LlamaIndex Overview

LlamaIndex

LlamaIndex is a data framework designed to fully leverage LLMs (Language Models) pre-trained on vast amounts of public data.

Why LlamaIndex?

LLMs offer a natural language interface between humans and inferred data. Most models are pre-trained on vast amounts of publicly available data such as Wikipedia, mailing lists, textbooks, and source code. LlamaIndex plays a pivotal role in supplementing these models with private or domain-specific data.

How LlamaIndex Helps?

• Data Connectors: Imports existing data from its native source or format, including APIs, PDFs, SQL, etc.
• Data Index: Structures data into a format that's easy for LLMs to consume.
• Engine: Provides natural language access to the data.

Who is LlamaIndex for?

LlamaIndex offers tools for beginners, advanced users, and everyone in between.

Starting with LlamaIndex

To start using LlamaIndex, simply install it using pip.
Full installation instructions and starter tutorials are also provided.

pip install llama-index

Ecosystem

You can download or contribute to LlamaIndex at:

• Github
• PyPi

Join the LlamaIndex community for assistance or to suggest features:

• Twitter
• Discord

There are related projects such as LlamaHub and LlamaLab.

Quick Code Guide

1. Setting up OpenAI API key:
   

   import os
   os.environ["OPENAI_API_KEY"] = 'YOUR_OPENAI_API_KEY'
   

2. Loading data and creating index:
   

   from llama_index import VectorStoreIndex, SimpleDirectoryReader
   documents = SimpleDirectoryReader('data').load_data()
   index = VectorStoreIndex.from_documents(documents)
   

3. Executing a query:
   

   query_engine = index.as_query_engine()
   query_engine.query("<question_text>?")
   

4. Persisting data:
   

   index.storage_context.persist()
   

5. Loading data from disk:
   

   from llama_index import StorageContext, load_index_from_storage
   storage_context = StorageContext.from_defaults(persist_dir='./storage')
   index = load_index_from_storage(storage_context)
   

Dependencies
The main third-party package requirements for LlamaIndex are tiktoken, openai, and langchain. All requirements can be installed with:

pip install -r requirements.txt