D-Net

package eu.dnetlib.data.mapreduce.hbase.lodExport.preprocessing;

import eu.dnetlib.data.mapreduce.hbase.lodExport.utils.configuration.LodConfiguration;

import eu.dnetlib.data.mapreduce.hbase.lodExport.utils.configuration.Properties;

import org.apache.commons.lang.StringUtils;

import org.apache.hadoop.io.LongWritable;

import org.apache.hadoop.io.Text;

import org.apache.hadoop.mapreduce.Mapper;

import org.apache.log4j.Logger;

import java.io.IOException;

/**

 * Mapper Class that reads HBASE contents and prepares them for the StatsDB

 * export

 */

/*

-> Parse LOD dump files

        Process lod input files and divide by entity type (both source and target)

        Transform to id, array of [ properties] format

        Store to HDFS

        For -> Multiple outputs and inputs

        Multiple inputs: all source and target datasets and their corresponding mappings

        M/O: separate output files for each dataset: mark records so that they are written to the correct one

*/

/*

<http://lod.openaire.eu/data/result/doajarticles::89217af00809a91acc15a416e56b3782> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://www.eurocris.org/ontologies/cerif/1.3#ResultEntity> .

<http://lod.openaire.eu/data/result/doajarticles::89217af00809a91acc15a416e56b3782> <http://www.eurocris.org/ontologies/cerif/1.3#name> "Une nouvelle anomalie a allure h r ditaire chez des agneaux  it khouzistans /it" .

*/

public class SourceMapper extends Mapper<LongWritable, Text, Text, Text> {

    private static final String OA_PREFIX = "OA";

    private static final String TAB_CHARACTER = "\t";

    private static final String DOUBLE_TAB = "\t.\t";

    private static final String BACKSLASH = "\"";

    private static final String SLASH = "/";

    private static final String SPACE = "\\s";

    private Logger log = Logger.getLogger(SourceMapper.class);

    private LodConfiguration lodConfiguration;

    public static enum SOURCE_COUNTERS {

        SOURCE_ENTITIES, TOTAL_ENTITIES

    }

    @Override

    protected void setup(Context context) throws IOException, InterruptedException {

        lodConfiguration = new LodConfiguration();

        lodConfiguration.load(context.getConfiguration().get(Properties.LOD_SOURCE_MAPPINGS));

    }

    @Override

    protected void map(final LongWritable keyIn, final Text result, final Context context) throws IOException {

        try {

            context.getCounter(SOURCE_COUNTERS.TOTAL_ENTITIES).increment(1);

            //get ID - output source_recordID so we can group by id and get all props of a record

            String[] inputRecord = parseInputRecord(result);

            String[] fields = StringUtils.join(inputRecord, BACKSLASH).split(TAB_CHARACTER);

            if (fields.length < 2) {

                log.error("Not a valid record");

                return;

            }

            String subject = fields[0];

            //here addd all  fields as array props and append "\t"

            // between them so we can write directly to output

            //DO NOT enter id- we'll get it from key output

            //extract entity type from subject

            String type = extractType(fields[0]);

            if (lodConfiguration.entityExists(type)) {

                String record = buildRecord(fields, subject);

                // write out type,source_ID as key, and rest of props as value

                writeKeyValuePair(context, type, subject, record);

            }

        } catch (Exception e) {

            log.error("Error while preprocessing", e);

        }

    }

    private String extractType(String value) throws Exception {

        String[] tmp = value.split(SLASH);

        if (tmp.length < 5) {

            log.error("Not a valid record");

            throw new Exception("Error while reading record: missing subject");

        }

        return tmp[4];

    }

    private String buildRecord(String[] fields, String subject) {

        StringBuilder value = new StringBuilder();

        for (int i = 1; i < fields.length - 1; i += 2) {

            String field = fields[i];

            String fieldValue = fields[i + 1];

            if (lodConfiguration.isValidField(field)) {

                value.append(subject).append(TAB_CHARACTER).append(field).append(TAB_CHARACTER).append(fieldValue).append(DOUBLE_TAB);

            }

        }

        return value.toString();

    }

    private void writeKeyValuePair(Context context, String type, String subject, String value) throws IOException, InterruptedException {

        if (value.length() > 0) {

            Text key = new Text(OA_PREFIX + "," + type + "," + subject);

            context.write(key, new Text());

            context.getCounter(SOURCE_COUNTERS.SOURCE_ENTITIES).increment(1);

        }

    }

    private String[] parseInputRecord(Text result) {

        String[] inputParts = result.toString().split(BACKSLASH);

        for (int i = 0; i < inputParts.length; i += 2) {

            inputParts[i] = inputParts[i].replaceAll(SPACE, TAB_CHARACTER);

        }

        return inputParts;

    }

    private static String cleanInput(Text result) {

        String resulString = result.toString().replace("<", "").replace(">", "");

        int ind = resulString.lastIndexOf(".");

        if (ind >= 0) {

            resulString = resulString.substring(0, ind);

        }

        return resulString;

    }

    @Override

    protected void cleanup(Context context) throws IOException, InterruptedException {

        super.cleanup(context);

    }

}