Update app.py

app.py

@@ -6,10 +6,11 @@ import pandas as pd
 import gradio as gr
 
 # -----------------------------
-# STRONGEST AVAILABLE PUBLIC MODEL (WORKING)
+# WORKING PUBLIC AI DETECTOR
 # -----------------------------
 MODEL_NAME = "openai-community/roberta-base-openai-detector"
 
+
 # -----------------------------
 # LOAD MODEL
 # -----------------------------
@@ -21,11 +22,10 @@ model.to(device).eval()
 
 
 # -----------------------------
-# PARAGRAPH SPLITTER
+# LINE SPLITTER  (SAFE, FIXED)
 # -----------------------------
-def paragraph_split(text):
-    paragraphs = [p.strip() for p in text.split("\n") if p.strip()]
-    return paragraphs
+def line_split(text):
+    return [l.strip() for l in text.split("\n") if l.strip()]
 
 
 # -----------------------------
@@ -35,81 +35,71 @@ def classify_text(text):
     if not text.strip():
         return "⚠️ Please enter some text.", None, None
 
-    paragraphs = paragraph_split(text)
-    if not paragraphs:
-        return "⚠️ No paragraphs detected.", None, None
+    lines = line_split(text)
+    if not lines:
+        return "⚠️ No content detected.", None, None
 
-    # Tokenize paragraphs
+    # Tokenize line by line → SAFE
     inputs = tokenizer(
-        paragraphs,
+        lines,
         return_tensors="pt",
         padding=True,
-        truncation=True,
-        max_length=model.config.max_position_embeddings
+        truncation=True, 
+        max_length=512  # SAFE for RoBERTa
     ).to(device)
 
-    # Predict
     with torch.no_grad():
         logits = model(**inputs).logits
         probs = F.softmax(logits, dim=-1).cpu()
         preds = torch.argmax(probs, dim=-1).cpu()
 
-    # -----------------------------
-    # BUILD RESULTS
-    # -----------------------------
     results = []
-    highlighted_paragraphs = []
-
-    for i, p in enumerate(paragraphs):
-        pred_label = preds[i].item()
-        confidence = probs[i, pred_label].item()
+    highlighted_lines = []
 
-        label = "AI" if pred_label == 1 else "Human"  
-        # NOTE: roberta-base-openai-detector → label 1 = Fake (AI), 0 = Real (Human)
+    for i, line in enumerate(lines):
+        pred = preds[i].item()
+        conf = probs[i, pred].item()
 
-        conf_text = f"{confidence:.2f}"
+        # For this model: 1 = AI, 0 = Human
+        label = "AI" if pred == 1 else "Human"
+        conf_text = f"{conf:.2f}"
 
-        results.append([p, label, conf_text])
+        results.append([line, label, conf_text])
 
         if label == "AI":
-            highlighted_paragraphs.append(
-                f"<p style='color:red; font-weight:bold; margin-bottom:10px'>{p}</p>"
-            )
+            highlighted_lines.append(f"<p style='color:red; font-weight:bold'>{line}</p>")
         else:
-            highlighted_paragraphs.append(
-                f"<p style='color:green; font-weight:bold; margin-bottom:10px'>{p}</p>"
-            )
+            highlighted_lines.append(f"<p style='color:green; font-weight:bold'>{line}</p>")
 
     # -----------------------------
-    # DOCUMENT LEVEL SCORE
+    # DOCUMENT AI SCORE
     # -----------------------------
     avg = torch.mean(probs, dim=0)
-    ai_likelihood = avg[1].item() * 100   # class 1 = AI
+    ai_percent = avg[1].item() * 100  # class 1 = AI
 
-    highlighted_html = "\n".join(highlighted_paragraphs)
-    df = pd.DataFrame(results, columns=["Paragraph", "Classification", "Confidence"])
+    highlighted_html = "\n".join(highlighted_lines)
+    df = pd.DataFrame(results, columns=["Line", "Classification", "Confidence"])
 
-    return f"⚖️ Document AI Likelihood: {ai_likelihood:.1f}%", highlighted_html, df
+    return f"⚖️ Document AI Likelihood: {ai_percent:.1f}%", highlighted_html, df
 
 
 # -----------------------------
 # GRADIO UI
 # -----------------------------
 with gr.Blocks() as demo:
-    gr.Markdown("## 🧠 Writenix Advanced AI Detection (Paragraph-Level)")
+    gr.Markdown("## 🧠 Writenix AI Detector (Line-Level, Stable Version)")
 
-    with gr.Row():
-        text_input = gr.Textbox(
-            label="Enter text",
-            lines=14,
-            placeholder="Paste your essay, article, or content here…"
-        )
+    text_input = gr.Textbox(
+        label="Enter text",
+        lines=14,
+        placeholder="Paste your essay, article, or content here…"
+    )
 
     classify_btn = gr.Button("🚀 Detect AI")
 
     ai_score = gr.Label(label="Overall AI Likelihood")
     highlighted = gr.HTML()
-    table = gr.Dataframe(headers=["Paragraph", "Classification", "Confidence"], wrap=True)
+    table = gr.Dataframe(headers=["Line", "Classification", "Confidence"], wrap=True)
 
     classify_btn.click(classify_text, inputs=text_input, outputs=[ai_score, highlighted, table])