Multi-Document Agents Pack

This LlamaPack provides an example of our multi-document agents.

This specific template shows the e2e process of building this. Given a set of documents, the pack will build our multi-document agents architecture.

• setup a document agent over agent doc (capable of QA and summarization)
• setup a top-level agent over doc agents
• During query-time, do "tool retrieval" to return the set of relevant candidate documents, and then do retrieval within each document.

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 MultiDocumentAgentsPack --download-dir ./multi_doc_agents_pack

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

Code Usage

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

from llama_index.core.llama_pack import download_llama_pack

# download and install dependencies
MultiDocumentAgentsPack = download_llama_pack(
    "MultiDocumentAgentsPack", "./multi_doc_agents_pack"
)

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

Then, you can set up the pack like so:

# imagine documents on different cities
docs = ...

doc_titles = ["Toronto", "Seattle", "Houston", "Chicago"]
doc_descriptions = [
    "<Toronto description>",
    "<Seattle description>",
    "<Houston description>",
    "<Chicago description>",
]

# create the pack
# get documents from any data loader
multi_doc_agents_pack = MultiDocumentAgentsPack(
    docs, doc_titles, doc_descriptions
)

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

response = multi_doc_agents_pack.run(
    "Tell me the demographics of Houston, and then compare with the demographics of Chicago"
)

You can also use modules individually.

# get the top-level agent
top_agent = multi_doc_agents_pack.top_agent

# get the object index (which indexes all document agents, can return top-k
# most relevant document agents as tools given user query)
obj_index = multi_doc_agents_pack.obj_index

# get document agents
doc_agents = multi_doc_agents_pack.agents