PDF-Insight: Streamlit Retrieval-Augmented Generation (RAG) 🗞️

This repository contains a Streamlit application designed for performing retrieval-augmented generation on PDF documents using large language models and embeddings from HuggingFace. The application allows users to upload a PDF, ingest its content into a vector store, and query the document using natural language.

Features

• PDF Upload and Display: Upload and view PDF files within the application.
• Data Ingestion: Ingest PDF content into a persistent storage context for efficient querying.
• Query Handling: Perform natural language queries on the ingested PDF content.
• Language Detection: Detects the language of the query to ensure accurate processing.

Setup

Prerequisites

• Python 3.8+
• Streamlit
• HuggingFace Transformers
• Llama Index
• dotenv
• langdetect

Installation

1. Clone the repository: bash git clone https://github.com/Armanasq/RAG-PDF-Streamli.git cd your-repo

2. Install dependencies: bash pip install -r requirements.txt

3. Set up environment variables: Create a .env file in the root directory and add your HuggingFace API token: env HUGGINGFACE_TOKEN=your_huggingface_api_token

Usage

1. Run the Streamlit application: bash streamlit run streamlitRAG.py

2. Upload a PDF:

   • Use the file uploader in the Streamlit interface to upload a PDF.

3. Ingest PDF Content:

   • The content of the uploaded PDF will be ingested and stored for querying.

4. Query the PDF Content:

   • Enter a query related to the PDF content in the text input box.
   • The application will return relevant information based on the content of the PDF.

Code Overview

Environment Setup

• Load Environment Variables: python load_dotenv() hf_token = os.getenv("HUGGINGFACE_TOKEN")

Llama Index Configuration

• Configure Llama Index to use GPU and HuggingFace models: python Settings.llm = HuggingFaceInferenceAPI( model_name="mistralai/Mistral-7B-Instruct-v0.3", tokenizer_name="mistralai/Mistral-7B-Instruct-v0.3", context_window=5000, max_new_tokens=1024, generate_kwargs={"temperature": 0.1}, device=0 # Use GPU 0, set to -1 for CPU ) Settings.embed_model = HuggingFaceEmbedding( model_name="BAAI/bge-small-en-v1.5", device=0 # Use GPU 0, set to -1 for CPU )

Data Handling

• PDF Display Function: python def display_pdf(file_path): with open(file_path, "rb") as f: base64_pdf = base64.b64encode(f.read()).decode('utf-8') pdf_display = f'<iframe src="data:application/pdf;base64,{base64_pdf}" width="100%" height="600" type="application/pdf"></iframe>' return pdf_display

• Data Ingestion: python def data_ingestion(): documents = SimpleDirectoryReader(DATA_DIR).load_data() storage_context = StorageContext.from_defaults() index = VectorStoreIndex.from_documents(documents) index.storage_context.persist(persist_dir=PERSIST_DIR)

Query Handling

• Handle Query Function: ```python def handlequery(query, lang): storagecontext = StorageContext.fromdefaults(persistdir=PERSISTDIR) index = loadindexfromstorage(storagecontext) chattextqamsgs = [ ( "user", """You are a Q&A assistant. Your main goal is to provide answers as accurately as possible, based on the instructions and context you have been given. If a question does not match the provided context or is outside the scope of the document, kindly advise the user to ask questions within the context of the document. Context: {contextstr} Question: {querystr} """ ) ] textqatemplate = ChatPromptTemplate.frommessages(chattextqamsgs)

  query_engine = index.as_query_engine(text_qa_template=text_qa_template)
  answer = query_engine.query(query)
  
  if hasattr(answer, 'response'):
      return answer.response, lang
  elif isinstance(answer, dict) and 'response' in answer:
      return answer['response'], lang
  else:
      return "Sorry, I couldn't find an answer.", lang
  

  ```

Streamlit UI

• Streamlit Interface: ```python st.title("(PDF) Information and Inference 🗞️") st.markdown("## Retrieval-Augmented Generation") st.markdown("Start chat ...🚀")

  uploadedfile = st.fileuploader("Upload your PDF file", type="pdf") query = st.textinput("Ask me anything about the content of the PDF:") chathistory = st.empty()

  if uploadedfile: pdfdisplay = processfile(uploadedfile) st.markdown(pdfdisplay, unsafeallow_html=True)

  if query: lang = detect(query) response, lang = handlequery(query, lang) chathistory.text_area("Chat History", f"User: {query}\nAssistant: {response}", height=300) ```

Contributing

Contributions are welcome. Please fork the repository and submit a pull request for any feature requests or bug fixes.

License

This project is licensed under the MIT License. See the LICENSE file for details.