【技术笔记】Python调用neo4j接口实现数据操作

在知识图谱工作处理中，难免会接触到图数据库，而neo4j在这个领域排名领先的厂商，因此在这里整理一些自己在工作中常见的数据处理函数。此处主要针对Python开发语言的同仁们。欢迎copy。

1. 写入数据

直接代码奉上：

import json
from neo4j import GraphDatabase

def add_symptom_info(tx, symptom_info):
    """
    description: 添加症状数据
    :param tx: 事务
    :param symptom_info: 症状信息字典结构为{ "症状名": {"features":["aaa", "bbb"], "degrees":["ccc"], "triggers":[], "frequencies":[]}}
    :return:
    """
    for symptom_name, object_dict in symptom_info.items():
        # 分别获取症状属性
        features = object_dict.get("features") if object_dict.get("features") else []
        degrees = object_dict.get("degrees") if object_dict.get("degrees") else []
        triggers = object_dict.get("triggers") if object_dict.get("triggers") else []
        frequencies = object_dict.get("frequencies") if object_dict.get("frequencies") else []
        tx.run("MERGE (s:Symptom {name: $name, "
               "features: $features, "
               "degrees: $degrees, "
               "triggers: $triggers, "
               "frequencies: $frequencies}) ",
               name=symptom_name,
               features=features,
               degrees=degrees,
               triggers=triggers,
               frequencies=frequencies)


def add_disease_info(tx, disease_info):
    """
    description: 增加疾病信息
    :param tx: 事务
    :param disease_info: 疾病信息
    """
    disease_name = disease_info.get("name")
    genetic = disease_info.get("genetic")
    infectivity = disease_info.get("infectivity")
    # 获取当前疾病对应症状
    symptoms = disease_info.get("symptoms")
    # 疾病发病部位
    part_list = disease_info.get("part")
    # 插入疾病数据
    tx.run("MERGE (a:Disease {name: $name, "
               "genetic: $genetic, "
               "cost: $cost, "
               "infectivity: $infectivity})",
           name=disease_name,
           genetic=genetic,
           infectivity=infectivity)
           
    # 更新式插入发病部位并建立关系
    for part in part_list:
        part = part.strip()
        tx.run("MATCH (d:Disease {name: $disease_name}) "
               "MERGE (a:Part {name: $name}) "
               "MERGE (a)<-[:发病部位]-(d)",
               name=part,
               disease_name=disease_name)
               
    # 查询症状并建立关联
    for symptom in symptoms:
        symptom = symptom.strip()
        tx.run("MATCH (d:Disease {name: $disease_name}) "
               "MATCH (s:Symptom {name: $name}) "
               "MERGE (s)-[:病症]-(d)",
               name=symptom,
               disease_name=disease_name)



def import_data_to_kg():
    """
    description: 导入数据至知识图谱
    """
    driver = GraphDatabase.driver("bolt://localhost:7687", auth=("username", "password"))
    with driver.session() as session:
    	# 增加症状信息
        symptom_info_dict = json.load(open("symptom_info_list_dict.json"))
        session.write_transaction(add_symptom_info, symptom_info_dict)
        
		# 增加疾病信息
        for disease_info in json.load(open("disease_info_list.json")):
            session.write_transaction(add_disease_info, disease_info)

---

2. 查询数据

同样直接上代码：

import json
from neo4j import GraphDatabase

def load_disease_list(tx):
    """
    description: 加载疾病信息
    :param tx:
    """
    disease_info_list = []
    for record in tx.run("MATCH (d:Disease) RETURN d ORDER BY d.name"):
        disease_info = record["d"]
        # 疾病名称
        disease_name = disease_info.get("name")
        # 疾病别名
        disease_alias = disease_info.get("alias")
        # 发病部位
        part = disease_info.get("part")
        # 多发人群
        crowd = disease_info.get("crowd")
        # 治疗方案
        therapies = disease_info.get("therapies")
        # 是否具有传染性
        infectivity = disease_info.get("infectivity")
        # 是否遗传
        genetic = disease_info.get("genetic")
        disease_dict = {
            "name": disease_name,
            "genetic": genetic,
            "therapies": therapies,
            "crowd": crowd,
            "part": part,
            "alias": disease_alias,
            "infectivity": infectivity
        }
        disease_info_list.append(disease_dict)
    # 输出数据至json文件中
    json.dump(disease_info_list,
              open("disease_info_list.json", mode="w", encoding="utf8"),
              ensure_ascii=False,
              indent=4)



def load_node_data():
    """
    description: 获取知识图谱节点数据
    """
    driver = GraphDatabase.driver("bolt://localhost:7687", auth=("username", "password"))
    with driver.session() as session:
        # 获取疾病信息
        session.read_transaction(load_disease_list)

---

参考

• 官方github地址->【Neo4j Bolt Driver for Python】