import gradio as gr
from transformers import pipeline
from PIL import Image

# Initialize the pipeline with the image captioning model
caption_pipeline = pipeline("image-to-text", model="Salesforce/blip-image-captioning-large")

# Initialize the pipeline for emotion classification
emotion_pipeline = pipeline("image-classification", model="RickyIG/emotion_face_image_classification_v3")

# Initialize the pipeline for object detection
object_pipeline = pipeline("object-detection", model="facebook/detr-resnet-50")

def generate_caption_emotion_and_objects(image):
    # Process the image for captioning
    caption_result = caption_pipeline(image)
    caption = caption_result[0]["generated_text"]

    # Process the image for emotion classification
    emotion_result = emotion_pipeline(image)
    emotions = ", ".join([f"{res['label']}: {res['score']:.2f}" for res in emotion_result])

    # Process the image for object detection
    object_result = object_pipeline(image)
    objects = ", ".join([f"{obj['label']}: {obj['score']:.2f}" for obj in object_result])

    # Combine results
    combined_result = f"Caption: {caption}\nEmotions: {emotions}\nObjects: {objects}"
    return combined_result

# Setup the Gradio interface
interface = gr.Interface(fn=generate_caption_emotion_and_objects,
                         inputs=gr.components.Image(type="pil", label="Upload an Image"),
                         outputs=gr.components.Textbox(label="Generated Caption, Emotions, and Objects Detected"))
interface.launch()