llama-index-packs-auto-merging-retriever 0.5.0

llama-index packs auto_merging_retriever integration

Auto Merging Retriever Pack

This LlamaPack provides an example of our auto-merging retriever.

This specific template shows the e2e process of building this. It loads a document, builds a hierarchical node graph (with bigger parent nodes and smaller child nodes).

Check out the notebook here.

CLI Usage

You can download llamapacks directly using llamaindex-cli, which comes installed with the llama-index python package:

llamaindex-cli download-llamapack AutoMergingRetrieverPack --download-dir ./auto_merging_retriever_pack

You can then inspect the files at ./auto_merging_retriever_pack and use them as a template for your own project.

Code Usage

You can download the pack to a the ./auto_merging_retriever_pack directory:

from llama_index.core.llama_pack import download_llama_pack

# download and install dependencies
AutoMergingRetrieverPack = download_llama_pack(
    "AutoMergingRetrieverPack", "./auto_merging_retriever_pack"
)

From here, you can use the pack, or inspect and modify the pack in ./auto_merging_retriever_pack.

Then, you can set up the pack like so:

# create the pack
# get documents from any data loader
auto_merging_retriever_pack = AutoMergingRetrieverPack(
    documents,
)

The run() function is a light wrapper around query_engine.query().

response = auto_merging_retriever_pack.run(
    "Tell me about what the author did growing up."
)

You can also use modules individually.

# get the node parser
node_parser = auto_merging_retriever_pack.node_parser

# get the retriever
retriever = auto_merging_retriever_pack.retriever

# get the query engine
query_engine = auto_merging_retriever_pack.query_engine