Java 获取热搜并生成图片

news2024/10/11 10:11:05

效果图如下：
在这里插入图片描述

第一步获取热搜

  public List<String> getHotNews4(Integer size) {
        if (size <= 0 || StringUtils.isEmpty(size)) {
            return null;
        }
        try {
            //set 转list
            return new ArrayList<>(getHotNews(size));
        } catch (Exception e) {
            logger.error("抓取热点排行榜异常：", e);
            return null;
        }
    }

    public Set<String> getHotNews(int size) {
        try {
            List<NewsModel> list1 = grabBaiduHotNews();
            List<NewsModel> list2 = grabWeiBoHotNews();
            List<NewsModel> list3 = getDouYinHotNews();
            //40%的权重是百度，30%的权重是微博 ,30%的权重是抖音
            int l1 = size * 40 / 100;
            int l2 = size * 30 / 100;
            int l3 = size - l1 - l2;
            Set<String> set = new HashSet<>();
            if (list1.size() < l1) {
                l1 = list1.size();
                l2 = size - l1 - l2;
            }
            if (list2.size() < l2) {
                l2 = list2.size();
                l3 = size - l1 - l2;
            }
            if (list3.size() < l3) {
                l3 = list3.size();
            }
            int j = 0;
            for (j = 0; j < l1; j++) {
                set.add(list1.get(j).getTitle());
                // sb.append(list1.get(i).getTitle()).append("。");
            }
            for (int i = 0; i < l2; i++) {
                set.add(list2.get(i).getTitle());
            }
            for (int i = 0; i < l3; i++) {
                set.add(list3.get(i).getTitle());
            }
            if (set.size() < size) {
                int n = size - set.size();
                if ((list1.size() - l1) < n) {
                    n = list1.size() - l1;
                }
                for (int i = 0; i < n; i++) {
                    set.add(list1.get(++j).getTitle());
                }
            }
            return set;
        } catch (Exception e) {
            logger.error("抓取热点排行榜异常：", e);
            return null;
        }
    }

    /**
     * 抓取百度热点排行榜
     *
     * @return
     */
    public List<NewsModel> grabBaiduHotNews() {
        String url = "https://top.baidu.com/board?tab=realtime&sa=fyb_realtime_31065";
        List<NewsModel> list = new ArrayList<>();
        try {
            Document doc = Jsoup.connect(url).get();
            //标题
            Elements titles = doc.select(".c-single-text-ellipsis");
            //图片
            Elements imgs = doc.select(".category-wrap_iQLoo .index_1Ew5p").next("img");
            //内容
            Elements contents = doc.select(".hot-desc_1m_jR.large_nSuFU");
            //推荐图
            Elements urls = doc.select(".category-wrap_iQLoo a.img-wrapper_29V76");
            //热搜指数
            Elements levels = doc.select(".hot-index_1Bl1a");
            for (int i = 0; i < levels.size(); i++) {
                NewsModel o = new NewsModel();
                o.setTitle(titles.get(i).text().trim());
                o.setImg(imgs.get(i).attr("src"));
                o.setContent(contents.get(i).text().replaceAll("查看更多>", "").trim());
                o.setUrl(urls.get(i).attr("href"));
//                o.setLevel(levels.get(i).text().trim());
                list.add(o);
            }
            return list;

        } catch (IOException e) {
            logger.error("抓取百度热点排行榜异常：" + e.getMessage());
        }
        return null;
    }

    /**
     * 抓取微博热搜榜
     */
    public List<NewsModel> grabWeiBoHotNews() {
        String url = "https://weibo.com/ajax/statuses/hot_band";
        String s = HttpUtil.get(url);
        WeiBoHot weiBoHot = JSON.parseObject(s, WeiBoHot.class);
        List<NewsModel> list = new ArrayList<>();
        for (WeiBoHot.WeiBo weiBo : weiBoHot.getData().getBand_list()) {
            NewsModel o = new NewsModel();
            o.setTitle(weiBo.getNote());
            o.setImg(weiBo.getMblog());
            o.setContent(weiBo.getWord());
            o.setUrl(weiBo.getWord_scheme());
            list.add(o);
        }
        return list;
    }

    public List<NewsModel> getDouYinHotNews() {
        try {

            String s = HttpUtils.get("https://www.iesdouyin.com/web/api/" +
                    "v2/hotsearch/billboard/word/?reflow_source=reflow_page");

            List<NewsModel> list = new ArrayList<>();
            JSONArray jsonArray = JSON.parseObject(s).getJSONArray("word_list");
            for (int i = 0; i < jsonArray.size(); i++) {
                JSONObject jsonObject = jsonArray.getJSONObject(i);
                list.add(new NewsModel(jsonObject.getString("word")));
            }

            return list;
        } catch (Exception e) {
            logger.error("", e);
        }
        //防止空指针
        return new ArrayList<>();
    }

第二步生成图片



public class WordCodeUtil {

   private static final Logger logger = LoggerFactory.getLogger(WordCodeUtil.class);

    public static WordCloud getWordCode(List<String> words) {
        words.remove(words.size() - 1);
        FrequencyAnalyzer frequencyAnalyzer = new FrequencyAnalyzer();
        frequencyAnalyzer.setWordFrequenciesToReturn(600);

        frequencyAnalyzer.setMinWordLength(3);
        // 引入中文解析器
        frequencyAnalyzer.setWordTokenizer(new ChineseWordTokenizer());
        final List<WordFrequency> wordFrequencyList = frequencyAnalyzer.load(words);
        int i = words.size();
        for (String s : words) {
            wordFrequencyList.add(new WordFrequency(s, i--));
        }
        // 设置图片分辨率
        Dimension dimension = new Dimension(940, 400);
        // 创建词云对象
        WordCloud wordCloud = new WordCloud(dimension, CollisionMode.PIXEL_PERFECT);
        // 加载字体文件
        try (BufferedInputStream bis = getFileFromResource("LXGWWenKaiMono-Bold.ttf")) {
            Font font = Font.createFont(Font.TRUETYPE_FONT, bis);
            wordCloud.setKumoFont(new KumoFont(font));
        } catch (IOException e) {
           logger.error("加载字体文件失败！",e);
        } catch (FontFormatException e) {
            throw new RuntimeException(e);
        }

        // 设置词云样式
        wordCloud.setPadding(3);
        wordCloud.setColorPalette(new ColorPalette(
                new Color(0xed1941), new Color(0xf26522), new Color(0x845538),
                new Color(0x8a5d19), new Color(0x7f7522), new Color(0x5c7a29),
                new Color(0x1d953f), new Color(0x007d65), new Color(0x65c294)
        ));
        wordCloud.setBackground(new RectangleBackground(dimension));
        wordCloud.setFontScalar(new SqrtFontScalar(12, 45));
        wordCloud.setBackgroundColor(new Color(255, 255, 255));

        // 生成词云
        wordCloud.build(wordFrequencyList);


        return wordCloud;



    }

    //返回Base64到览器
    public static String getWordCloud2(List<String> words){
        return BufferedImageToBase64(getWordCode(words).getBufferedImage());
    }

    public static void getWordCloud3(List<String> words ,
                                     HttpServletResponse response) throws IOException {
       WordCloud wordCloud = getWordCode(words);
       wordCloud.writeToStreamAsPNG(response.getOutputStream());
       response.flushBuffer();
    }


    /**
     * 方法六：使用Hutool的ResourceUtil
     * 备注：jar包可用
     *
     * @param fileName
     * @return
     * @throws IOException
     */
    public static BufferedInputStream getFileFromResource(String fileName) throws IOException {
        List<URL> resources = ResourceUtil.getResources(fileName);
        URL resource = resources.get(0);
        return new BufferedInputStream(resource.openStream());
    }



    /**
     * BufferedImage 编码转换为 base64
     * @param bufferedImage
     * @return
     */
    private static String BufferedImageToBase64(BufferedImage bufferedImage) {
        ByteArrayOutputStream bao = new ByteArrayOutputStream();//io流
        try {
            ImageIO.write(bufferedImage, "png", bao);//写入流中
        } catch (IOException e) {
            logger.error(e.getMessage(),e);
        }
        byte[] bytes = Base64.getEncoder().encode(bao.toByteArray());
        String base64 = new String(bytes);
        base64 = base64.replaceAll("\n", "").replaceAll("\r", "");//删除 \r\n
        return "data:image/png;base64," + base64;
    }
}