RAG with LangChain 🦜🔗¶

In [1]:

# requirements for this example:
%pip install -qq docling docling-core python-dotenv langchain-text-splitters langchain-huggingface langchain-milvus

# requirements for this example: %pip install -qq docling docling-core python-dotenv langchain-text-splitters langchain-huggingface langchain-milvus

Note: you may need to restart the kernel to use updated packages.

In [2]:

import os

from dotenv import load_dotenv

load_dotenv()

import os from dotenv import load_dotenv load_dotenv()

Out[2]:

True

Setup¶

Loader and splitter¶

Below we set up:

• a Loader which will be used to create LangChain documents, and
• a splitter, which will be used to split these documents

In [3]:

from typing import Iterator

from langchain_core.document_loaders import BaseLoader
from langchain_core.documents import Document as LCDocument

from docling.document_converter import DocumentConverter

class DoclingPDFLoader(BaseLoader):

    def __init__(self, file_path: str | list[str]) -> None:
        self._file_paths = file_path if isinstance(file_path, list) else [file_path]
        self._converter = DocumentConverter()

    def lazy_load(self) -> Iterator[LCDocument]:
        for source in self._file_paths:
            dl_doc = self._converter.convert(source).document
            text = dl_doc.export_to_markdown()
            yield LCDocument(page_content=text)

from typing import Iterator from langchain_core.document_loaders import BaseLoader from langchain_core.documents import Document as LCDocument from docling.document_converter import DocumentConverter class DoclingPDFLoader(BaseLoader): def __init__(self, file_path: str | list[str]) -> None: self._file_paths = file_path if isinstance(file_path, list) else [file_path] self._converter = DocumentConverter() def lazy_load(self) -> Iterator[LCDocument]: for source in self._file_paths: dl_doc = self._converter.convert(source).document text = dl_doc.export_to_markdown() yield LCDocument(page_content=text)

In [4]:

FILE_PATH = "https://raw.githubusercontent.com/DS4SD/docling/main/tests/data/2206.01062.pdf"  # DocLayNet paper

FILE_PATH = "https://raw.githubusercontent.com/DS4SD/docling/main/tests/data/2206.01062.pdf" # DocLayNet paper

In [5]:

from langchain_text_splitters import RecursiveCharacterTextSplitter

loader = DoclingPDFLoader(file_path=FILE_PATH)
text_splitter = RecursiveCharacterTextSplitter(
    chunk_size=1000,
    chunk_overlap=200,
)

from langchain_text_splitters import RecursiveCharacterTextSplitter loader = DoclingPDFLoader(file_path=FILE_PATH) text_splitter = RecursiveCharacterTextSplitter( chunk_size=1000, chunk_overlap=200, )

We now used the above-defined objects to get the document splits:

In [6]:

docs = loader.load()
splits = text_splitter.split_documents(docs)

docs = loader.load() splits = text_splitter.split_documents(docs)

Embeddings¶

In [7]:

from langchain_huggingface.embeddings import HuggingFaceEmbeddings

HF_EMBED_MODEL_ID = "BAAI/bge-small-en-v1.5"
embeddings = HuggingFaceEmbeddings(model_name=HF_EMBED_MODEL_ID)

from langchain_huggingface.embeddings import HuggingFaceEmbeddings HF_EMBED_MODEL_ID = "BAAI/bge-small-en-v1.5" embeddings = HuggingFaceEmbeddings(model_name=HF_EMBED_MODEL_ID)

Vector store¶

In [8]:

from tempfile import TemporaryDirectory

from langchain_milvus import Milvus

MILVUS_URI = os.environ.get(
    "MILVUS_URI", f"{(tmp_dir := TemporaryDirectory()).name}/milvus_demo.db"
)

vectorstore = Milvus.from_documents(
    splits,
    embeddings,
    connection_args={"uri": MILVUS_URI},
    drop_old=True,
)

from tempfile import TemporaryDirectory from langchain_milvus import Milvus MILVUS_URI = os.environ.get( "MILVUS_URI", f"{(tmp_dir := TemporaryDirectory()).name}/milvus_demo.db" ) vectorstore = Milvus.from_documents( splits, embeddings, connection_args={"uri": MILVUS_URI}, drop_old=True, )

LLM¶

In [9]:

from langchain_huggingface import HuggingFaceEndpoint

HF_API_KEY = os.environ.get("HF_API_KEY")
HF_LLM_MODEL_ID = "mistralai/Mistral-7B-Instruct-v0.3"

llm = HuggingFaceEndpoint(
    repo_id=HF_LLM_MODEL_ID,
    huggingfacehub_api_token=HF_API_KEY,
)

from langchain_huggingface import HuggingFaceEndpoint HF_API_KEY = os.environ.get("HF_API_KEY") HF_LLM_MODEL_ID = "mistralai/Mistral-7B-Instruct-v0.3" llm = HuggingFaceEndpoint( repo_id=HF_LLM_MODEL_ID, huggingfacehub_api_token=HF_API_KEY, )

The token has not been saved to the git credentials helper. Pass `add_to_git_credential=True` in this function directly or `--add-to-git-credential` if using via `huggingface-cli` if you want to set the git credential as well.
Token is valid (permission: write).
Your token has been saved to /Users/pva/.cache/huggingface/token
Login successful

RAG¶

In [10]:

from typing import Iterable

from langchain_core.documents import Document as LCDocument
from langchain_core.output_parsers import StrOutputParser
from langchain_core.prompts import PromptTemplate
from langchain_core.runnables import RunnablePassthrough


def format_docs(docs: Iterable[LCDocument]):
    return "\n\n".join(doc.page_content for doc in docs)


retriever = vectorstore.as_retriever()

prompt = PromptTemplate.from_template(
    "Context information is below.\n---------------------\n{context}\n---------------------\nGiven the context information and not prior knowledge, answer the query.\nQuery: {question}\nAnswer:\n"
)

rag_chain = (
    {"context": retriever | format_docs, "question": RunnablePassthrough()}
    | prompt
    | llm
    | StrOutputParser()
)

from typing import Iterable from langchain_core.documents import Document as LCDocument from langchain_core.output_parsers import StrOutputParser from langchain_core.prompts import PromptTemplate from langchain_core.runnables import RunnablePassthrough def format_docs(docs: Iterable[LCDocument]): return "\n\n".join(doc.page_content for doc in docs) retriever = vectorstore.as_retriever() prompt = PromptTemplate.from_template( "Context information is below.\n---------------------\n{context}\n---------------------\nGiven the context information and not prior knowledge, answer the query.\nQuery: {question}\nAnswer:\n" ) rag_chain = ( {"context": retriever | format_docs, "question": RunnablePassthrough()} | prompt | llm | StrOutputParser() )

In [11]:

rag_chain.invoke("How many pages were human annotated for DocLayNet?")

rag_chain.invoke("How many pages were human annotated for DocLayNet?")

Out[11]:

'- 80,863 pages were human annotated for DocLayNet.'