1、36壁纸网爬虫实现并保存到磁盘然后持久化到数据库

xjs-business/xjs-business-common/src/main/java/com/xjs/consts/RedisConst.java

@@ -44,6 +44,10 @@ public class RedisConst {
      */
     public static final String FORECAST_WEATHER = "weather:forecast";
 
+    /**
+     * 爬虫记录循环次数常量信息
+     */
+    public static final String REPTILE_COUNT= "reptile:count";
 
     //-------------------有效时间-----------------------
     public static final Integer TRAN_DICT_EXPIRE = 1;   //小时

xjs-business/xjs-business-common/src/main/java/com/xjs/consts/ReptileConst.java

@@ -31,6 +31,9 @@ public class ReptileConst {
     public static final String BOSS_JOB_URL= "https://www.zhipin.com";
 
 
+    public static final String _36_WALLPAPER_URL= "https://www.3gbizhi.com/";
+
+
 
 
 

xjs-business/xjs-business-webmagic/src/main/java/com/xjs/_36wallpaper/controller/_36wallpaperController.java

@@ -0,0 +1,33 @@
+package com.xjs._36wallpaper.controller;
+
+import com.ruoyi.common.core.domain.R;
+import com.xjs._36wallpaper.task._36wallpaperTask;
+import io.swagger.annotations.Api;
+import io.swagger.annotations.ApiOperation;
+import org.springframework.beans.factory.annotation.Autowired;
+import org.springframework.web.bind.annotation.GetMapping;
+import org.springframework.web.bind.annotation.RequestMapping;
+import org.springframework.web.bind.annotation.RestController;
+
+/**
+ * 36壁纸网爬虫controller
+ * @author xiejs
+ * @since 2022-02-20
+ */
+@RestController
+@RequestMapping("_36wallpaper")
+@Api(tags = "爬虫模块-36壁纸网")
+public class _36wallpaperController {
+
+    @Autowired
+    private _36wallpaperTask wallpaperTask;
+
+
+    //----------------------远程rpc调用---------------------------
+    @GetMapping("taskForPRC")
+    @ApiOperation("供定时任务服务RPC远程调用")
+    public R _36wallpaperControllerTaskForPRC() {
+        Long count = wallpaperTask.reptileWallpaper();
+        return R.ok(count);
+    }
+}

xjs-business/xjs-business-webmagic/src/main/java/com/xjs/_36wallpaper/mapper/_36wallpaperMapper.java

@@ -0,0 +1,18 @@
+package com.xjs._36wallpaper.mapper;
+
+import com.baomidou.mybatisplus.core.mapper.BaseMapper;
+import com.xjs._36wallpaper.pojo._36wallpaper;
+
+/**
+ * 36壁纸网mapper
+ * @author xiejs
+ * @since 2022-02-20
+ */
+public interface _36wallpaperMapper extends BaseMapper<_36wallpaper> {
+
+    /**
+     * 删除重复数据
+     * @return int
+     */
+    int deleteRepeatData();
+}

xjs-business/xjs-business-webmagic/src/main/java/com/xjs/_36wallpaper/pojo/_36wallpaper.java

@@ -0,0 +1,48 @@
+package com.xjs._36wallpaper.pojo;
+
+import com.baomidou.mybatisplus.annotation.FieldFill;
+import com.baomidou.mybatisplus.annotation.TableField;
+import com.baomidou.mybatisplus.annotation.TableId;
+import com.baomidou.mybatisplus.annotation.TableName;
+import com.ruoyi.common.core.annotation.Excel;
+import lombok.Data;
+
+import java.io.Serializable;
+import java.util.Date;
+
+/**
+ * 36壁纸网图片数据实体类
+ * @author xiejs
+ * @since 2022-02-20
+ */
+@Data
+@TableName("webmagic_36wallpaper")
+public class _36wallpaper implements Serializable {
+
+    private static final long serialVersionUID = 1L;
+
+    /** 主键 */
+    @TableId("id")
+    private Long id;
+
+    /** 照片存放地址 */
+    @Excel(name = "照片存放地址")
+    private String pictureUrl;
+
+    /** 照片名称 */
+    @Excel(name = "照片名称")
+    private String pictureName;
+
+    /** 照片类型 */
+    @Excel(name = "照片类型")
+    private String type;
+
+    /** 照片标签(多个用 , 分割) */
+    @Excel(name = "照片标签(多个用 , 分割)")
+    private String label;
+
+    /** 创建时间 */
+    @Excel(name = "创建时间",dateFormat = "yyyy-MM-dd HH:mm:ss")
+    @TableField(fill = FieldFill.INSERT)
+    private Date createTime;
+}

xjs-business/xjs-business-webmagic/src/main/java/com/xjs/_36wallpaper/service/_36wallpaperService.java

@@ -0,0 +1,18 @@
+package com.xjs._36wallpaper.service;
+
+import com.baomidou.mybatisplus.extension.service.IService;
+import com.xjs._36wallpaper.pojo._36wallpaper;
+
+/**
+ * 36壁纸网service接口
+ * @author xiejs
+ * @since 2022-02-20
+ */
+public interface _36wallpaperService extends IService<_36wallpaper> {
+
+    /**
+     * 删除重复数据
+     * @return int
+     */
+    int deleteRepeatData();
+}

xjs-business/xjs-business-webmagic/src/main/java/com/xjs/_36wallpaper/service/impl/_36wallpaperServiceImpl.java

@@ -0,0 +1,27 @@
+package com.xjs._36wallpaper.service.impl;
+
+import com.baomidou.mybatisplus.extension.service.impl.ServiceImpl;
+import com.xjs._36wallpaper.mapper._36wallpaperMapper;
+import com.xjs._36wallpaper.pojo._36wallpaper;
+import com.xjs._36wallpaper.service._36wallpaperService;
+import org.springframework.stereotype.Service;
+
+import javax.annotation.Resource;
+
+/**
+ * 36壁纸网service实现
+ * @author xiejs
+ * @since 2022-02-20
+ */
+@Service
+public class _36wallpaperServiceImpl extends ServiceImpl<_36wallpaperMapper, _36wallpaper> implements _36wallpaperService {
+
+
+    @Resource
+    private _36wallpaperMapper wallpaperMapper;
+
+    @Override
+    public int deleteRepeatData() {
+        return wallpaperMapper.deleteRepeatData();
+    }
+}

xjs-business/xjs-business-webmagic/src/main/java/com/xjs/_36wallpaper/task/_36wallpaperTask.java

@@ -0,0 +1,33 @@
+package com.xjs._36wallpaper.task;
+
+import com.xjs._36wallpaper.webmagic._36wallpaperProcessor;
+import com.xjs.annotation.ReptileLog;
+import lombok.extern.log4j.Log4j2;
+import org.springframework.beans.factory.annotation.Autowired;
+import org.springframework.stereotype.Component;
+
+import static com.xjs.consts.ReptileConst._36_WALLPAPER_URL;
+
+/**
+ * 36壁纸网爬虫任务
+ * @author xiejs
+ * @since 2022-02-20
+ */
+@Component
+@Log4j2
+public class _36wallpaperTask {
+
+    @Autowired
+    private _36wallpaperProcessor wallpaperProcessor;
+
+
+    /**
+     * 提供定时任务调取
+     * @return 循环次数
+     */
+    @ReptileLog(name = "36壁纸网", url = _36_WALLPAPER_URL)
+    public Long reptileWallpaper() {
+        return wallpaperProcessor.run();
+    }
+
+}

xjs-business/xjs-business-webmagic/src/main/java/com/xjs/_36wallpaper/webmagic/_36wallpaperProcessor.java

@@ -1,9 +1,29 @@
 package com.xjs._36wallpaper.webmagic;
 
+import com.ruoyi.common.redis.service.RedisService;
+import com.xjs._36wallpaper.pojo._36wallpaper;
+import com.xjs._36wallpaper.service._36wallpaperService;
+import lombok.extern.log4j.Log4j2;
+import org.jsoup.Jsoup;
+import org.jsoup.nodes.Document;
+import org.springframework.beans.factory.annotation.Autowired;
+import org.springframework.stereotype.Component;
 import us.codecraft.webmagic.Page;
 import us.codecraft.webmagic.Site;
 import us.codecraft.webmagic.Spider;
 import us.codecraft.webmagic.processor.PageProcessor;
+import us.codecraft.webmagic.scheduler.BloomFilterDuplicateRemover;
+import us.codecraft.webmagic.scheduler.QueueScheduler;
+import us.codecraft.webmagic.selector.Selectable;
+
+import java.io.*;
+import java.net.URL;
+import java.util.ArrayList;
+import java.util.List;
+import java.util.concurrent.TimeUnit;
+
+import static com.xjs.consts.RedisConst.REPTILE_COUNT;
+import static com.xjs.consts.ReptileConst._36_WALLPAPER_URL;
 
 /**
  * 36壁纸网站 爬虫处理
@@ -11,23 +31,45 @@ import us.codecraft.webmagic.processor.PageProcessor;
  * @author xiejs
  * @since 2022-02-19
  */
+@Log4j2
+@Component
 public class _36wallpaperProcessor implements PageProcessor {
 
-    private static final String URL = "https://www.3gbizhi.com/";
+    /**
+     * 是否全网爬虫
+     */
+    private static boolean init = false;
+
 
+    private static RedisService redisService;
 
-    private Site site = Site.me()
-            .setCharset("uft8")//设置字符编码
-            .setTimeOut(10000)//设置超时时间
-            .setRetrySleepTime(2000)//设置重试间隔时间
-            .setCycleRetryTimes(3)//设置重试次数
-            .setSleepTime(10)//设置两个页面之间的间隔时间
-        ;
+    @Autowired
+    public void setRedisService(RedisService redisService) {
+        _36wallpaperProcessor.redisService = redisService;
+    }
 
-    public void setSite(Site site) {
-        this.site = site;
+
+    private static _36wallpaperService wallpaperService;
+
+    @Autowired
+    public void setWallpaperService(_36wallpaperService wallpaperService) {
+        _36wallpaperProcessor.wallpaperService = wallpaperService;
     }
 
+    /**
+     * 图片保存到磁盘的路径
+     */
+    private static final String path = "D:\\Dev\\WebCrawler\\36wallpaper";
+
+
+    private Site site = Site.me()
+            .setCharset("utf8")//设置字符编码
+            .setTimeOut(2000)//设置超时时间
+            .setRetrySleepTime(200)//设置重试间隔时间
+            .setCycleRetryTimes(6)//设置重试次数
+            .setSleepTime(1)//设置两个页面之间的间隔时间
+            ;
+
     //解析页面
     @Override
     public void process(Page page) {
@@ -44,12 +86,87 @@ public class _36wallpaperProcessor implements PageProcessor {
         //第三种写法:正则表达式
         page.putField("url_regex", page.getHtml().css(".tmenu li a").regex(".*图片*.*").all());*/
 
-        // todo 爬取36壁纸图片
+        try {
+            //获取链接
+            page.addTargetRequests(page.getHtml().css(".topmenuc .tmenu li").links().all());
 
+            //爬取所有分页---后期这两段可以注释
+            //当init为true时全网都爬
+            if (init) {
+                page.addTargetRequests(page.getHtml().css("#pageNum a").links().all());
+            }
 
-        //获取链接
-        page.addTargetRequests(page.getHtml().css(".tmenu li").links().all());
-        page.putField("url", page.getHtml().css(".imgw").links().all());
+            page.addTargetRequests(page.getHtml().css(".contlistw > .cl > li > a").links().all());
+
+
+            ArrayList<_36wallpaper> wallpapers = new ArrayList<>();
+            List<Selectable> bodyNote = page.getHtml().css("body").nodes();
+            for (Selectable body : bodyNote) {
+                _36wallpaper wallpaper = new _36wallpaper();
+
+                //爬取图片分类
+                String titleHtml = body.css(".catpos a:nth-child(3)").get();
+
+                String title = "null";
+                if (titleHtml != null) {
+                    Document titleDom = Jsoup.parse(titleHtml);
+                    title = titleDom.text();
+                    wallpaper.setType(title);
+                }
+
+
+                //爬取图片名称
+                String html = body.css(".showtitle h2").get();
+                String pictureName = "null";
+                if (html != null) {
+                    pictureName = Jsoup.parse(html).text();
+                    wallpaper.setPictureName(pictureName);
+                }
+
+
+                //爬取图片路径
+                String link = body.css(".morew").links().get();
+                wallpaper.setPictureUrl(link);
+
+
+                //保存到磁盘
+                if (link != null) {
+                    String thisPath = path + File.separator + title;
+                    downloadPicture(link, thisPath, pictureName + ".jpg");
+                }
+
+                //爬取图片标签
+                List<String> tagList = body.css(".showcontw > .showtaglistw a").all();
+                StringBuilder bf = new StringBuilder();
+
+                for (int i = 0; i < tagList.size(); i++) {
+                    String text = Jsoup.parse(tagList.get(i)).text();
+                    if (i != tagList.size() - 1) {
+                        bf.append(text + ",");
+                    } else {
+                        bf.append(text);
+                    }
+                }
+
+                wallpaper.setLabel(bf.toString());
+
+                if (wallpaper.getPictureUrl() != null) {
+                    wallpapers.add(wallpaper);
+                }
+            }
+
+            //持久化
+            wallpaperService.saveBatch(wallpapers, 25);
+
+            //循环次数存入redis中
+            Integer count = redisService.getCacheObject(REPTILE_COUNT);
+            redisService.setCacheObject(REPTILE_COUNT, count+1);
+
+        } catch (Exception e) {
+            e.printStackTrace();
+        } finally {
+            redisService.expire(REPTILE_COUNT, 1, TimeUnit.HOURS);
+        }
 
     }
 
@@ -61,14 +178,69 @@ public class _36wallpaperProcessor implements PageProcessor {
 
     /**
      * 执行爬虫
+     *
+     * @return 返回循环次数
      */
-    public void run() {
-        Spider.create(new _36wallpaperProcessor()).addUrl(URL).thread(5).runAsync();
+    public Long run() {
+        Spider.create(new _36wallpaperProcessor()).addUrl(_36_WALLPAPER_URL).thread(20)
+                .setScheduler(new QueueScheduler().setDuplicateRemover(new BloomFilterDuplicateRemover(110000)))
+                .run();
+
+        //删除重复数据
+        int count = wallpaperService.deleteRepeatData();
+        log.info("36壁纸删除重复数据数：" + count);
+
+        //从redis中获取循环次数
+        Integer cache = redisService.getCacheObject(REPTILE_COUNT);
+        redisService.deleteObject(REPTILE_COUNT);
+
+        if (cache != null) {
+            return Long.valueOf(cache);
+        }
+        return 0L;
+    }
+
+
+    //链接url下载图片
+    private void downloadPicture(String urlList, String path, String fileName) {
+        java.net.URL url = null;
+        try {
+            url = new URL(urlList);
+            DataInputStream dataInputStream = new DataInputStream(url.openStream());
+
+            File file = new File(path);
+            if (!file.exists()) {
+                file.mkdir();
+            }
+
+            //替换\ 防止报错
+            if (fileName.contains("/") || fileName.contains("\\")) {
+                fileName = fileName.replace("/", "-");
+            }
+
+            String absolutePath = file.getAbsolutePath() + File.separator + fileName;
+
+            FileOutputStream fileOutputStream = new FileOutputStream(absolutePath);
+            ByteArrayOutputStream output = new ByteArrayOutputStream();
+
+            byte[] buffer = new byte[1024];
+            int length;
+
+            while ((length = dataInputStream.read(buffer)) > 0) {
+                output.write(buffer, 0, length);
+            }
+
+            fileOutputStream.write(output.toByteArray());
+            dataInputStream.close();
+            fileOutputStream.close();
+        } catch (IOException e) {
+            e.printStackTrace();
+        }
     }
 
 
     public static void main(String[] args) {
-        Spider.create(new _36wallpaperProcessor()).addUrl("https://www.3gbizhi.com/").thread(5).run();
+        Spider.create(new _36wallpaperProcessor()).addUrl(_36_WALLPAPER_URL).thread(15).run();
 
     }
 }

xjs-business/xjs-business-webmagic/src/main/resources/mapper/webmagic/_36wallpaperMapper.xml

@@ -0,0 +1,17 @@
+<?xml version="1.0" encoding="UTF-8" ?>
+<!DOCTYPE mapper
+        PUBLIC "-//mybatis.org//DTD Mapper 3.0//EN"
+        "http://mybatis.org/dtd/mybatis-3-mapper.dtd">
+<mapper namespace="com.xjs._36wallpaper.mapper._36wallpaperMapper">
+
+
+
+    <delete id="deleteRepeatData">
+        delete from webmagic_36wallpaper where id not in	(
+            SELECT
+                t.min_id
+            FROM
+                    ( SELECT min( id ) AS min_id FROM webmagic_36wallpaper GROUP BY picture_url ) AS t
+        )
+    </delete>
+</mapper>