Skip to content

Dependency Installation Guide

This guide explains how to install GraphBit with different dependency configurations based on your specific needs and use cases.

Overview

GraphBit uses a modular dependency system that allows you to install only the components you need:

  • Core dependencies: Essential packages required for basic GraphBit functionality
  • Optional dependencies: Additional packages grouped by functionality (vector databases, benchmarking tools, or everything graphbit provides support for)
  • Extras groups: Predefined collections of optional dependencies for common use cases

Core Installation (Minimal Dependencies)

The core installation includes only the essential dependencies needed for basic GraphBit functionality.

Installation Commands

pip install graphbit

poetry add graphbit

uv add graphbit

Core Installation Use Cases

  • Lightweight deployments where you only need basic workflow functionality
  • Custom integrations where you'll add your own databases
  • Minimal Docker images for production environments
  • Development environments where you want to add dependencies incrementally

Optional Dependency Groups (Extras)

GraphBit provides several extras groups that bundle related optional dependencies for specific use cases.

Available Extras Groups

benchmark - Benchmarking and Performance Testing

Includes comprehensive tools for performance analysis and AI framework comparisons.

What's included: - Visualization: matplotlib, seaborn, pandas, numpy - Performance monitoring: psutil, memory-profiler, tabulate - AI frameworks: anthropic, openai, langchain, langchain-openai, langchain-anthropic, langchain-ollama, langgraph, llama-index, llama-index-llms-anthropic, llama-index-cli, crewai, pydantic-ai - CLI tools: click

Installation:

pip install "graphbit[benchmark]"

poetry add "graphbit[benchmark]"

uv add "graphbit[benchmark]"

Use cases: - Performance benchmarking against other AI frameworks - Analyzing workflow execution metrics - Comparing different LLM providers - Research and development environments

full - Complete Installation

Includes all available optional dependencies for maximum functionality.

What's included: - All dependencies from benchmark group - All vector database connectors - All AI providers currently supported by GraphBit. - All utility libraries

Installation:

pip install "graphbit[full]"

poetry add "graphbit[full]"

uv add "graphbit[full]"

Use cases: - Development environments where you need access to all features - Prototyping with multiple AI providers and databases - Educational or training environments - One-stop installation for comprehensive GraphBit usage

Vector Database Extras

chromadb - ChromaDB Integration
pip install "graphbit[chromadb]"

poetry add "graphbit[chromadb]"

uv add "graphbit[chromadb]"

Includes: chromadb Use case: Document similarity search, RAG applications

pinecone - Pinecone Integration
pip install "graphbit[pinecone]"

poetry add "graphbit[pinecone]"

uv add "graphbit[pinecone]"

Includes: pinecone Use case: Managed vector database for production applications

qdrant - Qdrant Integration
pip install "graphbit[qdrant]"

poetry add "graphbit[qdrant]"

uv add "graphbit[qdrant]"

Includes: qdrant-client Use case: Open-source vector database with advanced filtering

weaviate - Weaviate Integration
pip install "graphbit[weaviate]"

poetry add "graphbit[weaviate]"

uv add "graphbit[weaviate]"

Includes: weaviate-client Use case: Knowledge graphs and semantic search

milvus - Milvus Integration
pip install "graphbit[milvus]"

poetry add "graphbit[milvus]"

uv add "graphbit[milvus]"

Includes: pymilvus Use case: Large-scale vector similarity search

faiss - FAISS Integration
pip install "graphbit[faiss]"

poetry add "graphbit[faiss]"

uv add "graphbit[faiss]"

Includes: faiss-cpu, numpy Use case: High-performance similarity search and clustering

elasticsearch - Elasticsearch Integration
pip install "graphbit[elasticsearch]"

poetry add "graphbit[elasticsearch]"

uv add "graphbit[elasticsearch]"

Includes: elasticsearch Use case: Full-text search with vector capabilities

Database Extras

pgvector - PostgreSQL with pgvector
pip install "graphbit[pgvector]"

poetry add "graphbit[pgvector]"

uv add "graphbit[pgvector]"

Includes: psycopg2 Use case: PostgreSQL with vector extensions

mariadb - MariaDB Integration
pip install "graphbit[mariadb]"

poetry add "graphbit[mariadb]"

uv add "graphbit[mariadb]"

Includes: mariadb, numpy Use case: MariaDB with vector capabilities

mongodb - MongoDB Integration
pip install "graphbit[mongodb]"

poetry add "graphbit[mongodb]"

uv add "graphbit[mongodb]"

Includes: pymongo Use case: Document database with vector search

db2 - IBM Db2 Integration
pip install "graphbit[db2]"

poetry add "graphbit[db2]"

uv add "graphbit[db2]"

Includes: ibm-db, numpy Use case: Enterprise IBM Db2 database

Cloud Provider Extras

boto3 - Amazon Web Services
pip install "graphbit[boto3]"

poetry add "graphbit[boto3]"

uv add "graphbit[boto3]"

Includes: boto3 Use case: AWS services integration

astradb - DataStax Astra DB
pip install "graphbit[astradb]"

poetry add "graphbit[astradb]"

uv add "graphbit[astradb]"

Includes: astrapy Use case: Managed Cassandra with vector search

Multiple Extras Installation

You can install multiple extras groups simultaneously:

# Install multiple specific extras
pip install "graphbit[chromadb,boto3]"

# Install benchmark tools with specific database
pip install "graphbit[benchmark,pgvector]"

# Install full AI stack with vector databases
pip install "graphbit[benchmark,chromadb,pinecone,qdrant]"

# Install multiple specific extras
poetry add "graphbit[chromadb,boto3]"

# Install benchmark tools with specific database
poetry add "graphbit[benchmark,pgvector]"

# Install full AI stack with vector databases
poetry add "graphbit[benchmark,chromadb,pinecone,qdrant]"

# Install multiple specific extras
uv add "graphbit[chromadb,boto3]"

# Install benchmark tools with specific database
uv add "graphbit[benchmark,pgvector]"

# Install full AI stack with vector databases
uv add "graphbit[benchmark,chromadb,pinecone,qdrant]"

Installation Scenarios by Use Case

Scenario 1:

# Minimal installation for simple workflows
pip install graphbit

# Minimal installation for simple workflows
poetry add graphbit

# Minimal installation for simple workflows
uv add graphbit

Scenario 2: RAG Application Development

# Vector database
pip install "graphbit[chromadb]"

# Alternative with managed vector DB
pip install "graphbit[pinecone]"

# Vector database
poetry add "graphbit[chromadb]"

# Alternative with managed vector DB
poetry add "graphbit[pinecone]"

# Vector database
uv add "graphbit[chromadb]"

# Alternative with managed vector DB
uv add "graphbit[pinecone]"

Scenario 3: Enterprise Production Environment

# PostgreSQL + Boto3
pip install "graphbit[pgvector,boto3]"

# IBM enterprise stack
pip install "graphbit[db2,boto3]"

# PostgreSQL + Boto3
poetry add "graphbit[pgvector,boto3]"

# IBM enterprise stack
poetry add "graphbit[db2,boto3]"

# PostgreSQL + Boto3
uv add "graphbit[pgvector,boto3]"

# IBM enterprise stack
uv add "graphbit[db2,boto3]"

Scenario 4: Research and Benchmarking

# Full benchmarking suite
pip install "graphbit[benchmark]"

# Comprehensive research environment
pip install "graphbit[full]"

# Full benchmarking suite
poetry add "graphbit[benchmark]"

# Comprehensive research environment
poetry add "graphbit[full]"

# Full benchmarking suite
uv add "graphbit[benchmark]"

# Comprehensive research environment
uv add "graphbit[full]"

Scenario 5: Multi-Cloud Development

# Multiple vector databases for testing
pip install "graphbit[chromadb,pinecone,qdrant,weaviate]"

# Multiple vector databases for testing
poetry add "graphbit[chromadb,pinecone,qdrant,weaviate]"

# Multiple vector databases for testing
uv add "graphbit[chromadb,pinecone,qdrant,weaviate]"

Dependency Comparison Table

Extra Group AI Providers Vector DBs Visualization Performance Cloud Size
Core ✅ Built-in ❌ ❌ ❌ ❌ Minimal
benchmark ✅ All + External Frameworks ❌ ✅ Full ✅ Full ❌ Large
chromadb ❌ ChromaDB ❌ ❌ ❌ Small
pinecone ❌ Pinecone ❌ ❌ ❌ Small
pgvector ❌ PostgreSQL ❌ ❌ ❌ Small
boto3 ❌ ❌ ❌ ❌ AWS Small
full ✅ All ✅ All ✅ Full ✅ Full ✅ All Very Large

Development vs Production Recommendations

Development Environment

# Option 1: Start minimal, add as needed
pip install graphbit
pip install "graphbit[chromadb]"  # Add vector DB when needed

# Option 2: Full development stack
pip install "graphbit[benchmark]"  # Includes AI providers + tools

# Option 3: Complete development environment
pip install "graphbit[full]"  # Everything available

# Option 1: Start minimal, add as needed
poetry add graphbit
poetry add "graphbit[chromadb]"  # Add vector DB when needed

# Option 2: Full development stack
poetry add "graphbit[benchmark]"  # Includes AI providers + tools

# Option 3: Complete development environment
poetry add "graphbit[full]"  # Everything available

# Option 1: Start minimal, add as needed
uv add graphbit
uv add "graphbit[chromadb]"  # Add vector DB when needed

# Option 2: Full development stack
uv add "graphbit[benchmark]"  # Includes AI providers + tools

# Option 3: Complete development environment
uv add "graphbit[full]"  # Everything available

Production Environment

# Minimal production deployment
pip install "graphbit[pgvector,boto3]"  # Only what you need

# Specific use case
pip install "graphbit[pinecone]"  # Managed services only

# Minimal production deployment
poetry add "graphbit[pgvector,boto3]"  # Only what you need

# Specific use case
poetry add "graphbit[pinecone]"  # Managed services only

# Minimal production deployment
uv add "graphbit[pgvector,boto3]"  # Only what you need

# Specific use case
uv add "graphbit[pinecone]"  # Managed services only

Troubleshooting Installation Issues

Common Issues and Solutions

1. Large Installation Size

Problem: pip install "graphbit[full]" downloads too many packages Solution: Use specific extras instead

# Instead of full, use specific combinations
pip install "graphbit[chromadb,boto3]"

# Instead of full, use specific combinations
poetry add "graphbit[chromadb,boto3]"

# Instead of full, use specific combinations
uv add "graphbit[chromadb,boto3]"

2. Dependency Conflicts

Problem: Version conflicts between optional dependencies Solution: Use virtual environments and specific versions

# Create isolated environment
python -m venv graphbit-env
source graphbit-env/bin/activate  # Linux/macOS
pip install "graphbit[your-extras]"

# Create isolated environment
python -m venv graphbit-env
source graphbit-env/bin/activate  # Linux/macOS
poetry add "graphbit[your-extras]"

# Create isolated environment
python -m venv graphbit-env
source graphbit-env/bin/activate  # Linux/macOS
uv add "graphbit[your-extras]"

3. Database Driver Installation Failures

Problem: Native database drivers fail to compile Solutions:

PostgreSQL (psycopg2):

# Install system dependencies first
# Ubuntu/Debian
sudo apt-get install libpq-dev python3-dev

# macOS
brew install postgresql

# Then install GraphBit
pip install "graphbit[pgvector]"

# Install system dependencies first
# Ubuntu/Debian
sudo apt-get install libpq-dev python3-dev

# macOS
brew install postgresql

# Then install GraphBit
poetry add "graphbit[pgvector]"

# Install system dependencies first
# Ubuntu/Debian
sudo apt-get install libpq-dev python3-dev

# macOS
brew install postgresql

# Then install GraphBit
uv add "graphbit[pgvector]"

IBM Db2:

# Requires IBM Db2 client libraries
# Follow IBM Db2 client installation guide first
pip install "graphbit[db2]"

# Requires IBM Db2 client libraries
# Follow IBM Db2 client installation guide first
poetry add "graphbit[db2]"

# Requires IBM Db2 client libraries
# Follow IBM Db2 client installation guide first
uv add "graphbit[db2]"

4. Memory Issues During Installation

Problem: Installation runs out of memory Solution: Install dependencies separately

# Install heavy dependencies first
pip install numpy pandas matplotlib
pip install "graphbit[benchmark]"

# Install heavy dependencies first
poetry add numpy pandas matplotlib
poetry add "graphbit[benchmark]"

# Install heavy dependencies first
uv add numpy pandas matplotlib
uv add "graphbit[benchmark]"

5. Network/Proxy Issues

Problem: Cannot download packages from PyPI Solution: Configure pip for proxy/private PyPI

# Use proxy
pip install "graphbit[extras]"

# Use private PyPI
pip install -i "graphbit[extras]"

# Use proxy
poetry add "graphbit[extras]"

# Use private PyPI
poetry add -i "graphbit[extras]"

# Use proxy
uv add "graphbit[extras]"

# Use private PyPI
uv add -i "graphbit[extras]"

Verification Commands

After installation, verify your setup:

from graphbit import init, version, health_check

# Initialize and check system
init()
print(f"GraphBit version: {version()}")

# Check system health
health = health_check()
print(f"System healthy: {health['overall_healthy']}")

Getting Help

If you encounter installation issues:

  1. Check system requirements: Ensure Python 3.10+ and sufficient disk space
  2. Update pip: pip install --upgrade pip
  3. Use virtual environments: Isolate GraphBit dependencies
  4. Check the Installation Guide for system-specific instructions
  5. Search GitHub Issues for similar problems
  6. Create a new issue with:
  7. Your operating system and Python version
  8. Complete error message
  9. Installation command used
  10. Output of pip list and python --version

Next Steps

Once you have GraphBit installed with the appropriate dependencies:

  1. Basic usage: Follow the Quick Start Tutorial
  2. Configuration: Set up your API keys and environment
  3. Examples: Explore use case examples for your specific installation
  4. Development: See the Contributing Guide for development setup

Choose the installation method that best fits your use case. You can always add more extras later as your needs evolve!