app.py

import torch
import gradio as gr
from transformers import AutoTokenizer, ViTImageProcessor, VisionEncoderDecoderModel
import zipfile
import os
import csv
from PIL import Image

device = 'cpu'
model_name="NourFakih/Vit-GPT2-COCO2017Flickr-40k-05"
# Load the pretrained model, feature extractor, and tokenizer
model = VisionEncoderDecoderModel.from_pretrained(model_name).to(device)
feature_extractor = ViTImageProcessor.from_pretrained(model_name)
tokenizer = AutoTokenizer.from_pretrained(model_name)

def predict(image, max_length=64, num_beams=4):
    # Process the input image
    image = image.convert('RGB')
    pixel_values = feature_extractor(images=image, return_tensors="pt").pixel_values.to(device)
    
    # Generate the caption
    caption_ids = model.generate(pixel_values, max_length=max_length, num_beams=num_beams)[0]
    
    # Decode and clean the generated caption
    caption = tokenizer.decode(caption_ids, skip_special_tokens=True)
    return caption

def process_images(image_files):
    captions = []
    for image_file in image_files:
        try:
            # Open and verify the image
            with Image.open(image_file) as img:
                caption = predict(img)
                captions.append((os.path.basename(image_file), caption))
        except Exception as e:
            print(f"Skipping file {image_file}: {e}")
    
    # Save the results to a CSV file
    csv_file_path = 'image_captions.csv'
    with open(csv_file_path, mode='w', newline='') as file:
        writer = csv.writer(file)
        writer.writerow(['Image', 'Caption'])
        writer.writerows(captions)
    
    return csv_file_path

def process_zip_files(zip_file_paths):
    # Create a directory to extract images
    extract_dir = 'extracted_images'
    os.makedirs(extract_dir, exist_ok=True)
    
    captions = []
    for zip_file_path in zip_file_paths:
        with zipfile.ZipFile(zip_file_path, 'r') as zip_ref:
            zip_ref.extractall(extract_dir)
        
        # Verify extracted files and process images
        for root, dirs, files in os.walk(extract_dir):
            for file in files:
                file_path = os.path.join(root, file)
                try:
                    # Open and verify the image
                    with Image.open(file_path) as img:
                        caption = predict(img)
                        captions.append((file, caption))
                except Exception as e:
                    print(f"Skipping file {file}: {e}")
    
    # Save the results to a CSV file
    csv_file_path = 'zip_image_captions.csv'
    with open(csv_file_path, mode='w', newline='') as file:
        writer = csv.writer(file)
        writer.writerow(['Image Name', 'Caption'])
        writer.writerows(captions)
    
    return csv_file_path

def gr_process(zip_files, image_files):
    if not zip_files and not image_files:
        raise ValueError("At least one of zip_files or image_files must be provided.")
    elif zip_files:
        zip_file_paths = [zip_file.name for zip_file in zip_files]
        return process_zip_files(zip_file_paths)
    elif image_files:
        image_file_paths = [image_file.name for image_file in image_files]
        return process_images(image_file_paths)

def combine_csv_files(file1, file2, output_file='combined_captions.csv'):
    with open(output_file, mode='w', newline='') as outfile:
        writer = csv.writer(outfile)
        writer.writerow(['Image Name', 'Caption'])
        
        for file in [file1, file2]:
            if os.path.exists(file):
                with open(file, mode='r') as infile:
                    reader = csv.reader(infile)
                    next(reader)  # Skip header row
                    for row in reader:
                        writer.writerow(row)
    
    return output_file

css = '''
h1#title {
  text-align: center;
}
h3#header {
  text-align: center;
}
img#overview {
  max-width: 800px;
  max-height: 600px;
}
img#style-image {
  max-width: 1000px;
  max-height: 600px;
}
.gr-image {
  max-width: 150px;  /* Set a small box for the image */
  max-height: 150px;
}
'''

demo = gr.Blocks(css=css)

with demo:
    gr.Markdown('''<h1 id="title">Image Caption 🖼️</h1>''')
    gr.Markdown('''Made by : No. Fa.''')
    
    zip_files = gr.State([])
    image_files = gr.State([])

    with gr.Row():
        with gr.Column(scale=1):
            new_zip_files = gr.File(label="Upload Zip Files", type="filepath", file_count="multiple")
            generate_zip_captions_btn = gr.Button("Generate Zip Captions")
            new_image_files = gr.File(label="Upload Images", type="filepath", file_count="multiple")
            generate_image_captions_btn = gr.Button("Generate Image Captions")
        with gr.Column(scale=3):
            output_zip_file = gr.File(label="Download Zip Captions")
            output_image_file = gr.File(label="Download Image Captions")
            combined_file = gr.File(label="Download Combined Captions")
    combine_files_btn = gr.Button("Combine CSV Files")

   
    generate_zip_captions_btn.click(fn=gr_process, inputs=new_zip_files, outputs=output_zip_file)
    generate_image_captions_btn.click(fn=gr_process, inputs=image_files, outputs=output_image_file)
    combine_files_btn.click(fn=combine_csv_files, inputs=[output_zip_file, output_image_file], outputs=combined_file)

demo.launch()