# 声明命名空间前缀
@prefix ex: <http://example.org/> .
@prefix schema: <http://schema.org/> .

# 三元组：主语 谓语 宾语
ex:Einstein
  schema:name      "Albert Einstein" ;
  schema:birthDate "1879-03-14" ;
  ex:proposed     ex:TheoryOfRelativity .

// 创建节点和关系
CREATE (e:Person {
  name: "Albert Einstein",
  born: 1879
})
CREATE (t:Theory {
  name: "Theory of Relativity",
  year: 1905
})
CREATE (e)-[:PROPOSED {
  year: 1905
}]->(t)

// 查询：找到爱因斯坦提出的所有理论
MATCH (p:Person {name: "Einstein"})
      -[:PROPOSED]->(t:Theory)
RETURN t.name, t.year

# 定义类层次
ex:Mammal rdfs:subClassOf ex:Animal .

# 定义对称关系（A是B的兄弟 → B也是A的兄弟）
ex:isSiblingOf
  a owl:SymmetricProperty .

# 定义传递关系（A是B祖先，B是C祖先 → A是C祖先）
ex:isAncestorOf
  a owl:TransitiveProperty .

SELECT ?person ?award
WHERE {
  ?person rdf:type    ex:Scientist .
  ?person ex:wonAward ?award .
  ?award  rdf:type    ex:NobelPrize .
  ?person ex:bornIn   ?country .
  FILTER(?country = "Germany")
}
ORDER BY ?person

import spacy

# 加载预训练的 Transformer 模型
nlp = spacy.load("en_core_web_trf")

text = "Elon Musk founded SpaceX in 2002."
doc = nlp(text)

for ent in doc.ents:
    print(ent.text, ent.label_)

# 输出：
# Elon Musk  PERSON
# SpaceX     ORG
# 2002       DATE

from transformers import pipeline

# 使用 REBEL 端到端关系抽取模型
re_pipe = pipeline(
    "text2text-generation",
    model="Babelscape/rebel-large"
)

text = "Elon Musk founded SpaceX in 2002."
result = re_pipe(text, max_length=512)

# 解析输出的三元组：
# subject:  Elon Musk
# relation: founded by
# object:   SpaceX

from pykeen.pipeline import pipeline

# 训练 TransE 模型
result = pipeline(
    model='TransE',
    dataset='FB15k237',
    epochs=100,
    embedding_dim=128
)

# 链接预测：给定主语和谓语，预测宾语
model = result.model
scores = model.predict_t(
    h='Paris',
    r='capital_of'
)
# 返回最可能的宾语：France（置信度最高）

from langchain_community.graphs import Neo4jGraph
from langchain.chains import GraphCypherQAChain
from langchain_openai import ChatOpenAI

# 1. 连接 Neo4j 知识图谱
graph = Neo4jGraph(
    url="bolt://localhost:7687",
    username="neo4j",
    password="password"
)

# 2. 初始化 LLM
llm = ChatOpenAI(
    model="gpt-4o",
    temperature=0
)

# 3. 构建 GraphRAG Chain
chain = GraphCypherQAChain.from_llm(
    llm=llm,
    graph=graph,
    verbose=True
)

# 4. 自然语言问答
result = chain.invoke(
    "Who are the scientists that won Nobel Prize and collaborated with Turing Award winners?"
)
print(result["result"])

# 安装
pip install graphrag

# 初始化项目
python -m graphrag.index --init \
  --root ./my-project

# 将文档放入 input/ 目录后，构建图谱
python -m graphrag.index \
  --root ./my-project

# 全局查询（适合宏观问题）
python -m graphrag.query \
  --root ./my-project \
  --method global \
  --query "What are the main themes?"

# 局部查询（适合具体实体问题）
python -m graphrag.query \
  --root ./my-project \
  --method local \
  --query "Who is the key person?"