188 lines
6.6 KiB
Python
188 lines
6.6 KiB
Python
"""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 datetime import datetime
|
||
|
||
from app.core.errors import LinkExpiredError, 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
|
||
|
||
_TIME_FMT = "%Y-%m-%d %H:%M:%S"
|
||
|
||
|
||
class KbPublicService:
|
||
def __init__(self, session: Session) -> None:
|
||
self._session = session
|
||
self._kb_repo = KnowledgeBaseRepository(session)
|
||
self._doc_repo = DocumentRepository(session)
|
||
|
||
def get_kb_by_token(self, token: str) -> KnowledgeBase:
|
||
"""通过 token 获取知识库。
|
||
|
||
不存在/禁用/删除 → 404；存在但已过期 → 410（LinkExpiredError）。
|
||
"""
|
||
token_hash = hash_token(token)
|
||
kb = self._kb_repo.get_by_token_hash(token_hash)
|
||
if kb is None or not kb.enabled or kb.status == "DELETED":
|
||
raise NotFoundError("知识库不存在。")
|
||
|
||
if kb.token_expires_at:
|
||
try:
|
||
expires = datetime.strptime(kb.token_expires_at, _TIME_FMT)
|
||
except ValueError:
|
||
expires = None
|
||
if expires is not None and datetime.now() > expires:
|
||
raise LinkExpiredError()
|
||
return kb
|
||
|
||
@staticmethod
|
||
def is_expired(kb: KnowledgeBase) -> bool:
|
||
"""判断知识库链接是否已过期（管理端展示用）。"""
|
||
if not kb.token_expires_at:
|
||
return False
|
||
try:
|
||
expires = datetime.strptime(kb.token_expires_at, _TIME_FMT)
|
||
except ValueError:
|
||
return False
|
||
return datetime.now() > expires
|
||
|
||
def get_category_tree(self, kb: KnowledgeBase) -> list[dict]:
|
||
"""获取目录树（含文档数量），排除已删除目录。"""
|
||
stmt = (
|
||
select(DocumentCategory)
|
||
.where(
|
||
DocumentCategory.knowledge_base_id == kb.id,
|
||
DocumentCategory.deleted_at.is_(None),
|
||
)
|
||
.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, *, 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）。支持按分类过滤。"""
|
||
return self._doc_repo.list_by_knowledge_base(
|
||
kb.id, category_id=category_id, page=page, page_size=page_size
|
||
)
|
||
|
||
def get_document_by_token(self, kb: KnowledgeBase, doc_token: str) -> Document:
|
||
"""通过 token 获取单个文档。"""
|
||
doc_token_hash = hash_token(doc_token)
|
||
doc = self._doc_repo.get_by_doc_token_hash(doc_token_hash)
|
||
if doc is None or doc.knowledge_base_id != kb.id or doc.status != "READY":
|
||
raise NotFoundError("文档不存在。")
|
||
return doc
|
||
|
||
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
|
||
|
||
# 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）。"""
|
||
conditions = [
|
||
Document.knowledge_base_id == kb.id,
|
||
Document.status == "READY",
|
||
]
|
||
|
||
like_pattern = f"%{query}%"
|
||
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)
|
||
|
||
count_stmt = select(func.count()).select_from(Document).where(*conditions)
|
||
total = self._session.scalar(count_stmt) or 0
|
||
|
||
stmt = (
|
||
select(Document)
|
||
.where(*conditions)
|
||
.order_by(Document.created_at.desc())
|
||
.offset((page - 1) * page_size)
|
||
.limit(page_size)
|
||
)
|
||
docs = list(self._session.scalars(stmt).all())
|
||
|
||
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,
|
||
"description": doc.description,
|
||
"keywords": doc.keywords,
|
||
"file_type": doc.file_ext,
|
||
"updated_at": doc.updated_at,
|
||
"url_hint": doc.doc_token_hint,
|
||
"category_path": category_path,
|
||
})
|
||
|
||
return results, total
|