今日已结

2026-01-21 16:28:53 +08:00
parent 7bb698c9c0
commit 6f1c990d51
7 changed files with 481 additions and 0 deletions
--- a/pom.xml
+++ b/pom.xml
@@ -78,6 +78,18 @@
            <artifactId>tess4j</artifactId>
            <version>5.8.0</version>
        </dependency>-->
+        <!-- pom.xml -->
+        <dependency>
+            <groupId>org.springframework.boot</groupId>
+            <artifactId>spring-boot-starter-cache</artifactId>
+        </dependency>
+
+        <!-- 使用Caffeine作为缓存实现 -->
+        <dependency>
+            <groupId>com.github.ben-manes.caffeine</groupId>
+            <artifactId>caffeine</artifactId>
+            <version>3.1.8</version>
+        </dependency>
    </dependencies>
    <build>
        <plugins>
--- a/src/main/java/com/tem/bocai/config/CacheConfig.java
+++ b/src/main/java/com/tem/bocai/config/CacheConfig.java
@@ -0,0 +1,42 @@
+package com.tem.bocai.config;
+
+
+import com.github.benmanes.caffeine.cache.Caffeine;
+import org.springframework.boot.autoconfigure.condition.ConditionalOnMissingBean;
+import org.springframework.cache.CacheManager;
+import org.springframework.cache.annotation.EnableCaching;
+import org.springframework.cache.caffeine.CaffeineCacheManager;
+import org.springframework.context.annotation.Bean;
+import org.springframework.context.annotation.Configuration;
+import org.springframework.context.annotation.Primary;
+
+import java.util.concurrent.TimeUnit;
+
+@Configuration
+@EnableCaching  // 启用缓存支持
+public class CacheConfig {
+
+    /**
+     * 主缓存管理器 - 用于token缓存（19分钟过期）
+     */
+    @Bean
+    @Primary  // 标记为主缓存管理器
+    public CacheManager cacheManager() {
+        CaffeineCacheManager cacheManager = new CaffeineCacheManager();
+        cacheManager.setCaffeine(Caffeine.newBuilder()
+                .expireAfterWrite(19, TimeUnit.MINUTES)  // 19分钟过期
+                .maximumSize(100)                        // 最大缓存数量
+                .recordStats()                           // 记录统计信息
+        );
+        return cacheManager;
+    }
+
+    /**
+     * 备用缓存管理器 - 如果没有配置Caffeine，使用ConcurrentMap
+     */
+    @Bean
+    @ConditionalOnMissingBean(CacheManager.class)
+    public CacheManager fallbackCacheManager() {
+        return new org.springframework.cache.concurrent.ConcurrentMapCacheManager();
+    }
+}
--- a/src/main/java/com/tem/bocai/controller/LoginCrawler.java
+++ b/src/main/java/com/tem/bocai/controller/LoginCrawler.java
@@ -24,6 +24,12 @@ public class LoginCrawler {
        return ResponseEntity.ok(result);
    }

+    @GetMapping("/ocr/completedToday")
+    public ResponseEntity<String> completedToday() throws IOException, TesseractException {
+        String result = loginService.completedToday();
+        return ResponseEntity.ok(result);
+    }
+
 }


--- a/src/main/java/com/tem/bocai/service/LoginService.java
+++ b/src/main/java/com/tem/bocai/service/LoginService.java
@@ -7,4 +7,8 @@ public interface LoginService {

    //获取token
    String getToken(String username, String password, String loginUrl);
+
+    //获取token
+    String completedToday();
+
 }
--- a/src/main/java/com/tem/bocai/service/impl/LoginServiceImpl.java
+++ b/src/main/java/com/tem/bocai/service/impl/LoginServiceImpl.java
@@ -1,6 +1,7 @@
 package com.tem.bocai.service.impl;

 import com.tem.bocai.service.LoginService;
+import com.tem.bocai.util.CompletedTodayCrawler;
 import com.tem.bocai.util.LotteryDataPipeline;
 import com.tem.bocai.util.LotteryWebMagicCrawler;
 import com.tem.bocai.util.TokenCacheService;
@@ -107,6 +108,25 @@ public class LoginServiceImpl implements LoginService {
        return "";
    }

+    @Override
+    public String completedToday() {
+        String token = tokenCacheService.getToken();
+        System.out.println("得到token = " + token);
+        if (token != null && !token.isEmpty()) {
+            // 2. 创建爬虫实例，传入token
+            CompletedTodayCrawler crawler = new CompletedTodayCrawler(token);
+
+            // 4. 执行爬虫
+            String url = "https://4701268539-esh.qdk63ayw8g.com/member/bets?settled=true";
+
+            Spider.create(crawler)
+                    .addUrl(url)
+                    .thread(1)
+                    .run();
+        }
+        return "";
+    }
+
    /**
     * 单次登录尝试
     */
--- a/src/main/java/com/tem/bocai/util/CompletedTodayCrawler.java
+++ b/src/main/java/com/tem/bocai/util/CompletedTodayCrawler.java
@@ -0,0 +1,396 @@
+package com.tem.bocai.util;
+
+
+import org.jsoup.Jsoup;
+import org.jsoup.nodes.Document;
+import org.jsoup.nodes.Element;
+import org.jsoup.select.Elements;
+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.selector.Html;
+
+import java.io.File;
+import java.text.SimpleDateFormat;
+import java.util.*;
+import java.util.regex.Matcher;
+import java.util.regex.Pattern;
+
+public class CompletedTodayCrawler implements PageProcessor {
+
+    private final String token;
+    private Site site;
+    private static final SimpleDateFormat dateFormat = new SimpleDateFormat("yyyy-MM-dd HH:mm:ss");
+
+    public CompletedTodayCrawler(String token) {
+        this.token = token;
+        initSite();
+    }
+
+    private void initSite() {
+        site = Site.me()
+                .setRetryTimes(3)
+                .setSleepTime(2000) // 增加等待时间
+                .setTimeOut(15000)  // 增加超时时间
+                .setUserAgent("Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/141.0.0.0 Safari/537.36")
+                .addHeader("Accept", "text/html,application/xhtml+xml,application/xml;q=0.9,image/avif,image/webp,image/apng,*/*;q=0.8")
+                .addHeader("Accept-Language", "zh-CN,zh;q=0.9,en;q=0.8")
+                .addHeader("Accept-Encoding", "gzip, deflate, br")
+                .addHeader("Connection", "keep-alive")
+                .addHeader("Upgrade-Insecure-Requests", "1")
+                .addHeader("Sec-Fetch-Dest", "document")
+                .addHeader("Sec-Fetch-Mode", "navigate")
+                .addHeader("Sec-Fetch-Site", "same-origin")
+                .addHeader("Sec-Fetch-User", "?1");
+
+        // 设置cookie
+        if (token != null && !token.isEmpty()) {
+            site.addHeader("cookie", "token=" + token);
+        }
+    }
+
+    @Override
+    public void process(Page page) {
+        String url = page.getUrl().toString();
+        System.out.println("处理页面: " + url);
+
+        Html html = page.getHtml();
+        String content = html.toString();
+
+        // 打印一些基本信息
+        System.out.println("页面标题: " + html.xpath("//title/text()").get());
+        System.out.println("页面大小: " + content.length() + " 字符");
+
+        // 检查是否有"暂无数据"提示
+        if (content.contains("暂无数据")) {
+            System.out.println("警告: 页面显示'暂无数据'");
+            return;
+        }
+        // 解析注单数据
+        List<Map<String, Object>> betList = parseBetHtml(content);
+        if (betList.isEmpty()) {
+            System.out.println("未解析到注单数据");
+
+            // 尝试从其他可能的位置解析
+            extractDebugInfo(html);
+        } else {
+            System.out.println("解析到 " + betList.size() + " 条注单数据");
+
+            // 打印部分数据示例
+            printSampleData(betList);
+        }
+
+        // 保存原始HTML用于调试
+        saveHtmlForDebug(content, url);
+    }
+
+    /**
+     * 解析注单HTML数据
+     */
+    private List<Map<String, Object>> parseBetHtml(String htmlContent) {
+        List<Map<String, Object>> betList = new ArrayList<>();
+
+        try {
+            Document doc = Jsoup.parse(htmlContent);
+
+            // 查找注单表格
+            Element table = doc.selectFirst("table.list");
+            if (table == null) {
+                System.out.println("未找到注单表格");
+                return betList;
+            }
+
+            // 查找表头
+            Elements headers = table.select("thead th");
+            List<String> headerList = new ArrayList<>();
+            for (Element header : headers) {
+                headerList.add(header.text().trim());
+            }
+            System.out.println("表头信息: " + headerList);
+
+            // 查找数据行（跳过表头）
+            Elements rows = table.select("tbody tr");
+
+            for (Element row : rows) {
+                // 跳过"暂无数据"的行
+                if (row.select("td.nodata").size() > 0) {
+                    continue;
+                }
+
+                Map<String, Object> betData = new HashMap<>();
+                Elements cells = row.select("td");
+
+                // 按列解析数据
+                for (int i = 0; i < cells.size() && i < headerList.size(); i++) {
+                    String header = headerList.get(i);
+                    String value = cells.get(i).text().trim();
+
+                    // 根据表头映射到对应的字段名
+                    switch (header) {
+                        case "注单号":
+                            betData.put("bet_id", value);
+                            break;
+                        case "时间":
+                            betData.put("time", value);
+                            break;
+                        case "类型":
+                            betData.put("type", value);
+                            break;
+                        case "玩法":
+                            betData.put("game_type", value);
+                            break;
+                        case "盘":
+                            betData.put("plate", value);
+                            break;
+                        case "下注金额":
+                            betData.put("bet_amount", parseAmount(value));
+                            break;
+                        case "退水(%)":
+                            betData.put("rebate_rate", parseRate(value));
+                            break;
+                        case "结果":
+                            betData.put("result", parseResult(value));
+                            betData.put("result_amount", parseResultAmount(value));
+                            break;
+                        default:
+                            betData.put(header, value);
+                    }
+                }
+
+                // 添加额外信息
+                if (!betData.isEmpty()) {
+                    betData.put("parse_time", dateFormat.format(new Date()));
+                    betData.put("source", "completed_today");
+
+                    // 提取期数信息（从玩法中提取）
+                    extractPeriodInfo(betData);
+
+                    betList.add(betData);
+                }
+            }
+
+        } catch (Exception e) {
+            System.err.println("解析HTML时出错: " + e.getMessage());
+            e.printStackTrace();
+        }
+
+        return betList;
+    }
+
+    /**
+     * 从玩法中提取期数信息
+     */
+    private void extractPeriodInfo(Map<String, Object> betData) {
+        try {
+            Object gameTypeObj = betData.get("game_type");
+            if (gameTypeObj instanceof String) {
+                String gameType = (String) gameTypeObj;
+
+                // 尝试匹配期数模式，如"2024001", "001", "期号2024001"等
+                Pattern pattern = Pattern.compile("(\\d{7})|期[号码]?(\\d{3,7})|(\\d{3,4})期");
+                Matcher matcher = pattern.matcher(gameType);
+
+                if (matcher.find()) {
+                    for (int i = 1; i <= matcher.groupCount(); i++) {
+                        if (matcher.group(i) != null) {
+                            betData.put("period", matcher.group(i));
+                            break;
+                        }
+                    }
+                }
+            }
+        } catch (Exception e) {
+            // 忽略提取错误
+        }
+    }
+
+    /**
+     * 解析金额（去除货币符号）
+     */
+    private Double parseAmount(String amountStr) {
+        try {
+            if (amountStr == null || amountStr.isEmpty()) {
+                return 0.0;
+            }
+            // 移除非数字字符（保留小数点和负号）
+            String cleaned = amountStr.replaceAll("[^\\d.-]", "");
+            return cleaned.isEmpty() ? 0.0 : Double.parseDouble(cleaned);
+        } catch (Exception e) {
+            return 0.0;
+        }
+    }
+
+    /**
+     * 解析退水率
+     */
+    private Double parseRate(String rateStr) {
+        try {
+            if (rateStr == null || rateStr.isEmpty()) {
+                return 0.0;
+            }
+            // 移除百分号
+            String cleaned = rateStr.replace("%", "").trim();
+            return cleaned.isEmpty() ? 0.0 : Double.parseDouble(cleaned) / 100;
+        } catch (Exception e) {
+            return 0.0;
+        }
+    }
+
+    /**
+     * 解析结果状态
+     */
+    private String parseResult(String resultStr) {
+        if (resultStr == null) {
+            return "未知";
+        }
+
+        if (resultStr.contains("赢") || resultStr.contains("+")) {
+            return "赢";
+        } else if (resultStr.contains("输") || resultStr.contains("-")) {
+            return "输";
+        } else if (resultStr.contains("和") || resultStr.contains("0")) {
+            return "和";
+        } else if (resultStr.contains("取消")) {
+            return "取消";
+        } else {
+            return "未知";
+        }
+    }
+
+    /**
+     * 解析结果金额
+     */
+    private Double parseResultAmount(String resultStr) {
+        try {
+            // 提取数字部分（包含负号）
+            Pattern pattern = Pattern.compile("[-+]?\\d+\\.?\\d*");
+            Matcher matcher = pattern.matcher(resultStr);
+
+            if (matcher.find()) {
+                return Double.parseDouble(matcher.group());
+            }
+            return 0.0;
+        } catch (Exception e) {
+            return 0.0;
+        }
+    }
+
+    /**
+     * 转换数据结构以适应数据库
+     */
+    private List<Map<String, Object>> convertForDatabase(List<Map<String, Object>> betList) {
+        List<Map<String, Object>> dbData = new ArrayList<>();
+
+        for (Map<String, Object> bet : betList) {
+            Map<String, Object> dbRecord = new HashMap<>();
+
+            dbRecord.put("id", bet.get("bet_id"));
+            dbRecord.put("bet_id", bet.get("bet_id"));
+            dbRecord.put("period", bet.get("period"));
+            dbRecord.put("bet_time", bet.get("time"));
+            dbRecord.put("game_type", bet.get("game_type"));
+            dbRecord.put("plate", bet.get("plate"));
+            dbRecord.put("bet_amount", bet.get("bet_amount"));
+            dbRecord.put("rebate_rate", bet.get("rebate_rate"));
+            dbRecord.put("result", bet.get("result"));
+            dbRecord.put("result_amount", bet.get("result_amount"));
+            dbRecord.put("parse_time", bet.get("parse_time"));
+            dbRecord.put("source", bet.get("source"));
+
+            dbData.add(dbRecord);
+        }
+
+        return dbData;
+    }
+
+
+    /**
+     * 保存HTML用于调试
+     */
+    private void saveHtmlForDebug(String content, String url) {
+        try {
+            String safeUrl = url.replaceAll("[^a-zA-Z0-9]", "_");
+            String fileName = "debug_" + safeUrl + "_" + System.currentTimeMillis() + ".html";
+            String filePath = "output/debug/" + fileName;
+
+            File directory = new File("output/debug");
+            if (!directory.exists()) {
+                directory.mkdirs();
+            }
+
+            File outputFile = new File(filePath);
+            java.nio.file.Files.write(outputFile.toPath(), content.getBytes());
+
+            System.out.println("调试HTML已保存: " + outputFile.getAbsolutePath());
+
+        } catch (Exception e) {
+            System.err.println("保存调试HTML失败: " + e.getMessage());
+        }
+    }
+
+    /**
+     * 提取调试信息
+     */
+    private void extractDebugInfo(Html html) {
+        System.out.println("\n=== 调试信息 ===");
+
+        // 检查所有表格
+        List<String> tables = html.xpath("//table/@class").all();
+        System.out.println("所有表格class: " + tables);
+
+        // 检查所有tr
+        int trCount = html.xpath("//tr").all().size();
+        System.out.println("TR数量: " + trCount);
+
+        // 检查所有td
+        int tdCount = html.xpath("//td").all().size();
+        System.out.println("TD数量: " + tdCount);
+
+        // 检查cookie相关元素
+        String cookieScript = html.xpath("//script[contains(text(), 'token')]/text()").get();
+        if (cookieScript != null && cookieScript.contains("token")) {
+            System.out.println("发现token相关脚本");
+        }
+
+        // 检查是否有JavaScript重定向
+        String redirectScript = html.xpath("//script[contains(text(), 'location.href') or contains(text(), 'window.location')]/text()").get();
+        if (redirectScript != null) {
+            System.out.println("发现重定向脚本: " + redirectScript.substring(0, Math.min(100, redirectScript.length())));
+        }
+    }
+
+    /**
+     * 打印示例数据
+     */
+    private void printSampleData(List<Map<String, Object>> betList) {
+        System.out.println("\n=== 前3条数据示例 ===");
+        int count = Math.min(3, betList.size());
+        for (int i = 0; i < count; i++) {
+            Map<String, Object> bet = betList.get(i);
+            System.out.printf("注单%d: ID=%s, 时间=%s, 金额=%.2f, 结果=%s, 金额=%.2f%n",
+                    i + 1,
+                    bet.get("bet_id"),
+                    bet.get("time"),
+                    bet.get("bet_amount"),
+                    bet.get("result"),
+                    bet.get("result_amount"));
+        }
+    }
+
+    @Override
+    public Site getSite() {
+        return site;
+    }
+
+    public static void main(String[] args) {
+        String url = "https://4701268539-esh.qdk63ayw8g.com/member/bets?settled=true";
+        // 创建爬虫
+        Spider.create(new CompletedTodayCrawler(""))
+                .addUrl(url) // 添加起始URL
+                .thread(1) // 线程数
+                .run(); // 开始爬取
+    }
+
+
+}
--- a/src/main/java/com/tem/bocai/util/LotteryWebMagicCrawler.java
+++ b/src/main/java/com/tem/bocai/util/LotteryWebMagicCrawler.java
@@ -23,6 +23,7 @@ import java.io.File;
 import java.io.IOException;
 import java.util.*;

+//开奖的历史结果
 public class LotteryWebMagicCrawler implements PageProcessor {

    private final String token;