RAGatouille

RAGatouille makes it as simple as can be to use ColBERT! ColBERT is a fast and accurate retrieval model, enabling scalable BERT-based search over large text collections in tens of milliseconds.

There are multiple ways that we can use RAGatouille.

Setup

The integration lives in the ragatouille package.

pip install -U ragatouille

from ragatouille import RAGPretrainedModel

RAG = RAGPretrainedModel.from_pretrained("colbert-ir/colbertv2.0")

[Jan 10, 10:53:28] Loading segmented_maxsim_cpp extension (set COLBERT_LOAD_TORCH_EXTENSION_VERBOSE=True for more info)...
``````output
/Users/harrisonchase/.pyenv/versions/3.10.1/envs/langchain/lib/python3.10/site-packages/torch/cuda/amp/grad_scaler.py:125: UserWarning: torch.cuda.amp.GradScaler is enabled, but CUDA is not available.  Disabling.
  warnings.warn(

Retriever

We can use RAGatouille as a retriever. For more information on this, see the RAGatouille Retriever

Document Compressor

We can also use RAGatouille off-the-shelf as a reranker. This will allow us to use ColBERT to rerank retrieved results from any generic retriever. The benefits of this are that we can do this on top of any existing index, so that we don't need to create a new idex. We can do this by using the document compressor abstraction in LangChain.

Setup Vanilla Retriever

First, let's set up a vanilla retriever as an example.

import requests
from langchain_community.vectorstores import FAISS
from langchain_openai import OpenAIEmbeddings
from langchain_text_splitters import RecursiveCharacterTextSplitter


def get_wikipedia_page(title: str):
    """
    Retrieve the full text content of a Wikipedia page.

    :param title: str - Title of the Wikipedia page.
    :return: str - Full text content of the page as raw string.
    """
    # Wikipedia API endpoint
    URL = "https://en.wikipedia.org/w/api.php"

    # Parameters for the API request
    params = {
        "action": "query",
        "format": "json",
        "titles": title,
        "prop": "extracts",
        "explaintext": True,
    }

    # Custom User-Agent header to comply with Wikipedia's best practices
    headers = {"User-Agent": "RAGatouille_tutorial/0.0.1 (ben@clavie.eu)"}

    response = requests.get(URL, params=params, headers=headers)
    data = response.json()

    # Extracting page content
    page = next(iter(data["query"]["pages"].values()))
    return page["extract"] if "extract" in page else None


text = get_wikipedia_page("Hayao_Miyazaki")
text_splitter = RecursiveCharacterTextSplitter(chunk_size=500, chunk_overlap=0)
texts = text_splitter.create_documents([text])

retriever = FAISS.from_documents(texts, OpenAIEmbeddings()).as_retriever(
    search_kwargs={"k": 10}
)

docs = retriever.invoke("What animation studio did Miyazaki found")
docs[0]

Document(page_content='collaborative projects. In April 1984, Miyazaki opened his own office in Suginami Ward, naming it Nibariki.')

We can see that the result isn't super relevant to the question asked

Using ColBERT as a reranker

from langchain.retrievers import ContextualCompressionRetriever

compression_retriever = ContextualCompressionRetriever(
    base_compressor=RAG.as_langchain_document_compressor(), base_retriever=retriever
)

compressed_docs = compression_retriever.invoke(
    "What animation studio did Miyazaki found"
)

/Users/harrisonchase/.pyenv/versions/3.10.1/envs/langchain/lib/python3.10/site-packages/torch/amp/autocast_mode.py:250: UserWarning: User provided device_type of 'cuda', but CUDA is not available. Disabling
  warnings.warn(

compressed_docs[0]

Document(page_content='In June 1985, Miyazaki, Takahata, Tokuma and Suzuki founded the animation production company Studio Ghibli, with funding from Tokuma Shoten. Studio Ghibli\'s first film, Laputa: Castle in the Sky (1986), employed the same production crew of Nausicaä. Miyazaki\'s designs for the film\'s setting were inspired by Greek architecture and "European urbanistic templates". Some of the architecture in the film was also inspired by a Welsh mining town; Miyazaki witnessed the mining strike upon his first', metadata={'relevance_score': 26.5194149017334})

This answer is much more relevant!