import os import gradio as gr import pdfplumber from openai import OpenAI from langdetect import detect client = OpenAI(api_key=os.environ["OPENAI_API_KEY"]) SUMMARY_STYLES = ["Child-Friendly", "Academic", "Tweet"] LANGUAGES = ["English", "Turkish", "French", "German", "Spanish", "Arabic"] LANGUAGE_CODES = { "English": "en", "Turkish": "tr", "French": "fr", "German": "de", "Spanish": "es", "Arabic": "ar" } QUIZ_TITLES = { "English": "### 📘 Quiz Questions:", "Turkish": "### 📘 Test Soruları:", "French": "### 📘 Questions à Choix Multiples :", "German": "### 📘 Quizfragen:", "Spanish": "### 📘 Preguntas del cuestionario:", "Arabic": "### 📘 أسئلة الاختيار من متعدد:" } PROMPT_TEMPLATES = { "English": "Please summarize the following text in a {style} style and keep it within approximately {char_limit} characters.\nThen provide 5 relevant keywords.\n\nTEXT:\n{text}", "Turkish": "Lütfen aşağıdaki metni {style} tarzında ve yaklaşık {char_limit} karakter olacak şekilde özetle.\nArdından 5 anahtar kelime ver.\n\nMETİN:\n{text}", "French": "Veuillez résumer le texte suivant dans un style {style}, en environ {char_limit} caractères.\nFournissez ensuite 5 mots-clés pertinents.\n\nTEXTE :\n{text}", "German": "Fassen Sie den folgenden Text im {style}-Stil mit etwa {char_limit} Zeichen zusammen.\nGeben Sie anschließend 5 relevante Schlüsselwörter an.\n\nTEXT:\n{text}", "Spanish": "Resume el siguiente texto en un estilo {style} y con un límite de aproximadamente {char_limit} caracteres.\nLuego proporciona 5 palabras clave relevantes.\n\nTEXTO:\n{text}", "Arabic": "الرجاء تلخيص النص التالي بأسلوب {style}، على ألا يتجاوز {char_limit} حرفًا.\nثم قدم 5 كلمات مفتاحية مهمة.\n\nالنص:\n{text}" } QUIZ_PROMPTS = { "English": "Based on the text below, generate 2 multiple choice questions (each with 4 options A-D):\n\n{text}", "Turkish": "Aşağıdaki metne dayanarak, 4 seçenekli (A, B, C, D) 2 adet çoktan seçmeli soru oluştur:\n\n{text}", "French": "Sur la base du texte ci-dessous, générez 2 questions à choix multiples (4 options A à D) :\n\n{text}", "German": "Erstelle basierend auf dem folgenden Text 2 Multiple-Choice-Fragen (mit jeweils 4 Optionen A–D):\n\n{text}", "Spanish": "Con base en el siguiente texto, genera 2 preguntas de opción múltiple (cada una con 4 opciones A–D):\n\n{text}", "Arabic": "قم بإنشاء سؤالين اختيار من متعدد استنادًا إلى النص أدناه، مع أربعة خيارات لكل سؤال (أ، ب، ج، د):\n\n{text}" } def extract_text_from_pdf(file): try: with pdfplumber.open(file.name) as pdf: text = "" for page in pdf.pages: page_text = page.extract_text() if page_text: text += page_text + "\n" return text.strip() if text.strip() else None, None except Exception as e: return None, str(e) def translate_text(text, target_lang): prompt = f"Translate the following text into {target_lang}:\n\n{text}" response = client.chat.completions.create( model="gpt-3.5-turbo", messages=[{"role": "user", "content": prompt}], temperature=0.3, max_tokens=1000 ) return response.choices[0].message.content def replace_arabic_choices(quiz_text): return quiz_text.replace("A)", "أ.").replace("B)", "ب.").replace("C)", "ج.").replace("D)", "د.") def generate_summary_and_keywords(text, summary_lang, style, char_limit): template = PROMPT_TEMPLATES[summary_lang] prompt = template.format(style=style, char_limit=char_limit, text=text[:2000]) response = client.chat.completions.create( model="gpt-3.5-turbo", messages=[{"role": "user", "content": prompt}], temperature=0.7, max_tokens=700 ) return response.choices[0].message.content def generate_quiz(summary_text, summary_lang): prompt = QUIZ_PROMPTS[summary_lang].format(text=summary_text) response = client.chat.completions.create( model="gpt-3.5-turbo", messages=[{"role": "user", "content": prompt}], temperature=0.7, max_tokens=500 ) result = response.choices[0].message.content return replace_arabic_choices(result) if summary_lang == "Arabic" else result def process(text_input, pdf_file, summary_lang, style, char_limit, make_quiz): if not char_limit.isdigit(): return "⚠️ Please enter a numeric character limit." text, error = extract_text_from_pdf(pdf_file) if pdf_file else (text_input.strip(), None) if error or not text: return f"⚠️ Error: {error or 'No valid text provided.'}" detected = detect(text) if detected != LANGUAGE_CODES[summary_lang]: text = translate_text(text, summary_lang) summary = generate_summary_and_keywords(text, summary_lang, style, char_limit) if make_quiz: quiz = generate_quiz(summary, summary_lang) return summary + "\n\n" + QUIZ_TITLES[summary_lang] + "\n" + quiz return summary with gr.Blocks(css=""" .big-file-upload .file-wrap, .big-file-upload .wrap, .big-file-upload .upload-box, .big-file-upload .dropbox { min-height: 258px !important; max-height: 258px !important; height: 258px !important; } .big-textbox textarea { min-height: 210px !important; max-height: 210px !important; height: 210px !important; } """) as demo: gr.Markdown("## 🌍 Multilingual Summarizer + Quiz Generator") with gr.Accordion("📘 View README / Usage Guide", open=False): gr.Markdown(""" This application allows you to upload a PDF or paste text, select your preferred summary language, and receive: - ✂️ A clear summary - 🏷️ An auto-generated title - 🔑 5 relevant keywords - 🌐 If the content language and summary language differ, the app will auto-translate before summarizing Powered by OpenAI GPT-3.5 and Gradio. """) with gr.Row(): summary_lang = gr.Dropdown(LANGUAGES, value="English", label="Summary Language") summary_style = gr.Dropdown(SUMMARY_STYLES, value="Academic", label="Summary Style") char_limit = gr.Textbox(label="Character Limit", value="300") make_quiz = gr.Checkbox(label="Generate Quiz Questions", value=True) with gr.Row(): with gr.Column(scale=2): text_input = gr.Textbox(lines=15, label="Text Input", elem_classes="big-textbox") with gr.Column(scale=2): pdf_file = gr.File(label="Or Upload PDF", elem_classes="big-file-upload") output = gr.Textbox(label="Output", lines=8) run_btn = gr.Button("Summarize") run_btn.click( inputs=[summary_lang, summary_style, char_limit, text_input, pdf_file, make_quiz], outputs=output ) demo.launch()