This repository provides end-to-end example notebooks for utilizing Weaviate's features and integrations. It targets developers and researchers looking to implement vector search, hybrid search, generative search, and various data platform and LLM framework integrations. The benefit is accelerated development and exploration of Weaviate's capabilities through practical, runnable code.

How It Works

The repository is structured into three main categories: Integrations, Weaviate Features, and Weaviate Services. Integrations showcase Weaviate's synergy with other technologies like LLM frameworks (LangChain, LlamaIndex), data platforms (Databricks, Airbyte), and cloud providers (AWS, Google, NVIDIA). Weaviate Features cover core functionalities such as similarity, hybrid, and generative search, reranking, media search, and multi-tenancy. Weaviate Services demonstrate built-in agents and embedding generation.

Quick Start & Requirements

• Install: Notebooks typically require installing the weaviate-client Python package.
• Prerequisites: Python environment, Weaviate instance (local or cloud), and potentially API keys for various services (e.g., OpenAI, Cohere). Specific model providers or data sources may have additional dependencies.
• Resources: Running notebooks will depend on the complexity of the example, potentially requiring significant compute for embedding generation or large dataset processing.
• Links: Weaviate Integrations Documentation

Highlighted Details

• Extensive coverage of LLM and Agent Frameworks including LangChain, LlamaIndex, CrewAI, and DSPy.
• Examples for data platforms like Databricks, Confluent, and Spark, as well as data ingestion tools like Airbyte.
• Demonstrations of advanced Weaviate features: multi-vector embeddings, product quantization, and generative search (RAG).
• Integration examples with operations and evaluation tools such as Arize, DeepEval, and Weights & Biases.

Maintenance & Community

This is an ongoing project with frequent updates. Contributions are welcomed via GitHub issues.

Licensing & Compatibility

The repository itself is not explicitly licensed in the provided README. However, it demonstrates usage of Weaviate, which is typically Apache 2.0 licensed. Compatibility for commercial use depends on the licenses of the integrated third-party services and the underlying Weaviate license.

Limitations & Caveats

The README states this is an ongoing project, implying potential for frequent changes and evolving examples. Specific notebook requirements and setup complexity can vary significantly.