WebMagic+Spring Boot爬取网易云音乐评论

关于WebMagic

WebMagic是一个简单灵活的Java爬虫框架。基于WebMagic，你可以快速开发出一个高效、易维护的爬虫。
特性：
简单的API，可快速上手
模块化的结构，可轻松扩展
提供多线程和分布式支持
官网：http://webmagic.io/
中文: http://webmagic.io/docs/zh/
English: http://webmagic.io/docs/en
Javadocs: http://webmagic.io/apidocs/

WebMagic由四个组件(PageProcessor、Pipeline、Downloader、Scheduler)构成

• PageProcessor
  PageProcessor负责解析页面，抽取有用信息，以及发现新的链接。
• Pipeline
  Pipeline负责抽取结果的处理，包括计算、持久化到文件、数据库等.。
• Downloader
  Downloader负责从互联网上下载页面，以便后续处理。WebMagic默认使用了Apache HttpClient作为下载工具。
• Scheduler
  Scheduler负责管理待抓取的URL，以及一些去重的工作。WebMagic默认提供了JDK的内存队列来管理URL，并用集合来进行去重。也支持使用Redis进行分布式管理。

项目结构

代码实现

Maven依赖

    <dependency>
        <groupId>us.codecraft</groupId>
        <artifactId>webmagic-core</artifactId>
        <version>0.7.2</version>
    </dependency>
    <dependency>
        <groupId>us.codecraft</groupId>
        <artifactId>webmagic-extension</artifactId>
        <version>0.7.2</version>
    </dependency>

配置数据库

具体代码

@Component
public class NetEaseMusicPageProcessor implements PageProcessor {
    // 正则表达式\\. \\转义java中的\ \.转义正则中的.
    // 主域名
    public static final String BASE_URL = "http://music.163.com/";
    // 匹配专辑URL
    public static final String ALBUM_URL = "http://music\\.163\\.com/playlist\\?id=\\d+";
    // 匹配歌曲URL
    public static final String MUSIC_URL = "http://music\\.163\\.com/song\\?id=\\d+";
    // 初始地址, 褐言喜欢的音乐id 148174530
    public static final String START_URL = "http://music.163.com/playlist?id=148174530";
    public static final int ONE_PAGE = 20;

    private Site site = Site.me()
            .setDomain("http://music.163.com")
            .setSleepTime(1000)
            .setRetryTimes(30)
            .setCharset("utf-8")
            .setTimeOut(30000)
            .setUserAgent("Mozilla/5.0 (Macintosh; Intel Mac OS X 10_7_2) AppleWebKit/537.31 (KHTML, like Gecko) Chrome/26.0.1410.65 Safari/537.31");

    @Override
    public Site getSite() {
        return site;
    }

    @Autowired
    MusicService mMusicService;

    @Override
    public void process(Page page) {
        // 根据URL判断页面类型
        if (page.getUrl().regex(ALBUM_URL).match()) {
            System.out.println("歌曲总数----->" + page.getHtml().xpath("//span[@id='playlist-track-count']/text()").toString());
            // 爬取歌曲URl加入队列
            page.addTargetRequests(page.getHtml().xpath("//div[@id=\"song-list-pre-cache\"]").links().regex(MUSIC_URL).all());
        } else {
            String url = page.getUrl().toString();
            Music music = new Music();
            // 单独对AJAX请求获取评论数, 使用JSON解析返回结果
            String songId = url.substring(url.indexOf("id=") + 3);
            int commentCount = getComment(page, songId, 0);
            // music 保存到数据库
            music.setSongId(songId);
            music.setCommentCount(commentCount);
            music.setTitle(page.getHtml().xpath("//em[@class='f-ff2']/text()").toString());
            music.setAuthor(page.getHtml().xpath("//p[@class='des s-fc4']/span/a/text()").toString());
            music.setAlbum(page.getHtml().xpath("//p[@class='des s-fc4']/a/text()").toString());
            music.setURL(url);
            //page.putField("music", music);
            mMusicService.addMusic(music);
        }
    }

    private int getComment(Page page, String songId, int offset) {
        int commentCount;
        String s = NetEaseMusicUtils.crawlAjaxUrl(songId, offset);

        if (s.contains("503 Service Temporarily Unavailable")) {
            commentCount = -1;
        } else {
            JSONObject jsonObject = JSON.parseObject(s);
            commentCount = (Integer) JSONPath.eval(jsonObject, "$.total");
            for (; offset < commentCount; offset = offset + ONE_PAGE) {
                JSONObject obj = JSON.parseObject(NetEaseMusicUtils.crawlAjaxUrl(songId, offset));
                List<Integer> commentIds = (List<Integer>) JSONPath.eval(obj, "$.comments.commentId");
                List<String> contents = (List<String>) JSONPath.eval(obj, "$.comments.content");
                List<Integer> likedCounts = (List<Integer>) JSONPath.eval(obj, "$.comments.likedCount");
                List<String> nicknames = (List<String>) JSONPath.eval(obj, "$.comments.user.nickname");
                List<Long> times = (List<Long>) JSONPath.eval(obj, "$.comments.time");
                List<Comment> comments = new ArrayList<>();
                for (int i = 0; i < contents.size(); i++) {
                    // 保存到数据库
                    Comment comment = new Comment();
                    comment.setCommentId(commentIds.get(i));
                    comment.setSongId(songId);
                    comment.setContent(NetEaseMusicUtils.filterEmoji(contents.get(i)));
                    comment.setLikedCount(likedCounts.get(i));
                    comment.setNickname(nicknames.get(i));
                    comment.setTime(NetEaseMusicUtils.stampToDate(times.get(i)));
                    comments.add(comment);
                    mMusicService.addComment(comment);
                }

                try {
                    Thread.sleep(1000);
                } catch (InterruptedException e) {
                    e.printStackTrace();
                }
            }
        }

        return commentCount;
    }


    public void start(NetEaseMusicPageProcessor processor, NetEaseMusicPipeline netEaseMusicPipeline) {

        long start = System.currentTimeMillis();
        Spider.create(processor)
                .addUrl(START_URL)
//              .addPipeline(netEaseMusicPipeline)
                .run();
        long end = System.currentTimeMillis();
        System.out.println("爬虫结束,耗时--->" + NetEaseMusicUtils.parseMillisecone(end - start));

    }
}

遇到的主要一个问题就是评论获取，网易对其进行了加密，参考 平胸小仙女的知乎回答的python版本去获取评论数据的。具体实现在NetEaseMusicUtils.java中。

public class NetEaseMusicUtils {
    public static String crawlAjaxUrl(String songId, int offset) {

        CloseableHttpClient httpclient = HttpClients.createDefault();
        CloseableHttpResponse response = null;
        String first_param = "{rid:\"\", offset:\"offset_param\", total:\"true\", limit:\"20\", csrf_token:\"\"}";
        first_param = first_param.replace("offset_param", offset + "");
        //first_param = first_param.replace("limit_param", ONE_PAGE + "");
        try {
            // 参数加密
            // 16位随机字符串，直接FFF
            // String secKey = new BigInteger(100, new SecureRandom()).toString(32).substring(0, 16);
            String secKey = "FFFFFFFFFFFFFFFF";
            // 两遍ASE加密
            String encText = NetEaseMusicUtils.aesEncrypt(aesEncrypt(first_param, "0CoJUm6Qyw8W8jud"), secKey);
            //
            String encSecKey = rsaEncrypt();

            HttpPost httpPost = new HttpPost("http://music.163.com/weapi/v1/resource/comments/R_SO_4_" + songId + "/?csrf_token=");
            httpPost.addHeader("Referer", NetEaseMusicPageProcessor.BASE_URL);

            List<NameValuePair> ls = new ArrayList<NameValuePair>();
            ls.add(new BasicNameValuePair("params", encText));
            ls.add(new BasicNameValuePair("encSecKey", encSecKey));

            UrlEncodedFormEntity paramEntity = new UrlEncodedFormEntity(ls, "utf-8");
            httpPost.setEntity(paramEntity);

            response = httpclient.execute(httpPost);
            HttpEntity entity = response.getEntity();

            if (entity != null) {
                return EntityUtils.toString(entity, "utf-8");
            }

        } catch (Exception e) {
            e.printStackTrace();
        } finally {
            try {
                response.close();
                httpclient.close();
            } catch (IOException e) {
                e.printStackTrace();
            }
        }

        return "";
    }

    /**
     * ASE-128-CBC加密模式可以需要16位
     *
     * @param src 加密内容
     * @param key 密钥
     * @return
     */
    public static String aesEncrypt(String src, String key) throws Exception {
        String encodingFormat = "UTF-8";
        String iv = "0102030405060708";

        Cipher cipher = Cipher.getInstance("AES/CBC/PKCS5Padding");
        byte[] raw = key.getBytes();
        SecretKeySpec secretKeySpec = new SecretKeySpec(raw, "AES");
        IvParameterSpec ivParameterSpec = new IvParameterSpec(iv.getBytes());
        // 使用CBC模式，需要一个向量vi，增加加密算法强度
        cipher.init(Cipher.ENCRYPT_MODE, secretKeySpec, ivParameterSpec);
        byte[] encrypted = cipher.doFinal(src.getBytes(encodingFormat));
        return new BASE64Encoder().encode(encrypted);

    }

    public static String rsaEncrypt() {
        String secKey = "257348aecb5e556c066de214e531faadd1c55d814f9be95fd06d6bff9f4c7a41f831f6394d5a3fd2e3881736d94a02ca919d952872e7d0a50ebfa1769a7a62d512f5f1ca21aec60bc3819a9c3ffca5eca9a0dba6d6f7249b06f5965ecfff3695b54e1c28f3f624750ed39e7de08fc8493242e26dbc4484a01c76f739e135637c";
        return secKey;
    }

    public static String parseMillisecone(long millisecond) {
        String time = null;
        try {
            long yushu_day = millisecond % (1000 * 60 * 60 * 24);
            long yushu_hour = (millisecond % (1000 * 60 * 60 * 24))
                    % (1000 * 60 * 60);
            long yushu_minute = millisecond % (1000 * 60 * 60 * 24)
                    % (1000 * 60 * 60) % (1000 * 60);
            @SuppressWarnings("unused")
            long yushu_second = millisecond % (1000 * 60 * 60 * 24)
                    % (1000 * 60 * 60) % (1000 * 60) % 1000;
            if (yushu_day == 0) {
                return (millisecond / (1000 * 60 * 60 * 24)) + "天";
            } else {
                if (yushu_hour == 0) {
                    return (millisecond / (1000 * 60 * 60 * 24)) + "天"
                            + (yushu_day / (1000 * 60 * 60)) + "时";
                } else {
                    if (yushu_minute == 0) {
                        return (millisecond / (1000 * 60 * 60 * 24)) + "天"
                                + (yushu_day / (1000 * 60 * 60)) + "时"
                                + (yushu_hour / (1000 * 60)) + "分";
                    } else {
                        return (millisecond / (1000 * 60 * 60 * 24)) + "天"
                                + (yushu_day / (1000 * 60 * 60)) + "时"
                                + (yushu_hour / (1000 * 60)) + "分"
                                + (yushu_minute / 1000) + "秒";

                    }

                }

            }

        } catch (Exception e) {
            e.printStackTrace();
        }
        return time;
    }

    /*
     * 将时间戳转换为时间
     */
    public static String stampToDate(long s) {
        String res;
        SimpleDateFormat simpleDateFormat = new SimpleDateFormat("yyyy-MM-dd HH:mm:ss");
        long lt = s;
        Date date = new Date(lt);
        res = simpleDateFormat.format(date);
        return res;
    }

    /**
     * 将emoji表情替换成*
     *
     * @param source
     * @return 过滤后的字符串
     */
    public static String filterEmoji(String source) {
        if (StringUtils.isNotBlank(source)) {
            return source.replaceAll("[\\ud800\\udc00-\\udbff\\udfff\\ud800-\\udfff]", "*");
        } else {
            return source;
        }
    }
}

展示

最后

运行NetEaseMusicApplication.java
localhost:8888/...

之前评论每页50条爬的，爬了4万多条就503了，改成每页20条和配置了代理， 配置代理ProxyProvider

项目地址：https://github.com/EzioL/neteasemusic