import gradio as gr
from transformers import BlipProcessor, BlipForConditionalGeneration
from transformers import DetrImageProcessor, DetrForObjectDetection
import torch
from PIL import Image, ImageDraw

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

# Load DETR model for object detection (Detectron)
detr_processor = DetrImageProcessor.from_pretrained("facebook/detr-resnet-50")
detr_model = DetrForObjectDetection.from_pretrained("facebook/detr-resnet-50")

# List of objects for dynamic description
objects_of_interest = ["tree", "water", "mountain", "beach"]

def generate_caption(image):
    # Process the image for caption generation
    inputs = processor(images=image, return_tensors="pt")
    out = model.generate(**inputs)
    caption = processor.decode(out[0], skip_special_tokens=True)

    # Object Detection: Detect objects in the image
    inputs = detr_processor(images=image, return_tensors="pt")
    outputs = detr_model(**inputs)

    # Get detected objects and their labels
    target_sizes = torch.tensor([image.size[::-1]])
    results = detr_processor.post_process_object_detection(outputs, target_sizes=target_sizes, threshold=0.9)[0]

    detected_objects = []
    for score, label in zip(results["scores"], results["labels"]):
        if label.item() == 23:  # label for "tree"
            detected_objects.append("trees")
        if label.item() == 8:  # label for "water"
            detected_objects.append("water")
        if label.item() == 72:  # label for "mountain"
            detected_objects.append("mountains")

    # Custom dynamic description based on detected objects
    description = "This image includes "
    if detected_objects:
        description += ", ".join(detected_objects)
    else:
        description += "various elements of nature."

    description += ". It provides a beautiful view that invites relaxation and exploration."

    return caption + "\n" + description

# Gradio Interface
iface = gr.Interface(fn=generate_caption, 
                     inputs=gr.Image(type="pil"), 
                     outputs=gr.Textbox(), 
                     title="Dynamic Image Caption Generator", 
                     description="Upload any image and get a detailed description of its contents.")

if __name__ == "__main__":
    iface.launch()