部分功能完善

src/main/java/com/wbq/fujiati/AmericanElection.java

@@ -0,0 +1,159 @@
+package com.wbq.fujiati;
+
+import com.google.gson.JsonArray;
+import com.google.gson.JsonObject;
+import com.google.gson.JsonParser;
+import com.wbq.zuoye.GetDanmu;
+import org.apache.http.client.methods.CloseableHttpResponse;
+import org.apache.http.client.methods.HttpGet;
+import org.apache.http.impl.client.CloseableHttpClient;
+import org.apache.http.impl.client.HttpClients;
+import org.apache.http.util.EntityUtils;
+import org.jsoup.Jsoup;
+import org.jsoup.nodes.Document;
+import org.jsoup.nodes.Element;
+import org.jsoup.select.Elements;
+
+import java.io.IOException;
+import java.util.*;
+import java.util.regex.Matcher;
+import java.util.regex.Pattern;
+
+public class AmericanElection {
+    public static void main(String[] args) throws IOException {
+        String Url = "https://search.bilibili.com/all?vt=08336694&keyword=%E7%BE%8E%E5%9B%BD%E5%A4%A7%E9%80%89&from_source=webtop_search&spm_id_from=333.1007&search_source=5";
+        List<String> urls = getTopUrls(Url);
+        Map<String, Integer> danmuFrequecy = new HashMap<>();
+        Integer num = 0;
+
+        for (String url : urls) {
+            String BVid = getBVid(url);
+            String apiUrl = "https://api.bilibili.com/x/player/pagelist?bvid=" + BVid;
+            String cid = null;
+
+            try (CloseableHttpClient httpClient = HttpClients.createDefault()) {
+                HttpGet request = new HttpGet(apiUrl);
+                request.setHeader("User-Agent", "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/92.0.4515.107 Safari/537.36");
+                try (CloseableHttpResponse response = httpClient.execute(request)) {
+                    String jsonResponse = EntityUtils.toString(response.getEntity(), "UTF-8");
+                    JsonObject jsonObject = JsonParser.parseString(jsonResponse).getAsJsonObject();
+                    JsonArray dataArray = jsonObject.getAsJsonArray("data");
+                    if (dataArray.size() > 0) {
+                        JsonObject firstPage = dataArray.get(0).getAsJsonObject();
+                        cid = firstPage.get("cid").getAsString();
+                        System.out.println("Found cid: " + cid);
+                    } else {
+                        System.out.println("No data found.");
+                    }
+                }
+            } catch (IOException e) {
+                e.printStackTrace();
+            }
+            //获取整个页面
+            Document document = Jsoup.connect("http://commen" +
+                    "t.bilibili.com/" + cid + ".xml").get();
+            //获取所有的d标签   也就是存放弹幕的标签
+            Elements d = document.getElementsByTag("d");
+            //遍历所有的d标签
+            for(Element element: d){
+                String s = element.text();
+                num++;
+                danmuFrequecy.put(s, danmuFrequecy.getOrDefault(s, 0) + 1);
+            }
+        }
+
+        Map<String, Integer> candidatePopularity = new HashMap<>();
+
+        candidatePopularity.put("特朗普", 0);
+        candidatePopularity.put("拜登", 0);
+        candidatePopularity.put("哈里斯", 0);
+        Map<String, Integer> zhichi = new HashMap<>();
+        int ha = 0;
+        int dong = 0;
+        int deng = 0;
+
+        for (Map.Entry<String, Integer> entry : danmuFrequecy.entrySet()) {
+            if (entry.getKey().contains("特朗普") || entry.getKey().contains("川建国") || entry.getKey().contains("川普") || entry.getKey().contains("懂王")) {
+                candidatePopularity.put("特朗普", candidatePopularity.get("特朗普") + entry.getValue());
+                //System.out.println(entry.getKey());
+            } else if (entry.getKey().contains("拜登") || entry.getKey().contains("老登") || entry.getKey().contains("稀宗")) {
+                candidatePopularity.put("拜登", candidatePopularity.get("拜登") + entry.getValue());
+                //System.out.println(entry.getKey());
+            } else if (entry.getKey().contains("哈里斯")) {
+                candidatePopularity.put("哈里斯", candidatePopularity.get("哈里斯") + entry.getValue());
+                //System.out.println(entry.getKey());
+            }
+            else if (entry.getKey().contains("支持") && (entry.getKey().contains("哈") || entry.getKey().contains("懂王") || entry.getKey().contains("登") || entry.getKey().contains("普"))) {
+                zhichi.put(entry.getKey(), entry.getValue());
+                if (entry.getKey().contains("哈")) ha++;
+                else if (entry.getKey().contains("懂")) dong++;
+                else if (entry.getKey().contains("登")) deng++;
+            }
+        }
+
+        // 排序
+        // 将 HashMap 的条目放入 List 中
+        List<Map.Entry<String, Integer>> entryList = new ArrayList<>(candidatePopularity.entrySet());
+
+        Collections.sort(entryList, new Comparator<Map.Entry<String, Integer>>() {
+            @Override
+            public int compare(Map.Entry<String, Integer> o1, Map.Entry<String, Integer> o2) {
+                return o2.getValue().compareTo(o1.getValue());
+            }
+        });
+
+        //输出结果
+        System.out.println("排序如下：");
+        for(Map.Entry<String, Integer> entry : entryList) {
+            String s = entry.getKey() + ": " + entry.getValue();
+            System.out.println(s);
+        }
+
+        zhichi.forEach((value, key)->{
+            System.out.println(value + ":" + key);
+        });
+
+        System.out.println(ha);
+        System.out.println(dong);
+        System.out.println(deng);
+    }
+    public static String getBVid(String url) {
+        String bv = extractBV(url);
+        if (bv != null) {
+            System.out.println("Extracted BV: " + bv);
+            return bv;
+        } else {
+            System.out.println("BV not found in the URL.");
+            return null;
+        }
+    }
+
+    public static String extractBV(String url) {
+        String bvPattern = "BV[0-9A-Za-z]+";
+        Pattern pattern = Pattern.compile(bvPattern);
+        Matcher matcher = pattern.matcher(url);
+        if (matcher.find()) {
+            return matcher.group();
+        }
+        return null;
+    }
+    public static List<String> getTopUrls(String url) {
+        Set<String> urlSet = new HashSet<>();
+        try {
+            Document document = Jsoup.connect(url)
+                    .userAgent("Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/92.0.4515.107 Safari/537.36")
+                    .get();
+            Elements videoElements = document.select("a[href^=//www.bilibili.com/video/]");
+            for (Element videoElement : videoElements) {
+                String videoUrl = videoElement.attr("href");
+                urlSet.add(videoUrl);
+            }
+        }
+        catch (IOException e) {
+            e.printStackTrace();
+            //HHHhhhh
+            return null;
+        }
+        return new ArrayList<>(urlSet);
+    }
+}

src/main/java/com/wbq/fujiati/DanmuAnalyze_Popularity.java

@@ -0,0 +1,67 @@
+package com.wbq.fujiati;
+
+import com.wbq.zuoye.GetDanmu;
+
+import java.io.FileOutputStream;
+import java.io.IOException;
+import java.util.*;
+
+public class DanmuAnalyze_Popularity {
+    public static void main(String[] args) throws IOException {
+        Map<String, Integer> danmuFrequency = GetDanmu.getDanmuFrequency();
+        Map<String, Integer> athletePopularity = new HashMap<>();
+
+        athletePopularity.put("马龙", 0);
+        athletePopularity.put("樊振东", 0);
+        athletePopularity.put("王楚钦", 0);
+        athletePopularity.put("张继科", 0);
+        athletePopularity.put("许昕", 0);
+        athletePopularity.put("孙颖莎", 0);
+        athletePopularity.put("陈梦", 0);
+
+
+        for (Map.Entry<String, Integer> entry : danmuFrequency.entrySet()) {
+            if (entry.getKey().contains("马龙") || entry.getKey().contains("龙队")) {
+                athletePopularity.put("马龙", athletePopularity.get("马龙")+entry.getValue());
+            }
+            else if (entry.getKey().contains("樊振东") || entry.getKey().contains("小胖")) {
+                athletePopularity.put("樊振东", athletePopularity.get("樊振东")+entry.getValue());
+            }
+            else if (entry.getKey().contains("王楚钦") || entry.getKey().contains("大头")) {
+                athletePopularity.put("王楚钦", athletePopularity.get("王楚钦")+entry.getValue());
+            }
+            else if (entry.getKey().contains("张继科") || entry.getKey().contains("牢科")) {
+                athletePopularity.put("张继科", athletePopularity.get("张继科")+entry.getValue());
+            }
+            else if (entry.getKey().contains("许昕") || entry.getKey().contains("人民艺术家")) {
+                athletePopularity.put("许昕", athletePopularity.get("许昕")+entry.getValue());
+            }
+            else if (entry.getKey().contains("孙颖莎") || entry.getKey().contains("莎莎")) {
+                athletePopularity.put("孙颖莎", athletePopularity.get("孙颖莎")+entry.getValue());
+            }
+            else if (entry.getKey().contains("陈梦") || entry.getKey().contains("大梦")) {
+                athletePopularity.put("陈梦", athletePopularity.get("陈梦")+entry.getValue());
+            }
+        }
+        // 将 HashMap 的条目放入 List 中
+        List<Map.Entry<String, Integer>> entryList = new ArrayList<>(athletePopularity.entrySet());
+
+        Collections.sort(entryList, new Comparator<Map.Entry<String, Integer>>() {
+            @Override
+            public int compare(Map.Entry<String, Integer> o1, Map.Entry<String, Integer> o2) {
+                return o2.getValue().compareTo(o1.getValue());
+            }
+        });
+
+        String filePath = "C:\\Users\\21080\\IdeaProjects\\Software_Engineering\\Crawler\\src\\main\\resources\\InnovativeContent.txt";
+        FileOutputStream fileOutputStream = new FileOutputStream(filePath, true);
+        //输出结果
+        System.out.println("排序如下：");
+        for(Map.Entry<String, Integer> entry : entryList) {
+            String s = entry.getKey() + ": " + entry.getValue();
+            System.out.println(s);
+            fileOutputStream.write(s.getBytes());
+        }
+        fileOutputStream.close();
+    }
+}

src/main/java/com/wbq/zuoye/BilibiliCidFetcher.java

@@ -1,41 +1,50 @@
 package com.wbq.zuoye;
 
-import com.google.gson.JsonArray;
-import com.google.gson.JsonObject;
-import com.google.gson.JsonParser;
 import org.apache.http.client.methods.CloseableHttpResponse;
 import org.apache.http.client.methods.HttpGet;
 import org.apache.http.impl.client.CloseableHttpClient;
 import org.apache.http.impl.client.HttpClients;
 import org.apache.http.util.EntityUtils;
+import com.google.gson.JsonArray;
+import com.google.gson.JsonObject;
+import com.google.gson.JsonParser;
 
 import java.io.IOException;
 
 public class BilibiliCidFetcher {
+    private static final CloseableHttpClient httpClient = HttpClients.createDefault();
+
     public static String getCid(String bvid) {
+        if (bvid == null || bvid.isEmpty()) {
+            throw new IllegalArgumentException("bvid cannot be null or empty");
+        }
+
         String apiUrl = "https://api.bilibili.com/x/player/pagelist?bvid=" + bvid;
+        String cid = null;
+
+        HttpGet request = new HttpGet(apiUrl);
+        request.setHeader("User-Agent", "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/92.0.4515.107 Safari/537.36");
+
+        try (CloseableHttpResponse response = httpClient.execute(request)) {
+            String jsonResponse = EntityUtils.toString(response.getEntity(), "UTF-8");
+            JsonArray dataArray = JsonParser.parseString(jsonResponse).getAsJsonObject().getAsJsonArray("data");
 
-        try (CloseableHttpClient httpClient = HttpClients.createDefault()) {
-            HttpGet request = new HttpGet(apiUrl);
-            request.setHeader("User-Agent", "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/92.0.4515.107 Safari/537.36");
-            try (CloseableHttpResponse response = httpClient.execute(request)) {
-                String jsonResponse = EntityUtils.toString(response.getEntity(), "UTF-8");
-                JsonObject jsonObject = JsonParser.parseString(jsonResponse).getAsJsonObject();
-                JsonArray dataArray = jsonObject.getAsJsonArray("data");
-
-                if (dataArray.size() > 0) {
-                    JsonObject firstPage = dataArray.get(0).getAsJsonObject();
-                    String cid = firstPage.get("cid").getAsString();
-                    System.out.println("Found cid: " + cid);
-                    return cid;
-                } else {
-                    System.out.println("No data found.");
-                    return null;
-                }
+            if (dataArray.size() > 0) {
+                cid = dataArray.get(0).getAsJsonObject().get("cid").getAsString();
+                System.out.println("Found cid: " + cid);
+            } else {
+                System.out.println("No data found.");
             }
         } catch (IOException e) {
+            // 使用日志框架记录错误
             e.printStackTrace();
-            return null;
         }
+
+        return cid;
+    }
+
+    // 关闭 httpClient 的方法
+    public static void close() throws IOException {
+        httpClient.close();
     }
-}
+}

src/main/java/com/wbq/zuoye/DanmuAnalyze.java

@@ -1,20 +1,31 @@
 package com.wbq.zuoye;
 
-import com.wbq.demo.GetDanmu;
-
 import java.io.FileOutputStream;
 import java.io.IOException;
 import java.util.*;
 
 public class DanmuAnalyze {
     public static void main(String[] args) throws IOException {
+        long startTime_getDanmu = System.nanoTime();
         Map<String, Integer> danmuFrequency = GetDanmu.getDanmuFrequency();
+        long endTime_getDanmu = System.nanoTime();
+        System.out.println("duration_getDanmu: " + ((double)startTime_getDanmu
+                - (double)endTime_getDanmu)/-1_000_000 + "ms\n");
+
+
+        long startTime_Statistics = System.nanoTime();
         List<Map.Entry<String, Integer>> AIStrings = new ArrayList<>();
         for (Map.Entry<String, Integer> entry : danmuFrequency.entrySet()) {
             if (entry.getKey().contains("AI") || entry.getKey().contains("ai")) {
                 AIStrings.add(entry);
             }
         }
+        long endTime_Statistics = System.nanoTime();
+        System.out.println("duration_Statistics: " + ((double)startTime_Statistics
+                - (double)endTime_Statistics)/-1_000_000 + "ms\n");
+
+
+        long startTime_sort = System.nanoTime();
         // 按照频数排序
         Collections.sort(AIStrings, new Comparator<Map.Entry<String, Integer>>() {
             @Override
@@ -22,6 +33,10 @@ public class DanmuAnalyze {
                 return o2.getValue().compareTo(o1.getValue());
             }
         });
+        long endTime_sort = System.nanoTime();
+        System.out.println("duration_sort: " + ((double)startTime_sort
+                - (double)endTime_sort)/-1_000_000 + "ms\n");
+
 
         String filePath = "C:\\Users\\21080\\IdeaProjects\\Software_Engineering\\Crawler\\src\\main\\resources\\Analyze.txt";
         FileOutputStream fileOutputStream = new FileOutputStream(filePath, true);
@@ -32,5 +47,6 @@ public class DanmuAnalyze {
             System.out.println(s);
             fileOutputStream.write(s.getBytes());
         }
+        fileOutputStream.close();
     }
 }

src/main/java/com/wbq/zuoye/GetDanmu.java

@@ -1,8 +1,5 @@
 package com.wbq.zuoye;
 
-import com.wbq.demo.BVExtractor;
-import com.wbq.demo.BilibiliCidFetcher;
-import com.wbq.demo.TopUrls;
 import org.jsoup.Jsoup;
 import org.jsoup.nodes.Document;
 import org.jsoup.nodes.Element;
@@ -33,9 +30,15 @@ public class GetDanmu {
         // 存储弹幕
         Map<String, Integer> danmuFrequecy = new HashMap<>();
         int count = 0;
+        //获取File
+        File file = new File("C:\\Users\\21080\\IdeaProjects\\Software_Engineering\\Crawler\\src\\main\\resources\\Danmu.txt");
+        if(!file.exists()){
+            file.createNewFile();
+        }
+        FileOutputStream fileOutputStream=new FileOutputStream(file);
         for (String burl : bilibiliUrls) {
             List<String> urlList = TopUrls.getTopUrls(burl);
-           /* System.out.println(urlList);*/
+            /* System.out.println(urlList);*/
             for (String url : urlList) {
                 System.out.println(++count);
                 if (count > 300)
@@ -47,12 +50,6 @@ public class GetDanmu {
                         "t.bilibili.com/" + cid + ".xml").get();
                 //获取所有的d标签   也就是存放弹幕的标签
                 Elements d = document.getElementsByTag("d");
-                //获取File
-                File file = new File("C:\\Users\\21080\\IdeaProjects\\Software_Engineering\\Crawler\\src\\main\\resources\\Danmu.txt");
-                if(!file.exists()){
-                    file.createNewFile();
-                }
-                FileOutputStream fileOutputStream=new FileOutputStream(file, true);
                 //遍历所有的d标签
                 for(Element element: d){
                     //获取每条弹幕换行并添加到txt文件中
@@ -60,9 +57,9 @@ public class GetDanmu {
                     fileOutputStream.write((s+"\r\n").getBytes());
                     danmuFrequecy.put(s, danmuFrequecy.getOrDefault(s, 0) + 1);
                 }
-                fileOutputStream.close();
             }
         }
+        fileOutputStream.close();
         return danmuFrequecy;
     }
 }

src/main/resources/InnovativeContent.txt

@@ -0,0 +1 @@
+孙颖莎: 1205樊振东: 699张继科: 621马龙: 334王楚钦: 286陈梦: 195许昕: 65孙颖莎: 1238樊振东: 686张继科: 621马龙: 327王楚钦: 279陈梦: 239许昕: 66孙颖莎: 1659樊振东: 1291张继科: 645马龙: 627陈梦: 413王楚钦: 401许昕: 118

src/test/java/Demo04.java

@@ -1,5 +1,3 @@
-package com.wbq.demo;
-
 import java.io.BufferedWriter;
 import java.io.FileWriter;
 import java.io.IOException;

src/test/java/demo01.java

@@ -1,5 +1,3 @@
-package com.wbq.demo;
-
 import java.util.LinkedHashMap;
 import java.util.Map;
 import java.util.regex.Matcher;

src/test/java/demo02.java

@@ -1,5 +1,3 @@
-package com.wbq.demo;
-
 import org.jsoup.Jsoup;
 import org.jsoup.nodes.Document;
 import org.jsoup.nodes.Element;

src/test/java/demo03.java

@@ -0,0 +1,21 @@
+import java.io.IOException;
+
+public class demo03 {
+    public static void main(String[] args) {
+        long startTime2 = System.nanoTime();
+        System.out.println("程序已暂停，请按任意键继续...");
+        try {
+            System.in.read(); // 等待用户输入  
+        } catch (IOException e) {
+            e.printStackTrace();
+        }
+        long startTime1 = System.nanoTime();
+        System.out.println("程序继续执行...");
+        long endTime1 = System.nanoTime();
+        long endTime2 = System.nanoTime();
+        double duration1 = (double)(endTime1 - startTime1)/1_000_000;
+        double duration2 = (double)(endTime2 - startTime2)/1_000_000;
+        System.out.println("执行时间1: " + duration1 + " 毫秒");
+        System.out.println("执行时间2: " + duration2 + " 毫秒");
+    }
+}

target/classes/InnovativeContent.txt

@@ -0,0 +1 @@
+孙颖莎: 1205樊振东: 699张继科: 621马龙: 334王楚钦: 286陈梦: 195许昕: 65孙颖莎: 1238樊振东: 686张继科: 621马龙: 327王楚钦: 279陈梦: 239许昕: 66孙颖莎: 1659樊振东: 1291张继科: 645马龙: 627陈梦: 413王楚钦: 401许昕: 118

target/classes/WordCLoudPic.html

@@ -0,0 +1,94 @@
+
+<html lang="en">
+<head>
+    <meta charset="UTF-8">
+    <meta name="viewport" content="width=device-width, initial-scale=1.0">
+    <title>词云图示例</title>
+    <script src="https://cdn.jsdelivr.net/npm/echarts/dist/echarts.min.js"></script>
+    <script src="https://cdn.jsdelivr.net/npm/echarts-wordcloud/dist/echarts-wordcloud.min.js"></script>
+    <style>
+        body {
+            font-family: "Microsoft YaHei", sans-serif;
+            background-color: #f5f5f5;
+            display: flex;
+            justify-content: center;
+            align-items: center;
+            height: 100vh;
+            margin: 0;
+        }
+        #wordCloud {
+            width: 800px;
+            height: 600px;
+            background-color: white;
+            border: 1px solid #ccc;
+        }
+    </style>
+</head>
+<body>
+<div id="wordCloud"></div>
+<script>
+    var myChart = echarts.init(document.getElementById('wordCloud'));
+
+    // 生成随机颜色
+    function getRandomColor() {
+        var letters = '0123456789ABCDEF';
+        var color = '#';
+        for (var i = 0; i < 6; i++) {
+            color += letters[Math.floor(Math.random() * 16)];
+        }
+        return color;
+    }
+
+    // 词云数据，添加随机颜色
+    var data = [
+        { name: '该内容疑似使用AI技术合成，请谨慎甄别', value: 4, color: getRandomColor() },
+        { name: 'ai读', value: 2, color: getRandomColor() },
+        { name: 'AI', value: 2, color: getRandomColor() },
+        { name: '这些都ai的。很明显', value: 2, color: getRandomColor() },
+        { name: '醒醒！AI图谁都能做得很好看！', value: 2, color: getRandomColor() },
+        { name: '等一下这个地方的闪烁是AI作画吗？', value: 2, color: getRandomColor() },
+        { name: '一眼AI', value: 2, color: getRandomColor() },
+        { name: '确实很AI', value: 1, color: getRandomColor() },
+        { name: '就是ai，一眼就看出来了', value: 1, color: getRandomColor() },
+        { name: 'ai', value: 1, color: getRandomColor() },
+        { name: '就是AI', value: 1, color: getRandomColor() },
+        { name: '你到底在说什么 能不能用AI', value: 1, color: getRandomColor() },
+        { name: '要不用AI配音吧', value: 1, color: getRandomColor() },
+        { name: '最新 ai', value: 1, color: getRandomColor() },
+        { name: '日韩日韩，ai设计的', value: 1, color: getRandomColor() },
+        { name: 'AI人', value: 1, color: getRandomColor() },
+        { name: '这个像ai', value: 1, color: getRandomColor() },
+        { name: 'ai画图我也行', value: 1, color: getRandomColor() },
+        { name: '构思ai龙来了', value: 1, color: getRandomColor() },
+        { name: '第二史，全是ai，拼多多那个最史', value: 1, color: getRandomColor() },
+        { name: '本来就是AI，看起来有点别扭', value: 1, color: getRandomColor() },
+        { name: '怎么感觉像AI做的？', value: 1, color: getRandomColor() },
+        { name: '求up主用AI配音吧', value: 1, color: getRandomColor() },
+        { name: 'ai使用者', value: 1, color: getRandomColor() },
+        { name: '一股ai味儿', value: 1, color: getRandomColor() },
+        { name: '不知道为什么感觉有一股AI味儿。。。', value: 1, color: getRandomColor() },
+        { name: '不要什么都ai，这是ai的爹', value: 1, color: getRandomColor() },
+        { name: 'AI感明显，以后应该很厉害', value: 1, color: getRandomColor() },
+        { name: '能不能口齿清晰一点 不行ai配音啊。说的好难受啊', value: 1, color: getRandomColor() }
+    ];
+
+    var option = {
+        series: [{
+            type: 'wordCloud',
+            gridSize: 20,
+            shape: 'circle',
+            textStyle: {
+                normal: {
+                    color: function (params) {
+                        return params.data.color; // 使用每个词条的颜色
+                    }
+                },
+            },
+            data: data
+        }]
+    };
+
+    myChart.setOption(option);
+</script>
+</body>
+</html>