知识图谱示例

本章节提供了 HiRAG 与 TRAG 两种方法的知识图谱构建和使用的示例。

两种方法均由：实体-关系三元组抽取、知识图谱构建、基于图的回答三个阶段构成。其中，实体-关系三元组抽取提供了CommonKG、GraphRAG两种三元组抽取方法。

构建实体关系三元组：

将 MinerU 处理后的结果提取实体关系三元组：

from src.data_processor.knowledge_graph.entity_relation_extractor import entity_relation_extractor

# 根据MinerU生成的文件得到三元组
mineru_path = "src/resources/pdf"

# GraphRAG方法提取
output_path = "src/resources/temp/knowledge_graph/graphrag"
entity_relation_extractor(mineru_path, output_path, method="graphrag")

from src.data_processor.knowledge_graph.entity_relation_extractor import entity_relation_extractor

# 根据MinerU生成的文件得到三元组
mineru_path = "src/resources/pdf"

# CommonKG方法提取
output_path = "src/resources/temp/knowledge_graph/commonkg"
entity_relation_extractor(mineru_path, output_path, method="CommonKG")

参数说明：

mineru_path: MinerU解析后的PDF文件路径
output_path: 三元组数据保存路径
method: 提取方法，可选 “graphrag” 或 “CommonKG”，任选一种即可

构建知识图谱

根据提取的三元组数据构建知识图谱：

HiRAG 构建

from src.data_processor.knowledge_graph.graph_builder import graph_builder

# 实体关系三元组数据构建hirag
data_path = "src/resources/temp/knowledge_graph/graphrag"
working_dir = "src/resources/temp/knowledge_graph/hirag"
graph_builder(data_path, working_dir, method="hirag")

LeanRAG 构建

from src.data_processor.knowledge_graph.graph_builder import graph_builder

# 实体关系三元组数据构建leanrag
data_path = "src/resources/temp/knowledge_graph/graphrag"
working_dir = "src/resources/temp/knowledge_graph/leanrag"
graph_builder(data_path, working_dir, method="leanrag")

参数说明：

data_path: 三元组数据路径
working_dir: 构建好的知识图谱保存路径
method: 构建方法，可选 “hirag” 或 “leanrag”

查询知识图谱

对构建好的知识图谱进行查询：

HiRAG 查询

from src.data_processor.knowledge_graph.query_graph import query_graph

query = "Which leadership positions changed at Datalogic in the reporting period?"
working_dir = "src/resources/temp/knowledge_graph/hirag"
result = query_graph(query, working_dir, method="hirag")
print(result)

LeanRAG 查询

from src.data_processor.knowledge_graph.query_graph import query_graph

query = "Which leadership positions changed at Datalogic in the reporting period?"
working_dir = "src/resources/temp/knowledge_graph/leanrag"
result = query_graph(query, working_dir, method="leanrag")
print(result)

参数说明：

query: 查询问题
working_dir: 知识图谱保存路径
method: 查询方法，与构建方法对应