Codex
56 lines
2.6 KiB
Python
56 lines
2.6 KiB
Python
"""
|
||
知识库模型
|
||
"""
|
||
from sqlalchemy import Column, String, Text, Integer, Float, Boolean, JSON
|
||
|
||
from app.base_model import BaseModel
|
||
|
||
|
||
class KnowledgeBase(BaseModel):
|
||
"""
|
||
知识库
|
||
|
||
管理文档集合，配置分块策略和检索参数
|
||
"""
|
||
__tablename__ = "ai_knowledge_base"
|
||
|
||
application_id = Column(String(21), nullable=True, index=True, comment="所属应用ID（逻辑外键关联core_application）")
|
||
is_global = Column(Boolean, default=False, comment="是否在子应用中可见")
|
||
name = Column(String(100), nullable=False, comment="知识库名称")
|
||
code = Column(String(100), nullable=False, unique=True, comment="知识库编码")
|
||
description = Column(Text, nullable=True, comment="描述")
|
||
icon = Column(String(50), default="", comment="图标")
|
||
|
||
# Embedding 配置
|
||
embedding_model_id = Column(String(21), nullable=True, comment="Embedding 模型ID（逻辑外键关联ai_llm_model）")
|
||
embedding_dimensions = Column(Integer, default=1536, comment="向量维度")
|
||
|
||
# 分块策略
|
||
chunk_strategy = Column(String(20), default="recursive", comment="分块策略: recursive/semantic/markdown/fixed")
|
||
chunk_size = Column(Integer, default=500, comment="分块大小（字符数）")
|
||
chunk_overlap = Column(Integer, default=50, comment="分块重叠大小（字符数）")
|
||
separator = Column(String(50), nullable=True, comment="自定义分隔符")
|
||
|
||
# 检索配置
|
||
retrieval_mode = Column(String(20), default="hybrid", comment="检索模式: vector/fulltext/hybrid")
|
||
top_k = Column(Integer, default=5, comment="检索返回数量")
|
||
score_threshold = Column(Float, default=0.5, comment="相似度阈值（0-1）")
|
||
rerank_enabled = Column(Boolean, default=False, comment="是否启用重排序")
|
||
rerank_model_id = Column(String(21), nullable=True, comment="重排序模型ID")
|
||
retrieval_weight = Column(Float, default=1.0, comment="检索权重（多知识库检索时的加权系数，0.1-10.0）")
|
||
|
||
# 预处理规则
|
||
process_rules = Column(JSON, nullable=True, comment="预处理规则（清洗配置）")
|
||
|
||
# 索引模式
|
||
indexing_technique = Column(String(20), default="high_quality", comment="索引模式: high_quality/economy")
|
||
|
||
# 统计
|
||
document_count = Column(Integer, default=0, comment="文档数量")
|
||
segment_count = Column(Integer, default=0, comment="分段数量")
|
||
total_token_count = Column(Integer, default=0, comment="总 Token 数")
|
||
total_char_count = Column(Integer, default=0, comment="总字符数")
|
||
|
||
# 状态
|
||
status = Column(String(20), default="active", comment="状态: active/disabled")
|