File size: 1,642 Bytes
c33d1d0 018fb30 e5633a7 842c848 018fb30 f7493dd e5633a7 c33d1d0 1f5e9cb 018fb30 38e2fac cbed288 18cb8f3 842c848 142d17f 6976271 ea07eae c33d1d0 6976271 142d17f 68b31c9 018fb30 037c950 68b31c9 018fb30 037c950 842c848 018fb30 c33d1d0 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 |
import os
import gradio as gr
from dotenv import load_dotenv
from langchain.vectorstores.faiss import FAISS
from langchain.embeddings import HuggingFaceBgeEmbeddings
from langchain.document_loaders import PyPDFLoader
from langchain.text_splitter import CharacterTextSplitter
# Load environment variables
load_dotenv()
# Load and process the PDF files
loader = PyPDFLoader("./new_papers/ALiBi.pdf")
documents = loader.load()
# Split the documents into chunks and embed them using HuggingFaceBgeEmbeddings
text_splitter = CharacterTextSplitter(chunk_size=100, chunk_overlap=0)
vdocuments = text_splitter.split_documents(documents)
# Extract the text from the Document objects
docs_text = [doc.text for doc in vdocuments]
model = "BAAI/bge-base-en-v1.5"
encode_kwargs = {
"normalize_embeddings": True
} # set True to compute cosine similarity
embeddings = HuggingFaceBgeEmbeddings(
model_name=model, encode_kwargs=encode_kwargs, model_kwargs={"device": "cpu"}
)
# Create FAISS vector store for API embeddings
api_db = FAISS.from_texts(texts=docs_text, embedding=embeddings)
# Define the PDF retrieval function
def pdf_retrieval(query):
# Run the query through the retriever
response = api_db.similarity_search(query)
return response
# Create Gradio interface for the API retriever
api_tool = gr.Interface(
fn=pdf_retrieval,
inputs=[gr.Textbox()],
outputs=gr.Textbox(),
live=True,
title="API PDF Retrieval Tool",
description="This tool indexes PDF documents and retrieves relevant answers based on a given query (HuggingFaceBgeEmbeddings).",
)
# Launch the Gradio interface
api_tool.launch()
|