import gradio as gr
from transformers import BlipProcessor, BlipForConditionalGeneration
from deep_translator import GoogleTranslator
from PIL import Image

processor = BlipProcessor.from_pretrained("Salesforce/blip-image-captioning-base")
model = BlipForConditionalGeneration.from_pretrained("Salesforce/blip-image-captioning-base")

def caption_image(image):
    if image is None:
        return " من فضلك ارفع صورة"
    
    inputs = processor(image, return_tensors="pt")
    output = model.generate(**inputs, max_new_tokens=50)
    caption_en = processor.decode(output[0], skip_special_tokens=True)
    caption_ar = GoogleTranslator(source='en', target='ar').translate(caption_en)
    
    return f"** الوصف:** {caption_ar}"

demo = gr.Interface(
    fn=caption_image,
    inputs=gr.Image(label="ارفع الصوره التي تريد وصفها", type="pil"),
    outputs=gr.Markdown(label="الوصف"),
    title="مولّد أوصاف الصور",
   
)

demo.launch()