Apache AGE

Apache AGE 是一個 PostgreSQL 擴充功能，提供圖形資料庫功能。AGE 是 A Graph Extension 的縮寫，靈感來自 Bitnine 的 PostgreSQL 10 分支 AgensGraph，後者是一個多模型資料庫。此專案的目標是建立單一儲存空間，可以處理關聯式和圖形模型資料，以便使用者可以使用標準 ANSI SQL 以及開放原始碼的圖形查詢語言 openCypher。Apache AGE 儲存的資料元素是節點、連接它們的邊緣，以及節點和邊緣的屬性。

此筆記本展示如何使用 LLM 提供自然語言介面，來查詢可以使用 Cypher 查詢語言的圖形資料庫。

Cypher 是一種宣告式圖形查詢語言，可在屬性圖形中進行具表現力且有效率的資料查詢。

設定

您需要安裝 AGE 擴充功能的執行中 Postgre 實例。測試的一個選項是使用官方 AGE Docker 映像執行 Docker 容器。您可以執行以下腳本來執行本地 Docker 容器

docker run \
    --name age  \
    -p 5432:5432 \
    -e POSTGRES_USER=postgresUser \
    -e POSTGRES_PASSWORD=postgresPW \
    -e POSTGRES_DB=postgresDB \
    -d \
    apache/age

關於在 Docker 中執行的其他說明，請參閱此處。

from langchain_community.graphs.age_graph import AGEGraph
from langchain_neo4j import GraphCypherQAChain
from langchain_openai import ChatOpenAI

API 參考：AGEGraph | GraphCypherQAChain | ChatOpenAI

conf = {
    "database": "postgresDB",
    "user": "postgresUser",
    "password": "postgresPW",
    "host": "localhost",
    "port": 5432,
}

graph = AGEGraph(graph_name="age_test", conf=conf)

植入資料庫

假設您的資料庫是空的，您可以使用 Cypher 查詢語言來填入資料。以下 Cypher 陳述是等冪的，這表示如果您執行一次或多次，資料庫資訊將會相同。

graph.query(
    """
MERGE (m:Movie {name:"Top Gun"})
WITH m
UNWIND ["Tom Cruise", "Val Kilmer", "Anthony Edwards", "Meg Ryan"] AS actor
MERGE (a:Actor {name:actor})
MERGE (a)-[:ACTED_IN]->(m)
"""
)

[]

重新整理圖形結構描述資訊

如果資料庫的結構描述變更，您可以重新整理產生 Cypher 陳述所需的結構描述資訊。

graph.refresh_schema()

print(graph.schema)

        Node properties are the following:
        [{'properties': [{'property': 'name', 'type': 'STRING'}], 'labels': 'Actor'}, {'properties': [{'property': 'property_a', 'type': 'STRING'}], 'labels': 'LabelA'}, {'properties': [], 'labels': 'LabelB'}, {'properties': [], 'labels': 'LabelC'}, {'properties': [{'property': 'name', 'type': 'STRING'}], 'labels': 'Movie'}]
        Relationship properties are the following:
        [{'properties': [], 'type': 'ACTED_IN'}, {'properties': [{'property': 'rel_prop', 'type': 'STRING'}], 'type': 'REL_TYPE'}]
        The relationships are the following:
        ['(:`Actor`)-[:`ACTED_IN`]->(:`Movie`)', '(:`LabelA`)-[:`REL_TYPE`]->(:`LabelB`)', '(:`LabelA`)-[:`REL_TYPE`]->(:`LabelC`)']

查詢圖形

我們現在可以使用圖形 Cypher QA 鏈來詢問關於圖形的問題

chain = GraphCypherQAChain.from_llm(
    ChatOpenAI(temperature=0), graph=graph, verbose=True, allow_dangerous_requests=True
)

chain.invoke("Who played in Top Gun?")

[1m> Entering new GraphCypherQAChain chain...[0m
``````output
Generated Cypher:
[32;1m[1;3mMATCH (a:Actor)-[:ACTED_IN]->(m:Movie)
WHERE m.name = 'Top Gun'
RETURN a.name[0m
Full Context:
[32;1m[1;3m[{'name': 'Tom Cruise'}, {'name': 'Val Kilmer'}, {'name': 'Anthony Edwards'}, {'name': 'Meg Ryan'}][0m

[1m> Finished chain.[0m

{'query': 'Who played in Top Gun?',
 'result': 'Tom Cruise, Val Kilmer, Anthony Edwards, Meg Ryan played in Top Gun.'}

限制結果數量

您可以使用 top_k 參數來限制 Cypher QA 鏈的結果數量。預設值為 10。

chain = GraphCypherQAChain.from_llm(
    ChatOpenAI(temperature=0),
    graph=graph,
    verbose=True,
    top_k=2,
    allow_dangerous_requests=True,
)

chain.invoke("Who played in Top Gun?")

[1m> Entering new GraphCypherQAChain chain...[0m
Generated Cypher:
[32;1m[1;3mMATCH (a:Actor)-[:ACTED_IN]->(m:Movie {name: 'Top Gun'})
RETURN a.name[0m
Full Context:
[32;1m[1;3m[{'name': 'Tom Cruise'}, {'name': 'Val Kilmer'}][0m

[1m> Finished chain.[0m

{'query': 'Who played in Top Gun?',
 'result': 'Tom Cruise, Val Kilmer played in Top Gun.'}

傳回中繼結果

您可以使用 return_intermediate_steps 參數從 Cypher QA 鏈傳回中繼步驟

chain = GraphCypherQAChain.from_llm(
    ChatOpenAI(temperature=0),
    graph=graph,
    verbose=True,
    return_intermediate_steps=True,
    allow_dangerous_requests=True,
)

result = chain("Who played in Top Gun?")
print(f"Intermediate steps: {result['intermediate_steps']}")
print(f"Final answer: {result['result']}")

[1m> Entering new GraphCypherQAChain chain...[0m
Generated Cypher:
[32;1m[1;3mMATCH (a:Actor)-[:ACTED_IN]->(m:Movie)
WHERE m.name = 'Top Gun'
RETURN a.name[0m
Full Context:
[32;1m[1;3m[{'name': 'Tom Cruise'}, {'name': 'Val Kilmer'}, {'name': 'Anthony Edwards'}, {'name': 'Meg Ryan'}][0m

[1m> Finished chain.[0m
Intermediate steps: [{'query': "MATCH (a:Actor)-[:ACTED_IN]->(m:Movie)\nWHERE m.name = 'Top Gun'\nRETURN a.name"}, {'context': [{'name': 'Tom Cruise'}, {'name': 'Val Kilmer'}, {'name': 'Anthony Edwards'}, {'name': 'Meg Ryan'}]}]
Final answer: Tom Cruise, Val Kilmer, Anthony Edwards, Meg Ryan played in Top Gun.

傳回直接結果

您可以使用 return_direct 參數從 Cypher QA 鏈傳回直接結果

chain = GraphCypherQAChain.from_llm(
    ChatOpenAI(temperature=0),
    graph=graph,
    verbose=True,
    return_direct=True,
    allow_dangerous_requests=True,
)

chain.invoke("Who played in Top Gun?")

[1m> Entering new GraphCypherQAChain chain...[0m
Generated Cypher:
[32;1m[1;3mMATCH (a:Actor)-[:ACTED_IN]->(m:Movie {name: 'Top Gun'})
RETURN a.name[0m

[1m> Finished chain.[0m

{'query': 'Who played in Top Gun?',
 'result': [{'name': 'Tom Cruise'},
  {'name': 'Val Kilmer'},
  {'name': 'Anthony Edwards'},
  {'name': 'Meg Ryan'}]}

在 Cypher 產生提示中新增範例

您可以定義您希望 LLM 針對特定問題產生的 Cypher 陳述

from langchain_core.prompts.prompt import PromptTemplate

CYPHER_GENERATION_TEMPLATE = """Task:Generate Cypher statement to query a graph database.
Instructions:
Use only the provided relationship types and properties in the schema.
Do not use any other relationship types or properties that are not provided.
Schema:
{schema}
Note: Do not include any explanations or apologies in your responses.
Do not respond to any questions that might ask anything else than for you to construct a Cypher statement.
Do not include any text except the generated Cypher statement.
Examples: Here are a few examples of generated Cypher statements for particular questions:
# How many people played in Top Gun?
MATCH (m:Movie {{title:"Top Gun"}})<-[:ACTED_IN]-()
RETURN count(*) AS numberOfActors

The question is:
{question}"""

CYPHER_GENERATION_PROMPT = PromptTemplate(
    input_variables=["schema", "question"], template=CYPHER_GENERATION_TEMPLATE
)

chain = GraphCypherQAChain.from_llm(
    ChatOpenAI(temperature=0),
    graph=graph,
    verbose=True,
    cypher_prompt=CYPHER_GENERATION_PROMPT,
    allow_dangerous_requests=True,
)

API 參考：PromptTemplate

chain.invoke("How many people played in Top Gun?")

[1m> Entering new GraphCypherQAChain chain...[0m
``````output
Generated Cypher:
[32;1m[1;3mMATCH (:Movie {name:"Top Gun"})<-[:ACTED_IN]-(:Actor)
RETURN count(*) AS numberOfActors[0m
Full Context:
[32;1m[1;3m[{'numberofactors': 4}][0m

[1m> Finished chain.[0m

{'query': 'How many people played in Top Gun?',
 'result': "I don't know the answer."}

針對 Cypher 和答案產生使用不同的 LLM

您可以使用 cypher_llm 和 qa_llm 參數來定義不同的 LLM

chain = GraphCypherQAChain.from_llm(
    graph=graph,
    cypher_llm=ChatOpenAI(temperature=0, model="gpt-3.5-turbo"),
    qa_llm=ChatOpenAI(temperature=0, model="gpt-3.5-turbo-16k"),
    verbose=True,
    allow_dangerous_requests=True,
)

chain.invoke("Who played in Top Gun?")

[1m> Entering new GraphCypherQAChain chain...[0m
``````output
Generated Cypher:
[32;1m[1;3mMATCH (a:Actor)-[:ACTED_IN]->(m:Movie)
WHERE m.name = 'Top Gun'
RETURN a.name[0m
Full Context:
[32;1m[1;3m[{'name': 'Tom Cruise'}, {'name': 'Val Kilmer'}, {'name': 'Anthony Edwards'}, {'name': 'Meg Ryan'}][0m

[1m> Finished chain.[0m

{'query': 'Who played in Top Gun?',
 'result': 'Tom Cruise, Val Kilmer, Anthony Edwards, and Meg Ryan played in Top Gun.'}

忽略指定的節點和關係類型

您可以使用 include_types 或 exclude_types 在產生 Cypher 陳述時忽略圖形結構描述的部分。

chain = GraphCypherQAChain.from_llm(
    graph=graph,
    cypher_llm=ChatOpenAI(temperature=0, model="gpt-3.5-turbo"),
    qa_llm=ChatOpenAI(temperature=0, model="gpt-3.5-turbo-16k"),
    verbose=True,
    exclude_types=["Movie"],
    allow_dangerous_requests=True,
)

# Inspect graph schema
print(chain.graph_schema)

Node properties are the following:
Actor {name: STRING},LabelA {property_a: STRING},LabelB {},LabelC {}
Relationship properties are the following:
ACTED_IN {},REL_TYPE {rel_prop: STRING}
The relationships are the following:
(:LabelA)-[:REL_TYPE]->(:LabelB),(:LabelA)-[:REL_TYPE]->(:LabelC)

驗證產生的 Cypher 陳述

您可以使用 validate_cypher 參數來驗證和修正產生的 Cypher 陳述中的關係方向

chain = GraphCypherQAChain.from_llm(
    llm=ChatOpenAI(temperature=0, model="gpt-3.5-turbo"),
    graph=graph,
    verbose=True,
    validate_cypher=True,
    allow_dangerous_requests=True,
)

chain.invoke("Who played in Top Gun?")

[1m> Entering new GraphCypherQAChain chain...[0m
Generated Cypher:
[32;1m[1;3mMATCH (a:Actor)-[:ACTED_IN]->(m:Movie)
WHERE m.name = 'Top Gun'
RETURN a.name[0m
Full Context:
[32;1m[1;3m[{'name': 'Tom Cruise'}, {'name': 'Val Kilmer'}, {'name': 'Anthony Edwards'}, {'name': 'Meg Ryan'}][0m

[1m> Finished chain.[0m

{'query': 'Who played in Top Gun?',
 'result': 'Tom Cruise, Val Kilmer, Anthony Edwards, Meg Ryan played in Top Gun.'}

設定​

植入資料庫​

重新整理圖形結構描述資訊​

查詢圖形​

限制結果數量​

傳回中繼結果​

傳回直接結果​

在 Cypher 產生提示中新增範例​

針對 Cypher 和答案產生使用不同的 LLM​

忽略指定的節點和關係類型​

驗證產生的 Cypher 陳述​

此頁面是否對您有幫助？

設定