arangoimport 0.1.8

A high-performance tool for importing Neo4j JSONL graph data exports into ArangoDB

ArangoImport

A high-performance tool for importing Neo4j JSONL graph data exports into ArangoDB.

Features

• Import Neo4j database exports into ArangoDB
• Efficient parallel processing of large JSONL files
• Support for both local and Docker ArangoDB instances
• Dynamic memory management and batch sizing
• Connection pooling for optimal performance
• Progress tracking and detailed logging
• Available as both CLI tool and Python package

Installation

pip install arangoimport

Quick Start

1. Export your Neo4j database to JSONL:

   CALL apoc.export.json.all("path/to/export.jsonl", {useTypes: true})
   

2. Import into ArangoDB using either method:

   A. Command Line Interface (CLI)

   After installation, the arangoimport command is available in your terminal:

   # Show help and available options
   arangoimport --help
   
   # Import data with default settings (will prompt for password)
   arangoimport import-data /path/to/neo4j_export.jsonl
   
   # Import with custom settings
   arangoimport import-data /path/to/neo4j_export.jsonl \
       --db-name my_graph \
       --host arangodb.example.com \
       --port 8530 \
       --username graph_user
   

   B. Python API

   from arangoimport.connection import ArangoConfig
   from arangoimport.importer import parallel_load_data
   
   # Configure database connection
   db_config = ArangoConfig(
       host="localhost",
       port=8529,
       username="root",
       password="your_password",  # Or use ARANGO_PASSWORD env var
       db_name="db_name"
   )
   
   # Import the data
   nodes, edges = parallel_load_data(
       "path/to/neo4j_export.jsonl",
       dict(db_config),
       num_processes=None  # None means use (CPU count - 1)
   )
   
   print(f"Successfully imported {nodes:,} nodes and {edges:,} edges!")
   

Environment Variables

• ARANGO_PASSWORD: Database password (avoid hardcoding in scripts)
• ARANGO_USER: Username (default: root)

CLI Options

General Options

• --file <string>: The file to import ("-" for stdin)
• --type <string>: Input format (auto/csv/json/jsonl/tsv, default: auto)
• --collection <string>: Target collection name
• --create-collection <boolean>: Create collection if missing (default: false)
• --create-collection-type <string>: Collection type if created (document/edge, default: document)
• --create-database <boolean>: Create database if missing (default: false)
• --threads <uint32>: Number of parallel import threads (default: 32)
• --batch-size <uint64>: Data batch size in bytes (default: 8MB)
• --progress <boolean>: Show progress (default: true)

Server Connection

• --server.database <string>: Target database (default: "_system")
• --server.endpoint <string>: Server endpoint (default: "http+tcp://127.0.0.1:8529")
• --server.username <string>: Username (default: "root")
• --server.password <string>: Password (prompted if not provided)
• --server.authentication <boolean>: Require authentication (default: true)

Performance Options

• --auto-rate-limit <boolean>: Auto-adjust loading rate (default: false)
• --compress-transfer <boolean>: Compress data transfer (default: false)
• --max-errors <uint64>: Maximum errors before stopping (default: 20)
• --skip-validation <boolean>: Skip schema validation (default: false)

For a complete list of options, run:

arangoimport --help

Docker Support

When using Docker, ensure your ArangoDB container is running:

docker run -p 8529:8529 -e ARANGO_ROOT_PASSWORD=yourpassword arangodb:latest

Then import using either the CLI or Python API, pointing to the exposed port.

Performance Tuning

The importer automatically optimizes for:

• Available system memory
• CPU cores (uses CPU count - 1 by default)
• Network conditions

You can fine-tune performance with:

• --threads: Control parallel threads
• --batch-size: Adjust batch size
• --auto-rate-limit: Enable automatic rate limiting
• --compress-transfer: Enable data compression

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

This project is licensed under the MIT License - see the LICENSE file for details.