AstrBot 4.3.5__py3-none-any.whl → 4.5.0__py3-none-any.whl

astrbot/core/file_token_service.py

@@ -23,7 +23,12 @@ class FileTokenService:
         for token in expired_tokens:
             self.staged_files.pop(token, None)
 
-    async def register_file(self, file_path: str, timeout: float = None) -> str:
+    async def check_token_expired(self, file_token: str) -> bool:
+        async with self.lock:
+            await self._cleanup_expired_tokens()
+            return file_token not in self.staged_files
+
+    async def register_file(self, file_path: str, timeout: float | None = None) -> str:
         """向令牌服务注册一个文件。
 
         Args:

astrbot/core/initial_loader.py

@@ -41,10 +41,13 @@ class InitialLoader:
         self.dashboard_server = AstrBotDashboard(
             core_lifecycle, self.db, core_lifecycle.dashboard_shutdown_event, webui_dir
         )
-        task = asyncio.gather(
-            core_task, self.dashboard_server.run()
-        )  # 启动核心任务和仪表板服务器
 
+        coro = self.dashboard_server.run()
+        if coro:
+            # 启动核心任务和仪表板服务器
+            task = asyncio.gather(core_task, coro)
+        else:
+            task = core_task
         try:
             await task  # 整个AstrBot在这里运行
         except asyncio.CancelledError:

astrbot/core/knowledge_base/chunking/__init__.py

@@ -0,0 +1,11 @@
+"""
+文档分块模块
+"""
+
+from .base import BaseChunker
+from .fixed_size import FixedSizeChunker
+
+__all__ = [
+    "BaseChunker",
+    "FixedSizeChunker",
+]

astrbot/core/knowledge_base/chunking/base.py

@@ -0,0 +1,24 @@
+"""文档分块器基类
+
+定义了文档分块处理的抽象接口。
+"""
+
+from abc import ABC, abstractmethod
+
+
+class BaseChunker(ABC):
+    """分块器基类
+
+    所有分块器都应该继承此类并实现 chunk 方法。
+    """
+
+    @abstractmethod
+    async def chunk(self, text: str, **kwargs) -> list[str]:
+        """将文本分块
+
+        Args:
+            text: 输入文本
+
+        Returns:
+            list[str]: 分块后的文本列表
+        """

astrbot/core/knowledge_base/chunking/fixed_size.py

@@ -0,0 +1,57 @@
+"""固定大小分块器
+
+按照固定的字符数将文本分块,支持重叠区域。
+"""
+
+from .base import BaseChunker
+
+
+class FixedSizeChunker(BaseChunker):
+    """固定大小分块器
+
+    按照固定的字符数分块,并支持块之间的重叠。
+    """
+
+    def __init__(self, chunk_size: int = 512, chunk_overlap: int = 50):
+        """初始化分块器
+
+        Args:
+            chunk_size: 块的大小(字符数)
+            chunk_overlap: 块之间的重叠字符数
+        """
+        self.chunk_size = chunk_size
+        self.chunk_overlap = chunk_overlap
+
+    async def chunk(self, text: str, **kwargs) -> list[str]:
+        """固定大小分块
+
+        Args:
+            text: 输入文本
+            chunk_size: 每个文本块的最大大小
+            chunk_overlap: 每个文本块之间的重叠部分大小
+
+        Returns:
+            list[str]: 分块后的文本列表
+        """
+        chunk_size = kwargs.get("chunk_size", self.chunk_size)
+        chunk_overlap = kwargs.get("chunk_overlap", self.chunk_overlap)
+
+        chunks = []
+        start = 0
+        text_len = len(text)
+
+        while start < text_len:
+            end = start + chunk_size
+            chunk = text[start:end]
+
+            if chunk:
+                chunks.append(chunk)
+
+            # 移动窗口,保留重叠部分
+            start = end - chunk_overlap
+
+            # 防止无限循环: 如果重叠过大,直接移到end
+            if start >= end or chunk_overlap >= chunk_size:
+                start = end
+
+        return chunks

astrbot/core/knowledge_base/chunking/recursive.py

@@ -0,0 +1,155 @@
+from collections.abc import Callable
+from .base import BaseChunker
+
+
+class RecursiveCharacterChunker(BaseChunker):
+    def __init__(
+        self,
+        chunk_size: int = 500,
+        chunk_overlap: int = 100,
+        length_function: Callable[[str], int] = len,
+        is_separator_regex: bool = False,
+        separators: list[str] | None = None,
+    ):
+        """
+        初始化递归字符文本分割器
+
+        Args:
+            chunk_size: 每个文本块的最大大小
+            chunk_overlap: 每个文本块之间的重叠部分大小
+            length_function: 计算文本长度的函数
+            is_separator_regex: 分隔符是否为正则表达式
+            separators: 用于分割文本的分隔符列表，按优先级排序
+        """
+        self.chunk_size = chunk_size
+        self.chunk_overlap = chunk_overlap
+        self.length_function = length_function
+        self.is_separator_regex = is_separator_regex
+
+        # 默认分隔符列表，按优先级从高到低
+        self.separators = separators or [
+            "\n\n",  # 段落
+            "\n",  # 换行
+            "。",  # 中文句子
+            "，",  # 中文逗号
+            ". ",  # 句子
+            ", ",  # 逗号分隔
+            " ",  # 单词
+            "",  # 字符
+        ]
+
+    async def chunk(self, text: str, **kwargs) -> list[str]:
+        """
+        递归地将文本分割成块
+
+        Args:
+            text: 要分割的文本
+            chunk_size: 每个文本块的最大大小
+            chunk_overlap: 每个文本块之间的重叠部分大小
+
+        Returns:
+            分割后的文本块列表
+        """
+        if not text:
+            return []
+
+        overlap = kwargs.get("chunk_overlap", self.chunk_overlap)
+        chunk_size = kwargs.get("chunk_size", self.chunk_size)
+
+        text_length = self.length_function(text)
+        if text_length <= chunk_size:
+            return [text]
+
+        for separator in self.separators:
+            if separator == "":
+                return self._split_by_character(text, chunk_size, overlap)
+
+            if separator in text:
+                splits = text.split(separator)
+                # 重新添加分隔符（除了最后一个片段）
+                splits = [s + separator for s in splits[:-1]] + [splits[-1]]
+                splits = [s for s in splits if s]
+                if len(splits) == 1:
+                    continue
+
+                # 递归合并分割后的文本块
+                final_chunks = []
+                current_chunk = []
+                current_chunk_length = 0
+
+                for split in splits:
+                    split_length = self.length_function(split)
+
+                    # 如果单个分割部分已经超过了chunk_size，需要递归分割
+                    if split_length > chunk_size:
+                        # 先处理当前积累的块
+                        if current_chunk:
+                            combined_text = "".join(current_chunk)
+                            final_chunks.extend(
+                                await self.chunk(
+                                    combined_text,
+                                    chunk_size=chunk_size,
+                                    chunk_overlap=overlap,
+                                )
+                            )
+                            current_chunk = []
+                            current_chunk_length = 0
+
+                        # 递归分割过大的部分
+                        final_chunks.extend(
+                            await self.chunk(
+                                split, chunk_size=chunk_size, chunk_overlap=overlap
+                            )
+                        )
+                    # 如果添加这部分会使当前块超过chunk_size
+                    elif current_chunk_length + split_length > chunk_size:
+                        # 合并当前块并添加到结果中
+                        combined_text = "".join(current_chunk)
+                        final_chunks.append(combined_text)
+
+                        # 处理重叠部分
+                        overlap_start = max(0, len(combined_text) - overlap)
+                        if overlap_start > 0:
+                            overlap_text = combined_text[overlap_start:]
+                            current_chunk = [overlap_text, split]
+                            current_chunk_length = (
+                                self.length_function(overlap_text) + split_length
+                            )
+                        else:
+                            current_chunk = [split]
+                            current_chunk_length = split_length
+                    else:
+                        # 添加到当前块
+                        current_chunk.append(split)
+                        current_chunk_length += split_length
+
+                # 处理剩余的块
+                if current_chunk:
+                    final_chunks.append("".join(current_chunk))
+
+                return final_chunks
+
+        return [text]
+
+    def _split_by_character(
+        self, text: str, chunk_size: int | None = None, overlap: int | None = None
+    ) -> list[str]:
+        """
+        按字符级别分割文本
+
+        Args:
+            text: 要分割的文本
+
+        Returns:
+            分割后的文本块列表
+        """
+        chunk_size = chunk_size or self.chunk_size
+        overlap = overlap or self.chunk_overlap
+        result = []
+        for i in range(0, len(text), chunk_size - overlap):
+            end = min(i + chunk_size, len(text))
+            result.append(text[i:end])
+            if end == len(text):
+                break
+
+        return result

astrbot/core/knowledge_base/kb_db_sqlite.py

@@ -0,0 +1,299 @@
+from contextlib import asynccontextmanager
+from pathlib import Path
+
+from sqlmodel import col, desc
+from sqlalchemy import text, func, select, update, delete
+from sqlalchemy.ext.asyncio import AsyncSession, async_sessionmaker, create_async_engine
+
+from astrbot.core import logger
+from astrbot.core.knowledge_base.models import (
+    BaseKBModel,
+    KBDocument,
+    KBMedia,
+    KnowledgeBase,
+)
+from astrbot.core.db.vec_db.faiss_impl import FaissVecDB
+
+
+class KBSQLiteDatabase:
+    def __init__(self, db_path: str = "data/knowledge_base/kb.db") -> None:
+        """初始化知识库数据库
+
+        Args:
+            db_path: 数据库文件路径, 默认为 data/knowledge_base/kb.db
+        """
+        self.db_path = db_path
+        self.DATABASE_URL = f"sqlite+aiosqlite:///{db_path}"
+        self.inited = False
+
+        # 确保目录存在
+        Path(db_path).parent.mkdir(parents=True, exist_ok=True)
+
+        # 创建异步引擎
+        self.engine = create_async_engine(
+            self.DATABASE_URL,
+            echo=False,
+            pool_pre_ping=True,
+            pool_recycle=3600,
+        )
+
+        # 创建会话工厂
+        self.async_session = async_sessionmaker(
+            self.engine,
+            class_=AsyncSession,
+            expire_on_commit=False,
+        )
+
+    @asynccontextmanager
+    async def get_db(self):
+        """获取数据库会话
+
+        用法:
+            async with kb_db.get_db() as session:
+                # 执行数据库操作
+                result = await session.execute(stmt)
+        """
+        async with self.async_session() as session:
+            yield session
+
+    async def initialize(self) -> None:
+        """初始化数据库,创建表并配置 SQLite 参数"""
+        async with self.engine.begin() as conn:
+            # 创建所有知识库相关表
+            await conn.run_sync(BaseKBModel.metadata.create_all)
+
+            # 配置 SQLite 性能优化参数
+            await conn.execute(text("PRAGMA journal_mode=WAL"))
+            await conn.execute(text("PRAGMA synchronous=NORMAL"))
+            await conn.execute(text("PRAGMA cache_size=20000"))
+            await conn.execute(text("PRAGMA temp_store=MEMORY"))
+            await conn.execute(text("PRAGMA mmap_size=134217728"))
+            await conn.execute(text("PRAGMA optimize"))
+            await conn.commit()
+
+        self.inited = True
+
+    async def migrate_to_v1(self) -> None:
+        """执行知识库数据库 v1 迁移
+
+        创建所有必要的索引以优化查询性能
+        """
+        async with self.get_db() as session:
+            session: AsyncSession
+            async with session.begin():
+                # 创建知识库表索引
+                await session.execute(
+                    text(
+                        "CREATE INDEX IF NOT EXISTS idx_kb_kb_id "
+                        "ON knowledge_bases(kb_id)"
+                    )
+                )
+                await session.execute(
+                    text(
+                        "CREATE INDEX IF NOT EXISTS idx_kb_name "
+                        "ON knowledge_bases(kb_name)"
+                    )
+                )
+                await session.execute(
+                    text(
+                        "CREATE INDEX IF NOT EXISTS idx_kb_created_at "
+                        "ON knowledge_bases(created_at)"
+                    )
+                )
+
+                # 创建文档表索引
+                await session.execute(
+                    text(
+                        "CREATE INDEX IF NOT EXISTS idx_doc_doc_id "
+                        "ON kb_documents(doc_id)"
+                    )
+                )
+                await session.execute(
+                    text(
+                        "CREATE INDEX IF NOT EXISTS idx_doc_kb_id "
+                        "ON kb_documents(kb_id)"
+                    )
+                )
+                await session.execute(
+                    text(
+                        "CREATE INDEX IF NOT EXISTS idx_doc_name "
+                        "ON kb_documents(doc_name)"
+                    )
+                )
+                await session.execute(
+                    text(
+                        "CREATE INDEX IF NOT EXISTS idx_doc_type "
+                        "ON kb_documents(file_type)"
+                    )
+                )
+                await session.execute(
+                    text(
+                        "CREATE INDEX IF NOT EXISTS idx_doc_created_at "
+                        "ON kb_documents(created_at)"
+                    )
+                )
+
+                # 创建多媒体表索引
+                await session.execute(
+                    text(
+                        "CREATE INDEX IF NOT EXISTS idx_media_media_id "
+                        "ON kb_media(media_id)"
+                    )
+                )
+                await session.execute(
+                    text(
+                        "CREATE INDEX IF NOT EXISTS idx_media_doc_id "
+                        "ON kb_media(doc_id)"
+                    )
+                )
+                await session.execute(
+                    text(
+                        "CREATE INDEX IF NOT EXISTS idx_media_kb_id ON kb_media(kb_id)"
+                    )
+                )
+                await session.execute(
+                    text(
+                        "CREATE INDEX IF NOT EXISTS idx_media_type "
+                        "ON kb_media(media_type)"
+                    )
+                )
+
+                await session.commit()
+
+    async def close(self) -> None:
+        """关闭数据库连接"""
+        await self.engine.dispose()
+        logger.info(f"知识库数据库已关闭: {self.db_path}")
+
+    async def get_kb_by_id(self, kb_id: str) -> KnowledgeBase | None:
+        """根据 ID 获取知识库"""
+        async with self.get_db() as session:
+            stmt = select(KnowledgeBase).where(col(KnowledgeBase.kb_id) == kb_id)
+            result = await session.execute(stmt)
+            return result.scalar_one_or_none()
+
+    async def get_kb_by_name(self, kb_name: str) -> KnowledgeBase | None:
+        """根据名称获取知识库"""
+        async with self.get_db() as session:
+            stmt = select(KnowledgeBase).where(col(KnowledgeBase.kb_name) == kb_name)
+            result = await session.execute(stmt)
+            return result.scalar_one_or_none()
+
+    async def list_kbs(self, offset: int = 0, limit: int = 100) -> list[KnowledgeBase]:
+        """列出所有知识库"""
+        async with self.get_db() as session:
+            stmt = (
+                select(KnowledgeBase)
+                .offset(offset)
+                .limit(limit)
+                .order_by(desc(KnowledgeBase.created_at))
+            )
+            result = await session.execute(stmt)
+            return list(result.scalars().all())
+
+    async def count_kbs(self) -> int:
+        """统计知识库数量"""
+        async with self.get_db() as session:
+            stmt = select(func.count(col(KnowledgeBase.id)))
+            result = await session.execute(stmt)
+            return result.scalar() or 0
+
+    # ===== 文档查询 =====
+
+    async def get_document_by_id(self, doc_id: str) -> KBDocument | None:
+        """根据 ID 获取文档"""
+        async with self.get_db() as session:
+            stmt = select(KBDocument).where(col(KBDocument.doc_id) == doc_id)
+            result = await session.execute(stmt)
+            return result.scalar_one_or_none()
+
+    async def list_documents_by_kb(
+        self, kb_id: str, offset: int = 0, limit: int = 100
+    ) -> list[KBDocument]:
+        """列出知识库的所有文档"""
+        async with self.get_db() as session:
+            stmt = (
+                select(KBDocument)
+                .where(col(KBDocument.kb_id) == kb_id)
+                .offset(offset)
+                .limit(limit)
+                .order_by(desc(KBDocument.created_at))
+            )
+            result = await session.execute(stmt)
+            return list(result.scalars().all())
+
+    async def count_documents_by_kb(self, kb_id: str) -> int:
+        """统计知识库的文档数量"""
+        async with self.get_db() as session:
+            stmt = select(func.count(col(KBDocument.id))).where(
+                col(KBDocument.kb_id) == kb_id
+            )
+            result = await session.execute(stmt)
+            return result.scalar() or 0
+
+    async def get_document_with_metadata(self, doc_id: str) -> dict | None:
+        async with self.get_db() as session:
+            stmt = (
+                select(KBDocument, KnowledgeBase)
+                .join(KnowledgeBase, col(KBDocument.kb_id) == col(KnowledgeBase.kb_id))
+                .where(col(KBDocument.doc_id) == doc_id)
+            )
+            result = await session.execute(stmt)
+            row = result.first()
+
+            if not row:
+                return None
+
+            return {
+                "document": row[0],
+                "knowledge_base": row[1],
+            }
+
+    async def delete_document_by_id(self, doc_id: str, vec_db: FaissVecDB):
+        """删除单个文档及其相关数据"""
+        # 在知识库表中删除
+        async with self.get_db() as session:
+            async with session.begin():
+                # 删除文档记录
+                delete_stmt = delete(KBDocument).where(col(KBDocument.doc_id) == doc_id)
+                await session.execute(delete_stmt)
+                await session.commit()
+
+        # 在 vec db 中删除相关向量
+        await vec_db.delete_documents(metadata_filters={"kb_doc_id": doc_id})
+
+    # ===== 多媒体查询 =====
+
+    async def list_media_by_doc(self, doc_id: str) -> list[KBMedia]:
+        """列出文档的所有多媒体资源"""
+        async with self.get_db() as session:
+            stmt = select(KBMedia).where(col(KBMedia.doc_id) == doc_id)
+            result = await session.execute(stmt)
+            return list(result.scalars().all())
+
+    async def get_media_by_id(self, media_id: str) -> KBMedia | None:
+        """根据 ID 获取多媒体资源"""
+        async with self.get_db() as session:
+            stmt = select(KBMedia).where(col(KBMedia.media_id) == media_id)
+            result = await session.execute(stmt)
+            return result.scalar_one_or_none()
+
+    async def update_kb_stats(self, kb_id: str, vec_db: FaissVecDB) -> None:
+        """更新知识库统计信息"""
+        chunk_cnt = await vec_db.count_documents()
+
+        async with self.get_db() as session:
+            async with session.begin():
+                update_stmt = (
+                    update(KnowledgeBase)
+                    .where(col(KnowledgeBase.kb_id) == kb_id)
+                    .values(
+                        doc_count=select(func.count(col(KBDocument.id)))
+                        .where(col(KBDocument.kb_id) == kb_id)
+                        .scalar_subquery(),
+                        chunk_count=chunk_cnt,
+                    )
+                )
+
+                await session.execute(update_stmt)
+                await session.commit()