source: SRUAggregator/trunk/src/main/java/eu/clarin/sru/fcs/aggregator/search/Kwic.java @ 6065

package eu.clarin.sru.fcs.aggregator.search;

import eu.clarin.sru.client.fcs.DataViewHits;
import eu.clarin.sru.fcs.aggregator.app.Aggregator;
import eu.clarin.sru.fcs.aggregator.util.LanguagesISO693;
import java.util.ArrayList;
import java.util.List;
import org.apache.commons.lang.StringEscapeUtils;

/**
 * Represents keyword in context data view and information about its PID and
 * reference.
 *
 * @author Yana Panchenko
 */
public class Kwic {

        public static class TextFragment {

                String text;
                boolean isHit;

                public TextFragment(String text, boolean isHit) {
                        this.text = text;
                        this.isHit = isHit;
                }

                public String getText() {
                        return text;
                }

                public boolean isHit() {
                        return isHit;
                }

                @Override
                public String toString() {
                        return (isHit ? "[" : "") + text + (isHit ? "]" : "");
                }
        }

        private String pid;
        private String reference;
        private String language;
        private List<TextFragment> fragments = new ArrayList<TextFragment>();

        public Kwic(DataViewHits hits, String pid, String reference) {
                this.pid = pid;
                this.reference = reference;

                // warning: the client library doesn't unescape the xml
                // so the text can still contains &lt; and &amp; codes
                String str = hits.getText();

                int lastOffset = 0;
                for (int i = 0; i < hits.getHitCount(); i++) {
                        int[] offsets = hits.getHitOffsets(i);
                        if (lastOffset < offsets[0]) {
                                String text = StringEscapeUtils.unescapeXml(str.substring(lastOffset, offsets[0]));
                                fragments.add(new TextFragment(text, false));
                        }
                        if (offsets[0] < offsets[1]) {
                                String text = StringEscapeUtils.unescapeXml(str.substring(offsets[0], offsets[1]));
                                fragments.add(new TextFragment(text, true));
                        }
                        lastOffset = offsets[1];
                }
                if (lastOffset < str.length()) {
                        String text = StringEscapeUtils.unescapeXml(str.substring(lastOffset, str.length()));
                        fragments.add(new TextFragment(text, false));
                }

                String code_iso639_1 = Aggregator.getInstance().detectLanguage(str);
                language = code_iso639_1 == null ? null
                                : LanguagesISO693.getInstance().code_3ForCode(code_iso639_1);
        }

        public List<TextFragment> getFragments() {
                return fragments;
        }

        public String getPid() {
                return pid;
        }

        public String getReference() {
                return reference;
        }

        public String getLanguage() {
                return language;
        }

        @Deprecated
        public String getLeft() {
                StringBuilder sb = new StringBuilder();
                for (TextFragment tf : fragments) {
                        if (tf.isHit) {
                                break;
                        }
                        sb.append(tf.text);
                }
                return sb.toString();
        }

        @Deprecated
        public String getKeyword() {
                for (TextFragment tf : fragments) {
                        if (tf.isHit) {
                                return tf.text;
                        }
                }
                return "";
        }

        @Deprecated
        public String getRight() {
                StringBuilder sb = new StringBuilder();
                boolean pastHit = false;
                for (TextFragment tf : fragments) {
                        if (pastHit) {
                                sb.append(tf.text);
                        }
                        if (tf.isHit) {
                                pastHit = true;
                        }
                }
                return sb.toString();
        }
}