Fresh initial commit

.gitattributes

@@ -0,0 +1,2 @@
+*.gif filter=lfs diff=lfs merge=lfs -text
+*.pdf filter=lfs diff=lfs merge=lfs -text

.gitignore

@@ -0,0 +1 @@
+*.ipynb

LICENSE.md

@@ -0,0 +1,19 @@
+Copyright (c) 2025 Muzenda-K
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

README.md

@@ -0,0 +1,77 @@
+# PDF Chatbot with LLaMA
+
+![Python 3.12](https://img.shields.io/badge/python-3.12-blue?logo=python&logoColor=white)
+![License](https://img.shields.io/badge/license-MIT-green)
+![Streamlit](https://img.shields.io/badge/Streamlit-1.45.1-FF4B4B)
+![PyTorch](https://img.shields.io/badge/PyTorch-2.7.0-EE4C2C)
+![Transformers](https://img.shields.io/badge/Transformers-4.52.4-yellow)
+
+## Overview
+
+A powerful PDF chatbot application that allows users to upload PDF documents and ask questions about their content. Built with Streamlit for the frontend and leveraging LLaMA-based models for natural language processing, this application provides an intuitive interface for document-based question answering.
+
+## Features
+
+📄 Upload and process PDF documents
+
+💬 Chat interface for asking questions about document content
+
+⚡ Fast response generation using LLaMA-based models
+
+🧠 Context-aware answers based on document content
+
+🎨 Clean, user-friendly interface
+
+🔍 Sample PDF with demo questions included
+
+## Installation
+
+1. Clone the repository:
+
+```bash
+git clone https://github.com/Muzenda-K/PDF-Chatbot.git
+cd pdf-chatbot
+```
+
+2. Create and activate a virtual environment (recommended):
+
+```bash
+python -m venv venv
+source venv/bin/activate  # On Windows use `venv\Scripts\activate`
+```
+
+3. Install the required dependencies:
+
+```bash
+pip install -r requirements.txt
+```
+
+## Usage
+
+1. Run the Streamlit application:
+
+```bash
+streamlit run app.py
+```
+
+2. The application will open in your default browser at `http://localhost:8501`
+3. Either use the provided sample PDF or upload your own document
+4. Start asking questions about the document content
+
+## Project demo
+
+![Demo](demo.gif)
+
+## Contributing
+
+Contributions are welcome! Please follow these steps:
+
+1. Fork the repository
+2. Create your feature branch (git checkout -b feature/AmazingFeature)
+3. Commit your changes (git commit -m 'Add some AmazingFeature')
+4. Push to the branch (git push origin feature/AmazingFeature)
+5. Open a Pull Request
+
+## License
+
+Distributed under the MIT License.

app (1).py

@@ -0,0 +1,162 @@
+import streamlit as st
+from pdf_utils import extract_text, chunk_text
+from vector_store import build_index
+from tiny_llama import answer_query
+
+SAMPLE_PDF_PATH = "sample.pdf"
+SAMPLE_QUESTION = "What is this document about?"
+
+st.set_page_config(page_title="PDF Chatbot", page_icon="📄")
+st.title("📄 Chat with your PDF (LLaMA-based)")
+
+# ---------------------- Initialize Session State ----------------------
+if "messages" not in st.session_state:
+    st.session_state.messages = []
+if "index" not in st.session_state:
+    st.session_state.index = None
+if "chunks" not in st.session_state:
+    st.session_state.chunks = None
+if "pending_question" not in st.session_state:
+    st.session_state.pending_question = None
+if "pdf_name" not in st.session_state:
+    st.session_state.pdf_name = None
+if "using_sample" not in st.session_state:
+    st.session_state.using_sample = False
+if "sample_processed" not in st.session_state:
+    st.session_state.sample_processed = False
+
+# ---------------------- Sample PDF Load ----------------------
+if not st.session_state.index and not st.session_state.pdf_name and not st.session_state.sample_processed:
+    with st.spinner("Loading sample PDF and preparing demo..."):
+        try:
+            with open(SAMPLE_PDF_PATH, "rb") as f:
+                text = extract_text(f)
+                if text:
+                    chunks = chunk_text(text)
+                    index, _ = build_index(chunks)
+                    
+                    st.session_state.index = index
+                    st.session_state.chunks = chunks
+                    st.session_state.pdf_name = "Sample PDF"
+                    st.session_state.using_sample = True
+                    st.session_state.sample_processed = True
+                    st.session_state.messages = []
+                    
+                    # Add sample question
+                    st.session_state.messages.append({
+                        "role": "user", 
+                        "content": SAMPLE_QUESTION,
+                        "is_sample": True
+                    })
+                    
+                    # Generate actual answer from the sample PDF
+                    with st.spinner("Generating sample answer..."):
+                        answer = answer_query(SAMPLE_QUESTION, index, chunks)
+                        if not answer:
+                            answer = "I couldn't generate an answer from this document."
+                            
+                    st.session_state.messages.append({
+                        "role": "assistant", 
+                        "content": answer,
+                        "is_sample": True
+                    })
+                else:
+                    st.warning("Could not extract text from sample PDF.")
+        except FileNotFoundError:
+            st.warning("Sample PDF not found. Please upload your own.")
+        except Exception as e:
+            st.error(f"Error loading sample PDF: {str(e)}")
+
+# ---------------------- PDF Upload ----------------------
+uploaded = st.file_uploader("Upload your PDF", type=["pdf"])
+if uploaded is not None:
+    # Reset everything if uploading a new PDF
+    if st.session_state.pdf_name != uploaded.name:
+        with st.spinner("Processing uploaded PDF..."):
+            try:
+                text = extract_text(uploaded)
+                if text:
+                    chunks = chunk_text(text)
+                    index, _ = build_index(chunks)
+                    
+                    st.session_state.index = index
+                    st.session_state.chunks = chunks
+                    st.session_state.messages = []
+                    st.session_state.pdf_name = uploaded.name
+                    st.session_state.using_sample = False
+                    st.success(f"Uploaded: {uploaded.name}. You can now chat!")
+                else:
+                    st.warning("Could not extract text from uploaded PDF. It might be scanned or encrypted.")
+            except Exception as e:
+                st.error(f"Error processing uploaded PDF: {str(e)}")
+
+# ---------------------- Display Messages ----------------------
+if st.session_state.pdf_name:
+    st.subheader(f"Chatting with: {st.session_state.pdf_name}")
+    
+    for msg in st.session_state.messages:
+        role = "🧑 You" if msg["role"] == "user" else "🤖 Assistant"
+        
+        # Style sample messages differently
+        if msg.get("is_sample", False):
+            st.markdown(f"""
+            <div style="
+                background-color: #f0f2f6;
+                padding: 10px;
+                border-radius: 10px;
+                margin-bottom: 10px;
+            ">
+                <strong>{role}:</strong> {msg['content']}
+            </div>
+            """, unsafe_allow_html=True)
+        else:
+            st.markdown(f"**{role}:** {msg['content']}")
+
+# ---------------------- User Input ----------------------
+if st.session_state.index and st.session_state.pdf_name:
+    user_input = st.chat_input("Ask a question about this PDF")
+    if user_input:
+        st.session_state.messages.append({
+            "role": "user", 
+            "content": user_input,
+            "is_sample": False
+        })
+        st.session_state.pending_question = user_input
+        st.rerun()
+
+# ---------------------- Answer Generation ----------------------
+if st.session_state.pending_question and st.session_state.index:
+    with st.spinner("Thinking..."):
+        try:
+            answer = answer_query(
+                st.session_state.pending_question,
+                st.session_state.index,
+                st.session_state.chunks
+            )
+            if not answer:
+                answer = "Sorry, I couldn't generate an answer for that question."
+        except Exception as e:
+            answer = f"An error occurred while generating the answer: {str(e)}"
+    
+    st.session_state.messages.append({
+        "role": "assistant", 
+        "content": answer,
+        "is_sample": False
+    })
+    st.session_state.pending_question = None
+    st.rerun()
+
+# ---------------------- Help Text ----------------------
+if st.session_state.using_sample:
+    st.markdown("""
+    <div style="
+        background-color: #e6f7ff;
+        padding: 15px;
+        border-radius: 10px;
+        margin-top: 20px;
+    ">
+        ℹ️ <strong>How this works:</strong> This is a sample PDF demonstrating the chatbot. 
+        The question above was automatically generated from the sample document. 
+        Upload your own PDF to ask questions about your specific documents.
+    </div>
+    """, unsafe_allow_html=True)

chatpdf_app.py

@@ -0,0 +1,47 @@
+import streamlit as st
+from pdf_utils import extract_text, chunk_text
+from vector_store import build_index
+from tiny_llama import answer_query
+
+st.set_page_config(page_title="PDF Chatbot", page_icon="📄")
+st.title("📄 Chat with your PDF (LLaMA-based)")
+
+# Initialize session state
+if "messages" not in st.session_state:
+    st.session_state.messages = []
+if "index" not in st.session_state:
+    st.session_state.index = None
+if "chunks" not in st.session_state:
+    st.session_state.chunks = None
+
+uploaded = st.file_uploader("Upload a PDF", type=["pdf"])
+
+# Load and index PDF if uploaded
+if uploaded:
+    with st.spinner("Processing PDF..."):
+        text = extract_text(uploaded)
+        chunks = chunk_text(text)
+        index, _ = build_index(chunks)
+        st.session_state.index = index
+        st.session_state.chunks = chunks
+        st.success("PDF processed. Ask me anything!")
+
+# Display chat messages
+for msg in st.session_state.messages:
+    role = "🧑 You" if msg["role"] == "user" else "🤖 Assistant"
+    st.markdown(f"**{role}:** {msg['content']}")
+
+# Chat input
+if st.session_state.index:
+    question = st.chat_input("Ask a question about your PDF")
+    if question:
+        # Append user's question to chat history
+        st.session_state.messages.append({"role": "user", "content": question})
+
+        # Get answer from the LLM
+        with st.spinner("Thinking..."):
+            response = answer_query(question, st.session_state.index, st.session_state.chunks)
+        st.session_state.messages.append({"role": "assistant", "content": response})
+
+        # Rerun to show updated chat
+        st.rerun()

pdf_utils.py

@@ -0,0 +1,18 @@
+#!/usr/bin/env python
+# coding: utf-8
+
+# In[1]:
+
+
+import PyPDF2
+
+def extract_text(file_stream):
+    reader = PyPDF2.PdfReader(file_stream)
+    return "\n".join(page.extract_text() for page in reader.pages)
+
+def chunk_text(text, chunk_size=1000, overlap=200):
+    chunks = []
+    for i in range(0, len(text), chunk_size - overlap):
+        chunks.append(text[i:i+chunk_size])
+    return chunks
+

project enhancements.md

@@ -0,0 +1,10 @@
+🔧 6. Upgrades & Enhancements
+Use LLaMA‑3 (e.g., 8B) for better performance.
+
+Switch to LangChain RetrievalQA and Chroma or Qdrant for vector storage.
+
+Add chat history display, caching, UI enhancements (highlighting source text).
+
+Add file previews, multi-PDF support.
+
+Containerize with Docker.

requirements.txt

@@ -0,0 +1,48 @@
+# Core dependencies
+accelerate==1.7.0
+aiofiles==24.1.0
+aiohttp==3.12.11
+bitsandbytes==0.46.0
+dataclasses-json==0.6.7
+datasets==3.6.0
+faiss-cpu==1.11.0
+fastapi==0.115.12
+filelock==3.13.1
+frozenlist==1.6.2
+fsspec==2025.3.0
+gradio==5.33.0
+gradio_client==1.10.2
+huggingface-hub==0.32.4
+langchain==0.3.25
+langchain-community==0.3.24
+langchain-core==0.3.64
+langchain-text-splitters==0.3.8
+markdown-it-py==3.0.0
+multidict==6.4.4
+numpy==2.2.6
+openai==1.30.1
+orjson==3.10.18
+packaging==24.2
+peft==0.15.2
+PyPDF2==3.0.1
+python-dotenv==1.1.0
+python-multipart==0.0.20
+regex==2024.11.6
+requests==2.31.0
+safetensors==0.5.3
+sentence-transformers==4.1.0
+sentencepiece==0.2.0
+streamlit==1.45.1
+tenacity==9.1.2
+tokenizers==0.21.1
+torch==2.7.0
+torchaudio==2.7.0
+torchvision==0.22.0
+transformers==4.52.4
+triton==3.3.0
+typer==0.16.0
+typing_extensions==4.11.0
+uvicorn==0.34.3
+websockets==15.0.1
+xxhash==3.5.0
+yarl==1.20.0

sample.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ef0298c98084ec572463c1c9bc838471205afcb947c34ae31e92eb59d27bdebd
+size 416028

tiny_llama.py

@@ -0,0 +1,38 @@
+from transformers import AutoTokenizer, AutoModelForCausalLM
+import torch
+
+# Load TinyLlama model and tokenizer
+model_name = "TinyLlama/TinyLlama-1.1B-Chat-v1.0"
+tokenizer = AutoTokenizer.from_pretrained(model_name)
+model = AutoModelForCausalLM.from_pretrained(
+    model_name,
+    torch_dtype=torch.float16 if torch.cuda.is_available() else torch.float32,
+    device_map="auto"
+)
+
+def answer_query(question, index, chunks, top_k=3):
+    # Retrieve top-k most relevant chunks
+    docs = index.similarity_search(question, k=top_k)
+    context = "\n".join([doc.page_content for doc in docs])
+
+    # Construct prompt
+    prompt = f"<|system|>\nYou are a helpful assistant.\n<|user|>\n{context}\n\nQuestion: {question}\n<|assistant|>\n"
+
+    # Tokenize
+    inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
+
+    # Generate
+    with torch.no_grad():
+        outputs = model.generate(
+            **inputs,
+            max_new_tokens=256,
+            temperature=0.7,
+            top_p=0.9,
+            do_sample=True,
+            eos_token_id=tokenizer.eos_token_id
+        )
+
+    # Decode response
+    full_output = tokenizer.decode(outputs[0], skip_special_tokens=True)
+    response = full_output[len(prompt):].strip()
+    return response

vector_store.py

@@ -0,0 +1,19 @@
+# vector_store.py
+
+from langchain.vectorstores import FAISS
+from langchain.embeddings import HuggingFaceEmbeddings
+from langchain.docstore.document import Document
+import faiss
+
+# You can replace this with any sentence transformer you prefer
+def build_index(chunks):
+    # Convert string chunks to Document objects
+    documents = [Document(page_content=chunk) for chunk in chunks]
+
+    # Load a small sentence transformer model for embeddings
+    embedding_model = HuggingFaceEmbeddings(model_name="sentence-transformers/all-MiniLM-L6-v2")
+
+    # Create FAISS index wrapped with LangChain
+    vector_index = FAISS.from_documents(documents, embedding_model)
+
+    return vector_index, embedding_model