xieyinlu/ruoyi-ai
1
0
Fork 0
mirror of https://gitcode.com/ageerle/ruoyi-ai.git synced 2026-03-23 17:43:43 +08:00
Code Issues Packages Projects Releases Wiki Activity

恢复微信模块，优化知识库切片功能

Browse Source
This commit is contained in:
ageerle
2025-09-19 11:15:37 +08:00
parent fa5dc80a93
commit afc1272ff5
39 changed files with 1356 additions and 367 deletions
Download Patch File Download Diff File Expand all files Collapse all files

41
ruoyi-modules-api/ruoyi-knowledge-api/src/main/java/org/ruoyi/chain/loader/ExcelFileLoader.java Normal file
View File

@@ -0,0 +1,41 @@
package org.ruoyi.chain.loader;
import dev.langchain4j.data.document.Document;
import dev.langchain4j.data.document.parser.apache.tika.ApacheTikaDocumentParser;
import lombok.AllArgsConstructor;
import lombok.extern.slf4j.Slf4j;
import org.ruoyi.chain.split.TextSplitter;
import org.ruoyi.common.core.exception.UtilException;
import org.springframework.stereotype.Component;
import java.io.BufferedInputStream;
import java.io.IOException;
import java.io.InputStream;
import java.util.List;
@Component
@AllArgsConstructor
@Slf4j
public class ExcelFileLoader implements ResourceLoader {
private final TextSplitter textSplitter;
private static final int DEFAULT_BUFFER_SIZE = 8192;
@Override
public String getContent(InputStream inputStream) {
// 使用带缓冲的输入流包装（保持原流不自动关闭）
try (InputStream bufferedStream = new BufferedInputStream(inputStream, DEFAULT_BUFFER_SIZE)) {
ApacheTikaDocumentParser apacheTikaDocumentParser = new ApacheTikaDocumentParser();
Document document = apacheTikaDocumentParser.parse(bufferedStream);
return document.text();
} catch (IOException e) {
String errorMsg = "Excel文件流读取失败";
throw new UtilException(errorMsg, e);
} catch (RuntimeException e) {
String errorMsg = "Excel内容解析异常";
throw new UtilException(errorMsg, e);
}
}
@Override
public List<String> getChunkList(String content, String kid) {
return textSplitter.split(content, kid);
}
}

2
ruoyi-modules-api/ruoyi-knowledge-api/src/main/java/org/ruoyi/chain/loader/ResourceLoaderFactory.java
View File

@@ -36,6 +36,8 @@ public class ResourceLoaderFactory {
return new MarkDownFileLoader(markdownTextSplitter);
}else if (FileType.isCodeFile(fileType)) {
return new CodeFileLoader(codeTextSplitter);
} else if (FileType.isExcel(fileType)) {
return new ExcelFileLoader(excelTextSplitter);
}else {
return new TextFileLoader(characterTextSplitter);
}

17
ruoyi-modules-api/ruoyi-knowledge-api/src/main/java/org/ruoyi/chain/split/ExcelTextSplitter.java Normal file
View File

@@ -0,0 +1,17 @@
package org.ruoyi.chain.split;
import lombok.AllArgsConstructor;
import lombok.extern.slf4j.Slf4j;
import org.springframework.stereotype.Component;
import java.util.List;
@Component
@AllArgsConstructor
@Slf4j
public class ExcelTextSplitter implements TextSplitter{
@Override
public List<String> split(String content, String kid) {
return null;
}
}

23
ruoyi-modules-api/ruoyi-knowledge-api/src/main/java/org/ruoyi/service/DashscopeService.java
View File

@@ -1,23 +0,0 @@
package org.ruoyi.service;
import java.io.IOException;
/**
* @Description: 阿里百炼api
* @Date: 2025/6/4 下午2:24
*/
public interface DashscopeService {
/**
* 视觉推理（QVQ）
* @param imageUrl 图片可访问的地址
* @return
*/
String qvq(String imageUrl) throws IOException;
/**
* 视觉推理（QVQ） 使用本地文件（输入Base64编码或本地路径）
* @param localPath 图片文件的绝对路径
* @return
*/
String qvq4LocalPath(String localPath) throws IOException;
}

34
ruoyi-modules-api/ruoyi-knowledge-api/src/main/java/org/ruoyi/service/PdfImageExtractService.java
View File

@@ -1,34 +0,0 @@
/**
* 处理文件内容
*
* @param unzip Base64编码的图片数组
* @return 文件内容结果列表
* @throws IOException 如果API调用过程中发生错误
*/
List<PdfFileContentResult> dealFileContent(String[] unzip) throws IOException;
/**
*利用百炼接口处理文件内容
*
* @param imageUrl 传入图片地址
* @return 文件内容结果列表
* @throws IOException 如果API调用过程中发生错误
*/
List<PdfFileContentResult> dealFileContent4Dashscope(String imageUrl) throws IOException;
/**
* 利用百炼接口处理文件内容
*
* 视觉推理（QVQ） 使用本地文件（输入Base64编码或本地路径）
* @param localPath 图片文件的绝对路径
* @return
*/
List<PdfFileContentResult> dealFileContent4DashscopeBase64(String localPath)throws IOException;
/**
* 提取PDF中的图片并调用gpt-4o-mini,识别图片内容并返回
* @param file
* @return
* @throws IOException
*/
List<PdfFileContentResult> extractImages(MultipartFile file) throws IOException;
}

150
ruoyi-modules-api/ruoyi-knowledge-api/src/main/java/org/ruoyi/service/impl/DashscopeServiceImpl.java
View File

@@ -1,150 +0,0 @@
package org.ruoyi.service.impl;
import com.alibaba.dashscope.aigc.multimodalconversation.MultiModalConversation;
import com.alibaba.dashscope.aigc.multimodalconversation.MultiModalConversationParam;
import com.alibaba.dashscope.aigc.multimodalconversation.MultiModalConversationResult;
import com.alibaba.dashscope.common.MultiModalMessage;
import com.alibaba.dashscope.common.Role;
import io.reactivex.Flowable;
import java.io.IOException;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.Collections;
import java.util.HashMap;
import java.util.List;
import java.util.Map;
import java.util.Objects;
import lombok.extern.slf4j.Slf4j;
import org.ruoyi.domain.PdfFileContentResult;
import org.ruoyi.service.DashscopeService;
import org.springframework.beans.factory.annotation.Value;
import org.springframework.stereotype.Service;
/**
* @Description: 阿里百炼API
* @Date: 2025/6/4 下午2:28
*/
@Service
@Slf4j
public class DashscopeServiceImpl implements DashscopeService {
private boolean isFirstPrint;
@Value("${dashscope.model}")
private String serviceModel;
@Value("${dashscope.key}")
private String serviceKey;
/**
* 视觉推理（QVQ）
* @param imageUrl 图片可访问地址
* @return
*/
@Override
public String qvq(String imageUrl) throws IOException {
try {
// 构建多模态消息
MultiModalMessage userMessage = MultiModalMessage.builder()
.role(Role.USER.getValue())
.content(Arrays.asList(
Collections.singletonMap("text", "这张图片有什么"),
Collections.singletonMap("image", imageUrl)
))
.build();
// 构建请求参数
MultiModalConversationParam param = MultiModalConversationParam.builder()
.apiKey(serviceKey) // 使用配置文件中的API Key
.model(serviceModel)
.message(userMessage)
.build();
MultiModalConversation conv = new MultiModalConversation();
// 调用API
Flowable<MultiModalConversationResult> result = conv.streamCall(
param);
StringBuilder reasoningContent = new StringBuilder();
StringBuilder finalContent = new StringBuilder();
isFirstPrint = true;
result.blockingForEach(message -> handleGenerationResult(message, reasoningContent, finalContent));
return finalContent.toString().replaceAll("[\n\r\s]", "");
} catch (Exception e) {
log.error("调用百炼API失败: {}", e.getMessage(), e);
throw new IOException("百炼API调用失败: " + e.getMessage(), e);
}
}
/**
* 视觉推理（QVQ） 使用本地文件（输入Base64编码或本地路径）
* @param localPath 图片文件的绝对路径
* @return
*/
@Override
public String qvq4LocalPath(String localPath) throws IOException {
try {
// 构建多模态消息
String filePath = "file://"+ localPath;
log.info("filePath: {}", filePath);
MultiModalMessage userMessage = MultiModalMessage.builder().role(Role.USER.getValue())
.content(Arrays.asList(new HashMap<String, Object>(){{put("image", filePath);}},
new HashMap<String, Object>(){{put("text", "这张图片有什么");}})).build();
// 构建请求参数
MultiModalConversationParam param = MultiModalConversationParam.builder()
.apiKey(serviceKey) // 使用配置文件中的API Key
.model(serviceModel)
.message(userMessage)
.build();
MultiModalConversation conv = new MultiModalConversation();
// 调用API
Flowable<MultiModalConversationResult> result = conv.streamCall(
param);
StringBuilder reasoningContent = new StringBuilder();
StringBuilder finalContent = new StringBuilder();
isFirstPrint = true;
result.blockingForEach(message -> handleGenerationResult(message, reasoningContent, finalContent));
return finalContent.toString().replaceAll("[\n\r\s]", "");
} catch (Exception e) {
log.error("调用百炼API失败: {}", e.getMessage(), e);
throw new IOException("百炼API调用失败: " + e.getMessage(), e);
}
}
private void handleGenerationResult(MultiModalConversationResult message, StringBuilder reasoningContent, StringBuilder finalContent) {
String re = message.getOutput().getChoices().get(0).getMessage().getReasoningContent();
String reasoning = Objects.isNull(re) ? "" : re;
List<Map<String, Object>> content = message.getOutput().getChoices().get(0).getMessage()
.getContent();
if (!reasoning.isEmpty()) {
reasoningContent.append(reasoning);
}
if (Objects.nonNull(content) && !content.isEmpty()) {
Object text = content.get(0).get("text");
finalContent.append(text);
}
// 检查是否是最后一个响应
if (message.getOutput().getChoices().get(0).getFinishReason() != null) {
// 输出思考过程
if (reasoningContent.length() > 0) {
System.out.println("====================思考过程====================");
System.out.println(reasoningContent.toString());
}
// 输出完整回复
if (finalContent.length() > 0) {
System.out.println("====================完整回复====================");
System.out.println(finalContent.toString());
}
}
}
}

147
ruoyi-modules-api/ruoyi-knowledge-api/src/main/java/org/ruoyi/service/impl/PdfImageExtractServiceImpl.java
View File

@@ -1,147 +0,0 @@
package org.ruoyi.service.impl;
import java.io.IOException;
import java.util.ArrayList;
import java.util.List;
import java.util.concurrent.TimeUnit;
import lombok.Data;
import lombok.extern.slf4j.Slf4j;
import okhttp3.MediaType;
import okhttp3.MultipartBody;
import okhttp3.OkHttpClient;
import okhttp3.OkHttpClient.Builder;
import okhttp3.Request;
import okhttp3.RequestBody;
import okhttp3.Response;
import org.ruoyi.domain.PdfFileContentResult;
import org.ruoyi.service.DashscopeService;
import org.ruoyi.service.PdfImageExtractService;
import org.ruoyi.utils.ZipUtils;
import org.springframework.beans.factory.annotation.Autowired;
import org.springframework.beans.factory.annotation.Value;
import org.springframework.stereotype.Service;
import org.springframework.web.multipart.MultipartFile;
/**
* PDF图片提取服务实现类
*/
@Service
@Slf4j
@Data
public class PdfImageExtractServiceImpl implements PdfImageExtractService {
@Value("${pdf.extract.service.url}")
private String serviceUrl;
@Value("${pdf.extract.ai-api.url}")
private String aiApiUrl;
@Value("${pdf.extract.ai-api.key}")
private String aiApiKey;
private final OkHttpClient client = new Builder()
.connectTimeout(100, TimeUnit.SECONDS)
.readTimeout(150, TimeUnit.SECONDS)
.writeTimeout(150, TimeUnit.SECONDS)
.callTimeout(300, TimeUnit.SECONDS)
.build();
private static final MediaType JSON = MediaType.parse("application/json; charset=utf-8");
// @Override
public byte[] extractImages(MultipartFile pdfFile, String imageFormat, boolean allowDuplicates)
throws IOException {
// 构建multipart请求
RequestBody requestBody = new MultipartBody.Builder()
.setType(MultipartBody.FORM)
.addFormDataPart("fileInput", pdfFile.getOriginalFilename(),
RequestBody.create(MediaType.parse("application/pdf"), pdfFile.getBytes()))
.addFormDataPart("format", imageFormat)
.addFormDataPart("allowDuplicates", String.valueOf(allowDuplicates))
.build();
// 创建请求
Request request = new Request.Builder()
.url(serviceUrl + "/api/v1/misc/extract-images")
.post(requestBody)
.build();
// 执行请求
try (Response response = client.newCall(request).execute()) {
if (!response.isSuccessful()) {
throw new IOException("请求失败: " + response.code());
}
return response.body().bytes();
}
}
/**
* 处理文件内容
*
* @param unzip Base64编码的图片数组
* @return 文件内容结果列表
* @throws IOException 如果API调用过程中发生错误
*/
// @Override
public List<PdfFileContentResult> dealFileContent(String[] unzip) throws IOException {
List<PdfFileContentResult> results = new ArrayList<>();
int i = 0;
for (String base64Image : unzip) {
// 构建请求JSON
String requestJson = String.format("{"
+ "\"model\": \"gpt-4o\","
+ "\"stream\": false,"
+ "\"messages\": [{"
+ "\"role\": \"user\","
+ "\"content\": [{"
+ "\"type\": \"text\","
+ "\"text\": \"这张图片有什么\""
+ "}, {"
+ "\"type\": \"image_url\","
+ "\"image_url\": {"
+ "\"url\": \"%s\""
+ "}}"
+ "]}],"
+ "\"max_tokens\": 400"
+ "}", base64Image);
// 创建请求
Request request = new Request.Builder()
.url(aiApiUrl)
.addHeader("Authorization", "Bearer " + aiApiKey)
.post(RequestBody.create(JSON, requestJson))
.build();
// 执行请求
try {
log.info("=============call=" + ++i);
Response response = client.newCall(request).execute();
log.info("=============response=" + response);
if (!response.isSuccessful()) {
throw new IOException("API请求失败: " + response.code() + response.toString());
}
String responseBody = response.body().string();
log.info("=============responseBody=" + responseBody);
// 使用文件名（这里使用base64的前10个字符作为标识）和API返回内容创建结果对象
String filename = base64Image.substring(0, Math.min(base64Image.length(), 10));
results.add(new PdfFileContentResult(filename, responseBody));
} catch (Exception e) {
log.error(e.getMessage());
throw new RuntimeException(e);
}
}
return results;
}
// @Override
public List<PdfFileContentResult> extractImages(MultipartFile file) throws IOException {
String format = "png";
boolean allowDuplicates = true;
// 获取ZIP数据
byte[] zipData = this.extractImages(file, format, allowDuplicates);
// 解压文件并识别图片内容并返回
String[] unzip = ZipUtils.unzipForBase64(zipData);
//解析图片内容
return this.dealFileContent(unzip);
}
}