如何通过Gremlin-Python在Gremlin-Server导入CSV创建图？

阿华AIGC实验室

2026-5-26

别发愁！我来一步步带你用Gremlin-Python在Gremlin Server上搭建漫威超级英雄图——其实核心思路和你找到的Groovy版指南是相通的，只是把语法换成Python风格而已，跟着做就行。

用Gremlin-Python在Gremlin Server创建漫威超级英雄图

一、先搞定环境准备

确保Gremlin Server已经启动（这里以默认的TinkerGraph为例，用JanusGraph、Neo4j等其他图数据库也可以，操作逻辑一致）
安装Gremlin-Python依赖：pip install gremlinpython

二、先理清数据结构（以漫威IP为例）

先明确我们要创建的顶点和边类型：

顶点类型：Hero（超级英雄）、Movie（电影）、Team（英雄团队）
边类型：appeared_in（英雄出演电影）、member_of（英雄属于某团队）、teamed_with（英雄与其他英雄组队）

三、编写Python代码实现导入

1. 建立与Gremlin Server的连接

首先创建客户端连接，默认端口是8182：

from gremlin_python.driver import client, serializer

# 连接本地Gremlin Server
gremlin_client = client.Client(
    'ws://localhost:8182/gremlin',
    'g',
    message_serializer=serializer.GraphSONSerializersV2d0()
)

2. 导入顶点数据

我们先准备一批测试数据，用参数化查询的方式添加顶点（既安全又高效）：

# 超级英雄数据
heroes = [
    {"id": "iron_man", "name": "Tony Stark", "alias": "Iron Man", "universe": "Marvel"},
    {"id": "captain_america", "name": "Steve Rogers", "alias": "Captain America", "universe": "Marvel"},
    {"id": "black_widow", "name": "Natasha Romanoff", "alias": "Black Widow", "universe": "Marvel"}
]

# 电影数据
movies = [
    {"id": "avengers_1", "title": "The Avengers", "year": 2012},
    {"id": "captain_civil_war", "title": "Captain America: Civil War", "year": 2016}
]

# 英雄团队数据
teams = [
    {"id": "avengers", "name": "The Avengers", "founded_year": 1963}
]

# 批量添加英雄顶点
for hero in heroes:
    query = """
        g.addV('Hero')
          .property('id', $id)
          .property('name', $name)
          .property('alias', $alias)
          .property('universe', $universe)
    """
    gremlin_client.submit(query, hero).all().result()

# 批量添加电影顶点
for movie in movies:
    query = """
        g.addV('Movie')
          .property('id', $id)
          .property('title', $title)
          .property('year', $year)
    """
    gremlin_client.submit(query, movie).all().result()

# 添加团队顶点
for team in teams:
    query = """
        g.addV('Team')
          .property('id', $id)
          .property('name', $name)
          .property('founded_year', $founded_year)
    """
    gremlin_client.submit(query, team).all().result()

3. 导入边数据

顶点添加完成后，我们来建立顶点之间的关联：

# 英雄出演电影的关联
appearances = [
    {"hero_id": "iron_man", "movie_id": "avengers_1"},
    {"hero_id": "iron_man", "movie_id": "captain_civil_war"},
    {"hero_id": "captain_america", "movie_id": "avengers_1"},
    {"hero_id": "captain_america", "movie_id": "captain_civil_war"},
    {"hero_id": "black_widow", "movie_id": "avengers_1"},
    {"hero_id": "black_widow", "movie_id": "captain_civil_war"}
]

# 英雄加入团队的关联
memberships = [
    {"hero_id": "iron_man", "team_id": "avengers"},
    {"hero_id": "captain_america", "team_id": "avengers"},
    {"hero_id": "black_widow", "team_id": "avengers"}
]

# 添加appeared_in边
for appearance in appearances:
    query = """
        g.V($hero_id).as('h')
          .V($movie_id).as('m')
          .addE('appeared_in')
          .from('h').to('m')
    """
    gremlin_client.submit(query, appearance).all().result()

# 添加member_of边
for membership in memberships:
    query = """
        g.V($hero_id).as('h')
          .V($team_id).as('t')
          .addE('member_of')
          .from('h').to('t')
    """
    gremlin_client.submit(query, membership).all().result()

4. 验证导入结果

我们可以执行几个查询来确认数据是否正确导入：

# 查询所有复仇者成员的别名
result = gremlin_client.submit("""
    g.V('avengers').in('member_of').values('alias')
""").all().result()
print("复仇者成员：", result)

# 查询钢铁侠出演的所有电影
result = gremlin_client.submit("""
    g.V('iron_man').out('appeared_in').values('title')
""").all().result()
print("钢铁侠出演的电影：", result)