知识图谱之py2neo

qq_16740151

已于 2023-03-12 17:29:49 修改

阅读量1.6k

点赞数 3

分类专栏：知识图谱 python 文章标签：知识图谱 Powered by 金山文档

于 2023-03-12 10:55:03 首次发布

本文链接：https://blog.csdn.net/qq_16740151/article/details/129472662

版权

python 同时被 2 个专栏收录

11 篇文章 0 订阅

订阅专栏

知识图谱

3 篇文章 0 订阅

订阅专栏

py2neo介绍

介绍

Py2neo是一个客户端库和工具包，用于从Python应用程序和命令行使用Neo4j（Neo4j Graph Data Platform | Graph Database Management System）。该库同时支持 Bolt 和 HTTP，并提供高级 API、OGM、管理工具、交互式控制台、Pygments 的 Cypher 词法分析器以及许多其他功能。

安装

pip install py2neo

注：Neo4j也需要安装，地址https://neo4j.com/

py2neo基本使用

启动Neo4j

连接代码

from py2neo import Graph, Node, Relationship
 
graph = Graph("http://localhost:7474//browser", name='neo4j', auth=("neo4j", "12345678"))
print(graph.run("show databases"))

基于xsls文件的demo

demo效果

xls文件内容

sheet 结构

"关系"sheet内容如下

id	from	to	type
1	令狐冲	宁中则	师娘
2	仪琳	令狐冲	倾慕
3	任盈盈	令狐冲	爱慕
4	令狐冲	岳林珊	小师妹
5	岳林珊	岳不群	父亲
6	任盈盈	任我行	父亲
7	岳不群	宁中则	夫妻
#8	林平之	岳林珊	夫妻

下面的表格分别对应"恒山派"、"华山派"、"日月教"sheet中的内容（）

id	title
1	令狐冲
2	仪清
3	仪和
4	仪琳

id	title
1	岳不群
2	宁中则
3	岳林珊
4	林平之

id	title
1	任我行
2	任盈盈
3	向问天

demo依赖

pip install pandas
pip install openpyxl

Neo4j和py2neo也是需要的。

demo主要代码块

import os

import pandas as pd
from py2neo import Graph, Node, Relationship


sheetNameList = [
    {'sheet':'恒山派',     'type':'Node'},
    {'sheet':'华山派',     'type':'Node'},
    {'sheet':'日月教',     'type':'Node'},
    {'sheet':'关系',       'type':'Relationship'},
]

def __init__(self, filename=None):
    self.filename = filename

    self.nodeList             = []
    self.relationshipList     = []

    self._loadfile()
    self._buildGraph()

 def _loadfile(self):
    if not os.path.exists(self.filename):
        print('_loadfile', 'file not exist')
        return False

    for sheetName in sheetNameList:
        sheetData = pd.read_excel(self.filename, index_col=None, comment='#', sheet_name=sheetName['sheet'])

        if sheetName['type'] == 'Node':
            keys=[]
            for index, row in sheetData.iterrows():
                if not row['title']:
                    continue
                keys.append(row['title'])
            
            keyval = dict(zip(keys, keys))
            node = Node(sheetName['sheet'], name=sheetName['sheet'], **keyval)
            self.nodeList.append(node)

            for index, row in sheetData.iterrows():
                if not row['title']:
                    continue
                subNode = Node(sheetName['sheet'], name=row['title'])
                self.nodeList.append(subNode)
                rel = Relationship(node, '属性', subNode)
                self.relationshipList.append(rel)

        elif sheetName['type'] == 'Relationship':
            startNode = None
            endNode   = None
            for index, row in sheetData.iterrows():
                if not row['from'] or not row['to']:
                    continue

                for node in self.nodeList:
                    if node['name'] == row['from']:
                        startNode = node
                    elif node['name'] == row['to']:
                        endNode   = node
                    else:
                        pass
                rel = Relationship(startNode, row['type'], endNode)
                self.relationshipList.append(rel)
        else:
            pass

def _buildGraph(self):
    graph = Graph("http://localhost:7474//browser", name='neo4j', auth=("neo4j", "12345678"))
    graph.delete_all()

    for node in self.nodeList:
        graph.create(node)

    for rel in self.relationshipList:
        graph.merge(rel)
        pass

    graph.begin()