6

backend/app/services/doc_service.py

@@ -41,6 +41,7 @@ class DocumentService:
         kb_id: str,
         filename: str,
         content: bytes,
+        category_id: str | None = None,
     ) -> Document:
         """上传文档。
 
@@ -111,6 +112,7 @@ class DocumentService:
             doc_token_hash=doc_token_hash,
             doc_token_encrypted=doc_token_encrypted,
             doc_token_hint=doc_token_hint,
+            category_id=category_id,
         )
 
         # 更新存储路径中的 document_id
@@ -200,3 +202,87 @@ class DocumentService:
 
         processor = LocalDocumentProcessor(self._session)
         processor.process(doc.id)
+
+    def create_text(
+        self,
+        user: User,
+        kb_id: str,
+        title: str,
+        content: str,
+        content_format: str = "markdown",
+        category_id: str | None = None,
+    ) -> Document:
+        """创建文本内容文档（无需上传文件）。"""
+        # 校验 KB 归属
+        kb = self._kb_repo.get_by_id(kb_id)
+        if kb is None or kb.user_id != user.id or kb.status == "DELETED":
+            raise NotFoundError("知识库不存在。")
+
+        # 生成文档 token
+        doc_token = generate_token()
+        doc_token_hash = hash_token(doc_token)
+        doc_token_encrypted = encrypt_token(doc_token)
+        doc_token_hint = doc_token[-8:] if len(doc_token) >= 8 else doc_token
+
+        # 计算内容大小
+        content_size = len(content.encode("utf-8"))
+
+        # 配额检查
+        self._check_quota(user, content_size)
+
+        # 从内容提取摘要和关键词
+        summary = self._extract_summary(content)
+        keywords = self._extract_keywords(content)
+
+        # 创建文档记录
+        doc = Document(
+            knowledge_base_id=kb_id,
+            user_id=user.id,
+            category_id=category_id,
+            original_filename=title,
+            storage_path=None,  # 文本内容不存储文件
+            markdown_path=None,
+            content=content,
+            content_format=content_format,
+            file_size=content_size,
+            mime_type="text/markdown" if content_format == "markdown" else "text/plain",
+            file_ext=".md" if content_format == "markdown" else ".txt",
+            sha256=None,
+            doc_token_hash=doc_token_hash,
+            doc_token_encrypted=doc_token_encrypted,
+            doc_token_hint=doc_token_hint,
+            title=title,
+            content_summary=summary,
+            keywords=keywords,
+            status="READY",  # 文本内容直接就绪
+        )
+        self._session.add(doc)
+
+        # 扣减配额
+        self._deduct_quota(user, content_size)
+
+        self._session.commit()
+        return doc
+
+    def _extract_summary(self, content: str, max_len: int = 200) -> str:
+        """提取摘要。"""
+        import re
+
+        text = re.sub(r"[#*_`\[\]()>]", "", content)
+        text = re.sub(r"\s+", " ", text).strip()
+        if len(text) <= max_len:
+            return text
+        return text[:max_len] + "..."
+
+    def _extract_keywords(self, content: str, top_k: int = 10) -> str:
+        """提取关键词。"""
+        try:
+            import jieba.analyse
+            import re
+
+            text = re.sub(r"[#*_`\[\]()>]", "", content)
+            text = re.sub(r"\s+", " ", text).strip()
+            keywords = jieba.analyse.extract_tags(text, topK=top_k)
+            return ",".join(keywords)
+        except Exception:
+            return ""

backend/app/services/kb_public_service.py

@@ -1,14 +1,17 @@
 """公共知识库服务（Phase 10/11/12 统一数据来源）。
 
 HTML/MD/TXT/JSON/搜索全部通过此 Service 获取数据，不各自写查询逻辑。
+支持目录树结构和分类过滤。
 """
 
 from app.core.errors import NotFoundError
 from app.core.security import hash_token
 from app.models.document import Document
+from app.models.document_category import DocumentCategory
 from app.models.knowledge_base import KnowledgeBase
 from app.repositories.doc_repo import DocumentRepository
 from app.repositories.kb_repo import KnowledgeBaseRepository
+from sqlalchemy import func, select
 from sqlalchemy.orm import Session
 
 
@@ -26,12 +29,55 @@ class KbPublicService:
             raise NotFoundError("知识库不存在。")
         return kb
 
+    def get_category_tree(self, kb: KnowledgeBase) -> list[dict]:
+        """获取目录树（含文档数量）。"""
+        stmt = (
+            select(DocumentCategory)
+            .where(DocumentCategory.knowledge_base_id == kb.id)
+            .order_by(DocumentCategory.sort_order, DocumentCategory.name)
+        )
+        all_cats = list(self._session.scalars(stmt).all())
+
+        # 获取每个分类的文档数量
+        count_stmt = (
+            select(Document.category_id, func.count())
+            .where(
+                Document.knowledge_base_id == kb.id,
+                Document.status != "DELETED",
+            )
+            .group_by(Document.category_id)
+        )
+        doc_count_map = dict(self._session.execute(count_stmt).all())
+
+        # 构建树
+        cat_map = {}
+        for cat in all_cats:
+            cat_map[cat.id] = {
+                "id": cat.id,
+                "name": cat.name,
+                "path": cat.path or "/",
+                "is_folder": getattr(cat, 'is_folder', True),
+                "sort_order": cat.sort_order,
+                "doc_count": doc_count_map.get(cat.id, 0),
+                "children": [],
+            }
+
+        roots = []
+        for cat in all_cats:
+            node = cat_map[cat.id]
+            if cat.parent_id and cat.parent_id in cat_map:
+                cat_map[cat.parent_id]["children"].append(node)
+            else:
+                roots.append(node)
+
+        return roots
+
     def list_documents(
-        self, kb: KnowledgeBase, *, page: int = 1, page_size: int = 50
+        self, kb: KnowledgeBase, *, category_id: str | None = None, page: int = 1, page_size: int = 50
     ) -> tuple[list[Document], int]:
-        """获取知识库的文档列表（排除 DELETED）。"""
+        """获取知识库的文档列表（排除 DELETED）。支持按分类过滤。"""
         return self._doc_repo.list_by_knowledge_base(
-            kb.id, page=page, page_size=page_size
+            kb.id, category_id=category_id, page=page, page_size=page_size
         )
 
     def get_document_by_token(self, kb: KnowledgeBase, doc_token: str) -> Document:
@@ -42,34 +88,38 @@ class KbPublicService:
             raise NotFoundError("文档不存在。")
         return doc
 
-    def get_document_markdown(self, doc: Document) -> str:
-        """读取文档的 Markdown 内容。"""
-        from app.storage.local_storage import get_storage
+    def get_document_content(self, doc: Document) -> str:
+        """获取文档内容（优先直接内容，其次 Markdown 文件）。"""
+        # 直接文本内容
+        if doc.content:
+            return doc.content
 
-        if not doc.markdown_path:
-            return ""
-        storage = get_storage()
-        content = storage.read(doc.markdown_path)
-        return content.decode("utf-8")
+        # 从文件读取 Markdown
+        if doc.markdown_path:
+            from app.storage.local_storage import get_storage
+
+            storage = get_storage()
+            content = storage.read(doc.markdown_path)
+            return content.decode("utf-8")
+
+        return ""
 
     def search_documents(
         self, kb: KnowledgeBase, query: str, *, page: int = 1, page_size: int = 20
     ) -> tuple[list[dict], int]:
         """关键词搜索文档（Phase 12：SQLite LIKE / FTS5）。"""
-        # MVP 简化：使用 LIKE 搜索标题+描述+关键词
-        from sqlalchemy import func, or_, select
-
         conditions = [
             Document.knowledge_base_id == kb.id,
             Document.status == "READY",
         ]
 
         like_pattern = f"%{query}%"
-        search_condition = or_(
-            Document.title.like(like_pattern),
-            Document.description.like(like_pattern),
-            Document.keywords.like(like_pattern),
-            Document.content_summary.like(like_pattern),
+        search_condition = (
+            Document.title.like(like_pattern)
+            | Document.description.like(like_pattern)
+            | Document.keywords.like(like_pattern)
+            | Document.content_summary.like(like_pattern)
+            | Document.content.like(like_pattern)
         )
         conditions.append(search_condition)
 
@@ -87,6 +137,13 @@ class KbPublicService:
 
         results = []
         for doc in docs:
+            # 获取分类路径
+            category_path = ""
+            if doc.category_id:
+                cat = self._session.get(DocumentCategory, doc.category_id)
+                if cat:
+                    category_path = cat.path or ""
+
             results.append({
                 "id": doc.id,
                 "title": doc.title or doc.original_filename,
@@ -95,6 +152,7 @@ class KbPublicService:
                 "file_type": doc.file_ext,
                 "updated_at": doc.updated_at,
                 "url_hint": doc.doc_token_hint,
+                "category_path": category_path,
             })
 
         return results, total

backend/app/services/kb_service.py

@@ -1,6 +1,7 @@
-"""知识库服务：CRUD + Token 管理。"""
+"""知识库服务：CRUD + Token 管理 + 默认目录树。"""
 
 from app.core.errors import NotFoundError, PermissionDeniedError
+from app.models.document_category import DocumentCategory
 from app.models.knowledge_base import KnowledgeBase
 from app.models.user import User
 from app.repositories.kb_repo import KnowledgeBaseRepository
@@ -18,6 +19,7 @@ class KnowledgeBaseService:
         """创建知识库。返回 (kb, full_token)。
 
         full_token 仅此一次返回，用于构建完整 AI URL。
+        自动创建默认目录树结构。
         """
         token, token_hash, token_encrypted, token_hint = self._token_svc.create_token_pair()
         kb = self._kb_repo.create(
@@ -28,9 +30,46 @@ class KnowledgeBaseService:
             token_encrypted=token_encrypted,
             token_hint=token_hint,
         )
+        self._seed_default_categories(kb.id)
         self._session.commit()
         return kb, token
 
+    def _seed_default_categories(self, kb_id: str) -> None:
+        """创建默认目录树结构。"""
+        default_tree = [
+            ("01 公司层", ["公司基本信息", "经营理念", "四大价值", "A/M/B三态"]),
+            ("02 战略层", ["公司战略", "客户战略", "AI战略", "产品战略"]),
+            ("03 部门层", ["企划", "技术", "交付", "市场"]),
+            ("04 岗位/AI角色", []),
+            ("05 业务知识", ["价值创造", "价值传递", "价值交付", "价值支持"]),
+            ("06 资产库", ["文案", "SOP", "模板", "案例", "Prompt"]),
+        ]
+
+        for i, (folder_name, children) in enumerate(default_tree):
+            # 创建顶层文件夹
+            folder = DocumentCategory(
+                knowledge_base_id=kb_id,
+                parent_id=None,
+                name=folder_name,
+                path=f"/{folder_name}/",
+                is_folder=True,
+                sort_order=i,
+            )
+            self._session.add(folder)
+            self._session.flush()
+
+            # 创建子分类
+            for j, child_name in enumerate(children):
+                child = DocumentCategory(
+                    knowledge_base_id=kb_id,
+                    parent_id=folder.id,
+                    name=child_name,
+                    path=f"/{folder_name}/{child_name}/",
+                    is_folder=False,
+                    sort_order=j,
+                )
+                self._session.add(child)
+
     def get_or_404(self, kb_id: str, user: User) -> KnowledgeBase:
         """获取知识库，校验所有权。不存在或无权 → 404。"""
         kb = self._kb_repo.get_by_id(kb_id)