python操作neo4j创建知识图谱模板

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


class DataToNeo4j:
    def __init__(self):
        # 连接数据库
        link = Graph("http:localhost:7474", username="neo4j", password="rhino1qaz@wsx")
        self.graph = link
        # 建立节点
        self.buy = "buy"
        self.sell = "sell"
        self.graph.delete_all()  # 先清空数据库
        self.matcher = NodeMatcher(link)  # 然后定义一个matcher

    def create_node(self, node_buy_key, node_sell_key):
        """建立节点"""
        for name in node_buy_key:
            buy_node = Node(self.buy, name=name)
            self.graph.create(buy_node)
        for name in node_sell_key:
            sell_node = Node(self.sell, name=name)
            self.graph.create(sell_node)

    def create_relation(self, df_data):
        """建立联系"""
        for m in range(0, len(df_data)):
            # print(df_data['buy'][m], df_data['money'][m], df_data['sell'][m])
            rel = Relationship(self.matcher.match(self.buy).where("_.name=" + "'" + df_data['buy'][m] + "'").first(),
                               df_data['money'][m],
                               self.matcher.match(self.sell).where("_.name=" + "'" + df_data['sell'][m] + "'").first())
            self.graph.create(rel)


class Processor:
    def __init__(self, path):
        self.invoice_data = pd.read_excel(path, header=0)

    def node_extraction(self):
        # 取出购买方名称到list
        node_buy_key = self.invoice_data['购买方名称'].tolist()

        node_sell_key = self.invoice_data['销售方名称'].tolist()

        # 去除重复的发票名称
        node_buy_key = list(set(node_buy_key))
        node_sell_key = list(set(node_sell_key))

        return node_buy_key, node_sell_key

    def relation_extraction(self):
        """联系数据抽取"""

        links_dict = {}
        sell_list = []
        money_list = []
        buy_list = []

        for i in range(0, len(self.invoice_data)):
            money_list.append(self.invoice_data[self.invoice_data.columns[19]][i])  # 金额
            sell_list.append(self.invoice_data[self.invoice_data.columns[10]][i])  # 销售方方名称
            buy_list.append(self.invoice_data[self.invoice_data.columns[6]][i])  # 购买方名称

        # 将数据中int类型全部转成string
        sell_list = [str(i) for i in sell_list]
        buy_list = [str(i) for i in buy_list]
        money_list = [str(i) for i in money_list]

        # 整合数据，将三个list整合成一个dict
        links_dict['buy'] = buy_list
        links_dict['money'] = money_list
        links_dict['sell'] = sell_list
        # 将数据转成DataFrame
        df_data = pd.DataFrame(links_dict)
        return df_data


if __name__ == '__main__':
    # dataToNeo4j = DataToNeo4j()
    path = './Invoice_data_Demo.xls'
    processor = Processor(path)
    node_buy_key, node_sell_key = processor.node_extraction()
    print(node_buy_key)
    print(node_sell_key)
    df_data = processor.relation_extraction()
    dataToNeo4j = DataToNeo4j()
    dataToNeo4j.create_node(node_buy_key, node_sell_key)
    dataToNeo4j.create_relation(df_data)