de.julielab.geneexpbase.genemodel.DictionaryGeneIdRecord Maven / Gradle / Ivy

Go to download
package de.julielab.geneexpbase.genemodel;

import java.util.Arrays;
import java.util.List;
import java.util.stream.Collectors;
import java.util.stream.IntStream;
import java.util.stream.Stream;

/**
 * 
 * Represents an entry from a dictionary that matches gene mentions and assigns possible IDs.
 * 
 * The source of the match is a dictionary that lists gene names and, for each name, its possible IDs in different organisms.
 * The respective information is encoded into string form to be used as entity category by the gazetteer component:
 * 
 *  id1:tax1:5|id2:tax2:3
 * 
 * The format for each ID is the ID, the corresponding taxonomy ID and the priority level the gene name has respective
 * to the gene ID. Multiple such entries are separated by pipe symbols. The priority denotes the source of the gene name
 * for the specific gene ID. It is a number that ranges from -1 to 7. -1 indicates that the name is the official gene symbol
 * for the ID, 7 means that the name came from the BioThesaurus which sometimes introduces false positives due to rather general names.
 * 
 */
public class DictionaryGeneIdRecord {
    private String geneName;
    private final List recordItems;

    public DictionaryGeneIdRecord(String recordString) {
        // split the name from the ID record
        String[] split = recordString.split("\t");
        this.geneName = split[0];
        // now parse the individual elements of the record which are separated by pipe symbols
        recordItems = Arrays.stream(split[1].split("\\|")).map(DictionaryGeneIdRecordItem::new).collect(Collectors.toList());
    }

    public String getGeneName() {
        return geneName;
    }

    public void setGeneName(String geneName) {
        this.geneName = geneName;
    }

    public Stream getRecordItemsForTaxId(String tax) {
        return recordItems.stream().filter(r -> r.getTaxId().equals(tax));
    }

    public Stream getGeneIds() {
        return recordItems.stream().map(DictionaryGeneIdRecordItem::getGeneId);
    }

    public Stream getTaxIds() {
        return recordItems.stream().map(DictionaryGeneIdRecordItem::getTaxId);
    }

    public IntStream getPriorities() {
        return recordItems.stream().mapToInt(DictionaryGeneIdRecordItem::getPriority);
    }
}