🧠 Intelligent Document Q&A with LangChain, FAISS & HuggingFace

Ask questions from your PDF, DOCX, or TXT files using powerful open-source tools — all in Google Colab.

---

📌 Overview

This project allows you to upload a document and ask it natural language questions. Behind the scenes, it uses:

• 🧩 LangChain – to handle document loading, text splitting, and chaining logic
• 🔍 FAISS – to store and retrieve semantically similar chunks
• 🤗 HuggingFace Transformers – to run a local language model for question-answering
• ⚡ Google Colab – no setup required, runs in the cloud

---

🚀 Features

✅ Upload PDF, DOCX, or TXT
✅ Extract and chunk document contents
✅ Embed chunks using sentence‑transformers
✅ Store & search in a FAISS vector index
✅ Use a small, efficient language model (Flan‑T5) for answers
✅ Interactive Q&A loop

---

🛠️ Installation (Google Colab)

!pip install langchain langchain-community
!pip install faiss-cpu
!pip install pypdf python-docx
!pip install sentence-transformers
!pip install transformers

---

📤 Upload Your Document

from google.colab import files

uploaded = files.upload()
file_path = list(uploaded.keys())[0]
print("Uploaded:", file_path)

---

📄 Load & Split the Document

from langchain_community.document_loaders import PyPDFLoader, TextLoader, Docx2txtLoader
from langchain.text_splitter import RecursiveCharacterTextSplitter

# Choose loader based on file type
if file_path.endswith(".pdf"):
    loader = PyPDFLoader(file_path)
elif file_path.endswith(".docx"):
    loader = Docx2txtLoader(file_path)
else:
    loader = TextLoader(file_path)

docs = loader.load()

# Split into chunks
splitter = RecursiveCharacterTextSplitter(chunk_size=500, chunk_overlap=100)
documents = splitter.split_documents(docs)

print(f"Total Chunks: {len(documents)}")

---

🧠 Generate Embeddings & Build Vector Store

from langchain_community.embeddings import HuggingFaceEmbeddings
from langchain_community.vectorstores import FAISS

embeddings = HuggingFaceEmbeddings(model_name="sentence-transformers/all-MiniLM-L6-v2")
vectorstore = FAISS.from_documents(documents, embeddings)

---

🤖 Load Local Language Model

from transformers import pipeline
from langchain_community.llms import HuggingFacePipeline

flan_pipeline = pipeline(
    "text2text-generation",
    model="google/flan-t5-base",  # Swap to flan-t5-small for faster inference
    max_length=512
)

llm = HuggingFacePipeline(pipeline=flan_pipeline)

---

🔗 Set Up QA Chain

from langchain.chains import RetrievalQA

qa = RetrievalQA.from_chain_type(
    llm=llm,
    retriever=vectorstore.as_retriever(search_kwargs={"k": 3}),
    chain_type="stuff"
)

---

❓ Ask Questions

One-Off Query

query = "Give me a short summary of the document"
print(qa.run(query))

Interactive Q&A Loop

while True:
    q = input("Ask a question (or 'exit'): ")
    if q.lower() == "exit":
        break
    print("Answer:", qa.run(q))

---

🧪 Sample Output

Ask a question (or 'exit'): what is the importance of a healthy diet?
Answer: helps to protect against malnutrition in all its forms, as well as noncommunicable diseases (NCDs) such as diabetes, heart disease, stroke and cancer

---

📁 File Types Supported

• .pdf – parsed using PyPDFLoader
• .docx – parsed using Docx2txtLoader
• .txt – parsed using TextLoader

---