Update app.py

app.py

@@ -1,59 +1,95 @@
-import warnings
-warnings.filterwarnings('ignore')
-import transformers
 import gradio as gr
 from transformers import pipeline
-from TTS.api import TTS
+import scipy.io.wavfile
+import numpy as np
 
-#1 - Sentiment Analysis Tab:  
-sentiment =pipeline("text-classification", model="tabularisai/multilingual-sentiment-analysis")
+# --- تحميل النماذج (تم اختيار نماذج خفيفة وسريعة للنسخة المجانية) ---
 
-def get_sentiment(text):
-    result = sentiment(text)
-    label = result[0]['label']
-    score = result[0]['score']
-    return label, score
+# 1. Sentiment Analysis:
+# استبدلنا النموذج بنموذج DistilBERT خفيف وسريع ويدعم لغات متعددة
+sentiment_pipeline = pipeline("text-classification", model="lxyuan/distilbert-base-multilingual-cased-sentiments-student")
+
+# 2. Summarization:
+# استبدلنا Pegasus (الثقيل جداً) بنموذج DistilBART (سريع وممتاز للنسخة المجانية)
+summarizer_pipeline = pipeline("summarization", model="sshleifer/distilbart-cnn-12-6")
+
+# 3. Text-to-Speech:
+# استبدلنا مكتبة TTS الخارجية بنموذج مدمج داخل transformers لتوفير الذاكرة وتجنب الأخطاء
+# نستخدم نموذج MMS من فيسبوك للغة الإنجليزية (سريع جداً)
+tts_pipeline = pipeline("text-to-speech", model="facebook/mms-tts-eng")
+
+
+# --- الدوال (Functions) ---
 
-#2 - Summarization Tab:  
-summary_pipe = pipeline("summarization", model="google/pegasus-cnn_dailymail")
+def get_sentiment(text):
+    if not text:
+        return "Please enter text", "0.0"
+    try:
+        result = sentiment_pipeline(text)
+        label = result[0]['label']
+        score = round(result[0]['score'], 4)
+        return label, str(score)
+    except Exception as e:
+        return "Error", str(e)
 
 def get_summary(text):
-    max_len = int(len(text.split()) * 0.35)
-    max_len = max(5, min(max_len, 512))
-    min_len = int(len(text.split()) * 0.2)
-    
-    result = summary_pipe(text, max_length=max_len, min_length=min_len, do_sample=False)
-    return result[0]['summary_text']
+    if not text:
+        return "Please enter text to summarize."
+    try:
+        # حساب أطوال ذكية للملخص بناءً على النص المدخل
+        input_len = len(text.split())
+        max_len = max(10, int(input_len * 0.5))
+        min_len = max(5, int(input_len * 0.1))
+        
+        # التأكد من الحدود القصوى للنموذج
+        max_len = min(max_len, 100) 
 
-#3 - text-to-Speech Tab:
-tts_model= TTS("tts_models/en/ljspeech/tacotron2-DDC", gpu=False)
+        result = summarizer_pipeline(text, max_length=max_len, min_length=min_len, do_sample=False)
+        return result[0]['summary_text']
+    except Exception as e:
+        return f"Error in summarization: {str(e)}"
 
 def text_to_speech(text):
-    filename = "output.wav"
-    tts_model.tts_to_file(text=text, file_path=filename)
-    return filename
+    if not text:
+        return None
+    try:
+        speech = tts_pipeline(text)
+        # تحويل البيانات الصوتية لنسق يفهمه Gradio
+        audio_data = speech["audio"]
+        sampling_rate = speech["sampling_rate"]
+        return (sampling_rate, audio_data.T) # Transpose for Gradio compatibility
+    except Exception as e:
+        print(f"TTS Error: {e}")
+        return None
 
+# --- واجهة Gradio ---
 
-with gr.Blocks() as AI_applications:
+with gr.Blocks(theme=gr.themes.Soft()) as demo:
+    gr.Markdown("# 🚀 AI Multi-Tool (Lightweight Version)")
+    
     with gr.Tab("Sentiment Analysis"):
-        sentiment_input = gr.Textbox(label="Enter Text")
-        output_label = gr.Text(label="Sentiment Label")
-        output_score = gr.Text(label="Confidence Score")
-        analyze_button = gr.Button("Analyze Sentiment")
-        analyze_button.click(fn=get_sentiment,inputs=sentiment_input ,outputs=[output_label, output_score])
+        gr.Markdown("Analyze the emotion of the text (Positive/Negative).")
+        with gr.Row():
+            sent_input = gr.Textbox(label="Enter Text", placeholder="I love using AI models!", lines=2)
+            sent_btn = gr.Button("Analyze", variant="primary")
+        with gr.Row():
+            sent_label = gr.Textbox(label="Label")
+            sent_score = gr.Textbox(label="Confidence Score")
+        sent_btn.click(fn=get_sentiment, inputs=sent_input, outputs=[sent_label, sent_score])
 
     with gr.Tab("Summarization"):
-
-        summary_input  = gr.Textbox(label="Enter Long Text", lines=10, placeholder="Paste your text here...")
-        output_summary = gr.Textbox(label="Summary", lines=5)
-        summarize_btn  = gr.Button("Summarize")
-        summarize_btn .click(fn=get_summary, inputs=summary_input, outputs=output_summary)
+        gr.Markdown("Summarize long articles quickly.")
+        sum_input = gr.Textbox(label="Long Text", lines=8, placeholder="Paste your article here...")
+        sum_btn = gr.Button("Summarize", variant="primary")
+        sum_output = gr.Textbox(label="Summary", lines=4)
+        sum_btn.click(fn=get_summary, inputs=sum_input, outputs=sum_output)
 
     with gr.Tab("Text to Speech"):
+        gr.Markdown("Convert text to audio (English).")
+        tts_input = gr.Textbox(label="Text to Speak", lines=3, placeholder="Hello, how are you doing today?")
+        tts_btn = gr.Button("Generate Audio", variant="primary")
+        tts_output = gr.Audio(label="Output Audio")
+        tts_btn.click(fn=text_to_speech, inputs=tts_input, outputs=tts_output)
 
-        tts_input = gr.Textbox(label="Enter Text to Speak", lines=4, placeholder="Type something...")
-        output_audio = gr.Audio(label="Generated Speech", type="filepath")
-        generate_btn = gr.Button("Generate Speech")
-        generate_btn.click(fn=text_to_speech, inputs=tts_input , outputs=output_audio)
-
-AI_applications.launch()
+if __name__ == "__main__":
+    demo.launch()