Implementation:Run llama Llama index Core Public API

Knowledge Sources	Run_llama_Llama_index
Domains	LLM Framework, Public API
Last Updated	2026-02-11 19:00 GMT

Overview

This module defines the top-level public API surface of the llama-index-core package, re-exporting all major classes, functions, and utilities that users import from llama_index.core.

Description

The __init__.py file serves as the central entry point for the LlamaIndex core library. It aggregates and re-exports key components from various submodules to provide a convenient single-import experience. The file is organized into several logical sections:

• Version detection: Uses importlib.metadata to read the installed package version, falling back to "0.0.0" if the package is not formally installed (e.g., during development or testing).
• Index classes: Re-exports all major index types including VectorStoreIndex, SummaryIndex, TreeIndex, KeywordTableIndex, DocumentSummaryIndex, KnowledgeGraphIndex, PropertyGraphIndex, and their legacy GPT-prefixed aliases.
• Prompt classes: Exports PromptTemplate, ChatPromptTemplate, BasePromptTemplate, SelectorPromptTemplate, and the legacy Prompt alias.
• Storage and context: Exports StorageContext and ServiceContext along with their configuration functions.
• Settings: Exports the global Settings singleton for configuring LLM, embedding model, and other defaults.
• Utilities: Exports SQLDatabase, SQLDocumentContextBuilder, SimpleDirectoryReader, MockEmbedding, tokenizer utilities, and the global callback handler.
• Legacy aliases: Maintains SQLContextBuilder as a backward-compatible alias for SQLDocumentContextBuilder, and includes global mutable variables global_handler and global_tokenizer.

The __all__ list explicitly declares the public API consisting of approximately 50 symbols.

Usage

This module is the primary import target for users of the LlamaIndex library. All top-level imports such as from llama_index.core import VectorStoreIndex, Settings resolve through this file. It should not be modified unless adding or removing a public API symbol.

Code Reference

Source Location

• Repository: Run_llama_Llama_index
• File: llama-index-core/llama_index/core/__init__.py
• Lines: 1-161

Signature

# Module-level exports (no class/function signature; this is a package __init__)
__version__: str  # resolved from importlib.metadata
__all__: list  # ~50 public symbols

global_handler: Optional[BaseCallbackHandler] = None
global_tokenizer: Optional[Callable[[str], list]] = None

Import

from llama_index.core import (
    VectorStoreIndex,
    SummaryIndex,
    TreeIndex,
    Settings,
    StorageContext,
    ServiceContext,
    Document,
    SimpleDirectoryReader,
    load_index_from_storage,
)

I/O Contract

Inputs

Name	Type	Required	Description
N/A	N/A	N/A	This is a module-level init file; it does not accept direct inputs. It re-exports symbols from submodules.

Outputs

Name	Type	Description
__version__	str	The installed version of llama-index-core, or "0.0.0" if not installed.
__all__	list	List of approximately 50 public symbol names exported by this module.
global_handler	Optional[BaseCallbackHandler]	Global callback handler instance, initially None.
global_tokenizer	Optional[Callable]	Global tokenizer function, initially None.

Usage Examples

Basic Usage

from llama_index.core import VectorStoreIndex, SimpleDirectoryReader, Settings

# Configure global settings
Settings.llm = my_llm
Settings.embed_model = my_embed_model

# Load documents and build an index
documents = SimpleDirectoryReader("./data").load_data()
index = VectorStoreIndex.from_documents(documents)

# Query the index
query_engine = index.as_query_engine()
response = query_engine.query("What is the main topic?")
print(response)

Loading a Persisted Index

from llama_index.core import StorageContext, load_index_from_storage

storage_context = StorageContext.from_defaults(persist_dir="./storage")
index = load_index_from_storage(storage_context)

Related Pages

• Environment:Run_llama_Llama_index_Python_LlamaIndex_Core