Add main application files

README.md

@@ -1,12 +1,63 @@
 ---
-title: Book Chatbot Better
-emoji: 🐠
-colorFrom: green
-colorTo: purple
+title: NCTB Textbook Q&A Assistant
+emoji: 📚
+colorFrom: blue
+colorTo: green
 sdk: gradio
-sdk_version: 6.3.0
+sdk_version: 5.9.1
 app_file: app.py
 pinned: false
+license: mit
 ---
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+# 🎓 NCTB Textbook Q&A Assistant
+
+An intelligent question-answering system for NCTB Class 11-12 textbooks powered by RAG (Retrieval Augmented Generation).
+
+## 📖 Available Books
+- **English** (Class 11-12)
+- **Bangla** (বাংলা বই)
+
+## ✨ Features
+- ✅ Answers strictly from textbook content
+- ✅ Retrieves relevant passages from the books
+- ✅ Supports both English and Bangla
+- ✅ Educational and contextually accurate responses
+- ✅ Powered by Llama 3.1 via Groq
+
+## 🚀 Setup Instructions
+
+### For Hugging Face Spaces:
+1. Upload all files to your Space
+2. Go to **Settings** → **Repository Secrets**
+3. Add a new secret:
+   - **Name:** `GROQ_API_KEY`
+   - **Value:** Your Groq API key (get free at https://console.groq.com/)
+4. Save and wait for the Space to build (~2-3 minutes)
+
+## 🔧 Technology Stack
+- **Framework:** Gradio
+- **Embeddings:** sentence-transformers (all-MiniLM-L6-v2, bangla-bert-base)
+- **Vector Store:** FAISS
+- **LLM:** Llama 3.1 8B (via Groq)
+- **RAG Chain:** LangChain
+
+## 📝 Example Questions
+
+**English:**
+- What is literature?
+- What are the elements of a short story?
+- Explain the theme of the poem
+
+**Bangla:**
+- সাহিত্য কী?
+- কবিতার বৈশিষ্ট্য কী?
+- গল্পের উপাদান কী কী?
+
+## ⚠️ Important Notes
+- The system only answers from textbook content
+- If information is not in the book, it will say so
+- Responses are educational and curriculum-aligned
+
+## 📜 License
+MIT License - Free to use for educational purposes

app.py

@@ -0,0 +1,99 @@
+import gradio as gr
+import os
+from langchain_community.vectorstores import FAISS
+from langchain_community.embeddings import HuggingFaceEmbeddings
+from langchain_groq import ChatGroq
+
+print("Loading English index...")
+english_embeddings = HuggingFaceEmbeddings(model_name="sentence-transformers/all-MiniLM-L6-v2")
+english_db = FAISS.load_local("nctb_english_index", english_embeddings, allow_dangerous_deserialization=True)
+
+print("Loading Bangla index...")
+bangla_embeddings = HuggingFaceEmbeddings(model_name="sagorsarker/bangla-bert-base")
+bangla_db = FAISS.load_local("nctb_bangla_index", bangla_embeddings, allow_dangerous_deserialization=True)
+
+llm = ChatGroq(model="llama-3.1-8b-instant", temperature=0, api_key=os.getenv("GROQ_API_KEY"))
+
+def chat(question, book_choice):
+    if book_choice == "English (Class 11-12)":
+        db = english_db
+        system_instruction = """You are an expert teacher assistant for NCTB Class 11-12 English textbook.
+
+STRICT RULES:
+1. Answer ONLY using information from the provided context
+2. If the answer is not in the context, say "I couldn't find this information in the NCTB textbook."
+3. Quote relevant parts from the book when possible
+4. Keep answers clear and educational
+5. Do NOT use external knowledge - stick to the textbook content only"""
+        
+    else:
+        db = bangla_db
+        system_instruction = """তুমি একজন দক্ষ শিক্ষক সহায়ক যিনি এনসিটিবি বাংলা পাঠ্যপুস্তক থেকে প্রশ্নের উত্তর দাও।
+
+কঠোর নিয়ম:
+1. শুধুমাত্র প্রদত্ত প্রসঙ্গ (context) থেকে উত্তর দিতে হবে
+2. বইয়ের বাইরে কোনো তথ্য ব্যবহার করা যাবে না
+3. প্রাসঙ্গিক অংশ উদ্ধৃত করো
+4. উত্তর স্পষ্ট ও শিক্ষামূলক হতে হবে
+5. যদি উত্তর প্রসঙ্গে না থাকে, বলো "এই তথ্য পাঠ্যপুস্তকে পাওয়া যায়নি।" """
+    
+    # Retrieve relevant passages
+    retriever = db.as_retriever(search_kwargs={"k": 5})
+    docs = retriever.invoke(question)
+    
+    if not docs:
+        return "No relevant information found in the textbook."
+    
+    context = "\n\n---\n\n".join([f"[Passage {i+1}]\n{doc.page_content}" for i, doc in enumerate(docs)])
+    
+    prompt = f"""{system_instruction}
+
+Context from NCTB Textbook:
+{context}
+
+Student Question: {question}
+
+Answer (based strictly on the textbook):"""
+    
+    response = llm.invoke(prompt)
+    
+    if hasattr(response, 'content'):
+        return response.content
+    elif isinstance(response, str):
+        return response
+    else:
+        return str(response)
+
+# Enhanced Gradio interface
+iface = gr.Interface(
+    fn=chat,
+    inputs=[
+        gr.Textbox(
+            label="📝 Your Question", 
+            placeholder="Example: What is the definition of literature? / সাহিত্য কী?",
+            lines=3
+        ),
+        gr.Dropdown(
+            choices=["English (Class 11-12)", "Bangla (বাংলা বই)"], 
+            value="English (Class 11-12)", 
+            label="📚 Select Textbook"
+        )
+    ],
+    outputs=gr.Textbox(label="💡 Answer from Textbook", lines=10),
+    title="🎓 NCTB Textbook Q&A Assistant",
+    description="""Ask questions from NCTB Class 11-12 textbooks and get answers directly from the book content.
+    
+**Features:**
+- ✅ Answers strictly from textbook material
+- ✅ Supports both English and Bangla books
+- ✅ Educational and accurate responses""",
+    examples=[
+        ["What is literature?", "English (Class 11-12)"],
+        ["What are the elements of a short story?", "English (Class 11-12)"],
+        ["সাহিত্য কী?", "Bangla (বাংলা বই)"],
+        ["কবিতার বৈশিষ্ট্য কী?", "Bangla (বাংলা বই)"]
+    ],
+    theme="soft"
+)
+
+iface.launch()

gitignore

@@ -0,0 +1,5 @@
+__pycache__/
+*.py[cod]
+.env
+*.log
+.DS_Store

requirements.txt

@@ -0,0 +1,6 @@
+langchain
+langchain-community
+langchain-groq
+sentence-transformers
+faiss-cpu
+gradio