Update app.py

app.py

@@ -4,28 +4,33 @@ import os
 import torch
 import uvicorn
 import json
+
 from fastapi import FastAPI
 from pydantic import BaseModel
 from transformers import BertTokenizer, AutoModelForSequenceClassification
 from arabert.preprocess import ArabertPreprocessor
-from tabulate import tabulate  # مكتبة فخمة لبناء جداول منسقة في اللوغز السحابية
+from tabulate import tabulate  
+
 
 MODEL_REPO = "kkAsmaa/ChildShield"
 MODEL_NAME = "aubmindlab/bert-base-arabertv02-twitter"
 SUB_FOLDER = "ChildShield"
 HF_TOKEN = os.getenv("HF_TOKEN")
 
+
 print("🔄 Loading ChildShield Model Weights with Deep Window Auto-Logging Features...")
 tokenizer = BertTokenizer.from_pretrained(MODEL_NAME)
 model = AutoModelForSequenceClassification.from_pretrained(MODEL_REPO, token=HF_TOKEN, subfolder=SUB_FOLDER)
 model.eval()
 arabic_prep = ArabertPreprocessor(model_name=MODEL_NAME)
 
+
 app = FastAPI(title="ChildShield Backend API")
 
 class InputData(BaseModel):
     text: str
 
+
 def clean_obfuscation(text):
     text = str(text)
     text = re.sub(r'https?://\S+|www\.\S+|@\S+|#', '', text)
@@ -39,9 +44,10 @@ def clean_obfuscation(text):
 
 def full_preprocess(text):
     text_no_trickery = clean_obfuscation(text)
-    final_text = arabic_prep.preprocess(text_no_trickery)
+    final_text = arabic_prep.preprocess(text_no_trickery)   
     return final_text
 
+
 def predict_safety_api(text):
     """
     Arabic text classification gateway utilizing a custom sliding window configuration with 20 token overlap.
@@ -72,7 +78,7 @@ def predict_safety_api(text):
     highest_safe_prob = 0.0
     windows_analysis = []
     triggered_windows = []
-    windows_table_data = [] # مصفوفة لتجهيز الجدول النصي في اللوغز
+    windows_table_data = [] 
 
     for idx, win_ids in enumerate(windows):
         window_text = tokenizer.decode(win_ids, skip_special_tokens=True)
@@ -83,6 +89,7 @@ def predict_safety_api(text):
             padding="max_length",
             max_length=60
         )
+        
         with torch.no_grad():
             outputs = model(**inputs)
         probs = torch.softmax(outputs.logits, dim=-1).flatten().tolist()
@@ -99,13 +106,13 @@ def predict_safety_api(text):
             "prediction": prediction
         })
         
-        # تجهيز أسطر الجدول النصي الفخم لمحاكاة مخرجات الكونسول المطور
+        
         windows_table_data.append([
             f"Win {idx + 1}",
             window_text[:45] + "..." if len(window_text) > 45 else window_text,
             f"{safe_prob * 100:.2f}%",
             f"{unsafe_prob * 100:.2f}%",
-            f"❌ {prediction}" if prediction == "UNSAFE" else f"🍏 {prediction}"
+            f"❌ {prediction}" if prediction == "UNSAFE" else f"✅ {prediction}"
         ])
         
         if unsafe_prob > 0.50:
@@ -123,30 +130,33 @@ def predict_safety_api(text):
     safe_confidence_score = round(1.0 - highest_unsafe_prob, 4)
     final_confidence = unsafe_confidence_score if is_blocked else safe_confidence_score
 
-    # 🔥 نقل كامل مخرجات الكونسول الملونة والمنظمة لتطبع حياً داخل شاشة الـ Logs السوداء
-    alert_banner = "🚨 [BLOCK] CHILDSHIELD AI INFERENCE REPORT" if is_blocked else "🍏 [PASS] CHILDSHIELD AI INFERENCE REPORT"
+  
+    alert_banner = "🚨 [BLOCK] CHILDSHIELD AI INFERENCE REPORT" if is_blocked else "✅ [PASS] CHILDSHIELD AI INFERENCE REPORT"
     print(f"\n================ {alert_banner} ================")
-    print(f"📥 Received Original Text:\n\"{text.strip()}\"")
-    print(f"\n🧹 Preprocessed Cleaned Text:\n\"{cleaned_text}\"")
-    print(f"\n🔑 Total Page Tokens Count  : {total_tokens_count}")
-    print(f"🪟 Total Sliding Windows Run : {total_windows_count} Windows (Size: 60, Overlap: 20)")
-    print(f"🎯 Final Security Verdict    : {final_prediction}")
-    print(f"🏆 Model Decision Confidence : {formatted_confidence}")
-    print(f"🛑 Triggered Windows ID      : {triggered_windows}")
-    print("\n📊 --- Windows Detailed Semantic Analysis Table ---")
+    print(f" Received Original Text:\n\"{text.strip()}\"")
+    print(f"\n Preprocessed Cleaned Text:\n\"{cleaned_text}\"")
+    print(f"\n Total Page Tokens Count  : {total_tokens_count}")
+    print(f" Total Sliding Windows Run : {total_windows_count} Windows (Size: 60, Overlap: 20)")
+    print(f" Final Security Verdict    : {final_prediction}")
+    print(f" Model Decision Confidence : {formatted_confidence}")
+    print(f" Triggered Windows ID      : {triggered_windows}")
+    print("\n --- Windows Detailed Semantic Analysis Table ---")
     print(tabulate(windows_table_data, headers=["ID", "Window Text Preview", "Safe Prob", "Unsafe Prob", "Verdict"], tablefmt="grid"))
     print("========================================================================\n")
 
-    # 🎯 التحديث التوسعي العميق: تخزين وحفظ النصوص والنوافذ والنسب المئوية كاملة للمستقبل
+  
     try:
         log_file_path = "production_logs.txt"
-        # تحويل مصفوفة النوافذ كاملة لنص جيسون مضغوط في سطر واحد
+        
         windows_json_blob = json.dumps(windows_analysis, ensure_ascii=False)
         with open(log_file_path, "a", encoding="utf-8") as log_file:
             log_file.write(f"Verdict: {final_prediction} | Confidence: {formatted_confidence} | Tokens: {total_tokens_count} | Windows: {total_windows_count} | Text: {text.strip()} | DeepAnalysis: {windows_json_blob}\n")
+    
+    
     except Exception as e:
         print(f"⚠️ [Logging Warning] Could not write to log file: {e}")
 
+    
     return {
         "original_text": text,
         "cleaned_text": cleaned_text,
@@ -162,11 +172,13 @@ def predict_safety_api(text):
         "confidence": formatted_confidence 
     }  
 
+
 @app.post("/predict")
 def predict(data: InputData):
     result = predict_safety_api(data.text)
     return result
 
+
 gradio_interface = gr.Interface(    
     fn=predict_safety_api,
     inputs=gr.Textbox(lines=4, placeholder="Enter Arabic text to analyze..."),