger.enrichContext(oaf, cc, context);

        List<ResultProtos.Result.Context> contextList = oaf.getEntity().getResult().getMetadata().getContextList();

        assertNotNull(contextList);

        assertEquals(contextList.size(),1);

        assertEquals(contextList.get(0).getId(),"ni");

        assertEquals(contextList.get(0).getDataInfoCount(),1);

        assertEquals(contextList.get(0).getDataInfoList().get(0).getInferenceprovenance(),"bulktagging::community");

    }

    @Test

    public void testReadOaf() throws IOException {

        assertNotNull(getOaf("oaf_existingcontext_withprovenance.json"));

    }

    @Test

    public void testResultTaggerExistingContextWithProvenance() throws IOException{

        OafProtos.Oaf oaf = getOaf("oaf_existingcontext_withprovenance.json");

        List<ResultProtos.Result.Context> contextList = oaf.getEntity().getResult().getMetadata().getContextList();

        assertFalse(contextList.isEmpty());

        assertEquals(contextList.size(),1);

        assertEquals(contextList.get(0).getId(),"ni");

        assertEquals(contextList.get(0).getDataInfoCount(),1);

        assertEquals(contextList.get(0).getDataInfoList().get(0).getInferenceprovenance(),"bulktagging::community");

        oaf = resultTagger.enrichContext(oaf, cc, context);

        contextList = oaf.getEntity().getResult().getMetadata().getContextList();

        assertEquals(contextList.size(),1);

        assertEquals(contextList.get(0).getId(),"ni");

        assertEquals(contextList.get(0).getDataInfoCount(),1);

        assertEquals(contextList.get(0).getDataInfoList().get(0).getInferenceprovenance(),"bulktagging::community");

    }

    @Test

    public void testResultTaggerDatasource() throws IOException{

        OafProtos.Oaf oaf = getOaf("oaf_datasource.json");

        List<ResultProtos.Result.Context> contextList = oaf.getEntity().getResult().getMetadata().getContextList();

        assertTrue(contextList.isEmpty());

        oaf = resultTagger.enrichContext(oaf,cc,context);

        contextList = oaf.getEntity().getResult().getMetadata().getContextList();

        assertFalse(contextList.isEmpty());

        assertEquals(contextList.size(),2);

        Set<String> ids = new HashSet<>();

        ids.addAll(contextList.stream().map(c->{

            assertTrue(c.getDataInfoCount()== 1);

            return c.getId();

        }).collect(Collectors.toList()));

        assertTrue(ids.contains("aginfra"));

        assertTrue(ids.contains("ni"));

    }

    @Test

    public void testResultTaggerDatasource2() throws IOException{

        OafProtos.Oaf oaf = getOaf("oaf_datasource2.json");

        List<ResultProtos.Result.Context> contextList = oaf.getEntity().getResult().getMetadata().getContextList();

        assertTrue(contextList.isEmpty());

        oaf = resultTagger.enrichContext(oaf,cc,context);

        contextList = oaf.getEntity().getResult().getMetadata().getContextList();

        assertFalse(contextList.isEmpty());

        assertEquals(contextList.size(),3);

        Set<String> ids = new HashSet<>();

        ids.addAll(contextList.stream().map(c->{

            assertTrue(c.getDataInfoCount()== 1);

            return c.getId();

        }).collect(Collectors.toList()));

        assertTrue(ids.contains("aginfra"));

        assertTrue(ids.contains("ni"));

        assertTrue(ids.contains("mes"));

    }

    @Test

    public void testResultTaggerDatasource3() throws IOException{

        OafProtos.Oaf oaf = getOaf("oaf_existingcontext3.json");

        List<ResultProtos.Result.Context> contextList = oaf.getEntity().getResult().getMetadata().getContextList();

        assertFalse(contextList.isEmpty());

        oaf = resultTagger.enrichContext(oaf,cc,context);

        contextList = oaf.getEntity().getResult().getMetadata().getContextList();

        assertFalse(contextList.isEmpty());

        assertEquals(contextList.size(),1);

        assertTrue(contextList.get(0).getId().equals("ni"));

        assertTrue(contextList.get(0).getDataInfoCount() == 2);

        Set<String> provenance = new HashSet<>();

        final List<FieldTypeProtos.DataInfo> dataInfoList = contextList.get(0).getDataInfoList();

        provenance.addAll(dataInfoList.stream().map(d->d.getInferenceprovenance()).collect(Collectors.toList()));

        assertTrue(provenance.size()==2);

        assertTrue(provenance.contains("bulktagging::community"));

    }

    private OafProtos.Oaf getOaf(String oafjson) throws IOException {

        final String json = IOUtils.toString(getClass().getResourceAsStream(oafjson));

        final OafProtos.Oaf.Builder oaf = OafProtos.Oaf.newBuilder();

        JsonFormat.merge(json, oaf);

        return oaf.build();

    }

}