Spaces:
Sleeping
Sleeping
| import streamlit as st | |
| from PyPDF2 import PdfReader | |
| from io import BytesIO | |
| import os | |
| import tempfile | |
| # Fixed imports for LangChain | |
| from langchain_text_splitters import RecursiveCharacterTextSplitter | |
| from langchain_google_genai import GoogleGenerativeAIEmbeddings | |
| from langchain_community.vectorstores import FAISS | |
| from langchain_google_genai import ChatGoogleGenerativeAI | |
| # --- Get API key from Hugging Face Secrets --- | |
| GOOGLE_API_KEY = os.environ.get("GOOGLE_API_KEY", "") | |
| # Use temporary directory for Hugging Face Spaces | |
| TEMP_DIR = tempfile.gettempdir() | |
| FAISS_INDEX_PATH = os.path.join(TEMP_DIR, "faiss_index") | |
| # PDF file path | |
| PDF_FILE_PATH = "./slide.pdf" | |
| def get_pdf_text_from_file(pdf_path): | |
| """Read PDF from file path""" | |
| text = "" | |
| if not os.path.exists(pdf_path): | |
| raise FileNotFoundError(f"PDF file not found: {pdf_path}") | |
| pdf_reader = PdfReader(pdf_path) | |
| for page in pdf_reader.pages: | |
| page_text = page.extract_text() | |
| if page_text: | |
| text += page_text | |
| return text | |
| def get_text_chunks(text): | |
| text_splitter = RecursiveCharacterTextSplitter(chunk_size=10000, chunk_overlap=1000) | |
| return text_splitter.split_text(text) | |
| def get_vector_store(text_chunks, api_key): | |
| embeddings = GoogleGenerativeAIEmbeddings(model="models/embedding-001", google_api_key=api_key) | |
| vector_store = FAISS.from_texts(text_chunks, embedding=embeddings) | |
| vector_store.save_local(FAISS_INDEX_PATH) | |
| def get_response(user_question, api_key, chat_history): | |
| """Get response from the AI model with chat history context""" | |
| embeddings = GoogleGenerativeAIEmbeddings(model="models/embedding-001", google_api_key=api_key) | |
| new_db = FAISS.load_local(FAISS_INDEX_PATH, embeddings, allow_dangerous_deserialization=True) | |
| docs = new_db.similarity_search(user_question, k=4) | |
| # Format chat history for context | |
| history_text = "" | |
| for msg in chat_history[-10:]: # Keep last 10 messages for context | |
| role = "User" if msg["role"] == "user" else "Assistant" | |
| history_text += f"{role}: {msg['content']}\n" | |
| # Combine document contents | |
| context = "\n\n".join([doc.page_content for doc in docs]) | |
| prompt_template = f""" | |
| You are a helpful assistant for Antimicrobial Pharmacology. You answer questions based ONLY on the context provided from the PDF documents. | |
| IMPORTANT RULES: | |
| 1. Do not use any external knowledge or assumptions. | |
| 2. If the answer is not found in the context, reply with "I don't know based on the provided materials." | |
| 3. Be conversational and helpful. | |
| 4. When generating MCQs, create questions that test understanding of the material. | |
| 5. When checking answers, be encouraging and provide explanations. | |
| 6. IMPORTANT: When referencing information from the course materials, always say "your professor says" or "according to your professor" instead of "the text states", "the document states", "the text says", or similar phrases. This makes the learning experience more personal and connected to the course. | |
| Chat History: | |
| {history_text} | |
| Context from PDF: | |
| {context} | |
| Current Question: | |
| {user_question} | |
| Instructions: | |
| - If the user asks for a multiple choice question (MCQ), quiz, or test question: | |
| * Generate a question with 4 options (A, B, C, D) based ONLY on the context | |
| * Make sure the question tests important concepts from the material | |
| * Do NOT reveal the correct answer yet | |
| * Ask the user to select their answer | |
| - If the user provides an answer (like "A", "B", "C", "D" or the full answer text) AND there was a recent MCQ in the chat history: | |
| * Check if the answer is correct based on the context | |
| * If correct: Congratulate them and explain why it's correct, referencing what "your professor says" in the materials | |
| * If incorrect: Encourage them, reveal the correct answer, and explain why using "your professor says" to reference the source material | |
| - For regular questions: Answer based on the context provided, using "your professor says" when citing information | |
| Remember: Never use phrases like "the text states", "the document says", "according to the text", etc. Always use "your professor says" or "according to your professor" instead. | |
| Answer: | |
| """ | |
| model = ChatGoogleGenerativeAI(model="gemini-2.0-flash", temperature=0.3, google_api_key=api_key) | |
| response = model.invoke(prompt_template) | |
| return response.content | |
| def main(): | |
| st.set_page_config( | |
| page_title="Antimicrobial Pharmacology Chatbot", | |
| page_icon="💊", | |
| layout="wide" | |
| ) | |
| st.header("Antimicrobial Pharmacology Chatbot (RX24)") | |
| st.markdown("---") | |
| # Initialize session state | |
| if "api_entered" not in st.session_state: | |
| st.session_state["api_entered"] = False | |
| if "pdf_processed" not in st.session_state: | |
| st.session_state["pdf_processed"] = False | |
| if "messages" not in st.session_state: | |
| st.session_state["messages"] = [] | |
| # Check for API key | |
| api_key = GOOGLE_API_KEY | |
| # STEP 1: API Key handling | |
| if not st.session_state["api_entered"]: | |
| if not api_key: | |
| st.warning("Google API Key not found in environment variables.") | |
| st.info("Please add GOOGLE_API_KEY to your Hugging Face Space secrets or enter it below.") | |
| user_api_key = st.text_input( | |
| "Enter your Gemini API key", | |
| type="password", | |
| help="Get your API key from https://makersuite.google.com/app/apikey" | |
| ) | |
| if st.button("Continue", type="primary") and user_api_key: | |
| st.session_state["user_api_key"] = user_api_key | |
| st.session_state["api_entered"] = True | |
| st.rerun() | |
| st.stop() | |
| else: | |
| st.session_state["user_api_key"] = api_key | |
| st.session_state["api_entered"] = True | |
| api_key = st.session_state.get("user_api_key", "") | |
| # STEP 2: Auto-process PDF file | |
| if not st.session_state["pdf_processed"]: | |
| st.subheader("Processing Antimicrobial Pharmacology Course") | |
| with st.spinner("Processing Antimicrobial Pharmacology Course... This may take a moment."): | |
| try: | |
| raw_text = get_pdf_text_from_file(PDF_FILE_PATH) | |
| if not raw_text.strip(): | |
| st.error("No text could be extracted from the PDF. Please check your file.") | |
| st.stop() | |
| st.info(f"Processing: {PDF_FILE_PATH}") | |
| text_chunks = get_text_chunks(raw_text) | |
| get_vector_store(text_chunks, api_key) | |
| st.session_state["pdf_processed"] = True | |
| st.success("PDF processed successfully. You can now ask questions.") | |
| st.rerun() | |
| except FileNotFoundError as e: | |
| st.error(str(e)) | |
| st.stop() | |
| except Exception as e: | |
| st.error(f"Error processing PDF: {str(e)}") | |
| st.stop() | |
| # STEP 3: Chat Interface | |
| # Sidebar with options | |
| with st.sidebar: | |
| st.subheader("Options") | |
| st.info("Loaded: Antimicrobial Pharmacology Course") | |
| if st.button("Reprocess PDF"): | |
| st.session_state["pdf_processed"] = False | |
| st.rerun() | |
| if st.button("Clear Chat History"): | |
| st.session_state["messages"] = [] | |
| st.rerun() | |
| st.markdown("---") | |
| st.subheader("Quick Actions") | |
| st.markdown(""" | |
| Try asking: | |
| - "Give me a multiple choice question" | |
| - "Quiz me on antibiotics" | |
| - "Generate an MCQ about [topic]" | |
| - "What are the main topics?" | |
| """) | |
| st.markdown("---") | |
| st.subheader("How to use MCQs") | |
| st.markdown(""" | |
| 1. Ask for an MCQ (e.g., "Give me a quiz question") | |
| 2. The bot will generate a question with options A-D | |
| 3. Reply with your answer (e.g., "A" or "The answer is B") | |
| 4. Get feedback on whether you're correct | |
| """) | |
| # Main chat area | |
| st.subheader("Chat with your Pharmacology Assistant") | |
| # Display chat messages | |
| for message in st.session_state["messages"]: | |
| with st.chat_message(message["role"]): | |
| st.markdown(message["content"]) | |
| # Quick action buttons (only show if no messages yet) | |
| if len(st.session_state["messages"]) == 0: | |
| st.markdown("### Quick Start") | |
| col1, col2, col3 = st.columns(3) | |
| with col1: | |
| if st.button("Generate MCQ", use_container_width=True): | |
| quick_question = "Give me a multiple choice question to test my knowledge" | |
| st.session_state["messages"].append({"role": "user", "content": quick_question}) | |
| with st.spinner("Generating question..."): | |
| response = get_response(quick_question, api_key, st.session_state["messages"]) | |
| st.session_state["messages"].append({"role": "assistant", "content": response}) | |
| st.rerun() | |
| with col2: | |
| if st.button("Summarize Topics", use_container_width=True): | |
| quick_question = "What are the main topics covered in this material?" | |
| st.session_state["messages"].append({"role": "user", "content": quick_question}) | |
| with st.spinner("Analyzing..."): | |
| response = get_response(quick_question, api_key, st.session_state["messages"]) | |
| st.session_state["messages"].append({"role": "assistant", "content": response}) | |
| st.rerun() | |
| with col3: | |
| if st.button("How can you help?", use_container_width=True): | |
| quick_question = "What can you help me with regarding this pharmacology material?" | |
| st.session_state["messages"].append({"role": "user", "content": quick_question}) | |
| with st.spinner("Processing..."): | |
| response = get_response(quick_question, api_key, st.session_state["messages"]) | |
| st.session_state["messages"].append({"role": "assistant", "content": response}) | |
| st.rerun() | |
| # Chat input | |
| if user_question := st.chat_input("Ask a question or answer an MCQ..."): | |
| # Add user message to chat history | |
| st.session_state["messages"].append({"role": "user", "content": user_question}) | |
| # Display user message | |
| with st.chat_message("user"): | |
| st.markdown(user_question) | |
| # Generate and display assistant response | |
| with st.chat_message("assistant"): | |
| with st.spinner("Processing..."): | |
| try: | |
| response = get_response(user_question, api_key, st.session_state["messages"]) | |
| st.markdown(response) | |
| st.session_state["messages"].append({"role": "assistant", "content": response}) | |
| except Exception as e: | |
| error_msg = f"Error getting answer: {str(e)}" | |
| st.error(error_msg) | |
| st.session_state["messages"].append({"role": "assistant", "content": error_msg}) | |
| # Add footer | |
| st.markdown("---") | |
| st.markdown( | |
| """ | |
| <div style='text-align: center'> | |
| <small>Antimicrobial Pharmacology Chatbot - Powered by Gemini AI</small> | |
| </div> | |
| """, | |
| unsafe_allow_html=True | |
| ) | |
| if __name__ == "__main__": | |
| main() |