李想 / aiqasystem · 提交

转到一个项目

作者 lixiang
提交 66e4cdf7d50d231cf106d0dc755d03f9ea29c580 1 个父辈 92f3bfe6

问题库以及自定义问答

内嵌 并排对比
正在显示 11 个修改的文件 包含 1104 行增加 和 0 行删除
  1 +//package org.jeecg.modules.airag.app.config;
  2 +//
  3 +//import org.springframework.beans.factory.annotation.Qualifier;
  4 +//import org.springframework.boot.context.properties.ConfigurationProperties;
  5 +//import org.springframework.boot.jdbc.DataSourceBuilder;
  6 +//import org.springframework.context.annotation.Bean;
  7 +//import org.springframework.context.annotation.Configuration;
  8 +//import org.springframework.jdbc.core.JdbcTemplate;
  9 +//
  10 +//import javax.sql.DataSource;
  11 +//
  12 +//@Configuration
  13 +//public class PgVectorDataSourceConfig {
  14 +//
  15 +// @Bean(name = "pgVectorDataSource")
  16 +// @ConfigurationProperties(prefix = "spring.datasource.dynamic.datasource.pgvector")
  17 +// public DataSource pgVectorDataSource() {
  18 +// return DataSourceBuilder.create().build();
  19 +// }
  20 +//
  21 +// @Bean(name = "pgVectorJdbcTemplate")
  22 +// public JdbcTemplate pgVectorJdbcTemplate(
  23 +// @Qualifier("pgVectorDataSource") DataSource dataSource) {
  24 +// return new JdbcTemplate(dataSource);
  25 +// }
  26 +//}
  1 +//package org.jeecg.modules.airag.app.config;
  2 +//import dev.langchain4j.data.document.Metadata;
  3 +//import dev.langchain4j.data.embedding.Embedding;
  4 +//import dev.langchain4j.data.segment.TextSegment;
  5 +//import dev.langchain4j.store.embedding.EmbeddingMatch;
  6 +//import dev.langchain4j.store.embedding.EmbeddingSearchRequest;
  7 +//import dev.langchain4j.store.embedding.EmbeddingSearchResult;
  8 +//import dev.langchain4j.store.embedding.EmbeddingStore;
  9 +//import dev.langchain4j.store.embedding.filter.Filter;
  10 +//import dev.langchain4j.store.embedding.pgvector.PgVectorEmbeddingStore;
  11 +//import lombok.extern.log4j.Log4j2;
  12 +//import org.springframework.beans.factory.annotation.Autowired;
  13 +//import org.springframework.beans.factory.annotation.Qualifier;
  14 +//import org.springframework.jdbc.core.JdbcTemplate;
  15 +//import org.springframework.stereotype.Component;
  16 +//
  17 +//import javax.sql.DataSource;
  18 +//import java.util.ArrayList;
  19 +//import java.util.Collection;
  20 +//import java.util.List;
  21 +//import java.util.Map;
  22 +//import java.util.Collection;
  23 +//import java.util.List;
  24 +//import java.util.stream.Collectors;
  25 +//
  26 +//@Component
  27 +//@Log4j2
  28 +//public class PostgreEmbeddingStore implements EmbeddingStore<TextSegment> {
  29 +//
  30 +// @Autowired
  31 +// private PgVectorEmbeddingStore pgVectorEmbeddingStore;
  32 +//
  33 +// @Autowired
  34 +// private JdbcTemplate pgJdbcTemplate;
  35 +//
  36 +// @Autowired
  37 +// public PostgreEmbeddingStore(
  38 +// PgVectorEmbeddingStore pgVectorEmbeddingStore) {
  39 +// this.pgJdbcTemplate = pgJdbcTemplate;
  40 +// this.pgVectorEmbeddingStore = pgVectorEmbeddingStore;
  41 +// }
  42 +//
  43 +//
  44 +// @Override
  45 +// public String add(Embedding embedding) {
  46 +// return "";
  47 +// }
  48 +//
  49 +// @Override
  50 +// public void add(String id, Embedding embedding) {
  51 +//
  52 +// }
  53 +//
  54 +// @Override
  55 +// public String add(Embedding embedding, TextSegment textSegment) {
  56 +// return "";
  57 +// }
  58 +//
  59 +// @Override
  60 +// public List<String> addAll(List<Embedding> embeddings) {
  61 +// return List.of();
  62 +// }
  63 +//
  64 +// @Override
  65 +// public List<String> addAll(List<Embedding> embeddings, List<TextSegment> embedded) {
  66 +// return List.of();
  67 +// }
  68 +//
  69 +// @Override
  70 +// public void remove(String id) {
  71 +// EmbeddingStore.super.remove(id);
  72 +// }
  73 +//
  74 +// @Override
  75 +// public void removeAll(Collection<String> ids) {
  76 +// EmbeddingStore.super.removeAll(ids);
  77 +// }
  78 +//
  79 +// @Override
  80 +// public void removeAll(Filter filter) {
  81 +// EmbeddingStore.super.removeAll(filter);
  82 +// }
  83 +//
  84 +// @Override
  85 +// public void removeAll() {
  86 +// EmbeddingStore.super.removeAll();
  87 +// }
  88 +//
  89 +// @Override
  90 +// public EmbeddingSearchResult<TextSegment> search(EmbeddingSearchRequest request) {
  91 +// return EmbeddingStore.super.search(request);
  92 +// }
  93 +//
  94 +// @Override
  95 +// public List<EmbeddingMatch<TextSegment>> findRelevant(Embedding referenceEmbedding, int maxResults) {
  96 +// return findRelevant(referenceEmbedding, maxResults, 0.0);
  97 +// }
  98 +//
  99 +// @Override
  100 +// public List<EmbeddingMatch<TextSegment>> findRelevant(Embedding referenceEmbedding, int maxResults, double minScore) {
  101 +// try {
  102 +//// // 使用 PgVectorEmbeddingStore 进行查询
  103 +//// EmbeddingSearchRequest request = EmbeddingSearchRequest.builder()
  104 +//// .queryEmbedding(referenceEmbedding)
  105 +//// .maxResults(maxResults)
  106 +//// .minScore(minScore)
  107 +//// .build();
  108 +//
  109 +// // 构建带内存ID过滤的查询
  110 +// String sql = "SELECT id, content, metadata, embedding <=> ? AS distance " +
  111 +// "FROM embeddings " +
  112 +// "WHERE (1 - (embedding <=> ?)) >= ? " +
  113 +// "ORDER BY distance " +
  114 +// "LIMIT ?";
  115 +//
  116 +// List<Map<String, Object>> rows = pgJdbcTemplate.queryForList(
  117 +// sql,
  118 +// referenceEmbedding.vectorAsList(),
  119 +// referenceEmbedding.vectorAsList(),
  120 +// minScore,
  121 +// maxResults
  122 +// );
  123 +//
  124 +//
  125 +//
  126 +//// EmbeddingSearchResult<TextSegment> result = pgVectorEmbeddingStore.search(request);
  127 +////
  128 +////
  129 +// return convertToMatches(rows);
  130 +// } catch (Exception e) {
  131 +// log.error("向量查询失败", e);
  132 +// throw new RuntimeException("向量搜索失败: " + e.getMessage(), e);
  133 +// }
  134 +// }
  135 +//
  136 +// @Override
  137 +// public List<EmbeddingMatch<TextSegment>> findRelevant(Object memoryId, Embedding referenceEmbedding, int maxResults) {
  138 +// return findRelevant(memoryId, referenceEmbedding, maxResults, 0.0);
  139 +// }
  140 +//
  141 +// @Override
  142 +// public List<EmbeddingMatch<TextSegment>> findRelevant(Object memoryId, Embedding referenceEmbedding,
  143 +// int maxResults, double minScore) {
  144 +// try {
  145 +// // 构建带内存ID过滤的查询
  146 +// String sql = "SELECT id, content, metadata, embedding <=> ? AS distance " +
  147 +// "FROM embeddings " +
  148 +// "WHERE metadata->>'memory_id' = ? " +
  149 +// "AND (1 - (embedding <=> ?)) >= ? " +
  150 +// "ORDER BY distance " +
  151 +// "LIMIT ?";
  152 +//
  153 +// List<Map<String, Object>> rows = pgJdbcTemplate.queryForList(
  154 +// sql,
  155 +// referenceEmbedding.vectorAsList(),
  156 +// memoryId.toString(),
  157 +// referenceEmbedding.vectorAsList(),
  158 +// minScore,
  159 +// maxResults
  160 +// );
  161 +//
  162 +// return convertToMatches(rows);
  163 +// } catch (Exception e) {
  164 +// log.error("带内存ID的向量查询失败", e);
  165 +// throw new RuntimeException("带内存ID的向量搜索失败: " + e.getMessage(), e);
  166 +// }
  167 +// }
  168 +//
  169 +// private List<EmbeddingMatch<TextSegment>> convertToMatches(List<Map<String, Object>> rows) {
  170 +// List<EmbeddingMatch<TextSegment>> matches = new ArrayList<>();
  171 +// for (Map<String, Object> row : rows) {
  172 +// String id = row.get("id").toString();
  173 +// String content = (String) row.get("content");
  174 +//
  175 +// // 处理 Metadata
  176 +// Map<String, String> metadataMap = (Map<String, String>) row.get("metadata");
  177 +// Metadata metadata = Metadata.from(metadataMap);
  178 +//
  179 +// // 处理 Embedding 转换
  180 +// List<Float> embeddingList = (List<Float>) row.get("embedding");
  181 +// float[] embeddingArray = new float[embeddingList.size()];
  182 +// for (int i = 0; i < embeddingList.size(); i++) {
  183 +// embeddingArray[i] = embeddingList.get(i);
  184 +// }
  185 +// Embedding embedding = new Embedding(embeddingArray);
  186 +//
  187 +// double score = 1 - (double) row.get("distance");
  188 +// TextSegment textSegment = TextSegment.from(content, metadata);
  189 +//
  190 +// matches.add(new EmbeddingMatch<>(score, id, embedding, textSegment));
  191 +// }
  192 +// return matches;
  193 +// }
  194 +//}
  1 +//package org.jeecg.modules.airag.app.config;
  2 +//
  3 +//import dev.langchain4j.store.embedding.pgvector.PgVectorEmbeddingStore;
  4 +//import org.springframework.beans.factory.annotation.Value;
  5 +//import org.springframework.context.annotation.Bean;
  6 +//import org.springframework.context.annotation.Configuration;
  7 +//
  8 +//@Configuration
  9 +//public class VectorStoreConfig {
  10 +// @Value("${jeecg.ai-rag.embed-store.host}")
  11 +// private String host;
  12 +// @Value("${jeecg.ai-rag.embed-store.port}")
  13 +// private Integer port;
  14 +// @Value("${jeecg.ai-rag.embed-store.database}")
  15 +// private String database;
  16 +// @Value("${jeecg.ai-rag.embed-store.user}")
  17 +// private String user;
  18 +// @Value("${jeecg.ai-rag.embed-store.password}")
  19 +// private String password;
  20 +//// @Value("${spring.datasource.vector.url}")
  21 +//// private String url;
  22 +////
  23 +//// @Value("${spring.datasource.vector.username}")
  24 +//// private String username;
  25 +////
  26 +//// @Value("${spring.datasource.vector.password}")
  27 +//// private String password;
  28 +//
  29 +// @Bean
  30 +// public PgVectorEmbeddingStore pgVectorEmbeddingStore() {
  31 +// return PgVectorEmbeddingStore.builder()
  32 +// .host(host) // 如果使用独立参数
  33 +// .port(port)
  34 +// .user(user)
  35 +// .password(password)
  36 +// .database(database)
  37 +// .table("embeddings")
  38 +// .dimension(1536) // 根据你的模型维度设置
  39 +// .useIndex(true)
  40 +// .createTable(true)
  41 +// .build();
  42 +//
  43 +//// return PgVectorEmbeddingStore.builder()
  44 +//// .dataSource(dataSource) // 注入配置好的DataSource
  45 +//// .table("embeddings")
  46 +//// .dimension(1536)
  47 +//// .build();
  48 +// }
  49 +//}
@@ -42,6 +42,7 @@ import org.jeecg.modules.airag.app.service.IEmbeddingsService; @@ -42,6 +42,7 @@ import org.jeecg.modules.airag.app.service.IEmbeddingsService;
42 import org.jeecg.common.system.base.controller.JeecgController; 42 import org.jeecg.common.system.base.controller.JeecgController;
43 import org.jeecg.modules.airag.app.utils.AiModelUtils; 43 import org.jeecg.modules.airag.app.utils.AiModelUtils;
44 import org.jeecg.modules.airag.llm.entity.AiragModel; 44 import org.jeecg.modules.airag.llm.entity.AiragModel;
  45 +import org.jeecg.modules.airag.llm.handler.EmbeddingHandler;
45 import org.springframework.beans.factory.annotation.Autowired; 46 import org.springframework.beans.factory.annotation.Autowired;
46 import org.springframework.beans.factory.annotation.Qualifier; 47 import org.springframework.beans.factory.annotation.Qualifier;
47 import org.springframework.boot.jdbc.DataSourceBuilder; 48 import org.springframework.boot.jdbc.DataSourceBuilder;
@@ -69,6 +70,9 @@ public class EmbeddingsController { @@ -69,6 +70,9 @@ public class EmbeddingsController {
69 @Autowired 70 @Autowired
70 private AiModelUtils aiModelUtils; 71 private AiModelUtils aiModelUtils;
71 72
  73 + @Autowired
  74 + private EmbeddingHandler embeddingHandler;
  75 +
72 76
73 /** 77 /**
74 * 分页列表查询 78 * 分页列表查询
@@ -86,6 +90,13 @@ public class EmbeddingsController { @@ -86,6 +90,13 @@ public class EmbeddingsController {
86 @RequestParam(name="pageNo", defaultValue="1") Integer pageNo, 90 @RequestParam(name="pageNo", defaultValue="1") Integer pageNo,
87 @RequestParam(name="pageSize", defaultValue="10") Integer pageSize, 91 @RequestParam(name="pageSize", defaultValue="10") Integer pageSize,
88 HttpServletRequest req) throws NoSuchFieldException, IllegalAccessException, SQLException { 92 HttpServletRequest req) throws NoSuchFieldException, IllegalAccessException, SQLException {
  93 +// AiragModel airagModel = new AiragModel();
  94 +// airagModel.setId("1925730210204721154");
  95 +// airagModel.setProvider("OLLAMA");
  96 +// airagModel.setModelName("nomic-embed-text");
  97 +// airagModel.setBaseUrl("http://localhost:11434");
  98 +// EmbeddingStore<TextSegment> embedStore = embeddingHandler.getEmbedStore(airagModel);
  99 +// embeddingHandler.searchEmbedding()
89 Response<Embedding> embedding = aiModelUtils.getEmbedding("1925730210204721154", "33333"); 100 Response<Embedding> embedding = aiModelUtils.getEmbedding("1925730210204721154", "33333");
90 101
91 List<Embeddings> records = embeddingsService.findAll(); 102 List<Embeddings> records = embeddingsService.findAll();
  1 +package org.jeecg.modules.airag.app.controller;
  2 +
  3 +import org.jeecg.common.api.vo.Result;
  4 +import org.jeecg.modules.airag.app.entity.QuestionEmbedding;
  5 +import org.jeecg.modules.airag.app.service.IQuestionEmbeddingService;
  6 +import org.springframework.beans.factory.annotation.Autowired;
  7 +import org.springframework.transaction.annotation.Transactional;
  8 +import org.springframework.web.bind.annotation.*;
  9 +import org.springframework.web.multipart.MultipartFile;
  10 +
  11 +import java.util.List;
  12 +
  13 +@RestController
  14 +@RequestMapping("/question/embedding")
  15 +public class QuestionEmbeddingController {
  16 + @Autowired
  17 + private IQuestionEmbeddingService questionEmbeddingService;
  18 +
  19 + @GetMapping("/list")
  20 + public Result<List<QuestionEmbedding>> findAll() {
  21 + List<QuestionEmbedding> list = questionEmbeddingService.findAll();
  22 + return Result.OK(list);
  23 + }
  24 +
  25 + @GetMapping("/queryById")
  26 + public Result<QuestionEmbedding> findById(@RequestParam String id) {
  27 + QuestionEmbedding record = questionEmbeddingService.findById(id);
  28 + if (record == null) {
  29 + return Result.error("未找到对应数据");
  30 + }
  31 + return Result.OK(record);
  32 + }
  33 +
  34 + @PostMapping("/add")
  35 + public Result<String> insert(@RequestBody QuestionEmbedding record) {
  36 + int result = questionEmbeddingService.insert(record);
  37 + return result > 0 ? Result.OK("添加成功！") : Result.error("添加失败");
  38 + }
  39 +
  40 + @RequestMapping(value = "/edit", method = {RequestMethod.PUT,RequestMethod.POST})
  41 + public Result<String> update(@RequestBody QuestionEmbedding record) {
  42 + int result = questionEmbeddingService.update(record);
  43 + return result > 0 ? Result.OK("编辑成功！") : Result.error("编辑失败");
  44 + }
  45 +
  46 + @DeleteMapping("/delete")
  47 + public Result<String> deleteById(@RequestParam String id) {
  48 + int result = questionEmbeddingService.deleteById(id);
  49 + return result > 0 ? Result.OK("删除成功！") : Result.error("删除失败");
  50 + }
  51 +
  52 + @PostMapping("/uploadZip")
  53 + @Transactional(rollbackFor = {Exception.class})
  54 + public Result<?> uploadZip(@RequestParam("file") MultipartFile file) {
  55 + return questionEmbeddingService.processZipUpload(file);
  56 + }
  57 +
  58 +}
  1 +package org.jeecg.modules.airag.app.entity;
  2 +
  3 +import lombok.AllArgsConstructor;
  4 +import lombok.Data;
  5 +import lombok.NoArgsConstructor;
  6 +
  7 +import java.util.Map;
  8 +
  9 +@Data
  10 +@AllArgsConstructor
  11 +@NoArgsConstructor
  12 +public class QuestionEmbedding {
  13 + private String id;
  14 + private String text;
  15 + private String question;
  16 + private String answer;
  17 + private String metadata;
  18 + private float[] embedding;
  19 + private Double similarity;
  20 +}
  1 +package org.jeecg.modules.airag.app.mapper;
  2 +
  3 +import com.fasterxml.jackson.core.JsonProcessingException;
  4 +import com.fasterxml.jackson.core.type.TypeReference;
  5 +import com.fasterxml.jackson.databind.ObjectMapper;
  6 +import com.pgvector.PGvector;
  7 +import dev.langchain4j.data.embedding.Embedding;
  8 +import dev.langchain4j.model.output.Response;
  9 +import lombok.extern.slf4j.Slf4j;
  10 +import org.apache.commons.lang3.StringUtils;
  11 +import org.jeecg.modules.airag.app.entity.QuestionEmbedding;
  12 +import org.jeecg.modules.airag.app.utils.AiModelUtils;
  13 +import org.postgresql.util.PGobject;
  14 +import org.springframework.beans.factory.annotation.Autowired;
  15 +import org.springframework.stereotype.Component;
  16 +
  17 +import java.sql.*;
  18 +import java.util.*;
  19 +
  20 +@Component
  21 +@Slf4j
  22 +public class QuestionEmbeddingMapper {
  23 +
  24 + @Autowired
  25 + private AiModelUtils aiModelUtils;
  26 +
  27 + // PostgreSQL连接参数（应与项目配置一致）
  28 + private static final String URL = "jdbc:postgresql://192.168.100.103:5432/postgres";
  29 + private static final String USER = "postgres";
  30 + private static final String PASSWORD = "postgres";
  31 +
  32 + // 获取数据库连接
  33 + private Connection getConnection() throws SQLException {
  34 + return DriverManager.getConnection(URL, USER, PASSWORD);
  35 + }
  36 +
  37 + // 查询所有记录
  38 + public List<QuestionEmbedding> findAll() {
  39 + List<QuestionEmbedding> results = new ArrayList<>();
  40 + String sql = "SELECT * FROM question_embedding";
  41 +
  42 + try (Connection conn = getConnection();
  43 + PreparedStatement stmt = conn.prepareStatement(sql);
  44 + ResultSet rs = stmt.executeQuery()) {
  45 +
  46 + while (rs.next()) {
  47 + results.add(mapRowToQuestionEmbedding(rs));
  48 + }
  49 + } catch (SQLException e) {
  50 + log.error("查询所有记录失败", e);
  51 + throw new RuntimeException("查询数据时发生数据库错误", e);
  52 + }
  53 + return results;
  54 + }
  55 +
  56 + // 根据ID查询单个记录
  57 + public QuestionEmbedding findById(String id) {
  58 + String sql = "SELECT * FROM question_embedding WHERE id = ?";
  59 +
  60 + try (Connection conn = getConnection();
  61 + PreparedStatement stmt = conn.prepareStatement(sql)) {
  62 +
  63 + stmt.setString(1, id);
  64 + try (ResultSet rs = stmt.executeQuery()) {
  65 + if (rs.next()) {
  66 + return mapRowToQuestionEmbedding(rs);
  67 + }
  68 + }
  69 + } catch (SQLException e) {
  70 + log.error("根据ID查询记录失败, ID: {}", id, e);
  71 + throw new RuntimeException("根据ID查询时发生数据库错误", e);
  72 + }
  73 + return null;
  74 + }
  75 +
  76 + // 插入新记录
  77 + public int insert(QuestionEmbedding record) {
  78 + String sql = "INSERT INTO question_embedding (id, text, question, answer, metadata,embedding) VALUES (?, ?, ?, ?, ?::jsonb,?)";
  79 +
  80 + try (Connection conn = getConnection();
  81 + PreparedStatement stmt = conn.prepareStatement(sql)) {
  82 +
  83 + stmt.setString(1, UUID.randomUUID().toString());
  84 + stmt.setString(2, record.getText());
  85 + stmt.setString(3, record.getQuestion());
  86 + stmt.setString(4, record.getAnswer());
  87 + PGobject jsonObject = new PGobject();
  88 + jsonObject.setType("json");
  89 + jsonObject.setValue("{\"name\":\"John\", \"age\":30}");
  90 + stmt.setObject(5, jsonObject);
  91 + Response<Embedding> embedding = aiModelUtils.getEmbedding("1925730210204721154", record.getQuestion());
  92 + stmt.setObject(6, embedding.content().vector());
  93 + return stmt.executeUpdate();
  94 + } catch (SQLException e) {
  95 + log.error("插入记录失败: {}", record, e);
  96 + throw new RuntimeException("插入数据时发生数据库错误", e);
  97 + }
  98 + }
  99 +
  100 + // 更新记录
  101 + public int update(QuestionEmbedding record) {
  102 + String sql = "UPDATE question_embedding SET text = ?, question = ?, answer = ?, metadata = ?::jsonb ,embedding = ? WHERE id = ?";
  103 +
  104 + try (Connection conn = getConnection();
  105 + PreparedStatement stmt = conn.prepareStatement(sql)) {
  106 +
  107 + stmt.setString(1, record.getText());
  108 + stmt.setString(2, record.getQuestion());
  109 + stmt.setString(3, record.getAnswer());
  110 + PGobject jsonObject = new PGobject();
  111 + jsonObject.setType("json");
  112 + jsonObject.setValue("{\"name\":\"John\", \"age\":30}");
  113 + stmt.setObject(4, jsonObject);
  114 +
  115 + Response<Embedding> embedding = aiModelUtils.getEmbedding("1925730210204721154", record.getQuestion());
  116 + stmt.setObject(5, embedding.content().vector());
  117 +
  118 + stmt.setString(6, record.getId());
  119 +
  120 + return stmt.executeUpdate();
  121 + } catch (SQLException e) {
  122 + log.error("更新记录失败: {}", record, e);
  123 + throw new RuntimeException("更新数据时发生数据库错误", e);
  124 + }
  125 + }
  126 +
  127 +
  128 + /**
  129 + * 向量相似度查询 (基于问题文本的向量)
  130 + * @param question 问题文本
  131 + * @param limit 返回结果数量
  132 + * @param minSimilarity 最小相似度阈值(0-1)
  133 + * @return 相似问答列表（按相似度降序）
  134 + */
  135 + public List<QuestionEmbedding> similaritySearchByQuestion(String question, int limit, Double minSimilarity) {
  136 + List<QuestionEmbedding> results = new ArrayList<>();
  137 +
  138 + // 1. 参数校验
  139 + if (minSimilarity < 0 || minSimilarity > 1) {
  140 + throw new IllegalArgumentException("相似度阈值必须在0到1之间");
  141 + }
  142 +
  143 + // 2. 获取问题的嵌入向量
  144 + Response<Embedding> embeddingResponse = aiModelUtils.getEmbedding("1925730210204721154", question);
  145 + float[] queryVector = embeddingResponse.content().vector();
  146 +
  147 + // 3. 计算最大允许距离（1 - 相似度阈值）
  148 + double maxDistance = 1 - minSimilarity;
  149 +
  150 + // 4. 执行向量相似度查询
  151 + String sql = "SELECT *, embedding <-> ? AS distance " +
  152 + "FROM question_embedding " +
  153 + "WHERE embedding <-> ? < ? " + // 距离小于阈值
  154 + "ORDER BY distance ASC " + // 按距离升序
  155 + "LIMIT ?";
  156 +
  157 + try (Connection conn = getConnection();
  158 + PreparedStatement stmt = conn.prepareStatement(sql)) {
  159 +
  160 + // 设置参数
  161 + PGvector vector = new PGvector(queryVector);
  162 + stmt.setObject(1, vector);
  163 + stmt.setObject(2, vector);
  164 + stmt.setDouble(3, maxDistance);
  165 + stmt.setInt(4, limit);
  166 +
  167 + try (ResultSet rs = stmt.executeQuery()) {
  168 + while (rs.next()) {
  169 + QuestionEmbedding record = mapRowToQuestionEmbedding(rs);
  170 + // 计算相似度（1 - 距离）
  171 + double distance = rs.getDouble("distance");
  172 + double similarity = 1 - distance;
  173 + record.setSimilarity(similarity);
  174 + results.add(record);
  175 + }
  176 + }
  177 + } catch (SQLException e) {
  178 + log.error("向量相似度查询失败", e);
  179 + throw new RuntimeException("执行向量相似度查询时发生数据库错误", e);
  180 + }
  181 + return results;
  182 + }
  183 +
  184 + /**
  185 + * 向量相似度查询 (直接使用向量)
  186 + * @param vector 查询向量
  187 + * @param limit 返回结果数量
  188 + * @return 相似问答列表（按相似度降序）
  189 + */
  190 + public List<QuestionEmbedding> similaritySearch(float[] vector, int limit) {
  191 + List<QuestionEmbedding> results = new ArrayList<>();
  192 + String sql = "SELECT *, embedding <-> ? AS similarity " +
  193 + "FROM question_embedding " +
  194 + "ORDER BY similarity ASC " +
  195 + "LIMIT ?";
  196 +
  197 + try (Connection conn = getConnection();
  198 + PreparedStatement stmt = conn.prepareStatement(sql)) {
  199 +
  200 + stmt.setObject(1, new PGvector(vector));
  201 + stmt.setInt(2, limit);
  202 +
  203 + try (ResultSet rs = stmt.executeQuery()) {
  204 + while (rs.next()) {
  205 + QuestionEmbedding record = mapRowToQuestionEmbedding(rs);
  206 + double similarity = 1 - rs.getDouble("similarity");
  207 + record.setSimilarity(similarity);
  208 + results.add(record);
  209 + }
  210 + }
  211 + } catch (SQLException e) {
  212 + log.error("向量相似度查询失败", e);
  213 + throw new RuntimeException("执行向量相似度查询时发生数据库错误", e);
  214 + }
  215 + return results;
  216 + }
  217 +
  218 + // 根据ID删除记录
  219 + public int deleteById(String id) {
  220 + String sql = "DELETE FROM question_embedding WHERE id = ?";
  221 +
  222 + try (Connection conn = getConnection();
  223 + PreparedStatement stmt = conn.prepareStatement(sql)) {
  224 +
  225 + stmt.setString(1, id);
  226 + return stmt.executeUpdate();
  227 + } catch (SQLException e) {
  228 + log.error("删除记录失败, ID: {}", id, e);
  229 + throw new RuntimeException("删除数据时发生数据库错误", e);
  230 + }
  231 + }
  232 +
  233 + // 将ResultSet行映射为QuestionEmbedding对象
  234 + private QuestionEmbedding mapRowToQuestionEmbedding(ResultSet rs) throws SQLException {
  235 + QuestionEmbedding record = new QuestionEmbedding();
  236 + record.setId(rs.getString("id"));
  237 + record.setText(rs.getString("text"));
  238 + record.setQuestion(rs.getString("question"));
  239 + record.setAnswer(rs.getString("answer"));
  240 +
  241 + String metadataJson = rs.getString("metadata");
  242 + if (StringUtils.isNotBlank(metadataJson)) {
  243 + record.setMetadata("");
  244 + }
  245 +
  246 + return record;
  247 + }
  248 +
  249 + // 将Map转换为JSON字符串
  250 + private String toJson(Map<String, Object> map) {
  251 + try {
  252 + return new ObjectMapper().writeValueAsString(map);
  253 + } catch (JsonProcessingException e) {
  254 + log.error("元数据转换为JSON失败", e);
  255 + return "{}";
  256 + }
  257 + }
  258 +
  259 + // 将JSON字符串转换为Map
  260 + private Map<String, Object> fromJson(String json) {
  261 + try {
  262 + return new ObjectMapper().readValue(json, new TypeReference<Map<String, Object>>() {});
  263 + } catch (JsonProcessingException e) {
  264 + log.error("JSON转换为元数据失败", e);
  265 + return Collections.emptyMap();
  266 + }
  267 + }
  268 +}
  1 +package org.jeecg.modules.airag.app.service;
  2 +
  3 +import org.jeecg.common.api.vo.Result;
  4 +import org.jeecg.modules.airag.app.entity.QuestionEmbedding;
  5 +import org.springframework.web.multipart.MultipartFile;
  6 +
  7 +import java.util.List;
  8 +
  9 +public interface IQuestionEmbeddingService {
  10 + List<QuestionEmbedding> findAll();
  11 + QuestionEmbedding findById(String id);
  12 + int insert(QuestionEmbedding record);
  13 + int update(QuestionEmbedding record);
  14 + int deleteById(String id);
  15 + List<QuestionEmbedding> similaritySearchByQuestion(String question, int limit, Double minSimilarity);
  16 + List<QuestionEmbedding> similaritySearch(float[] vector, int limit);
  17 +
  18 + Result<?> processZipUpload(MultipartFile file);
  19 +}
  1 +package org.jeecg.modules.airag.app.service.impl;
  2 +
  3 +import org.apache.poi.hwpf.usermodel.CharacterRun;
  4 +import org.apache.poi.hwpf.HWPFDocument;
  5 +import org.apache.poi.hwpf.usermodel.Paragraph;
  6 +import org.apache.poi.hwpf.usermodel.Range;
  7 +import dev.langchain4j.data.document.Document;
  8 +import dev.langchain4j.data.document.DocumentSplitter;
  9 +import dev.langchain4j.data.document.loader.FileSystemDocumentLoader;
  10 +import dev.langchain4j.data.document.parser.TextDocumentParser;
  11 +import dev.langchain4j.data.document.splitter.DocumentByParagraphSplitter;
  12 +import dev.langchain4j.data.document.splitter.DocumentSplitters;
  13 +import dev.langchain4j.data.embedding.Embedding;
  14 +import dev.langchain4j.data.segment.TextSegment;
  15 +import dev.langchain4j.model.embedding.EmbeddingModel;
  16 +import dev.langchain4j.model.output.Response;
  17 +import org.apache.commons.io.FilenameUtils;
  18 +import org.apache.poi.xwpf.usermodel.IBodyElement;
  19 +import org.apache.poi.xwpf.usermodel.XWPFDocument;
  20 +import org.apache.poi.xwpf.usermodel.XWPFParagraph;
  21 +import org.apache.poi.xwpf.usermodel.XWPFTable;
  22 +import org.jeecg.common.api.vo.Result;
  23 +import org.jeecg.common.util.CommonUtils;
  24 +import org.jeecg.modules.airag.app.entity.QuestionEmbedding;
  25 +import org.jeecg.modules.airag.app.mapper.QuestionEmbeddingMapper;
  26 +import org.jeecg.modules.airag.app.service.IQuestionEmbeddingService;
  27 +import org.jeecg.modules.airag.app.utils.AiModelUtils;
  28 +import org.jeecg.modules.airag.common.handler.IAIChatHandler;
  29 +import org.slf4j.Logger;
  30 +import org.slf4j.LoggerFactory;
  31 +import org.springframework.beans.factory.annotation.Autowired;
  32 +import org.springframework.beans.factory.annotation.Value;
  33 +import org.springframework.stereotype.Service;
  34 +import org.springframework.web.multipart.MultipartFile;
  35 +
  36 +import java.io.File;
  37 +import java.io.FileInputStream;
  38 +import java.io.IOException;
  39 +import java.nio.file.Files;
  40 +import java.nio.file.Path;
  41 +import java.nio.file.Paths;
  42 +import java.nio.file.StandardCopyOption;
  43 +import java.util.*;
  44 +import java.util.regex.Pattern;
  45 +import java.util.stream.Collectors;
  46 +import java.util.zip.ZipEntry;
  47 +import java.util.zip.ZipInputStream;
  48 +
  49 +@Service
  50 +public class QuestionEmbeddingServiceImpl implements IQuestionEmbeddingService {
  51 +
  52 + private static final Logger log = LoggerFactory.getLogger(QuestionEmbeddingServiceImpl.class);
  53 +
  54 + @Autowired
  55 + private QuestionEmbeddingMapper questionEmbeddingMapper;
  56 +
  57 + @Autowired
  58 + private AiModelUtils aiModelUtils;
  59 +
  60 + @Autowired
  61 + private IAIChatHandler aiChatHandler;
  62 +
  63 + @Value("${jeecg.upload.path}")
  64 + private String uploadPath;
  65 +
  66 + private static final Set<String> ALLOWED_EXTENSIONS = Set.of("txt", "doc", "docx");
  67 + private static final Pattern SPECIAL_CHARS_PATTERN = Pattern.compile("[^a-zA-Z0-9\\u4e00-\\u9fa5\\s]");
  68 +
  69 + @Override
  70 + public List<QuestionEmbedding> findAll() {
  71 + return questionEmbeddingMapper.findAll();
  72 + }
  73 +
  74 + @Override
  75 + public QuestionEmbedding findById(String id) {
  76 + return questionEmbeddingMapper.findById(id);
  77 + }
  78 +
  79 + @Override
  80 + public int insert(QuestionEmbedding record) {
  81 + return questionEmbeddingMapper.insert(record);
  82 + }
  83 +
  84 + @Override
  85 + public int update(QuestionEmbedding record) {
  86 + return questionEmbeddingMapper.update(record);
  87 + }
  88 +
  89 + @Override
  90 + public int deleteById(String id) {
  91 + return questionEmbeddingMapper.deleteById(id);
  92 + }
  93 +
  94 + @Override
  95 + public List<QuestionEmbedding> similaritySearchByQuestion(String question, int limit, Double minSimilarity) {
  96 + return questionEmbeddingMapper.similaritySearchByQuestion(question, limit, minSimilarity);
  97 + }
  98 +
  99 + @Override
  100 + public List<QuestionEmbedding> similaritySearch(float[] vector, int limit) {
  101 + return questionEmbeddingMapper.similaritySearch(vector, limit);
  102 + }
  103 +
  104 + public Result<?> processZipUpload(MultipartFile zipFile) {
  105 + try {
  106 + Path tempDir = Files.createTempDirectory("zip_upload_");
  107 + List<Path> validFiles = extractAndFilterZip(zipFile, tempDir);
  108 +
  109 + if (validFiles.isEmpty()) {
  110 + return Result.error("ZIP文件中没有有效的TXT或Word文档");
  111 + }
  112 +
  113 + for (Path filePath : validFiles) {
  114 + processSingleFile(filePath);
  115 + }
  116 +
  117 + return Result.OK("文件上传和处理成功");
  118 + } catch (Exception e) {
  119 + log.error("处理ZIP文件上传失败", e);
  120 + return Result.error("处理ZIP文件失败: " + e.getMessage());
  121 + }
  122 + }
  123 +
  124 + private List<Path> extractAndFilterZip(MultipartFile zipFile, Path tempDir) throws IOException {
  125 + List<Path> validFiles = new ArrayList<>();
  126 +
  127 + try (ZipInputStream zipIn = new ZipInputStream(zipFile.getInputStream())) {
  128 + ZipEntry entry;
  129 + while ((entry = zipIn.getNextEntry()) != null) {
  130 + if (!entry.isDirectory()) {
  131 + String fileName = entry.getName();
  132 + String ext = FilenameUtils.getExtension(fileName).toLowerCase();
  133 +
  134 + if (ALLOWED_EXTENSIONS.contains(ext)) {
  135 + String safeFileName = new File(fileName).getName();
  136 + Path outputPath = tempDir.resolve(safeFileName);
  137 + Files.copy(zipIn, outputPath, StandardCopyOption.REPLACE_EXISTING);
  138 + validFiles.add(outputPath);
  139 + }
  140 + }
  141 + zipIn.closeEntry();
  142 + }
  143 + }
  144 + return validFiles;
  145 + }
  146 +
  147 + private void processSingleFile(Path filePath) throws Exception {
  148 + String originalFileName = filePath.getFileName().toString();
  149 + String fileExt = FilenameUtils.getExtension(originalFileName);
  150 + String newFileName = FilenameUtils.removeExtension(originalFileName) + "_" + UUID.randomUUID() + "." + fileExt;
  151 + Path targetPath = Paths.get(uploadPath, newFileName);
  152 + Files.move(filePath, targetPath, StandardCopyOption.REPLACE_EXISTING);
  153 +
  154 + List<String> segments;
  155 + if (fileExt.equalsIgnoreCase("txt")) {
  156 + String fileContent = readFileContent(targetPath);
  157 + String cleanedContent = cleanText(fileContent);
  158 + segments = splitTxtDocument(cleanedContent);
  159 + } else {
  160 + segments = splitWordDocument(targetPath.toString());
  161 + }
  162 +
  163 + saveSegmentsToDatabase(segments, originalFileName, newFileName);
  164 + }
  165 +
  166 + private String readFileContent(Path filePath) throws IOException {
  167 + return new String(Files.readAllBytes(filePath));
  168 + }
  169 +
  170 + private String cleanText(String text) {
  171 + text = SPECIAL_CHARS_PATTERN.matcher(text).replaceAll("");
  172 + return text.replaceAll("\\s+", " ").trim();
  173 + }
  174 +
  175 + private List<String> splitTxtDocument(String content) {
  176 + DocumentSplitter splitter = new DocumentByParagraphSplitter(1000, 200);
  177 + Document document = Document.from(content);
  178 + return splitter.split(document).stream()
  179 + .map(TextSegment::text)
  180 + .map(this::cleanText)
  181 + .collect(Collectors.toList());
  182 + }
  183 +
  184 + public List<String> splitWordDocument(String filePath) throws Exception {
  185 + List<String> result = new ArrayList<>();
  186 + String ext = FilenameUtils.getExtension(filePath).toLowerCase();
  187 + StringBuilder fullContent = new StringBuilder();
  188 + String fileName = new File(filePath).getName();
  189 + fileName = fileName.substring(0, fileName.lastIndexOf('.')); // 去掉后缀
  190 +
  191 + if (ext.equals("docx")) {
  192 + try (XWPFDocument doc = new XWPFDocument(new FileInputStream(filePath))) {
  193 + StringBuilder currentSection = new StringBuilder();
  194 + boolean isTableSection = false;
  195 +
  196 + for (IBodyElement element : doc.getBodyElements()) {
  197 + if (element instanceof XWPFParagraph) {
  198 + XWPFParagraph para = (XWPFParagraph) element;
  199 + String text = cleanText(para.getText());
  200 + fullContent.append(text).append("\n");
  201 +
  202 + if (isTableSection) {
  203 + result.add(currentSection.toString().trim());
  204 + currentSection = new StringBuilder();
  205 + isTableSection = false;
  206 + }
  207 +
  208 + String style = para.getStyle();
  209 + if (style != null && style.matches("Heading\\d")) {
  210 + if (currentSection.length() > 0) {
  211 + result.add(currentSection.toString().trim());
  212 + }
  213 + currentSection = new StringBuilder(text).append("\n");
  214 + } else {
  215 + currentSection.append(text).append("\n");
  216 + }
  217 + } else if (element instanceof XWPFTable) {
  218 + String tableContent = extractTableContent((XWPFTable) element);
  219 + fullContent.append(tableContent).append("\n");
  220 +
  221 + if (!isTableSection) {
  222 + if (currentSection.length() > 0) {
  223 + result.add(currentSection.toString().trim());
  224 + }
  225 + currentSection = new StringBuilder();
  226 + isTableSection = true;
  227 + }
  228 + currentSection.append(tableContent).append("\n");
  229 + }
  230 + }
  231 +
  232 + if (currentSection.length() > 0) {
  233 + result.add(currentSection.toString().trim());
  234 + }
  235 + }
  236 + } else if (ext.equals("doc")) {
  237 + try (HWPFDocument doc = new HWPFDocument(new FileInputStream(filePath))) {
  238 + Range range = doc.getRange();
  239 + StringBuilder currentSection = new StringBuilder();
  240 + boolean isTableSection = false;
  241 +
  242 + for (int i = 0; i < range.numParagraphs(); i++) {
  243 + Paragraph para = range.getParagraph(i);
  244 + String text = cleanText(para.text());
  245 + fullContent.append(text).append("\n");
  246 +
  247 + if (para.isInTable()) {
  248 + if (!isTableSection) {
  249 + if (currentSection.length() > 0) {
  250 + result.add(currentSection.toString().trim());
  251 + }
  252 + currentSection = new StringBuilder();
  253 + isTableSection = true;
  254 + }
  255 + currentSection.append(text).append("\n");
  256 + } else {
  257 + if (isTableSection) {
  258 + result.add(currentSection.toString().trim());
  259 + currentSection = new StringBuilder();
  260 + isTableSection = false;
  261 + }
  262 +
  263 + if (isHeading(para, range)) {
  264 + if (currentSection.length() > 0) {
  265 + result.add(currentSection.toString().trim());
  266 + }
  267 + currentSection = new StringBuilder(text).append("\n");
  268 + } else {
  269 + currentSection.append(text).append("\n");
  270 + }
  271 + }
  272 + }
  273 +
  274 + if (currentSection.length() > 0) {
  275 + result.add(currentSection.toString().trim());
  276 + }
  277 + }
  278 + }
  279 +
  280 + if (fullContent.length() < 1000) {
  281 + return Collections.singletonList(fileName + "\n" + fullContent.toString().trim());
  282 + }
  283 +
  284 + return result;
  285 + }
  286 +
  287 + private String extractTableContent(XWPFTable table) {
  288 + StringBuilder tableContent = new StringBuilder();
  289 + table.getRows().forEach(row -> {
  290 + row.getTableCells().forEach(cell -> {
  291 + tableContent.append("| ").append(cleanText(cell.getText())).append(" ");
  292 + });
  293 + tableContent.append("|\n");
  294 + });
  295 + return tableContent.toString();
  296 + }
  297 +
  298 + private static boolean isHeading(Paragraph para, Range range) {
  299 + int styleIndex = para.getStyleIndex();
  300 + if (styleIndex >= 1 && styleIndex <= 9) {
  301 + return true;
  302 + }
  303 +
  304 + try {
  305 + CharacterRun run = para.getCharacterRun(0);
  306 + if (run.isBold() || run.getFontSize() > 12) {
  307 + return true;
  308 + }
  309 + } catch (Exception e) {
  310 + log.warn("获取字符格式失败", e);
  311 + }
  312 +
  313 + String text = para.text().trim();
  314 + return text.toUpperCase().equals(text) &&
  315 + text.length() < 100 &&
  316 + !text.contains(".") &&
  317 + !text.contains("\t");
  318 + }
  319 +
  320 + private void saveSegmentsToDatabase(List<String> segments, String originalFileName, String storedFileName) {
  321 + if (segments.isEmpty()) {
  322 + return;
  323 + }
  324 +
  325 + String fileNameWithoutExt = originalFileName.substring(0, originalFileName.lastIndexOf('.'));
  326 + String question = segments.size() == 1 ? fileNameWithoutExt : null;
  327 +
  328 + for (String segment : segments) {
  329 + if (segment.trim().isEmpty()) {
  330 + continue;
  331 + }
  332 +
  333 + QuestionEmbedding record = new QuestionEmbedding();
  334 + record.setId(UUID.randomUUID().toString());
  335 +
  336 + if (question != null) {
  337 + record.setQuestion(question);
  338 + } else {
  339 + String firstLine = segment.lines().findFirst().orElse("未命名问题");
  340 + record.setQuestion(cleanText(firstLine));
  341 + }
  342 +
  343 + record.setAnswer(segment.trim());
  344 + record.setText("");
  345 + record.setMetadata("{\"fileName\":\"" + storedFileName + "\"}");
  346 +
  347 + Response<Embedding> embeddingResponse = aiModelUtils.getEmbedding("1925730210204721154", record.getQuestion());
  348 + record.setEmbedding(embeddingResponse.content().vector());
  349 +
  350 + questionEmbeddingMapper.insert(record);
  351 + }
  352 + }
  353 +}
  1 +package org.jeecg.modules.airag.zdyrag.controller;
  2 +
  3 +import dev.langchain4j.data.message.ChatMessage;
  4 +import dev.langchain4j.data.message.UserMessage;
  5 +import dev.langchain4j.service.TokenStream;
  6 +import dev.langchain4j.store.embedding.pgvector.PgVectorEmbeddingStore;
  7 +import io.swagger.v3.oas.annotations.Operation;
  8 +import org.jeecg.ai.handler.AIParams;
  9 +import org.jeecg.ai.handler.LLMHandler;
  10 +import org.jeecg.common.api.vo.Result;
  11 +import org.jeecg.modules.airag.app.entity.QuestionEmbedding;
  12 +import org.jeecg.modules.airag.app.service.IQuestionEmbeddingService;
  13 +import org.jeecg.modules.airag.common.handler.IAIChatHandler;
  14 +import org.jeecg.modules.airag.llm.handler.EmbeddingHandler;
  15 +import org.springframework.beans.factory.annotation.Autowired;
  16 +import org.springframework.stereotype.Component;
  17 +import org.springframework.web.bind.annotation.GetMapping;
  18 +import org.springframework.web.bind.annotation.RequestMapping;
  19 +import org.springframework.web.bind.annotation.RestController;
  20 +
  21 +import java.util.ArrayList;
  22 +import java.util.HashMap;
  23 +import java.util.List;
  24 +import java.util.Map;
  25 +
  26 +@RestController
  27 +@RequestMapping("/airag/zdyRag")
  28 +public class ZdyRagController {
  29 + @Autowired
  30 + private EmbeddingHandler embeddingHandler;
  31 + @Autowired
  32 + IAIChatHandler aiChatHandler;
  33 + @Autowired
  34 + LLMHandler llmHandler;
  35 + @Autowired
  36 + private IQuestionEmbeddingService questionEmbeddingService;
  37 +
  38 +
  39 + @Operation(summary = "send")
  40 + @GetMapping("send")
  41 + public Result<Map<String, Object>> send(String questionText) {
  42 + String knowId = "1926872137990148098";
  43 +// String text = "身份证丢失办理流程";
  44 + Integer topNumber = 3;
  45 + Double similarity = 0.8;
  46 + HashMap<String, Object> resMap = new HashMap<>();
  47 + //根据问题相似度进行查询
  48 + List<QuestionEmbedding> questionEmbeddings = questionEmbeddingService.similaritySearchByQuestion(questionText, 1,0.8);
  49 + for (QuestionEmbedding questionEmbedding : questionEmbeddings) {
  50 + resMap.put("question", questionEmbedding.getQuestion());
  51 + resMap.put("answer", questionEmbedding.getAnswer());
  52 + resMap.put("similarity", similarity);
  53 + System.out.println("questionEmbedding.getQuestion() = " + questionEmbedding.getQuestion());
  54 + System.out.println("questionEmbedding.getAnswer() = " + questionEmbedding.getAnswer());
  55 + System.out.println("questionEmbedding.getSimilarity() = " + questionEmbedding.getSimilarity());
  56 + System.out.println("-------------------------------------------------------------");
  57 + }
  58 + //返回问题库命中的问题
  59 + if (!questionEmbeddings.isEmpty()) {
  60 + return Result.OK(resMap);
  61 + }
  62 +
  63 + List<Map<String, Object>> maps = embeddingHandler.searchEmbedding(knowId, questionText, topNumber, similarity);
  64 + StringBuilder content = new StringBuilder();
  65 + for (Map<String, Object> map : maps) {
  66 + if (Double.parseDouble(map.get("score").toString()) > similarity){
  67 + System.out.println("score = " + map.get("score").toString());
  68 + System.out.println("content = " + map.get("content").toString());
  69 + content.append(map.get("content").toString()).append("\n");
  70 + }
  71 + }
  72 +
  73 + List<ChatMessage> messages = new ArrayList<>();
  74 + String questin = "请整理出与用户所提出的问题相关的信息，舍弃掉与问题无关的内容，进行整理，回答用户的问题" +
  75 + "问题如下：" + questionText +
  76 + "文本信息如下：" + content
  77 + ;
  78 +
  79 +
  80 + messages.add(new UserMessage("user", questin));
  81 +// AIParams aiParams = new AIParams();
  82 +// aiParams.setBaseUrl("http://localhost:11434");
  83 +// aiParams.setModelName("EntropyYue/chatglm3");
  84 +// aiParams.setProvider("OLLAMA");
  85 + String chat = aiChatHandler.completions("1926875898187878401", messages, null);
  86 + resMap.put("question", questionText);
  87 + resMap.put("answer", chat);
  88 + return Result.OK(resMap);
  89 + }
  90 +
  91 + public static void main(String[] args) {
  92 + List<ChatMessage> messages = new ArrayList<>();
  93 + messages.add(new UserMessage("user", "你好，你是谁？"));
  94 + LLMHandler llmHandler1 = new LLMHandler();
  95 + AIParams aiParams = new AIParams();
  96 + aiParams.setBaseUrl("http://localhost:11434");
  97 + aiParams.setModelName("EntropyYue/chatglm3");
  98 + aiParams.setProvider("OLLAMA");
  99 + TokenStream chat = llmHandler1.chat(messages, aiParams);
  100 + System.out.println("chat = " + chat);
  101 +
  102 + }
  103 +
  104 +}
@@ -188,6 +188,8 @@ mybatis-plus: @@ -188,6 +188,8 @@ mybatis-plus:
188 minidao: 188 minidao:
189 base-package: org.jeecg.modules.jmreport.*,org.jeecg.modules.drag.* 189 base-package: org.jeecg.modules.jmreport.*,org.jeecg.modules.drag.*
190 jeecg: 190 jeecg:
  191 + upload:
  192 + path: D:\\upload\\
191 # AI集成 193 # AI集成
192 ai-chat: 194 ai-chat:
193 enabled: true 195 enabled: true