# 🗂️ LlamaIndex 🦙
[](https://pypi.org/project/llama-index/)
[](https://github.com/jerryjliu/llama_index/graphs/contributors)
[](https://discord.gg/dGcwcsnxhU)
LlamaIndex (GPT Index) is a data framework for your LLM application.
PyPI:
- LlamaIndex: https://pypi.org/project/llama-index/.
- GPT Index (duplicate): https://pypi.org/project/gpt-index/.
LlamaIndex.TS (Typescript/Javascript): https://github.com/run-llama/LlamaIndexTS.
Documentation: https://docs.llamaindex.ai/en/stable/.
Twitter: https://twitter.com/llama_index.legacy.
Discord: https://discord.gg/dGcwcsnxhU.
### Ecosystem
- LlamaHub (community library of data loaders): https://llamahub.ai.
- LlamaLab (cutting-edge AGI projects using LlamaIndex): https://github.com/run-llama/llama-lab.
## 🚀 Overview
**NOTE**: This README is not updated as frequently as the documentation. Please check out the documentation above for the latest updates!
### Context
- LLMs are a phenomenal piece of technology for knowledge generation and reasoning. They are pre-trained on large amounts of publicly available data.
- How do we best augment LLMs with our own private data?
We need a comprehensive toolkit to help perform this data augmentation for LLMs.
### Proposed Solution
That's where **LlamaIndex** comes in. LlamaIndex is a "data framework" to help you build LLM apps. It provides the following tools:
- Offers **data connectors** to ingest your existing data sources and data formats (APIs, PDFs, docs, SQL, etc.).
- Provides ways to **structure your data** (indices, graphs) so that this data can be easily used with LLMs.
- Provides an **advanced retrieval/query interface over your data**: Feed in any LLM input prompt, get back retrieved context and knowledge-augmented output.
- Allows easy integrations with your outer application framework (e.g. with LangChain, Flask, Docker, ChatGPT, anything else).
LlamaIndex provides tools for both beginner users and advanced users. Our high-level API allows beginner users to use LlamaIndex to ingest and query their data in
5 lines of code. Our lower-level APIs allow advanced users to customize and extend any module (data connectors, indices, retrievers, query engines, reranking modules),
to fit their needs.
## 💡 Contributing
Interested in contributing? See our [Contribution Guide](CONTRIBUTING.md) for more details.
## 📄 Documentation
Full documentation can be found here: https://docs.llamaindex.ai/en/latest/.
Please check it out for the most up-to-date tutorials, how-to guides, references, and other resources!
## 💻 Example Usage
```
pip install llama-index
```
Examples are in the `examples` folder. Indices are in the `indices` folder (see list of indices below).
To build a simple vector store index using OpenAI:
```python
import os
os.environ["OPENAI_API_KEY"] = "YOUR_OPENAI_API_KEY"
from llama_index.legacy.legacy import VectorStoreIndex, SimpleDirectoryReader
documents = SimpleDirectoryReader("YOUR_DATA_DIRECTORY").load_data()
index = VectorStoreIndex.from_documents(documents)
```
To build a simple vector store index using non-OpenAI LLMs, e.g. Llama 2 hosted on [Replicate](https://replicate.com/), where you can easily create a free trial API token:
```python
import os
os.environ["REPLICATE_API_TOKEN"] = "YOUR_REPLICATE_API_TOKEN"
from llama_index.legacy.legacy.llms import Replicate
llama2_7b_chat = "meta/llama-2-7b-chat:8e6975e5ed6174911a6ff3d60540dfd4844201974602551e10e9e87ab143d81e"
llm = Replicate(
model=llama2_7b_chat,
temperature=0.01,
additional_kwargs={"top_p": 1, "max_new_tokens": 300},
)
# set tokenizer to match LLM
from llama_index.legacy.legacy import set_global_tokenizer
from transformers import AutoTokenizer
set_global_tokenizer(
AutoTokenizer.from_pretrained("NousResearch/Llama-2-7b-chat-hf").encode
)
from llama_index.legacy.legacy.embeddings import HuggingFaceEmbedding
from llama_index.legacy.legacy import ServiceContext
embed_model = HuggingFaceEmbedding(model_name="BAAI/bge-small-en-v1.5")
service_context = ServiceContext.from_defaults(
llm=llm, embed_model=embed_model
)
from llama_index.legacy.legacy import VectorStoreIndex, SimpleDirectoryReader
documents = SimpleDirectoryReader("YOUR_DATA_DIRECTORY").load_data()
index = VectorStoreIndex.from_documents(
documents, service_context=service_context
)
```
To query:
```python
query_engine = index.as_query_engine()
query_engine.query("YOUR_QUESTION")
```
By default, data is stored in-memory.
To persist to disk (under `./storage`):
```python
index.storage_context.persist()
```
To reload from disk:
```python
from llama_index.legacy.legacy import StorageContext, load_index_from_storage
# rebuild storage context
storage_context = StorageContext.from_defaults(persist_dir="./storage")
# load index
index = load_index_from_storage(storage_context)
```
## 🔧 Dependencies
The main third-party package requirements are `tiktoken`, `openai`, and `langchain`.
All requirements should be contained within the `setup.py` file.
To run the package locally without building the wheel, simply run:
```bash
pip install poetry
poetry install --with dev
```
## 📖 Citation
Reference to cite if you use LlamaIndex in a paper:
```
@software{Liu_LlamaIndex_2022,
author = {Liu, Jerry},
doi = {10.5281/zenodo.1234},
month = {11},
title = {{LlamaIndex}},
url = {https://github.com/jerryjliu/llama_index},
year = {2022}
}
```
Raw data
{
"_id": null,
"home_page": "https://llamaindex.ai",
"name": "llama-index-legacy",
"maintainer": "Andrei Fajardo",
"docs_url": null,
"requires_python": "<4.0,>=3.8.1",
"maintainer_email": "andrei@runllama.ai",
"keywords": "LLM, NLP, RAG, data, devtools, index, retrieval",
"author": "Jerry Liu",
"author_email": "jerry@llamaindex.ai",
"download_url": "https://files.pythonhosted.org/packages/35/93/b07dbbee6d1e57dc389d8e66e549dad9205cf569cf7f78fa8ef839598711/llama_index_legacy-0.9.48.post3.tar.gz",
"platform": null,
"description": "# \ud83d\uddc2\ufe0f LlamaIndex \ud83e\udd99\n\n[](https://pypi.org/project/llama-index/)\n[](https://github.com/jerryjliu/llama_index/graphs/contributors)\n[](https://discord.gg/dGcwcsnxhU)\n\nLlamaIndex (GPT Index) is a data framework for your LLM application.\n\nPyPI:\n\n- LlamaIndex: https://pypi.org/project/llama-index/.\n- GPT Index (duplicate): https://pypi.org/project/gpt-index/.\n\nLlamaIndex.TS (Typescript/Javascript): https://github.com/run-llama/LlamaIndexTS.\n\nDocumentation: https://docs.llamaindex.ai/en/stable/.\n\nTwitter: https://twitter.com/llama_index.legacy.\n\nDiscord: https://discord.gg/dGcwcsnxhU.\n\n### Ecosystem\n\n- LlamaHub (community library of data loaders): https://llamahub.ai.\n- LlamaLab (cutting-edge AGI projects using LlamaIndex): https://github.com/run-llama/llama-lab.\n\n## \ud83d\ude80 Overview\n\n**NOTE**: This README is not updated as frequently as the documentation. Please check out the documentation above for the latest updates!\n\n### Context\n\n- LLMs are a phenomenal piece of technology for knowledge generation and reasoning. They are pre-trained on large amounts of publicly available data.\n- How do we best augment LLMs with our own private data?\n\nWe need a comprehensive toolkit to help perform this data augmentation for LLMs.\n\n### Proposed Solution\n\nThat's where **LlamaIndex** comes in. LlamaIndex is a \"data framework\" to help you build LLM apps. It provides the following tools:\n\n- Offers **data connectors** to ingest your existing data sources and data formats (APIs, PDFs, docs, SQL, etc.).\n- Provides ways to **structure your data** (indices, graphs) so that this data can be easily used with LLMs.\n- Provides an **advanced retrieval/query interface over your data**: Feed in any LLM input prompt, get back retrieved context and knowledge-augmented output.\n- Allows easy integrations with your outer application framework (e.g. with LangChain, Flask, Docker, ChatGPT, anything else).\n\nLlamaIndex provides tools for both beginner users and advanced users. Our high-level API allows beginner users to use LlamaIndex to ingest and query their data in\n5 lines of code. Our lower-level APIs allow advanced users to customize and extend any module (data connectors, indices, retrievers, query engines, reranking modules),\nto fit their needs.\n\n## \ud83d\udca1 Contributing\n\nInterested in contributing? See our [Contribution Guide](CONTRIBUTING.md) for more details.\n\n## \ud83d\udcc4 Documentation\n\nFull documentation can be found here: https://docs.llamaindex.ai/en/latest/.\n\nPlease check it out for the most up-to-date tutorials, how-to guides, references, and other resources!\n\n## \ud83d\udcbb Example Usage\n\n```\npip install llama-index\n```\n\nExamples are in the `examples` folder. Indices are in the `indices` folder (see list of indices below).\n\nTo build a simple vector store index using OpenAI:\n\n```python\nimport os\n\nos.environ[\"OPENAI_API_KEY\"] = \"YOUR_OPENAI_API_KEY\"\n\nfrom llama_index.legacy.legacy import VectorStoreIndex, SimpleDirectoryReader\n\ndocuments = SimpleDirectoryReader(\"YOUR_DATA_DIRECTORY\").load_data()\nindex = VectorStoreIndex.from_documents(documents)\n```\n\nTo build a simple vector store index using non-OpenAI LLMs, e.g. Llama 2 hosted on [Replicate](https://replicate.com/), where you can easily create a free trial API token:\n\n```python\nimport os\n\nos.environ[\"REPLICATE_API_TOKEN\"] = \"YOUR_REPLICATE_API_TOKEN\"\n\nfrom llama_index.legacy.legacy.llms import Replicate\n\nllama2_7b_chat = \"meta/llama-2-7b-chat:8e6975e5ed6174911a6ff3d60540dfd4844201974602551e10e9e87ab143d81e\"\nllm = Replicate(\n model=llama2_7b_chat,\n temperature=0.01,\n additional_kwargs={\"top_p\": 1, \"max_new_tokens\": 300},\n)\n\n# set tokenizer to match LLM\nfrom llama_index.legacy.legacy import set_global_tokenizer\nfrom transformers import AutoTokenizer\n\nset_global_tokenizer(\n AutoTokenizer.from_pretrained(\"NousResearch/Llama-2-7b-chat-hf\").encode\n)\n\nfrom llama_index.legacy.legacy.embeddings import HuggingFaceEmbedding\nfrom llama_index.legacy.legacy import ServiceContext\n\nembed_model = HuggingFaceEmbedding(model_name=\"BAAI/bge-small-en-v1.5\")\nservice_context = ServiceContext.from_defaults(\n llm=llm, embed_model=embed_model\n)\n\nfrom llama_index.legacy.legacy import VectorStoreIndex, SimpleDirectoryReader\n\ndocuments = SimpleDirectoryReader(\"YOUR_DATA_DIRECTORY\").load_data()\nindex = VectorStoreIndex.from_documents(\n documents, service_context=service_context\n)\n```\n\nTo query:\n\n```python\nquery_engine = index.as_query_engine()\nquery_engine.query(\"YOUR_QUESTION\")\n```\n\nBy default, data is stored in-memory.\nTo persist to disk (under `./storage`):\n\n```python\nindex.storage_context.persist()\n```\n\nTo reload from disk:\n\n```python\nfrom llama_index.legacy.legacy import StorageContext, load_index_from_storage\n\n# rebuild storage context\nstorage_context = StorageContext.from_defaults(persist_dir=\"./storage\")\n# load index\nindex = load_index_from_storage(storage_context)\n```\n\n## \ud83d\udd27 Dependencies\n\nThe main third-party package requirements are `tiktoken`, `openai`, and `langchain`.\n\nAll requirements should be contained within the `setup.py` file.\nTo run the package locally without building the wheel, simply run:\n\n```bash\npip install poetry\npoetry install --with dev\n```\n\n## \ud83d\udcd6 Citation\n\nReference to cite if you use LlamaIndex in a paper:\n\n```\n@software{Liu_LlamaIndex_2022,\nauthor = {Liu, Jerry},\ndoi = {10.5281/zenodo.1234},\nmonth = {11},\ntitle = {{LlamaIndex}},\nurl = {https://github.com/jerryjliu/llama_index},\nyear = {2022}\n}\n```\n",
"bugtrack_url": null,
"license": "MIT",
"summary": "Interface between LLMs and your data",
"version": "0.9.48.post3",
"project_urls": {
"Documentation": "https://docs.llamaindex.ai/en/stable/",
"Homepage": "https://llamaindex.ai",
"Repository": "https://github.com/run-llama/llama_index"
},
"split_keywords": [
"llm",
" nlp",
" rag",
" data",
" devtools",
" index",
" retrieval"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "8440f8e42a3c569031f888c46bcdf12b79736d8728c77c4290ed348feaba79ea",
"md5": "21ffcfcebbf6cc721ede073e98b383c3",
"sha256": "04221320d84d96ba9ee3e21e5055bd8527cbd769e8f1c60cf0368ed907e012a2"
},
"downloads": -1,
"filename": "llama_index_legacy-0.9.48.post3-py3-none-any.whl",
"has_sig": false,
"md5_digest": "21ffcfcebbf6cc721ede073e98b383c3",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": "<4.0,>=3.8.1",
"size": 1200711,
"upload_time": "2024-08-18T16:35:21",
"upload_time_iso_8601": "2024-08-18T16:35:21.951067Z",
"url": "https://files.pythonhosted.org/packages/84/40/f8e42a3c569031f888c46bcdf12b79736d8728c77c4290ed348feaba79ea/llama_index_legacy-0.9.48.post3-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "3593b07dbbee6d1e57dc389d8e66e549dad9205cf569cf7f78fa8ef839598711",
"md5": "84983c9c8aec37f03e27e05c95e45c6c",
"sha256": "f6969f1085efb0abebd6367e46f3512020f3f6b9c086f458a519830dd61e8206"
},
"downloads": -1,
"filename": "llama_index_legacy-0.9.48.post3.tar.gz",
"has_sig": false,
"md5_digest": "84983c9c8aec37f03e27e05c95e45c6c",
"packagetype": "sdist",
"python_version": "source",
"requires_python": "<4.0,>=3.8.1",
"size": 781593,
"upload_time": "2024-08-18T16:35:23",
"upload_time_iso_8601": "2024-08-18T16:35:23.854042Z",
"url": "https://files.pythonhosted.org/packages/35/93/b07dbbee6d1e57dc389d8e66e549dad9205cf569cf7f78fa8ef839598711/llama_index_legacy-0.9.48.post3.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2024-08-18 16:35:23",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "run-llama",
"github_project": "llama_index",
"travis_ci": false,
"coveralls": false,
"github_actions": true,
"lcname": "llama-index-legacy"
}
JSON
