refactor(ai): 合并ai_document表到vector_store表，切换文档ID类型为字符串

- 删除旧的ai_document表及相关索引，新增vector_store表兼容Spring AI PgVectorStore结构
- 调整实体类AiDocument映射到vector_store表，使用字符串ID代替Long类型
- 修改Mapper接口及XML中所有ID相关SQL使用字符串类型，并替换表名为vector_store
- 修改服务接口与实现类，文档ID参数类型统一为字符串
- 处理文档分块时改用UUID生成chunk ID，确保唯一且格式正确
- 禁用Spring Ai PgVectorStore的自动schema初始化，使用手动创建的表结构
- 更新配置文件OpenAI模型API key及基础URL配置，支持多模型与聊天功能
- 优化日志输出，增加分块文档ID和父文档ID显示，方便调试追踪

src/main/java/cn/yinlihupo/service/ai/AiKnowledgeBaseService.java

@@ -49,12 +49,12 @@ public interface AiKnowledgeBaseService {
      *
      * @param docId 文档ID
      */
-    void processDocument(Long docId);
+    void processDocument(String docId);
 
     /**
      * 异步处理文档
      *
      * @param docId 文档ID
      */
-    void processDocumentAsync(Long docId);
+    void processDocumentAsync(String docId);
 }

src/main/java/cn/yinlihupo/service/ai/impl/AiKnowledgeBaseServiceImpl.java

@@ -55,7 +55,7 @@ public class AiKnowledgeBaseServiceImpl implements AiKnowledgeBaseService {
 
         // 4. 保存文档元数据
         AiDocument doc = new AiDocument();
-        doc.setDocId(docId);
+        doc.setId(docId);  // 设置标准UUID格式的ID
         doc.setProjectId(projectId);
         doc.setSourceType("upload");
         doc.setTitle(originalFilename);
@@ -75,7 +75,7 @@ public class AiKnowledgeBaseServiceImpl implements AiKnowledgeBaseService {
         // 5. 异步处理文档（解析、切片、向量化）
         documentProcessor.processDocumentAsync(doc.getId());
 
-        log.info("文件上传成功: {}, docId: {}", originalFilename, docId);
+        log.info("文件上传成功: {}, docId: {}", originalFilename, doc.getId());
 
         // 6. 返回VO
         return convertToVO(doc);
@@ -133,13 +133,13 @@ public class AiKnowledgeBaseServiceImpl implements AiKnowledgeBaseService {
     }
 
     @Override
-    public void processDocument(Long docId) {
+    public void processDocument(String docId) {
         documentProcessor.processDocument(docId);
     }
 
     @Override
     @Async
-    public void processDocumentAsync(Long docId) {
+    public void processDocumentAsync(String docId) {
         documentProcessor.processDocument(docId);
     }
 
@@ -197,7 +197,7 @@ public class AiKnowledgeBaseServiceImpl implements AiKnowledgeBaseService {
     private KbDocumentVO convertToVO(AiDocument doc) {
         KbDocumentVO vo = new KbDocumentVO();
         vo.setId(doc.getId());
-        vo.setDocId(doc.getDocId());
+        vo.setDocId(doc.getId());
         vo.setTitle(doc.getTitle());
         vo.setDocType(doc.getDocType());
         vo.setFileType(doc.getFileType());

src/main/java/cn/yinlihupo/service/ai/rag/DocumentProcessor.java

@@ -16,6 +16,7 @@ import java.io.InputStream;
 import java.util.Collections;
 import java.util.List;
 import java.util.Map;
+import java.util.UUID;
 import java.util.stream.Collectors;
 
 /**
@@ -41,7 +42,7 @@ public class DocumentProcessor {
      *
      * @param docId 文档ID
      */
-    public void processDocument(Long docId) {
+    public void processDocument(String docId) {
         AiDocument doc = documentMapper.selectById(docId);
         if (doc == null) {
             log.error("文档不存在: {}", docId);
@@ -91,7 +92,7 @@ public class DocumentProcessor {
      * @param docId 文档ID
      */
     @Async("documentTaskExecutor")
-    public void processDocumentAsync(Long docId) {
+    public void processDocumentAsync(String docId) {
         processDocument(docId);
     }
 
@@ -169,23 +170,25 @@ public class DocumentProcessor {
      * @param chunks    切片列表
      */
     private void storeChunks(AiDocument parentDoc, List<String> chunks) {
-        String docId = parentDoc.getDocId();
-        Long parentId = parentDoc.getId();
+        String parentId = parentDoc.getId();
 
         for (int i = 0; i < chunks.size(); i++) {
             String chunkContent = chunks.get(i);
+            // 使用UUID生成唯一的chunk ID，确保格式正确
+            String chunkId = UUID.randomUUID().toString();
 
             // 创建向量文档
             Document vectorDoc = new Document(
+                    chunkId,
                     chunkContent,
                     Map.of(
-                            "doc_id", docId.toString(),
-                            "project_id", parentDoc.getProjectId(),
-                            "timeline_node_id", parentDoc.getTimelineNodeId() != null ? parentDoc.getTimelineNodeId() : "",
+                            "project_id", parentDoc.getProjectId() != null ? parentDoc.getProjectId().toString() : "",
+                            "timeline_node_id", parentDoc.getTimelineNodeId() != null ? parentDoc.getTimelineNodeId().toString() : "",
                             "chunk_index", i,
                             "chunk_total", chunks.size(),
-                            "title", parentDoc.getTitle(),
-                            "source_type", parentDoc.getSourceType(),
+                            "chunk_parent_id", parentId,
+                            "title", parentDoc.getTitle() != null ? parentDoc.getTitle() : "",
+                            "source_type", parentDoc.getSourceType() != null ? parentDoc.getSourceType() : "",
                             "status", "active"
                     )
             );
@@ -199,7 +202,7 @@ public class DocumentProcessor {
                 documentMapper.updateById(parentDoc);
             }
 
-            log.debug("存储切片: {}/{}, docId: {}", i + 1, chunks.size(), docId);
+            log.debug("存储切片: {}/{}, parentId: {}, chunkId: {}", i + 1, chunks.size(), parentId, chunkId);
         }
     }