Spaces:
Running
Running
| from transformers import ( | |
| BartTokenizer, BartForConditionalGeneration, | |
| T5Tokenizer, T5ForConditionalGeneration, | |
| PegasusTokenizer, PegasusForConditionalGeneration, | |
| ) | |
| import gradio as gr | |
| import re | |
| from collections import Counter | |
| # MODEL REGISTRY | |
| MODEL_OPTIONS = { | |
| "BART (facebook/bart-large-cnn)": "bart", | |
| "T5 (t5-small)": "t5", | |
| "Pegasus (google/pegasus-xsum)": "pegasus", | |
| } | |
| # Lazy-load cache so we only download what the user picks | |
| _cache = {} | |
| def load_model(key): | |
| if key in _cache: | |
| return _cache[key] | |
| if key == "bart": | |
| tok = BartTokenizer.from_pretrained("facebook/bart-large-cnn") | |
| mdl = BartForConditionalGeneration.from_pretrained("facebook/bart-large-cnn") | |
| elif key == "t5": | |
| tok = T5Tokenizer.from_pretrained("t5-small") | |
| mdl = T5ForConditionalGeneration.from_pretrained("t5-small") | |
| elif key == "pegasus": | |
| tok = PegasusTokenizer.from_pretrained("google/pegasus-xsum") | |
| mdl = PegasusForConditionalGeneration.from_pretrained("google/pegasus-xsum") | |
| else: | |
| raise ValueError(f"Unknown model key: {key}") | |
| _cache[key] = (tok, mdl) | |
| return tok, mdl | |
| # KEYWORD EXTRACTION (TF-IDF style, no extra library needed) | |
| STOPWORDS = { | |
| "the","a","an","and","or","but","in","on","at","to","for","of","with", | |
| "is","are","was","were","be","been","being","have","has","had","do","does", | |
| "did","will","would","could","should","may","might","this","that","these", | |
| "those","it","its","i","we","you","he","she","they","them","their","our", | |
| "your","his","her","from","by","as","not","no","so","if","about","which", | |
| "who","whom","when","where","how","what","all","also","just","more","than", | |
| } | |
| def extract_keywords(text, top_n=5): | |
| words = re.findall(r'\b[a-zA-Z]{4,}\b', text.lower()) | |
| filtered = [w for w in words if w not in STOPWORDS] | |
| freq = Counter(filtered) | |
| top = freq.most_common(top_n) | |
| if not top: | |
| return "No keywords found." | |
| return " · ".join([f"🏷️ {w}" for w, _ in top]) | |
| # READABILITY SCORE (Flesch Reading Ease) | |
| def count_syllables(word): | |
| word = word.lower().strip(".,!?;:") | |
| vowels = "aeiouy" | |
| count = 0 | |
| prev_vowel = False | |
| for ch in word: | |
| is_v = ch in vowels | |
| if is_v and not prev_vowel: | |
| count += 1 | |
| prev_vowel = is_v | |
| if word.endswith("e") and count > 1: | |
| count -= 1 | |
| return max(1, count) | |
| def flesch_score(text): | |
| sentences = [s.strip() for s in re.split(r'[.!?]+', text) if s.strip()] | |
| words = re.findall(r'\b\w+\b', text) | |
| if not sentences or not words: | |
| return 0.0 | |
| syllables = sum(count_syllables(w) for w in words) | |
| score = ( | |
| 206.835 | |
| - 1.015 * (len(words) / len(sentences)) | |
| - 84.6 * (syllables / len(words)) | |
| ) | |
| return round(max(0, min(100, score)), 1) | |
| def grade_label(score): | |
| if score >= 90: return "Very Easy (Grade 5)" | |
| if score >= 80: return "Easy (Grade 6)" | |
| if score >= 70: return "Fairly Easy (Grade 7)" | |
| if score >= 60: return "Standard (Grade 8–9)" | |
| if score >= 50: return "Fairly Difficult (Grade 10–12)" | |
| if score >= 30: return "Difficult (College)" | |
| return "Very Difficult (Professional)" | |
| def readability_report(original, summarized): | |
| if not original.strip() or not summarized.strip(): | |
| return "—" | |
| o_score = flesch_score(original) | |
| s_score = flesch_score(summarized) | |
| arrow = "⬆️ Easier" if s_score > o_score else ("⬇️ Harder" if s_score < o_score else "➡️ Same") | |
| return ( | |
| f"📄 Original : {o_score} / 100 → {grade_label(o_score)}\n" | |
| f"📝 Summary : {s_score} / 100 → {grade_label(s_score)}\n" | |
| f"📊 Change : {arrow}" | |
| ) | |
| # CORE SUMMARIZE FUNCTION | |
| def run_summary(text, model_label, max_length): | |
| if not text.strip(): | |
| return "⚠️ Please enter some text.", "—", "—" | |
| key = MODEL_OPTIONS[model_label] | |
| tokenizer, model = load_model(key) | |
| input_text = ("summarize: " + text) if key == "t5" else text | |
| inputs = tokenizer.encode( | |
| input_text, | |
| return_tensors="pt", | |
| max_length=1024, | |
| truncation=True | |
| ) | |
| ids = model.generate( | |
| inputs, | |
| max_length=int(max_length), | |
| min_length=30, | |
| do_sample=False | |
| ) | |
| result = tokenizer.decode(ids[0], skip_special_tokens=True) | |
| keywords = extract_keywords(result) | |
| readability = readability_report(text, result) | |
| return result, keywords, readability | |
| def run_file_summary(file, model_label, max_length): | |
| if file is None: | |
| return "⚠️ Please upload a .txt file.", "—", "—" | |
| with open(file.name, "r", encoding="utf-8") as f: | |
| content = f.read() | |
| if not content.strip(): | |
| return "⚠️ The uploaded file appears to be empty.", "—", "—" | |
| return run_summary(content, model_label, max_length) | |
| # HELPERS | |
| def clear_text(): | |
| return "", "", "—", "—" | |
| def clear_file(): | |
| return None, "", "—", "—" | |
| def sync_style(choice): | |
| return {"Short (50)": 50, "Medium (150)": 150, "Detailed (300)": 300}.get(choice, 150) | |
| # CSS (same as original) | |
| css = """ | |
| @import url('https://fonts.googleapis.com/css2?family=Syne:wght@400;600;700&family=DM+Sans:wght@300;400;500&display=swap'); | |
| :root { | |
| --qd-bg: #0a0a0f; | |
| --qd-surface: #13131a; | |
| --qd-card: #1a1a24; | |
| --qd-accent: #7c6af7; | |
| --qd-accent2: #4ecdc4; | |
| --qd-text: #f0eff8; | |
| --qd-muted: #8887a0; | |
| --qd-border: #2a2a3a; | |
| } | |
| body, .gradio-container { | |
| background: var(--qd-bg) !important; | |
| font-family: 'DM Sans', sans-serif !important; | |
| color: var(--qd-text) !important; | |
| } | |
| #qd-header { text-align: center; padding: 2rem 1rem 1rem; } | |
| #qd-logo { | |
| font-family: 'Syne', sans-serif; | |
| font-size: 2.4rem; font-weight: 700; | |
| letter-spacing: -1px; color: var(--qd-text); | |
| } | |
| #qd-logo span { color: var(--qd-accent); } | |
| #qd-tagline { | |
| display: inline-block; margin-top: 6px; | |
| background: #1e1b3a; border: 1px solid var(--qd-accent); | |
| color: var(--qd-accent); font-size: 11px; | |
| padding: 3px 14px; border-radius: 20px; | |
| letter-spacing: 1.5px; font-weight: 500; | |
| } | |
| .tab-nav { | |
| background: var(--qd-surface) !important; | |
| border-radius: 12px !important; padding: 6px !important; | |
| border: none !important; gap: 6px !important; | |
| } | |
| .tab-nav button { | |
| background: transparent !important; border: none !important; | |
| border-radius: 8px !important; color: var(--qd-muted) !important; | |
| font-family: 'DM Sans', sans-serif !important; | |
| font-size: 13px !important; font-weight: 500 !important; | |
| padding: 10px 20px !important; transition: all 0.2s !important; | |
| } | |
| .tab-nav button.selected { background: var(--qd-accent) !important; color: #fff !important; } | |
| label span, .label-wrap span { | |
| font-family: 'DM Sans', sans-serif !important; | |
| font-size: 12px !important; font-weight: 500 !important; | |
| color: var(--qd-muted) !important; letter-spacing: 0.8px !important; | |
| text-transform: uppercase !important; | |
| } | |
| textarea, input[type="text"] { | |
| background: var(--qd-surface) !important; | |
| border: 1px solid var(--qd-border) !important; | |
| border-radius: 10px !important; color: var(--qd-text) !important; | |
| font-family: 'DM Sans', sans-serif !important; | |
| font-size: 14px !important; padding: 12px !important; | |
| transition: border 0.2s !important; | |
| } | |
| textarea:focus, input[type="text"]:focus { | |
| border-color: var(--qd-accent) !important; | |
| outline: none !important; | |
| box-shadow: 0 0 0 2px rgba(124,106,247,0.15) !important; | |
| } | |
| input[type="range"] { accent-color: var(--qd-accent) !important; } | |
| #summarize-btn, #file-summarize-btn { | |
| background: var(--qd-accent) !important; border: none !important; | |
| border-radius: 10px !important; color: #fff !important; | |
| font-family: 'Syne', sans-serif !important; | |
| font-size: 14px !important; font-weight: 600 !important; | |
| letter-spacing: 0.5px !important; padding: 13px !important; | |
| transition: opacity 0.2s !important; width: 100% !important; | |
| } | |
| #summarize-btn:hover, #file-summarize-btn:hover { opacity: 0.85 !important; } | |
| #copy-btn, #file-copy-btn { | |
| background: #1e1b3a !important; | |
| border: 1px solid var(--qd-accent) !important; | |
| color: var(--qd-accent) !important; | |
| border-radius: 8px !important; font-size: 13px !important; | |
| } | |
| #clear-btn, #file-clear-btn { | |
| background: #1a1015 !important; border: 1px solid #3a2a2a !important; | |
| color: #c07070 !important; border-radius: 8px !important; | |
| font-size: 13px !important; | |
| } | |
| #kw-box, #file-kw-box { | |
| background: #0f1a1a !important; | |
| border: 1px solid var(--qd-accent2) !important; | |
| border-radius: 10px !important; color: var(--qd-accent2) !important; | |
| font-size: 13px !important; font-weight: 500 !important; | |
| } | |
| #read-box, #file-read-box { | |
| background: #0f0f1a !important; | |
| border: 1px solid #534AB7 !important; | |
| border-radius: 10px !important; color: #b0a8f7 !important; | |
| font-size: 13px !important; font-family: monospace !important; | |
| } | |
| .gr-box, .gr-form, .gr-panel, .block { | |
| background: var(--qd-card) !important; | |
| border: 1px solid var(--qd-border) !important; | |
| border-radius: 16px !important; | |
| } | |
| footer { display: none !important; } | |
| """ | |
| # UI | |
| with gr.Blocks(css=css, theme=gr.themes.Base(), title="⚡ QuickDigest") as demo: | |
| gr.HTML(""" | |
| <div id="qd-header"> | |
| <div id="qd-logo">⚡ Quick<span>Digest</span></div> | |
| <div id="qd-tagline">AI POWERED · BART · T5 · PEGASUS</div> | |
| </div> | |
| """) | |
| with gr.Tabs(): | |
| # Tab 1: Text Summarizer | |
| with gr.Tab("✍️ Text Summarizer"): | |
| with gr.Row(): | |
| with gr.Column(): | |
| text_input = gr.Textbox( | |
| label="Input Text", | |
| placeholder="Paste your article, paragraph, or any text here...", | |
| lines=10 | |
| ) | |
| model_selector = gr.Radio( | |
| choices=list(MODEL_OPTIONS.keys()), | |
| value=list(MODEL_OPTIONS.keys())[0], | |
| label="🤖 Choose Model", | |
| interactive=True | |
| ) | |
| style_radio = gr.Radio( | |
| choices=["Short (50)", "Medium (150)", "Detailed (300)"], | |
| value="Medium (150)", | |
| label="Summary Style", | |
| interactive=True | |
| ) | |
| length_slider = gr.Slider( | |
| minimum=50, maximum=300, value=150, step=50, | |
| label="Token Length", interactive=True | |
| ) | |
| summarize_btn = gr.Button( | |
| "⚡ Summarize Now", | |
| elem_id="summarize-btn", variant="primary" | |
| ) | |
| with gr.Column(): | |
| text_output = gr.Textbox( | |
| label="📄 Summarized Output", | |
| lines=7, interactive=False | |
| ) | |
| with gr.Row(): | |
| copy_btn = gr.Button("📋 Copy Output", elem_id="copy-btn", size="sm") | |
| clear_btn = gr.Button("🗑️ Clear All", elem_id="clear-btn", size="sm") | |
| kw_output = gr.Textbox( | |
| label="🏷️ Top Keywords", | |
| lines=2, interactive=False, elem_id="kw-box" | |
| ) | |
| read_output = gr.Textbox( | |
| label="📊 Readability Score", | |
| lines=3, interactive=False, elem_id="read-box" | |
| ) | |
| style_radio.change(fn=sync_style, inputs=style_radio, outputs=length_slider) | |
| summarize_btn.click( | |
| fn=run_summary, | |
| inputs=[text_input, model_selector, length_slider], | |
| outputs=[text_output, kw_output, read_output] | |
| ) | |
| clear_btn.click(fn=clear_text, inputs=[], outputs=[text_input, text_output, kw_output, read_output]) | |
| copy_btn.click(fn=lambda x: x, inputs=[text_output], outputs=[text_output]) | |
| # Tab 2: File Summarizer | |
| with gr.Tab("📂 File Summarizer"): | |
| with gr.Row(): | |
| with gr.Column(): | |
| file_input = gr.File( | |
| label="Upload Text File (.txt)", | |
| file_types=[".txt"] | |
| ) | |
| file_model = gr.Radio( | |
| choices=list(MODEL_OPTIONS.keys()), | |
| value=list(MODEL_OPTIONS.keys())[0], | |
| label="🤖 Choose Model", | |
| interactive=True | |
| ) | |
| file_style = gr.Radio( | |
| choices=["Short (50)", "Medium (150)", "Detailed (300)"], | |
| value="Medium (150)", | |
| label="Summary Style", | |
| interactive=True | |
| ) | |
| file_slider = gr.Slider( | |
| minimum=50, maximum=300, value=150, step=50, | |
| label="Token Length", interactive=True | |
| ) | |
| file_btn = gr.Button( | |
| "📂 Upload & Summarize", | |
| elem_id="file-summarize-btn", variant="primary" | |
| ) | |
| with gr.Column(): | |
| file_output = gr.Textbox( | |
| label="📄 File Summary Output", | |
| lines=7, interactive=False | |
| ) | |
| with gr.Row(): | |
| file_copy_btn = gr.Button("📋 Copy Output", elem_id="file-copy-btn", size="sm") | |
| file_clear_btn = gr.Button("🗑️ Clear All", elem_id="file-clear-btn", size="sm") | |
| file_kw_output = gr.Textbox( | |
| label="🏷️ Top Keywords", | |
| lines=2, interactive=False, elem_id="file-kw-box" | |
| ) | |
| file_read_output = gr.Textbox( | |
| label="📊 Readability Score", | |
| lines=3, interactive=False, elem_id="file-read-box" | |
| ) | |
| file_style.change(fn=sync_style, inputs=file_style, outputs=file_slider) | |
| file_btn.click( | |
| fn=run_file_summary, | |
| inputs=[file_input, file_model, file_slider], | |
| outputs=[file_output, file_kw_output, file_read_output] | |
| ) | |
| file_clear_btn.click(fn=clear_file, inputs=[], outputs=[file_input, file_output, file_kw_output, file_read_output]) | |
| file_copy_btn.click(fn=lambda x: x, inputs=[file_output], outputs=[file_output]) | |
| gr.HTML(""" | |
| <div style="text-align:center;margin-top:2rem;font-size:11px;color:#8887a0;letter-spacing:1px;"> | |
| QUICKDIGEST · BART · T5 · PEGASUS · BY ABU SHADAB KHAN | |
| </div> | |
| """) | |
| if __name__ == "__main__": | |
| demo.launch() | |