Knowledge Graph - Neo4j Integration

Connect to Neo4j graph databases and execute Cypher queries for storing, querying, and managing knowledge graph data using the property graph model. Full support for transactions, indexes, bulk operations, and result mapping.

Neo4j Integration

Connect to Neo4j graph databases and execute Cypher queries for efficient knowledge graph management.

This skill enables seamless interaction with Neo4j graph databases using the official Python driver. It provides connection management, query execution, transaction support, and result mapping for the property graph model.

Quick Start

Use When

• Working with Neo4j-backed knowledge graphs
• Executing Cypher queries on graph data
• Creating or updating nodes and relationships
• Importing graph data into Neo4j
• Querying graph structures with complex patterns
• Managing graph database transactions
• Creating indexes for performance optimization
• Building production graph applications

Inputs

• Neo4j connection credentials (URI, username, password)
• Cypher queries with optional parameters
• Node/relationship definitions
• Bulk data for import
• Transaction context

Outputs

• Query results (nodes, relationships, properties)
• Execution statistics and metrics
• Success/failure status
• Record counts and performance data

Connection & Authentication

Supported Protocols

text

bolt://        - Unencrypted connection
neo4j://       - Standard connection (recommended)
neo4j+s://     - TLS-encrypted connection
neo4j+ssc://   - Self-signed certificate

Connection Configuration

python

config = {
    "uri": "neo4j://localhost:7687",
    "username": "neo4j",
    "password": "secure_password",
    "encrypted": True,
    "trust": "TRUST_ALL_CERTIFICATES"
}

Connection Pool

• Default pool size: 50 connections
• Configurable connection limits
• Automatic connection recycling
• Health checks for stale connections

Property Graph Model

Neo4j uses a property graph model with three core elements:

1. Nodes

Represent entities with labels and properties.

cypher

CREATE (p:Person {name: "Alice", age: 30, email: "[e-mail]"})
CREATE (c:Company {name: "TechCorp", industry: "Technology"})

Properties:

• Name: String identifier
• Properties: Key-value pairs
• Labels: Type classification (Person, Company, etc.)

2. Relationships

Connect nodes with typed, directed edges and properties.

cypher

CREATE (a:Person)-[:WORKS_AT {since: 2020}]->(c:Company)
CREATE (a:Person)-[:KNOWS {strength: 0.8}]->(b:Person)

Characteristics:

• Direction: Start node → End node
• Type: Uppercase name (WORKS_AT, KNOWS, etc.)
• Properties: Optional metadata
• Can be traversed in both directions with <-

3. Properties

Attributes on nodes and relationships.

text

String: "text value"
Integer: 42
Float: 3.14
Boolean: true/false
DateTime: timestamp
List: [1, 2, 3]

Core Cypher Query Patterns

MATCH - Find Data

cypher

MATCH (p:Person) WHERE p.age > 30 RETURN p.name, p.age

CREATE - Add Data

cypher

CREATE (p:Person {name: "Bob", age: 25}) RETURN p

MERGE - Create or Update

cypher

MERGE (p:Person {name: "Alice"}) SET p.age = 31 RETURN p

MATCH + CREATE - Add Relationships

cypher

MATCH (a:Person {name: "Alice"}), (b:Person {name: "Bob"})
CREATE (a)-[:KNOWS]->(b) RETURN a, b

DELETE - Remove Data

cypher

MATCH (p:Person {name: "Old Person"}) DELETE p

RETURN + ORDER BY + LIMIT

cypher

MATCH (p:Person) RETURN p ORDER BY p.age DESC LIMIT 10

Advanced Query Features

Aggregations

cypher

MATCH (p:Person) RETURN COUNT(p), AVG(p.age), MAX(p.age)

Collection Functions

cypher

MATCH (p:Person)-[:KNOWS]->(friends:Person)
RETURN p.name, COLLECT(friends.name) AS friend_list

Conditional Logic

cypher

MATCH (p:Person) 
RETURN p.name, CASE WHEN p.age > 30 THEN "Senior" ELSE "Junior" END AS level

Path Queries

cypher

MATCH path = (a:Person)-[:KNOWS*1..3]->(b:Person)
WHERE a.name = "Alice" AND b.name = "Bob"
RETURN path, LENGTH(path) AS hops

Graph Algorithms

cypher

MATCH (n:Person) WHERE exists(n.pagerank) RETURN n ORDER BY n.pagerank DESC

Transaction Management

Simple Transaction

cypher

BEGIN
CREATE (p:Person {name: "Alice"})
CREATE (c:Company {name: "TechCorp"})
COMMIT

Rollback on Error

cypher

BEGIN
CREATE (p:Person {name: "Alice"})
ROLLBACK

Properties

• Atomicity: All-or-nothing execution
• Consistency: Graph constraints maintained
• Isolation: ACID compliance
• Durability: Persistent storage

Indexes & Performance

Create Index

cypher

CREATE INDEX person_name FOR (p:Person) ON (p.name)
CREATE INDEX company_id FOR (c:Company) ON (c.id)

Index Types

• Range Index - Efficient for range queries
• Full-text Index - Text search capability
• Lookup Index - Universal index
• Unique Index - Constraint enforcement

Query Optimization

1. Use indexes on filtered properties - WHERE clauses
2. Avoid cartesian products - Join on common properties
3. Limit result sets - Use LIMIT clause
4. Batch imports - Load data in chunks
5. Profile queries - EXPLAIN/PROFILE for analysis

Bulk Operations

Import CSV

cypher

LOAD CSV WITH HEADERS FROM "file:///data.csv" AS row
CREATE (p:Person {name: row.name, age: toInteger(row.age)})

Batch Create

cypher

UNWIND $nodes AS node
CREATE (n {id: node.id, name: node.name})

Batch Update

cypher

UNWIND $updates AS update
MATCH (p:Person {id: update.id})
SET p.age = update.age

Result Mapping

Simple Results

cypher

MATCH (p:Person) RETURN p.name, p.age

Maps to:

python

[
    {"name": "Alice", "age": 30},
    {"name": "Bob", "age": 25}
]

Node Results

cypher

MATCH (p:Person) RETURN p

Maps to Python Node objects with:

• id: Node internal ID
• labels: List of labels
• properties: Dict of properties

Relationship Results

cypher

MATCH (a)-[r]->(b) RETURN r

Maps to Relationship objects with:

• id: Relationship internal ID
• type: Relationship type
• properties: Dict of properties
• start_node_id: Source node ID
• end_node_id: Target node ID

Error Handling

Common Errors

Error	Cause	Solution
Connection refused	Neo4j not running	Start Neo4j server
Authentication failed	Wrong credentials	Verify username/password
Syntax error	Invalid Cypher	Check query syntax
Constraint violation	Duplicate/invalid data	Check constraints
Timeout	Query too slow	Add indexes, optimize query
Out of memory	Too much data	Batch operations, paginate

Retry Logic

• Connection failures: Automatic retry with exponential backoff
• Transient errors: Configurable retry attempts
• Circuit breaker: Fail fast on persistent failures

Best Practices

✓ Use parameterized queries - Prevent injection attacks
✓ Create appropriate indexes - Improve query performance
✓ Batch large imports - Avoid memory exhaustion
✓ Use transactions - Ensure data consistency
✓ Profile queries - Identify performance bottlenecks
✓ Close connections - Prevent resource leaks
✓ Limit result sets - Avoid network overhead
✓ Normalize node names - Prevent duplicate nodes
✓ Document schemas - Maintain data governance
✓ Monitor database - Track performance metrics

Integration Points

This skill integrates with:

• GraphQL Graph Mapping - Expose Neo4j via GraphQL
• Graph Query Optimization - Optimize Cypher queries
• Schema Validation - Validate graph structure
• CSV Graph Loader - Import CSV to Neo4j
• Constraint Generator - Define database constraints
• REST API Wrapper - Expose Neo4j as REST API

Recommended Libraries

Neo4j Python Driver

• neo4j - Official driver (4.x/5.x)
• neomodel - Python ORM for Neo4j
• py2neo - Pythonic interface

Query Building

• cypher-dsl-python - Build Cypher programmatically
• ipython-cypher - Jupyter integration

Data Processing

• pandas - Data frame operations
• polars - Efficient data loading
• networkx - Graph analysis

Visualization

• graphistry - Interactive graph visualization
• pyvis - Network visualization
• neovis.js - Neo4j visualization

Related Skills

• RDF Triple Store Integration - Alternative graph database
• TigerGraph Connector - Distributed graph platform
• JanusGraph Connector - Scalable graph database
• GraphQL Graph Mapping - API layer on Neo4j
• Graph Query Optimization - Improve query performance

---

Version: 1.0.0