com.github.houbb.auto.summary.api.impl.AutoSummaryWeight Maven / Gradle / Ivy

Go to download
Show more of this group Show more artifacts with this name
Show all versions of auto-summary Show documentation
The auto-summary tool for java.
There is a newer version: 0.0.3
package com.github.houbb.auto.summary.api.impl;

import com.github.houbb.auto.summary.api.IAutoSummaryContext;
import com.github.houbb.auto.summary.api.IAutoSummaryResult;
import com.github.houbb.heaven.annotation.ThreadSafe;
import com.github.houbb.heaven.support.handler.IHandler;
import com.github.houbb.heaven.support.handler.IMapHandler;
import com.github.houbb.heaven.support.tuple.impl.Pair;
import com.github.houbb.heaven.util.guava.Guavas;
import com.github.houbb.heaven.util.lang.ObjectUtil;
import com.github.houbb.heaven.util.lang.StringUtil;
import com.github.houbb.heaven.util.util.CollectionUtil;
import com.github.houbb.heaven.util.util.MapUtil;
import com.github.houbb.keyword.api.IKeywordResult;
import com.github.houbb.keyword.api.impl.KeywordResult;
import com.github.houbb.keyword.model.KeywordFreqBean;
import com.github.houbb.keyword.util.KeywordHelper;

import java.util.*;

/**
 * 权重
 *  project: auto-summary-IAutoSummary 
 *  create on 2020/3/12 21:56 
 *
 * @author binbin.hou
 * @since 0.0.2
 */
@ThreadSafe
public class AutoSummaryWeight extends AbstractAutoSummary {

    @Override
    protected List getSummaryList(List sentenceList, List keywordResults,
                                                      IAutoSummaryContext context) {
        final int limit = context.limit();
        final Map freqMap = MapUtil.toMap(keywordResults, new IMapHandler() {
            @Override
            public String getKey(IKeywordResult o) {
                return o.word();
            }

            @Override
            public Double getValue(IKeywordResult o) {
                return o.freq();
            }
        });

        //TOP-K 算法
        List> pairs = Guavas.newArrayList();
        for(String sentence : sentenceList) {
            double rank = calcRank(sentence, freqMap);
            Pair pair = Pair.of(sentence, rank);
            pairs.add(pair);
        }

        // 排序
        Collections.sort(pairs, new Comparator>() {
            @Override
            public int compare(Pair o1, Pair o2) {
                return o1.getValueTwo().compareTo(o2.getValueTwo());
            }
        });

        //获取前几个
        return CollectionUtil.toList(pairs.subList(0, limit), new IHandler, IAutoSummaryResult>() {
            @Override
            public IAutoSummaryResult handle(Pair objects) {
                return AutoSummaryResult.newInstance().sentence(objects.getValueOne())
                        .weight(objects.getValueTwo());
            }
        });
    }

    /**
     * 计算评分
     * @param sentence 语句
     * @param freqMap 频率
     * @return 结果
     * @since 0.0.2
     */
    private double calcRank(final String sentence, final Map freqMap) {
        List keywords = KeywordHelper.keyword(sentence, Integer.MAX_VALUE);

        double result = 0.0;

        for(String keyword : keywords) {
            Double freq = freqMap.get(keyword);
            if(ObjectUtil.isNotNull(freq)) {
                result += freq;
            }
        }

        return result;
    }

}