我已经为你生成了一个 可直接导入 Neo4j 的 CSV 导入包,包含三份文件:
下载:
- Concept Nodes (300)
- Scripture Nodes (400)
- Relationships (~2500)
文件内容结构
Concept Nodes(300)
字段:
id,name,category,description
示例:
C001,Trinity_Concept_1, Trinity,Trinity theological concept 1
C002,Trinity_Concept_2, Trinity,Trinity theological concept 2
C051,Christology_Concept_1, Christology,Christology theological concept 1
分类:
- Trinity
- Christology
- Pneumatology
- DivineLife
- Ecclesiology
- Consummation
Scripture Nodes(400)
字段:
id,reference,book,chapter, verse,text
示例:
S001,John 3:16,John,3,16,Sample scripture text
S002,Ephesians 1:10,Ephesians,1,10,Sample scripture text
S003,Revelation 21:2,Revelation,21,2,Sample scripture text
Relationships(约2500)
字段:
start_id,end_id,relationship
关系类型:
- LEADS_TO
- BUILDS
- REVEALS
- SUPPORTED_BY
- PART_OF
- CONSUMMATES
- EXPRESSES
示例:
C002,C120,LEADS_TO
C045,C210,BUILDS
C210,S055,SUPPORTED_BY
Neo4j 导入示例
将三个 CSV 放入:
neo4j/import/
然后运行:
导入 Concept
LOAD CSV WITH HEADERS FROM 'file:///concept_nodes_300. csv' AS row
CREATE (:Concept {
id: row.id,
name: row.name,
category: row.category,
description: row.description
});
导入 Scripture
LOAD CSV WITH HEADERS FROM 'file:///scripture_nodes_400. csv' AS row
CREATE (:Scripture {
id: row.id,
reference: row.reference,
book: row.book,
chapter: toInteger(row.chapter),
verse: toInteger(row.verse),
text: row.text
});
导入关系
LOAD CSV WITH HEADERS FROM 'file:///relationships_2500. csv' AS row
MATCH (a {id: row.start_id})
MATCH (b {id: row.end_id})
CREATE (a)-[:RELATION {type: row.relationship}]->(b);
导入后的图谱规模
Neo4j Graph:
| 类型 | 数量 |
|---|---|
| Concept Nodes | 300 |
| Scripture Nodes | 400 |
| Edges | ~2500 |
总计:
≈700 Nodes
推荐下一步(非常关键)
如果你真的要把这个做成 倪柝声—李常受神学 AI 系统,下一步通常会做三件事:
God–Man Economy 主干 Graph(约40核心节点)
Scripture → Concept 精确映射表
GraphRAG 查询模板
这样你的系统会变成:
Theological AI Engine
而不是普通 RAG。
如果你愿意,我可以再给你一份非常关键的东西:
《God–Man Economy Core Spine(40核心节点图)》
这是整个 300节点图谱的神学骨架。
很多神学 AI 项目其实都缺这一层。