import os import tempfile # ✅ CRITICAL: Set environment variables BEFORE any other imports os.environ['TRANSFORMERS_CACHE'] = tempfile.gettempdir() os.environ['HF_HOME'] = tempfile.gettempdir() os.environ['TORCH_HOME'] = tempfile.gettempdir() os.environ['HF_DATASETS_CACHE'] = tempfile.gettempdir() os.environ['HUGGINGFACE_HUB_CACHE'] = tempfile.gettempdir() import streamlit as st # ✅ CRITICAL: set_page_config() MUST be called first, before ANY Streamlit commands st.set_page_config( page_title="AI Study Helper Pro - by Umaima Qureshi", page_icon="🧠", layout="wide", initial_sidebar_state="expanded" ) # Now import other libraries from transformers import pipeline, AutoTokenizer, AutoModelForSeq2SeqLM, AutoModelForQuestionAnswering, AutoModelForSequenceClassification from nltk.tokenize import sent_tokenize import base64 import torch import nltk @st.cache_resource def init_nltk(): """Initialize NLTK with writable directory""" nltk_data_dir = os.path.join(tempfile.gettempdir(), "nltk_data") os.makedirs(nltk_data_dir, exist_ok=True) if nltk_data_dir not in nltk.data.path: nltk.data.path.insert(0, nltk_data_dir) for pkg in ["punkt", "punkt_tab"]: try: nltk.data.find(f"tokenizers/{pkg}") except LookupError: try: nltk.download(pkg, download_dir=nltk_data_dir, quiet=True) except: pass # Continue if download fails return True # Initialize NLTK init_nltk() # Device detection DEVICE = 0 if torch.cuda.is_available() else -1 # Lazy model loading with proper cache handling - FIXED (No cache_dir in pipeline) @st.cache_resource def get_summarizer(): """Load summarization model - cache_dir handled by environment variables""" try: model = AutoModelForSeq2SeqLM.from_pretrained( "sshleifer/distilbart-cnn-12-6", cache_dir=tempfile.gettempdir() ) tokenizer = AutoTokenizer.from_pretrained( "sshleifer/distilbart-cnn-12-6", cache_dir=tempfile.gettempdir() ) summarizer = pipeline( "summarization", model=model, tokenizer=tokenizer, device=DEVICE ) return summarizer except Exception as e: st.error(f"Failed to load summarizer: {str(e)}") return None @st.cache_resource def get_qa(): """Load Q&A model - cache_dir handled by environment variables""" try: model = AutoModelForQuestionAnswering.from_pretrained( "distilbert-base-uncased-distilled-squad", cache_dir=tempfile.gettempdir() ) tokenizer = AutoTokenizer.from_pretrained( "distilbert-base-uncased-distilled-squad", cache_dir=tempfile.gettempdir() ) qa_pipeline = pipeline( "question-answering", model=model, tokenizer=tokenizer, device=DEVICE ) return qa_pipeline except Exception as e: st.error(f"Failed to load Q&A model: {str(e)}") return None @st.cache_resource def get_classifier(): """Load classifier model - cache_dir handled by environment variables""" try: model = AutoModelForSequenceClassification.from_pretrained( "typeform/distilbert-base-uncased-mnli", cache_dir=tempfile.gettempdir() ) tokenizer = AutoTokenizer.from_pretrained( "typeform/distilbert-base-uncased-mnli", cache_dir=tempfile.gettempdir() ) classifier = pipeline( "zero-shot-classification", model=model, tokenizer=tokenizer, device=DEVICE ) return classifier except Exception as e: st.error(f"Failed to load classifier: {str(e)}") return None @st.cache_resource def load_translator(model_name): """Load translation model - cache_dir handled by environment variables""" try: model = AutoModelForSeq2SeqLM.from_pretrained( model_name, cache_dir=tempfile.gettempdir() ) tokenizer = AutoTokenizer.from_pretrained( model_name, cache_dir=tempfile.gettempdir() ) translator = pipeline( "translation", model=model, tokenizer=tokenizer, device=DEVICE ) return translator except Exception as e: st.error(f"Failed to load translator: {str(e)}") return None def truncate_text(text, max_words=400): """Truncate text to maximum word count""" words = text.split() return (" ".join(words[:max_words]), len(words) > max_words) # ULTRA PREMIUM CSS - Glassmorphism + Animations st.markdown(""" """, unsafe_allow_html=True) # Hero Header st.markdown("""
🧠 AI Study Helper Pro
⚡ Supercharge Your Learning with Advanced AI Technology
""", unsafe_allow_html=True) # Add cache clear button (for troubleshooting) with st.expander("⚙ī¸ Settings", expanded=False): if st.button("🔄 Clear Model Cache (if you see errors)"): st.cache_resource.clear() st.success("✅ Cache cleared! Please refresh the page.") st.info("💡 This will reload all AI models on next use.") # Sidebar with st.sidebar: st.markdown("### đŸŽ¯ Dashboard") st.markdown("---") # Stats col1, col2 = st.columns(2) with col1: st.markdown('
📊 247 Processed
', unsafe_allow_html=True) with col2: st.markdown('
⚡ 2.3s Avg
', unsafe_allow_html=True) st.markdown("---") st.markdown("### ✨ Features") features = [ "📝 AI Summarization", "đŸ’Ŧ Smart Q&A", "đŸŽ¯ Quiz Generator", "🌍 Multi-Language", "🔑 Keyword Extraction", "💨 Lightning Fast" ] for feat in features: st.markdown(f"**{feat}**") st.markdown("---") st.markdown("### 👩‍đŸ’ģ Developer") st.markdown("**Umaima Qureshi**") st.markdown("[GitHub](https://github.com/Umaima122)") # Initialize session state for key in ["summary", "quiz", "translation", "keywords"]: if key not in st.session_state: st.session_state[key] = "" if key not in ["quiz", "keywords"] else [] # Tabs tab1, tab2, tab3, tab4, tab5, tab6 = st.tabs([ "📝 Summarize", "đŸ’Ŧ Q&A", "đŸŽ¯ Quiz", "🌍 Translate", "🔑 Keywords", "đŸ“Ĩ Download" ]) # ============================================ # TAB 1: SUMMARIZE # ============================================ with tab1: st.markdown("### 📝 Intelligent Summarization") text = st.text_area( "✍ī¸ Your notes or textbook:", value="", height=250, key="sum_txt", placeholder="Paste your content here and watch AI magic happen..." ) col1, col2, col3 = st.columns([1, 1, 1]) with col2: if st.button("✨ Generate Summary", key="sum_btn"): if not text.strip(): st.error("⚠ī¸ Please provide text to summarize") else: trunc, was_trunc = truncate_text(text, 400) if was_trunc: st.info("📊 Text optimized to 400 words for processing") if len(trunc.split()) < 20: st.error("⚠ī¸ Need at least 20 words to generate a meaningful summary") else: with st.spinner("🧠 AI is thinking..."): try: summarizer = get_summarizer() if summarizer: result = summarizer( trunc, max_length=130, min_length=30, do_sample=False ) summary = result[0]['summary_text'] st.markdown(f"""

📄 AI-Generated Summary

{summary}

{len(summary.split())} words ✓ Completed
""", unsafe_allow_html=True) st.session_state["summary"] = summary except Exception as e: st.error(f"❌ Error generating summary: {str(e)}") # ============================================ # TAB 2: Q&A # ============================================ with tab2: st.markdown("### đŸ’Ŧ Intelligent Q&A System") context = st.text_area( "📚 Context (Your notes):", value="", height=200, key="qa_ctx", placeholder="Paste your study material here..." ) question = st.text_input( "❓ Ask your question:", key="qa_q", placeholder="What would you like to know?" ) col1, col2, col3 = st.columns([1, 1, 1]) with col2: if st.button("🔍 Get Answer", key="qa_btn"): if not context.strip() or not question.strip(): st.error("⚠ī¸ Please provide both context and question") else: trunc_ctx, _ = truncate_text(context, 400) with st.spinner("🤔 Analyzing..."): try: qa_model = get_qa() if qa_model: result = qa_model(question=question, context=trunc_ctx) answer = result['answer'] confidence = result.get('score', 0) st.markdown(f"""

💡 AI Answer

{answer}

✓ Answer Found Confidence: {confidence:.1%}
""", unsafe_allow_html=True) except Exception as e: st.error(f"❌ Error finding answer: {str(e)}") # ============================================ # TAB 3: QUIZ - FIXED # ============================================ with tab3: st.markdown("### đŸŽ¯ AI Quiz Generator") quiz_ctx = st.text_area( "📖 Study material:", value="", height=200, key="quiz_ctx", placeholder="Paste content for quiz generation..." ) col1, col2, col3 = st.columns([1, 1, 1]) with col2: if st.button("🚀 Generate Quiz", key="quiz_btn"): if not quiz_ctx.strip(): st.error("⚠ī¸ Please provide text for quiz generation") else: trunc_quiz, _ = truncate_text(quiz_ctx, 200) with st.spinner("🎲 Creating questions..."): try: sentences = sent_tokenize(trunc_quiz)[:5] if len(sentences) == 0: st.warning("⚠ī¸ Could not extract sentences from the text") else: def get_first_words(text, max_words=12): """Get first N complete words from text""" words = text.split() if len(words) <= max_words: return text return ' '.join(words[:max_words]) questions = [f"What is the main concept in: '{get_first_words(s, 12)}'?" for s in sentences if len(s) > 10] if questions: st.markdown("

📝 Generated Quiz Questions

", unsafe_allow_html=True) st.markdown('
', unsafe_allow_html=True) for i, q in enumerate(questions, 1): st.markdown(f"""
Question {i}: {q}
""", unsafe_allow_html=True) st.markdown('
', unsafe_allow_html=True) st.markdown('', unsafe_allow_html=True) st.session_state["quiz"] = questions else: st.warning("⚠ī¸ Could not generate questions from the provided text") except Exception as e: st.error(f"❌ Error generating quiz: {str(e)}") # ============================================ # TAB 4: TRANSLATE # ============================================ with tab4: st.markdown("### 🌍 AI Translation") trans_text = st.text_area( "✍ī¸ Text to translate:", height=200, key="trans_txt", placeholder="Enter text to translate..." ) col1, col2 = st.columns(2) with col1: lang = st.selectbox( "đŸŽ¯ Target language:", ["French", "German", "Spanish", "Italian", "Hindi"] ) with col2: st.write("") st.write("") if st.button("🌐 Translate Now", key="trans_btn"): if not trans_text.strip(): st.error("⚠ī¸ Please provide text to translate") else: model_map = { "French": "Helsinki-NLP/opus-mt-en-fr", "German": "Helsinki-NLP/opus-mt-en-de", "Spanish": "Helsinki-NLP/opus-mt-en-es", "Italian": "Helsinki-NLP/opus-mt-en-it", "Hindi": "Helsinki-NLP/opus-mt-en-hi" } trunc_trans, _ = truncate_text(trans_text, 200) with st.spinner(f"🌍 Translating to {lang}..."): try: translator = load_translator(model_map[lang]) if translator: result = translator(trunc_trans, max_length=256) translation = result[0]['translation_text'] st.markdown(f"""

🌐 Translation ({lang})

{translation}

✓ Translated
""", unsafe_allow_html=True) st.session_state["translation"] = translation except Exception as e: st.error(f"❌ Translation Error: {str(e)}") # ============================================ # TAB 5: KEYWORDS - FIXED # ============================================ with tab5: st.markdown("### 🔑 AI Keyword Extraction") keyword_input = st.text_area( "📝 Text for analysis:", value="", height=200, key="kw_txt", placeholder="Paste text to extract key concepts..." ) col1, col2, col3 = st.columns([1, 1, 1]) with col2: if st.button("🔍 Extract Keywords", key="kw_btn"): if not keyword_input.strip(): st.error("⚠ī¸ Please provide text for keyword extraction") else: trunc_kw, _ = truncate_text(keyword_input, 200) with st.spinner("🔎 Analyzing concepts..."): try: classifier = get_classifier() if classifier: labels = ["technology", "science", "education", "health", "business", "finance", "medical", "engineering", "mathematics", "history"] result = classifier(trunc_kw, labels) keywords = [lbl for lbl, score in zip(result['labels'], result['scores']) if score > 0.3][:5] if keywords: st.markdown('
', unsafe_allow_html=True) st.markdown("

đŸŽ¯ Extracted Keywords

", unsafe_allow_html=True) st.markdown('
', unsafe_allow_html=True) kw_html = " ".join([ f"{kw}" for kw in keywords ]) st.markdown(kw_html, unsafe_allow_html=True) st.markdown('
', unsafe_allow_html=True) st.session_state["keywords"] = keywords else: st.info("ℹī¸ No strong keywords found. Try providing more detailed text.") except Exception as e: st.error(f"❌ Error extracting keywords: {str(e)}") # ============================================ # TAB 6: DOWNLOAD # ============================================ with tab6: st.markdown("### đŸ“Ĩ Download Results") def download_link(text, filename, emoji): """Generate download link for text content""" b64 = base64.b64encode(text.encode()).decode() return f""" {emoji} Download {filename} """ col1, col2 = st.columns(2) with col1: if st.session_state["summary"]: st.markdown(download_link(st.session_state["summary"], "summary.txt", "📄"), unsafe_allow_html=True) else: st.info("📄 Generate a summary first") if st.session_state["quiz"]: quiz_text = "\n\n".join([f"Question {i}: {q}" for i, q in enumerate(st.session_state["quiz"], 1)]) st.markdown(download_link(quiz_text, "quiz.txt", "đŸŽ¯"), unsafe_allow_html=True) else: st.info("đŸŽ¯ Generate a quiz first") with col2: if st.session_state["translation"]: st.markdown(download_link(st.session_state["translation"], "translation.txt", "🌍"), unsafe_allow_html=True) else: st.info("🌍 Translate text first") if st.session_state["keywords"]: keywords_text = "Extracted Keywords:\n\n" + "\n".join([f"- {kw}" for kw in st.session_state["keywords"]]) st.markdown(download_link(keywords_text, "keywords.txt", "🔑"), unsafe_allow_html=True) else: st.info("🔑 Extract keywords first") st.markdown("---") if not any([st.session_state["summary"], st.session_state["quiz"], st.session_state["translation"], st.session_state["keywords"]]): st.warning("ℹī¸ Generate content in other tabs to enable downloads") else: st.success("✅ Content ready for download! Click the buttons above.") # ============================================ # PREMIUM FOOTER # ============================================ st.markdown("""

Built with ❤ī¸ by Umaima Qureshi

Š 2025 AI Study Helper Pro. All Rights Reserved.

🔗 GitHub đŸ’ŧ LinkedIn

Powered by Transformers â€ĸ PyTorch â€ĸ Streamlit â€ĸ NLTK

""", unsafe_allow_html=True)