Update app.py

app.py

@@ -1,15 +1,11 @@
+import subprocess, sys
+subprocess.run([sys.executable, "-m", "spacy", "download", "en_core_web_sm"],
+               stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL)
+
 import gradio as gr
-import spacy
-import torch
-import subprocess
-import sys
+import spacy, torch
 from transformers import T5ForConditionalGeneration, AutoTokenizer
 
-subprocess.run(
-    [sys.executable, "-m", "spacy", "download", "en_core_web_sm"],
-    check=True
-)
-
 MODEL_ID = "Hamzasajjad38/t5-small-qg"
 tokenizer = AutoTokenizer.from_pretrained(MODEL_ID, use_fast=True)
 model     = T5ForConditionalGeneration.from_pretrained(MODEL_ID)
@@ -18,81 +14,54 @@ model     = model.to(device)
 nlp       = spacy.load("en_core_web_sm")
 
 def generate_questions(passage, num_questions, question_type):
-    if not passage.strip() or len(passage) < 30:
-        return "⚠️ Please enter a longer passage."
-
+    if not passage or len(passage.strip()) < 30:
+        return "Please enter a longer passage (at least 30 characters)."
     doc        = nlp(passage)
-    candidates = [ent.text for ent in doc.ents]
-
+    candidates = [e.text for e in doc.ents]
     if len(candidates) < num_questions:
-        candidates += [chunk.text for chunk in doc.noun_chunks]
-
+        candidates += [c.text for c in doc.noun_chunks]
     seen, unique = set(), []
     for c in candidates:
-        if c.lower() not in seen and len(c.split()) <= 5:
-            seen.add(c.lower())
+        cl = c.lower().strip()
+        if cl not in seen and 1 < len(c.split()) <= 5:
+            seen.add(cl)
             unique.append(c)
-    candidates = unique[:num_questions]
-
+    candidates = unique[:int(num_questions)]
     if not candidates:
-        return "⚠️ Could not extract candidates. Try a longer passage."
-
+        return "Could not extract answer candidates. Try a more detailed passage."
     lines = []
     for i, answer in enumerate(candidates):
         highlighted = passage.replace(answer, f"<hl> {answer} <hl>", 1)
-        input_text  = f"generate question: {highlighted}"
-        inputs  = tokenizer(input_text, return_tensors="pt",
-                            max_length=512, truncation=True).to(device)
-        outputs = model.generate(**inputs, max_new_tokens=64,
-                                 num_beams=4, early_stopping=True)
-        question = tokenizer.decode(outputs[0], skip_special_tokens=True)
-
+        inp  = tokenizer(f"generate question: {highlighted}",
+                         return_tensors="pt", max_length=512,
+                         truncation=True).to(device)
+        out  = model.generate(**inp, max_new_tokens=64,
+                              num_beams=4, early_stopping=True)
+        q    = tokenizer.decode(out[0], skip_special_tokens=True)
         if question_type == "True / False":
-            question = f"True or False: {question.rstrip('?')}?"
-
-        lines.append(f"Q{i+1}: {question}")
+            q = f"True or False: {q.rstrip('?')}?"
+        lines.append(f"Q{i+1}: {q}")
         lines.append(f"   Answer: {answer}\n")
-
     return "\n".join(lines)
 
-with gr.Blocks(title="Automatic Question Generator", theme=gr.themes.Soft()) as demo:
-    gr.Markdown("# 🧠 Automatic Question Generator\n**Fine-tuned T5-small on SQuAD 1.1**")
-
-    with gr.Row():
-        with gr.Column():
-            passage_input = gr.Textbox(
-                label="Input Passage",
-                placeholder="Paste any educational paragraph here...",
-                lines=8
-            )
-            num_q  = gr.Slider(1, 6, value=3, step=1, label="Number of questions")
-            q_type = gr.Dropdown(
-                ["Short answer", "True / False"],
-                value="Short answer",
-                label="Question type"
-            )
-            gen_btn = gr.Button("⚡ Generate Questions", variant="primary")
-
-        with gr.Column():
-            output_box = gr.Textbox(
-                label="Generated Questions",
-                lines=16,
-                interactive=False
-            )
-
-    gr.Examples(
-        examples=[
-            ["Photosynthesis is a process used by plants to convert light energy into chemical energy stored in glucose. It occurs inside chloroplasts using chlorophyll to absorb sunlight.", 3, "Short answer"],
-            ["The Industrial Revolution began in Britain in the late 18th century. The steam engine invented by James Watt transformed manufacturing and transportation.", 3, "Short answer"],
-            ["Machine learning is a subset of artificial intelligence where systems learn from data to improve performance without being explicitly programmed.", 2, "True / False"],
-        ],
-        inputs=[passage_input, num_q, q_type]
-    )
-
-    gen_btn.click(
-        fn=generate_questions,
-        inputs=[passage_input, num_q, q_type],
-        outputs=output_box
-    )
+demo = gr.Interface(
+    fn=generate_questions,
+    inputs=[
+        gr.Textbox(lines=8, label="Input Passage",
+                   placeholder="Paste any educational paragraph here..."),
+        gr.Slider(1, 5, value=3, step=1, label="Number of Questions"),
+        gr.Radio(["Short answer", "True / False"],
+                 value="Short answer", label="Question Type"),
+    ],
+    outputs=gr.Textbox(lines=14, label="Generated Questions"),
+    title="Automatic Question Generator",
+    description="Fine-tuned T5-small on SQuAD 1.1 — paste any passage to generate questions.",
+    examples=[
+        ["Photosynthesis is a process used by plants to convert light energy into chemical energy stored in glucose. It occurs inside chloroplasts using chlorophyll to absorb sunlight.", 3, "Short answer"],
+        ["The Industrial Revolution began in Britain in the late 18th century. The steam engine invented by James Watt transformed manufacturing and transportation across the country.", 3, "Short answer"],
+        ["Machine learning is a subset of artificial intelligence where systems learn from data to improve performance without being explicitly programmed.", 2, "True / False"],
+    ],
+    allow_flagging="never",
+)
 
-demo.launch(server_name="0.0.0.0", server_port=7860)
+demo.launch()