|
|
package com.dsideal.base.AI32K;
|
|
|
|
|
|
import com.dsideal.base.AI.Generator.PptGenerator;
|
|
|
import com.dsideal.base.DataEase.Model.DataEaseModel;
|
|
|
import com.dsideal.base.Util.LocalMysqlConnectUtil;
|
|
|
import com.dsideal.base.Util.PptAIKit;
|
|
|
import com.jfinal.plugin.activerecord.Db;
|
|
|
import com.jfinal.plugin.activerecord.Record;
|
|
|
import com.dsideal.base.Util.CallDeepSeek;
|
|
|
|
|
|
import java.io.FileWriter;
|
|
|
import java.io.IOException;
|
|
|
import java.nio.charset.StandardCharsets;
|
|
|
import java.nio.file.Files;
|
|
|
import java.nio.file.Paths;
|
|
|
import java.util.List;
|
|
|
import java.util.Map;
|
|
|
import java.util.Set;
|
|
|
import java.util.LinkedHashSet;
|
|
|
import java.util.ArrayList;
|
|
|
import java.util.concurrent.CountDownLatch;
|
|
|
import java.text.SimpleDateFormat;
|
|
|
import java.util.Date;
|
|
|
|
|
|
import java.util.Scanner;
|
|
|
|
|
|
import com.dsideal.base.AI.Generator.WordGenerator;
|
|
|
|
|
|
public class TestMax32K {
|
|
|
private static final int MAX_CHUNK_SIZE = 30000; // 调整为30K字符，充分利用32K输入限制
|
|
|
private static boolean dataSubmitted = false; // 标记数据是否已提交
|
|
|
private static final String[] regions = {"楚雄州", "文山州"}; // 或者根据需要设置具体的地区
|
|
|
|
|
|
public static void main(String[] args) {
|
|
|
Scanner scanner = new Scanner(System.in);
|
|
|
LocalMysqlConnectUtil.Init();
|
|
|
|
|
|
try {
|
|
|
// 第一步：提交所有数据块
|
|
|
System.out.println("开始提交数据到大模型...");
|
|
|
submitAllDataChunks();
|
|
|
dataSubmitted = true;
|
|
|
System.out.println("\n数据提交完成！");
|
|
|
|
|
|
// 第二步：询问用户想生成哪种格式
|
|
|
while (true) {
|
|
|
System.out.println("\n=== 报告生成选项 ===");
|
|
|
System.out.println("现在可以基于上传的数据生成以下格式的报告：");
|
|
|
System.out.println("1. WORD文档 (.docx)");
|
|
|
System.out.println("2. HTML网页 (.html)");
|
|
|
System.out.println("3. PPT演示文稿 (.pptx)");
|
|
|
System.out.println("4. 退出程序");
|
|
|
System.out.print("\n请选择您想生成的报告格式 (1-4): ");
|
|
|
|
|
|
String choice = scanner.nextLine().trim();
|
|
|
|
|
|
switch (choice) {
|
|
|
case "1":
|
|
|
generateWordReport();
|
|
|
break;
|
|
|
case "2":
|
|
|
generateHtmlReport();
|
|
|
break;
|
|
|
case "3":
|
|
|
generatePptReport();
|
|
|
break;
|
|
|
case "4":
|
|
|
System.out.println("程序退出。");
|
|
|
return;
|
|
|
default:
|
|
|
System.out.println("无效选择，请输入 1-4 之间的数字。");
|
|
|
}
|
|
|
}
|
|
|
|
|
|
} catch (Exception e) {
|
|
|
System.err.println("程序执行出错: " + e.getMessage());
|
|
|
e.printStackTrace();
|
|
|
} finally {
|
|
|
scanner.close();
|
|
|
}
|
|
|
}
|
|
|
|
|
|
/**
|
|
|
* 提交所有数据块到大模型
|
|
|
*/
|
|
|
private static void submitAllDataChunks() throws Exception {
|
|
|
if (dataSubmitted) {
|
|
|
System.out.println("数据已经提交过了，无需重复提交。");
|
|
|
return;
|
|
|
}
|
|
|
|
|
|
// 添加获取参数的代码
|
|
|
String sql = "select table_name as TABLE_NAME from core_dataset_table where dataset_group_id in (select id from core_dataset_group where pid='1036317909951057920')";
|
|
|
List<Record> tableList = Db.use(DataEaseModel.DB_NAME).find(sql);
|
|
|
|
|
|
// 修改这一行，传递正确的参数
|
|
|
String[] dataChunks = getDataChunks(regions, tableList);
|
|
|
System.out.println("总共需要提交 " + dataChunks.length + " 个数据块");
|
|
|
|
|
|
// 添加调试功能：保存所有数据块到文件
|
|
|
saveDataChunksToFile(dataChunks);
|
|
|
|
|
|
CountDownLatch latch = new CountDownLatch(dataChunks.length);
|
|
|
|
|
|
for (int i = 0; i < dataChunks.length; i++) {
|
|
|
final int chunkIndex = i;
|
|
|
final boolean isLastChunk = (i == dataChunks.length - 1);
|
|
|
|
|
|
String prompt;
|
|
|
if (isLastChunk) {
|
|
|
prompt = "这是最后一个数据块（第" + (chunkIndex + 1) + "/" + dataChunks.length + "个）。\n" +
|
|
|
"请确认已接收所有数据块，现在数据提交完成。请回复'数据接收完成'以确认。\n\n" +
|
|
|
"数据内容：\n" + dataChunks[chunkIndex];
|
|
|
} else {
|
|
|
prompt = "这是第" + (chunkIndex + 1) + "/" + dataChunks.length + "个数据块，后续还有更多数据。\n" +
|
|
|
"请接收此数据块，无需分析，等待所有数据提交完成。\n\n" +
|
|
|
"数据内容：\n" + dataChunks[chunkIndex];
|
|
|
}
|
|
|
|
|
|
// 保存每个提示词到单独文件（可选）
|
|
|
savePromptToFile(prompt, chunkIndex + 1);
|
|
|
|
|
|
CallDeepSeek.callDeepSeekStream(prompt, new CallDeepSeek.SSEListener() {
|
|
|
@Override
|
|
|
public void onData(String data) {
|
|
|
if (isLastChunk) {
|
|
|
System.out.println("\n大模型确认: " + data);
|
|
|
}
|
|
|
}
|
|
|
|
|
|
@Override
|
|
|
public void onComplete(String fullResponse) {
|
|
|
System.out.println("数据块 " + (chunkIndex + 1) + "/" + dataChunks.length + " 提交完成");
|
|
|
latch.countDown();
|
|
|
}
|
|
|
|
|
|
@Override
|
|
|
public void onError(String error) {
|
|
|
System.err.println("数据块 " + (chunkIndex + 1) + " 提交失败: " + error);
|
|
|
latch.countDown();
|
|
|
}
|
|
|
});
|
|
|
|
|
|
// 避免请求过于频繁
|
|
|
Thread.sleep(1000);
|
|
|
}
|
|
|
|
|
|
latch.await();
|
|
|
}
|
|
|
|
|
|
/**
|
|
|
* 保存所有数据块到文件进行调试
|
|
|
*/
|
|
|
private static void saveDataChunksToFile(String[] dataChunks) {
|
|
|
try {
|
|
|
String timestamp = new SimpleDateFormat("yyyyMMdd_HHmmss").format(new Date());
|
|
|
|
|
|
// 保存所有数据块到一个文件
|
|
|
StringBuilder allData = new StringBuilder();
|
|
|
allData.append("=== 提交给大模型的所有数据块 ===").append("\n");
|
|
|
allData.append("生成时间: ").append(timestamp).append("\n");
|
|
|
allData.append("总数据块数: ").append(dataChunks.length).append("\n");
|
|
|
allData.append("地区: ").append(String.join(", ", regions)).append("\n");
|
|
|
allData.append("\n");
|
|
|
|
|
|
for (int i = 0; i < dataChunks.length; i++) {
|
|
|
allData.append("\n").append("=".repeat(80)).append("\n");
|
|
|
allData.append("数据块 ").append(i + 1).append("/").append(dataChunks.length).append("\n");
|
|
|
allData.append("数据块大小: ").append(dataChunks[i].length()).append(" 字符\n");
|
|
|
allData.append("=".repeat(80)).append("\n");
|
|
|
allData.append(dataChunks[i]);
|
|
|
allData.append("\n");
|
|
|
}
|
|
|
|
|
|
String debugFileName = "debug_data_chunks_" + timestamp + ".txt";
|
|
|
String debugFilePath = "Log/" + debugFileName;
|
|
|
Files.write(Paths.get(debugFilePath), allData.toString().getBytes("UTF-8"));
|
|
|
|
|
|
System.out.println("调试文件已保存: " + debugFilePath);
|
|
|
System.out.println("文件包含 " + dataChunks.length + " 个数据块，总大小: " + allData.length() + " 字符");
|
|
|
|
|
|
// 另外保存每个数据块到单独文件
|
|
|
for (int i = 0; i < dataChunks.length; i++) {
|
|
|
String chunkFileName = "debug_chunk_" + (i + 1) + "_" + timestamp + ".txt";
|
|
|
String chunkFilePath = "Log/" + chunkFileName;
|
|
|
Files.write(Paths.get(chunkFilePath), dataChunks[i].getBytes("UTF-8"));
|
|
|
}
|
|
|
|
|
|
System.out.println("每个数据块也已保存为单独文件: Log/debug_chunk_*_" + timestamp + ".txt");
|
|
|
|
|
|
} catch (Exception e) {
|
|
|
System.err.println("保存调试文件失败: " + e.getMessage());
|
|
|
e.printStackTrace();
|
|
|
}
|
|
|
}
|
|
|
|
|
|
/**
|
|
|
* 保存提示词到文件（可选）
|
|
|
*/
|
|
|
private static void savePromptToFile(String prompt, int chunkIndex) {
|
|
|
try {
|
|
|
String timestamp = new SimpleDateFormat("yyyyMMdd_HHmmss").format(new Date());
|
|
|
String promptFileName = "debug_prompt_" + chunkIndex + "_" + timestamp + ".txt";
|
|
|
String promptFilePath = "Log/" + promptFileName;
|
|
|
Files.write(Paths.get(promptFilePath), prompt.getBytes("UTF-8"));
|
|
|
} catch (Exception e) {
|
|
|
System.err.println("保存提示词文件失败: " + e.getMessage());
|
|
|
}
|
|
|
}
|
|
|
|
|
|
/**
|
|
|
* 生成Word报告（分批次生成）
|
|
|
*/
|
|
|
private static void generateWordReport() {
|
|
|
if (!dataSubmitted) {
|
|
|
System.out.println("请先提交数据！");
|
|
|
return;
|
|
|
}
|
|
|
|
|
|
System.out.println("\n开始分批次生成Word报告...");
|
|
|
|
|
|
try {
|
|
|
// 动态生成地区名称字符串
|
|
|
String regionNames = String.join("与", regions);
|
|
|
String reportTitle = regionNames + "教育资源配置对比分析报告";
|
|
|
|
|
|
// 定义报告的4个部分
|
|
|
String[] reportSections = {
|
|
|
"摘要和数据概览",
|
|
|
generateRegionAnalysisSections(regions),
|
|
|
"对比分析",
|
|
|
"结论与建议"
|
|
|
};
|
|
|
|
|
|
StringBuilder fullReport = new StringBuilder();
|
|
|
fullReport.append("# ").append(reportTitle).append("\n\n");
|
|
|
|
|
|
// 分4次生成报告内容
|
|
|
for (int i = 0; i < 4; i++) {
|
|
|
System.out.println("正在生成第" + (i + 1) + "/4部分: " + getSectionTitle(i));
|
|
|
String sectionContent = generateReportSection(i, regionNames);
|
|
|
if (sectionContent != null && !sectionContent.trim().isEmpty()) {
|
|
|
fullReport.append(sectionContent).append("\n\n");
|
|
|
}
|
|
|
|
|
|
// 添加延迟避免API限制
|
|
|
if (i < 3) {
|
|
|
Thread.sleep(2000);
|
|
|
}
|
|
|
}
|
|
|
|
|
|
// 保存完整报告
|
|
|
saveCompleteReport(fullReport.toString(), reportTitle);
|
|
|
|
|
|
} catch (Exception e) {
|
|
|
System.err.println("生成Word报告时出错: " + e.getMessage());
|
|
|
}
|
|
|
}
|
|
|
|
|
|
/**
|
|
|
* 构建各部分的提示词（优化版）
|
|
|
*/
|
|
|
private static String buildSectionPrompt(int sectionIndex, String regionNames) {
|
|
|
String basePrompt = "基于之前提交的" + regionNames + "的完整教育数据，";
|
|
|
|
|
|
switch (sectionIndex) {
|
|
|
case 0: // 摘要和数据概览
|
|
|
return basePrompt + "请生成报告的摘要和数据概览部分。\n\n" +
|
|
|
"**严格格式要求：**\n" +
|
|
|
"1. 直接从## 摘要开始，绝对不要重复报告主标题\n" +
|
|
|
"2. 摘要内容2-3个段落，每段之间空一行\n" +
|
|
|
"3. 使用## 数据概览作为标题\n" +
|
|
|
"4. 数据概览用简洁的段落描述，不使用表格\n" +
|
|
|
"5. 段落间必须有空行\n" +
|
|
|
"6. 不要使用复杂的格式符号\n" +
|
|
|
"7. 控制在2500字符以内\n" +
|
|
|
"8. 必须基于实际数据，不得编造\n" +
|
|
|
"9. 【重要】报告主标题已存在，请勿重复添加任何形式的主标题\n" +
|
|
|
"10. 【示例格式】直接以下面格式开始：\n" +
|
|
|
"## 摘要\n" +
|
|
|
"根据楚雄州与文山州的教育数据分析...\n\n" +
|
|
|
"## 数据概览\n" +
|
|
|
"本次分析涵盖了...";
|
|
|
|
|
|
case 1: // 各地区分析
|
|
|
return basePrompt + "请生成各地区的详细分析部分。\n\n" +
|
|
|
"**严格格式要求：**\n" +
|
|
|
"1. 为每个地区使用## [地区名]教育资源分析作为标题\n" +
|
|
|
"2. 每个地区分析包含：\n" +
|
|
|
" - 学前教育情况（一个段落）\n" +
|
|
|
" - 义务教育情况（一个段落）\n" +
|
|
|
" - 高中教育情况（一个段落）\n" +
|
|
|
"3. 段落间用空行分隔\n" +
|
|
|
"4. 使用简单的Markdown表格展示关键数据（最多2个表格）\n" +
|
|
|
"5. 表格格式：| 指标 | 数值 | 说明 |\n" +
|
|
|
"6. 控制在3500字符以内";
|
|
|
|
|
|
case 2: // 对比分析
|
|
|
return basePrompt + "请生成地区间的对比分析部分。\n\n" +
|
|
|
"**严格格式要求：**\n" +
|
|
|
"1. 使用## 对比分析作为标题\n" +
|
|
|
"2. 分为以下子部分：\n" +
|
|
|
" ### 教育资源配置对比\n" +
|
|
|
" ### 教育质量差异分析\n" +
|
|
|
" ### 发展趋势对比\n" +
|
|
|
"3. 每个子部分2-3个段落\n" +
|
|
|
"4. 段落间用空行分隔\n" +
|
|
|
"5. 可使用一个对比表格，格式简洁\n" +
|
|
|
"6. 控制在3000字符以内";
|
|
|
|
|
|
case 3: // 结论与建议
|
|
|
return basePrompt + "请生成结论与建议部分。\n\n" +
|
|
|
"**严格格式要求：**\n" +
|
|
|
"1. 使用## 结论与建议作为主标题\n" +
|
|
|
"2. 分为两个子部分：\n" +
|
|
|
" ### 主要结论\n" +
|
|
|
" ### 政策建议\n" +
|
|
|
"3. 主要结论部分：\n" +
|
|
|
" - 使用1. 2. 3. 4.编号列表\n" +
|
|
|
" - 每条结论一个段落\n" +
|
|
|
" - 结论间用空行分隔\n" +
|
|
|
"4. 政策建议部分：\n" +
|
|
|
" - 使用1. 2. 3. 4. 5.编号列表\n" +
|
|
|
" - 每条建议包含具体措施\n" +
|
|
|
" - 建议间用空行分隔\n" +
|
|
|
"5. 不要使用过多的子标题\n" +
|
|
|
"6. 控制在2800字符以内\n" +
|
|
|
"7. 必须基于前面的数据分析";
|
|
|
|
|
|
default:
|
|
|
return basePrompt + "请生成报告内容。";
|
|
|
}
|
|
|
}
|
|
|
|
|
|
/**
|
|
|
* 格式验证和修正方法
|
|
|
*/
|
|
|
private static String validateAndFixFormat(String content) {
|
|
|
if (content == null || content.trim().isEmpty()) {
|
|
|
return content;
|
|
|
}
|
|
|
|
|
|
StringBuilder fixed = new StringBuilder();
|
|
|
String[] lines = content.split("\n");
|
|
|
|
|
|
for (int i = 0; i < lines.length; i++) {
|
|
|
String line = lines[i];
|
|
|
|
|
|
// 修正标题格式
|
|
|
if (line.trim().startsWith("#")) {
|
|
|
// 确保标题前后有空行
|
|
|
if (i > 0 && !lines[i - 1].trim().isEmpty()) {
|
|
|
fixed.append("\n");
|
|
|
}
|
|
|
fixed.append(line).append("\n\n");
|
|
|
}
|
|
|
// 修正列表格式
|
|
|
else if (line.trim().matches("^[0-9]+\\.") || line.trim().startsWith("-")) {
|
|
|
fixed.append(line).append("\n\n");
|
|
|
}
|
|
|
// 普通段落
|
|
|
else if (!line.trim().isEmpty()) {
|
|
|
fixed.append(line).append("\n\n");
|
|
|
}
|
|
|
// 保留空行但不重复
|
|
|
else if (line.trim().isEmpty() && i > 0 && !lines[i - 1].trim().isEmpty()) {
|
|
|
// 空行已经在上面添加了，这里跳过
|
|
|
}
|
|
|
}
|
|
|
|
|
|
return fixed.toString().trim();
|
|
|
}
|
|
|
|
|
|
/**
|
|
|
* 修改后的生成报告的单个部分方法
|
|
|
*/
|
|
|
private static String generateReportSection(int sectionIndex, String regionNames) {
|
|
|
try {
|
|
|
String prompt = buildSectionPrompt(sectionIndex, regionNames);
|
|
|
|
|
|
StringBuilder sectionContent = new StringBuilder();
|
|
|
CountDownLatch latch = new CountDownLatch(1);
|
|
|
|
|
|
CallDeepSeek.callDeepSeekStream(prompt, new CallDeepSeek.SSEListener() {
|
|
|
@Override
|
|
|
public void onData(String data) {
|
|
|
sectionContent.append(data);
|
|
|
System.out.print(".");
|
|
|
}
|
|
|
|
|
|
@Override
|
|
|
public void onComplete(String fullResponse) {
|
|
|
System.out.println(" 完成！");
|
|
|
latch.countDown();
|
|
|
}
|
|
|
|
|
|
@Override
|
|
|
public void onError(String error) {
|
|
|
System.err.println("\n生成第" + (sectionIndex + 1) + "部分失败: " + error);
|
|
|
latch.countDown();
|
|
|
}
|
|
|
});
|
|
|
|
|
|
latch.await();
|
|
|
|
|
|
// 格式验证和修正
|
|
|
String rawContent = sectionContent.toString();
|
|
|
String fixedContent = validateAndFixFormat(rawContent);
|
|
|
|
|
|
// 保存原始和修正后的内容用于调试
|
|
|
saveDebugContent(rawContent, fixedContent, sectionIndex);
|
|
|
|
|
|
return fixedContent;
|
|
|
|
|
|
} catch (Exception e) {
|
|
|
System.err.println("生成第" + (sectionIndex + 1) + "部分时出错: " + e.getMessage());
|
|
|
return "";
|
|
|
}
|
|
|
}
|
|
|
|
|
|
/**
|
|
|
* 保存调试内容
|
|
|
*/
|
|
|
private static void saveDebugContent(String rawContent, String fixedContent, int sectionIndex) {
|
|
|
try {
|
|
|
String timestamp = new SimpleDateFormat("yyyyMMdd_HHmmss").format(new Date());
|
|
|
|
|
|
// 保存原始内容
|
|
|
String rawFileName = "debug_section_" + (sectionIndex + 1) + "_raw_" + timestamp + ".txt";
|
|
|
Files.write(Paths.get("Doc/" + rawFileName), rawContent.getBytes("UTF-8"));
|
|
|
|
|
|
// 保存修正后内容
|
|
|
String fixedFileName = "debug_section_" + (sectionIndex + 1) + "_fixed_" + timestamp + ".txt";
|
|
|
Files.write(Paths.get("Doc/" + fixedFileName), fixedContent.getBytes("UTF-8"));
|
|
|
|
|
|
System.out.println("第" + (sectionIndex + 1) + "部分调试文件已保存");
|
|
|
|
|
|
} catch (Exception e) {
|
|
|
System.err.println("保存调试文件失败: " + e.getMessage());
|
|
|
}
|
|
|
}
|
|
|
|
|
|
/**
|
|
|
* 获取部分标题
|
|
|
*/
|
|
|
private static String getSectionTitle(int index) {
|
|
|
switch (index) {
|
|
|
case 0:
|
|
|
return "摘要和数据概览";
|
|
|
case 1:
|
|
|
return "各地区详细分析";
|
|
|
case 2:
|
|
|
return "对比分析";
|
|
|
case 3:
|
|
|
return "结论与建议";
|
|
|
default:
|
|
|
return "未知部分";
|
|
|
}
|
|
|
}
|
|
|
|
|
|
/**
|
|
|
* 生成地区分析部分的信息
|
|
|
*/
|
|
|
private static String generateRegionAnalysisSections(String[] regions) {
|
|
|
return "各地区分析（" + String.join("、", regions) + "）";
|
|
|
}
|
|
|
|
|
|
/**
|
|
|
* 保存完整报告
|
|
|
*/
|
|
|
private static void saveCompleteReport(String fullContent, String reportTitle) {
|
|
|
try {
|
|
|
String timestamp = new SimpleDateFormat("yyyyMMdd_HHmmss").format(new Date());
|
|
|
|
|
|
// 保存为文本文件用于调试
|
|
|
String txtFileName = "complete_report_" + timestamp + ".txt";
|
|
|
String txtFilePath = "Log/" + txtFileName;
|
|
|
Files.write(Paths.get(txtFilePath), fullContent.getBytes("UTF-8"));
|
|
|
System.out.println("完整报告文本已保存: " + txtFilePath);
|
|
|
|
|
|
// 生成Word文档
|
|
|
String docxFileName = "analysis_report_" + timestamp + ".docx";
|
|
|
WordGenerator.generateWordDocument(fullContent, docxFileName, regions);
|
|
|
System.out.println("Word报告已保存至: WebRoot/upload/" + docxFileName);
|
|
|
|
|
|
System.out.println("\n=== 报告生成完成 ===");
|
|
|
System.out.println("总字符数: " + fullContent.length());
|
|
|
System.out.println("报告标题: " + reportTitle);
|
|
|
|
|
|
} catch (Exception e) {
|
|
|
System.err.println("保存完整报告失败: " + e.getMessage());
|
|
|
}
|
|
|
}
|
|
|
|
|
|
/**
|
|
|
* 生成HTML报告
|
|
|
*/
|
|
|
private static void generateHtmlReport() throws IOException {
|
|
|
if (!dataSubmitted) {
|
|
|
System.out.println("请先提交数据！");
|
|
|
return;
|
|
|
}
|
|
|
|
|
|
System.out.println("\n开始分段生成HTML报告...");
|
|
|
|
|
|
StringBuilder fullHtmlContent = new StringBuilder();
|
|
|
|
|
|
// 第一段：HTML基础结构 + 前3个图表
|
|
|
generateHtmlPart1(fullHtmlContent);
|
|
|
|
|
|
// 第二段：中间3个图表
|
|
|
generateHtmlPart2(fullHtmlContent);
|
|
|
|
|
|
// 第三段：最后3个图表 + 结束标签
|
|
|
generateHtmlPart3(fullHtmlContent);
|
|
|
|
|
|
// 保存完整的HTML文件
|
|
|
saveCompleteHtmlFile(fullHtmlContent.toString());
|
|
|
}
|
|
|
|
|
|
private static void generateHtmlPart1(StringBuilder fullContent) {
|
|
|
String prompt1 = "生成HTML报告的第一部分，要求：\n" +
|
|
|
"1. 包含完整的HTML文档开头：<!DOCTYPE html>、<html>、<head>等\n" +
|
|
|
"2. 在<head>中引入ECharts：<script src=\"./echarts.min.js\"></script>\n" +
|
|
|
"3. 包含CSS样式\n" +
|
|
|
"4. 包含报告标题和摘要部分\n" +
|
|
|
"5. 生成前3个图表的完整代码：\n" +
|
|
|
" - 人口变化趋势图（折线图，id: populationTrend）\n" +
|
|
|
" - 教育规模分布图（柱状图，id: educationScale）\n" +
|
|
|
" - 城乡人口对比图（饼图，id: urbanRuralComparison）\n" +
|
|
|
"6. 每个图表包含div容器和完整的JavaScript初始化代码\n" +
|
|
|
"7. 使用window.addEventListener('DOMContentLoaded', function() {包装JavaScript\n" +
|
|
|
"8. 【重要】不要添加</body>和</html>结束标签，这是第一部分\n" +
|
|
|
"9. 基于之前提交的数据生成真实图表";
|
|
|
|
|
|
try {
|
|
|
StringBuilder part1Content = new StringBuilder();
|
|
|
CountDownLatch latch = new CountDownLatch(1);
|
|
|
|
|
|
CallDeepSeek.callDeepSeekStream(prompt1, new CallDeepSeek.SSEListener() {
|
|
|
@Override
|
|
|
public void onData(String data) {
|
|
|
part1Content.append(data);
|
|
|
System.out.print("."); // 显示进度
|
|
|
}
|
|
|
|
|
|
@Override
|
|
|
public void onComplete(String fullResponse) {
|
|
|
fullContent.append(fullResponse);
|
|
|
System.out.println("\n第一部分生成完成");
|
|
|
latch.countDown();
|
|
|
}
|
|
|
|
|
|
@Override
|
|
|
public void onError(String error) {
|
|
|
System.err.println("生成第一部分时出错: " + error);
|
|
|
latch.countDown();
|
|
|
}
|
|
|
});
|
|
|
|
|
|
latch.await();
|
|
|
} catch (Exception e) {
|
|
|
System.err.println("生成HTML第一部分失败: " + e.getMessage());
|
|
|
e.printStackTrace();
|
|
|
}
|
|
|
}
|
|
|
|
|
|
private static void generateHtmlPart2(StringBuilder fullContent) {
|
|
|
String prompt2 = "生成HTML报告的第二部分，要求：\n" +
|
|
|
"1. 【重要】不要包含HTML文档头部，直接从图表内容开始\n" +
|
|
|
"2. 生成中间3个图表的完整代码：\n" +
|
|
|
" - 学龄人口预测图（面积图，id: schoolAgePrediction）\n" +
|
|
|
" - 教育资源配置图（雷达图，id: educationResources）\n" +
|
|
|
" - 区域教育发展对比图（条形图，id: regionalComparison）\n" +
|
|
|
"3. 每个图表包含div容器和完整的JavaScript初始化代码\n" +
|
|
|
"4. JavaScript代码要与第一部分的DOMContentLoaded事件兼容\n" +
|
|
|
"5. 【重要】不要添加</body>和</html>结束标签\n" +
|
|
|
"6. 基于之前提交的数据生成真实图表";
|
|
|
|
|
|
try {
|
|
|
StringBuilder part2Content = new StringBuilder();
|
|
|
CountDownLatch latch = new CountDownLatch(1);
|
|
|
|
|
|
CallDeepSeek.callDeepSeekStream(prompt2, new CallDeepSeek.SSEListener() {
|
|
|
@Override
|
|
|
public void onData(String data) {
|
|
|
part2Content.append(data);
|
|
|
System.out.print("."); // 显示进度
|
|
|
}
|
|
|
|
|
|
@Override
|
|
|
public void onComplete(String fullResponse) {
|
|
|
fullContent.append(fullResponse);
|
|
|
System.out.println("\n第二部分生成完成");
|
|
|
latch.countDown();
|
|
|
}
|
|
|
|
|
|
@Override
|
|
|
public void onError(String error) {
|
|
|
System.err.println("生成第二部分时出错: " + error);
|
|
|
latch.countDown();
|
|
|
}
|
|
|
});
|
|
|
|
|
|
latch.await();
|
|
|
} catch (Exception e) {
|
|
|
System.err.println("生成HTML第二部分失败: " + e.getMessage());
|
|
|
e.printStackTrace();
|
|
|
}
|
|
|
}
|
|
|
|
|
|
private static void generateHtmlPart3(StringBuilder fullContent) {
|
|
|
String prompt3 = "生成HTML报告的第三部分，要求：\n" +
|
|
|
"1. 【重要】不要包含HTML文档头部，直接从图表内容开始\n" +
|
|
|
"2. 生成最后3个图表的完整代码：\n" +
|
|
|
" - 教育投入产出分析图（散点图，id: inputOutputAnalysis）\n" +
|
|
|
" - 师资力量分布图（热力图，id: teacherDistribution）\n" +
|
|
|
" - 综合发展指数图（仪表盘图，id: comprehensiveIndex）\n" +
|
|
|
"3. 每个图表包含div容器和完整的JavaScript初始化代码\n" +
|
|
|
"4. 添加结论和建议部分\n" +
|
|
|
"5. 添加窗口大小变化处理代码\n" +
|
|
|
"6. 【重要】必须以完整的</script></body></html>结束\n" +
|
|
|
"7. 基于之前提交的数据生成真实图表";
|
|
|
|
|
|
try {
|
|
|
StringBuilder part3Content = new StringBuilder();
|
|
|
CountDownLatch latch = new CountDownLatch(1);
|
|
|
|
|
|
CallDeepSeek.callDeepSeekStream(prompt3, new CallDeepSeek.SSEListener() {
|
|
|
@Override
|
|
|
public void onData(String data) {
|
|
|
part3Content.append(data);
|
|
|
System.out.print("."); // 显示进度
|
|
|
}
|
|
|
|
|
|
@Override
|
|
|
public void onComplete(String fullResponse) {
|
|
|
fullContent.append(fullResponse);
|
|
|
System.out.println("\n第三部分生成完成");
|
|
|
latch.countDown();
|
|
|
}
|
|
|
|
|
|
@Override
|
|
|
public void onError(String error) {
|
|
|
System.err.println("生成第三部分时出错: " + error);
|
|
|
latch.countDown();
|
|
|
}
|
|
|
});
|
|
|
|
|
|
latch.await();
|
|
|
} catch (Exception e) {
|
|
|
System.err.println("生成HTML第三部分失败: " + e.getMessage());
|
|
|
e.printStackTrace();
|
|
|
}
|
|
|
}
|
|
|
|
|
|
/**
|
|
|
* 生成PPT报告
|
|
|
*/
|
|
|
private static void generatePptReport() {
|
|
|
if (!dataSubmitted) {
|
|
|
System.out.println("请先提交数据！");
|
|
|
return;
|
|
|
}
|
|
|
|
|
|
System.out.println("\n开始生成PPT报告...");
|
|
|
|
|
|
try {
|
|
|
String prompt = "基于之前提交的所有数据，请生成一份PPT演示文稿的详细内容大纲。\n" +
|
|
|
"要求：\n" +
|
|
|
"1. 提供完整的PPT结构和每页内容\n" +
|
|
|
"2. 包含标题页、目录、数据分析、图表说明、结论建议等\n" +
|
|
|
"3. 每页PPT都要有明确的标题和要点\n" +
|
|
|
"4. 适合制作成专业的演示文稿\n" +
|
|
|
"5. 内容要简洁明了，重点突出";
|
|
|
|
|
|
StringBuilder pptContent = new StringBuilder();
|
|
|
CountDownLatch latch = new CountDownLatch(1);
|
|
|
|
|
|
CallDeepSeek.callDeepSeekStream(prompt, new CallDeepSeek.SSEListener() {
|
|
|
@Override
|
|
|
public void onData(String data) {
|
|
|
pptContent.append(data);
|
|
|
System.out.print("."); // 显示进度
|
|
|
}
|
|
|
|
|
|
@Override
|
|
|
public void onComplete(String fullResponse) {
|
|
|
System.out.println("\nPPT内容生成完成！");
|
|
|
|
|
|
// 调用PptGenerator生成PPT
|
|
|
try {
|
|
|
String timestamp = new SimpleDateFormat("yyyyMMdd_HHmmss").format(new Date());
|
|
|
String fileName = "analysis_report_" + timestamp + ".pptx";
|
|
|
|
|
|
// 使用PptGenerator生成PPT
|
|
|
String token = PptAIKit.createApiToken("dsideal", 1000);
|
|
|
String pptInfo = PptGenerator.generatePptFromMarkdown(pptContent.toString(), token);
|
|
|
|
|
|
// 保存PPT信息到文本文件
|
|
|
String infoFilePath = "WebRoot/upload/ppt_info_" + timestamp + ".txt";
|
|
|
if (pptInfo != null) {
|
|
|
Files.write(Paths.get(infoFilePath), pptInfo.getBytes("UTF-8"));
|
|
|
System.out.println("PPT生成信息已保存至: " + infoFilePath);
|
|
|
} else {
|
|
|
Files.write(Paths.get(infoFilePath), "PPT生成失败".getBytes("UTF-8"));
|
|
|
System.out.println("PPT生成失败，错误信息已保存至: " + infoFilePath);
|
|
|
}
|
|
|
|
|
|
} catch (Exception e) {
|
|
|
System.err.println("生成PPT失败: " + e.getMessage());
|
|
|
}
|
|
|
|
|
|
latch.countDown();
|
|
|
}
|
|
|
|
|
|
@Override
|
|
|
public void onError(String error) {
|
|
|
System.err.println("\n生成PPT报告失败: " + error);
|
|
|
latch.countDown();
|
|
|
}
|
|
|
});
|
|
|
|
|
|
latch.await();
|
|
|
|
|
|
} catch (Exception e) {
|
|
|
System.err.println("生成PPT报告时出错: " + e.getMessage());
|
|
|
}
|
|
|
}
|
|
|
|
|
|
/**
|
|
|
* 分割过大的单表数据
|
|
|
*/
|
|
|
private static List<String> splitLargeTable(Set<String> fieldNames,
|
|
|
List<Record> allTableData, int maxSize) {
|
|
|
List<String> chunks = new ArrayList<>();
|
|
|
StringBuilder currentTableChunk = new StringBuilder();
|
|
|
|
|
|
for (Record dataRecord : allTableData) {
|
|
|
Map<String, Object> columns = dataRecord.getColumns();
|
|
|
StringBuilder rowData = new StringBuilder();
|
|
|
rowData.append("[");
|
|
|
|
|
|
boolean first = true;
|
|
|
for (String fieldName : fieldNames) {
|
|
|
if (!first) rowData.append(",");
|
|
|
Object value = columns.get(fieldName);
|
|
|
if (value instanceof String) {
|
|
|
rowData.append("\"").append(value).append("\"");
|
|
|
} else {
|
|
|
rowData.append(value);
|
|
|
}
|
|
|
first = false;
|
|
|
}
|
|
|
rowData.append("]\n");
|
|
|
|
|
|
// 检查是否超过限制
|
|
|
if (currentTableChunk.length() + rowData.length() > maxSize) {
|
|
|
if (!currentTableChunk.isEmpty()) {
|
|
|
chunks.add(currentTableChunk.toString());
|
|
|
currentTableChunk = new StringBuilder();
|
|
|
}
|
|
|
}
|
|
|
currentTableChunk.append(rowData);
|
|
|
}
|
|
|
|
|
|
if (!currentTableChunk.isEmpty()) {
|
|
|
chunks.add(currentTableChunk.toString());
|
|
|
}
|
|
|
|
|
|
return chunks;
|
|
|
}
|
|
|
|
|
|
/**
|
|
|
* 提取数据分块逻辑为独立方法
|
|
|
*/
|
|
|
private static String[] getDataChunks(String[] regions, List<Record> tableList) {
|
|
|
List<String> dataChunks = new ArrayList<>();
|
|
|
StringBuilder currentChunk = new StringBuilder();
|
|
|
|
|
|
String header = "数据说明: 以下是云南省教育数据的压缩格式\n" +
|
|
|
"格式: 表名 -> 字段列表 -> 数据行(数组格式)\n" +
|
|
|
"地区范围: " + String.join(",", regions) + "\n\n";
|
|
|
currentChunk.append(header);
|
|
|
|
|
|
// 遍历所有相关数据表
|
|
|
for (Record record : tableList) {
|
|
|
String tableName = record.getStr("TABLE_NAME");
|
|
|
|
|
|
// 为当前表收集所有数据
|
|
|
List<Record> allTableData = new ArrayList<>();
|
|
|
Set<String> fieldNames = new LinkedHashSet<>();
|
|
|
|
|
|
// 为每个地区收集数据
|
|
|
for (String region : regions) {
|
|
|
String sql = "select * from `" + tableName + "` where `行政区划`=?";
|
|
|
List<Record> listContent = Db.use(DataEaseModel.DB_NAME).find(sql, region);
|
|
|
|
|
|
if (!listContent.isEmpty()) {
|
|
|
allTableData.addAll(listContent);
|
|
|
// 收集字段名（使用第一条记录的字段结构）
|
|
|
if (fieldNames.isEmpty()) {
|
|
|
fieldNames.addAll(listContent.get(0).getColumns().keySet());
|
|
|
}
|
|
|
}
|
|
|
}
|
|
|
|
|
|
if (!allTableData.isEmpty()) {
|
|
|
// 构建当前表的完整数据块
|
|
|
StringBuilder tableData = new StringBuilder();
|
|
|
tableData.append("\n表: ").append(tableName).append("\n");
|
|
|
tableData.append("字段: ").append(String.join(",", fieldNames)).append("\n");
|
|
|
|
|
|
// 输出压缩格式的数据
|
|
|
for (Record dataRecord : allTableData) {
|
|
|
Map<String, Object> columns = dataRecord.getColumns();
|
|
|
tableData.append("[");
|
|
|
|
|
|
boolean first = true;
|
|
|
for (String fieldName : fieldNames) {
|
|
|
if (!first) tableData.append(",");
|
|
|
Object value = columns.get(fieldName);
|
|
|
if (value instanceof String) {
|
|
|
tableData.append("\"").append(value).append("\"");
|
|
|
} else {
|
|
|
tableData.append(value);
|
|
|
}
|
|
|
first = false;
|
|
|
}
|
|
|
|
|
|
tableData.append("]\n");
|
|
|
}
|
|
|
|
|
|
// 检查是否需要分块
|
|
|
String tableDataStr = tableData.toString();
|
|
|
if (currentChunk.length() + tableDataStr.length() > MAX_CHUNK_SIZE) {
|
|
|
// 当前块已满，保存并开始新块
|
|
|
if (currentChunk.length() > header.length()) {
|
|
|
dataChunks.add(currentChunk.toString());
|
|
|
currentChunk = new StringBuilder();
|
|
|
currentChunk.append(header);
|
|
|
}
|
|
|
|
|
|
// 如果单个表数据超过限制，需要进一步分割
|
|
|
if (tableDataStr.length() > MAX_CHUNK_SIZE - header.length()) {
|
|
|
List<String> tableChunks = splitLargeTable(fieldNames, allTableData, MAX_CHUNK_SIZE - header.length());
|
|
|
for (int i = 0; i < tableChunks.size(); i++) {
|
|
|
StringBuilder chunkBuilder = new StringBuilder();
|
|
|
chunkBuilder.append(header);
|
|
|
chunkBuilder.append("\n[续] 表: ").append(tableName).append(" (第").append(i + 1).append("部分)\n");
|
|
|
chunkBuilder.append("字段: ").append(String.join(",", fieldNames)).append("\n");
|
|
|
chunkBuilder.append(tableChunks.get(i));
|
|
|
dataChunks.add(chunkBuilder.toString());
|
|
|
}
|
|
|
} else {
|
|
|
currentChunk.append(tableDataStr);
|
|
|
}
|
|
|
} else {
|
|
|
currentChunk.append(tableDataStr);
|
|
|
}
|
|
|
}
|
|
|
}
|
|
|
|
|
|
// 添加最后一个块
|
|
|
if (currentChunk.length() > header.length()) {
|
|
|
dataChunks.add(currentChunk.toString());
|
|
|
}
|
|
|
|
|
|
return dataChunks.toArray(new String[0]);
|
|
|
}
|
|
|
/**
|
|
|
* 保存完整的HTML文件
|
|
|
* @param htmlContent HTML内容
|
|
|
*/
|
|
|
private static void saveCompleteHtmlFile(String htmlContent) {
|
|
|
try {
|
|
|
// 生成文件名（带时间戳）
|
|
|
SimpleDateFormat sdf = new SimpleDateFormat("yyyyMMdd_HHmmss");
|
|
|
String timestamp = sdf.format(new Date());
|
|
|
String fileName = "analysis_report_" + timestamp + ".html";
|
|
|
String outputPath = "WebRoot/upload/" + fileName;
|
|
|
|
|
|
// 清理HTML内容
|
|
|
htmlContent = cleanHtmlContent(htmlContent);
|
|
|
|
|
|
// 保存HTML文件
|
|
|
try (FileWriter writer = new FileWriter(outputPath, StandardCharsets.UTF_8)) {
|
|
|
writer.write(htmlContent);
|
|
|
}
|
|
|
System.out.println("HTML文件保存成功: " + outputPath);
|
|
|
} catch (IOException e) {
|
|
|
System.err.println("保存HTML文件时出错: " + e.getMessage());
|
|
|
e.printStackTrace();
|
|
|
throw new RuntimeException("保存HTML文件失败", e);
|
|
|
}
|
|
|
}
|
|
|
|
|
|
/**
|
|
|
* 清理HTML内容
|
|
|
*/
|
|
|
private static String cleanHtmlContent(String htmlContent) {
|
|
|
// 移除可能的markdown代码块标记
|
|
|
htmlContent = htmlContent.replaceAll("```html\\s*", "");
|
|
|
htmlContent = htmlContent.replaceAll("```\\s*$", "");
|
|
|
htmlContent = htmlContent.replaceAll("^```\\s*", "");
|
|
|
|
|
|
// 移除DeepSeek可能添加的说明文本
|
|
|
htmlContent = htmlContent.replaceAll("Here's the complete HTML report.*?:", "");
|
|
|
htmlContent = htmlContent.replaceAll("Here's.*?HTML.*?:", "");
|
|
|
htmlContent = htmlContent.replaceAll("以下是.*?HTML.*?:", "");
|
|
|
|
|
|
// 查找真正的HTML文档开始位置
|
|
|
int doctypeIndex = htmlContent.toLowerCase().indexOf("<!doctype html>");
|
|
|
int htmlIndex = htmlContent.toLowerCase().indexOf("<html");
|
|
|
|
|
|
if (doctypeIndex != -1) {
|
|
|
// 从<!DOCTYPE html>开始截取
|
|
|
htmlContent = htmlContent.substring(doctypeIndex);
|
|
|
} else if (htmlIndex != -1) {
|
|
|
// 从<html开始截取，并添加DOCTYPE
|
|
|
htmlContent = "<!DOCTYPE html>\n" + htmlContent.substring(htmlIndex);
|
|
|
} else {
|
|
|
// 如果都没有找到，说明不是完整的HTML文档，添加基本结构
|
|
|
htmlContent = "<!DOCTYPE html>\n<html lang='zh-CN'>\n<head>\n<meta charset='UTF-8'>\n<title>教育分析报告</title>\n</head>\n<body>\n" + htmlContent + "\n</body>\n</html>";
|
|
|
}
|
|
|
|
|
|
// 移除可能存在的重复DOCTYPE和html标签
|
|
|
htmlContent = htmlContent.replaceAll("(?i)<!DOCTYPE html>\\s*<!DOCTYPE html>", "<!DOCTYPE html>");
|
|
|
htmlContent = htmlContent.replaceAll("(?i)<html[^>]*>\\s*<html[^>]*>", "<html lang='zh-CN'>");
|
|
|
|
|
|
return htmlContent.trim();
|
|
|
}
|
|
|
}
|
|
|
|
|
|
|
|
|
|
