app.py

import os
import cv2
import spaces
from PIL import Image
import gradio as gr
import numpy as np
import random
import base64
import requests
import json
import time

# Add a new function for text-to-image generation
def generate_garment_image(prompt):
    # This is a placeholder function. You'll need to implement actual text-to-image generation here.
    # For example, you might use a service like DALL-E, Stable Diffusion, or any other text-to-image model.
    # For now, we'll just return a placeholder image.
    placeholder_image = np.zeros((256, 256, 3), dtype=np.uint8)
    cv2.putText(placeholder_image, prompt, (10, 128), cv2.FONT_HERSHEY_SIMPLEX, 0.5, (255, 255, 255), 1)
    return placeholder_image

def tryon(person_img, garment_prompt, seed, randomize_seed):
    post_start_time = time.time()
    if person_img is None or garment_prompt == "":
        return None, None, "Empty image or prompt"
    if randomize_seed:
        seed = random.randint(0, MAX_SEED)
    
    # Generate garment image from prompt
    garment_img = generate_garment_image(garment_prompt)
    
    encoded_person_img = cv2.imencode('.jpg', cv2.cvtColor(person_img, cv2.COLOR_RGB2BGR))[1].tobytes()
    encoded_person_img = base64.b64encode(encoded_person_img).decode('utf-8')
    encoded_garment_img = cv2.imencode('.jpg', cv2.cvtColor(garment_img, cv2.COLOR_RGB2BGR))[1].tobytes()
    encoded_garment_img = base64.b64encode(encoded_garment_img).decode('utf-8')

    # Rest of the function remains the same
    # ...

def start_tryon(person_img, garment_prompt, seed, randomize_seed):
    start_time = time.time()
    if person_img is None or garment_prompt == "":
        return None, None, "Empty image or prompt"
    if randomize_seed:
        seed = random.randint(0, MAX_SEED)
    
    # Generate garment image from prompt
    garment_img = generate_garment_image(garment_prompt)
    
    encoded_person_img = cv2.imencode('.jpg', cv2.cvtColor(person_img, cv2.COLOR_RGB2BGR))[1].tobytes()
    encoded_person_img = base64.b64encode(encoded_person_img).decode('utf-8')
    encoded_garment_img = cv2.imencode('.jpg', cv2.cvtColor(garment_img, cv2.COLOR_RGB2BGR))[1].tobytes()
    encoded_garment_img = base64.b64encode(encoded_garment_img).decode('utf-8')

    # Rest of the function remains the same
    # ...

MAX_SEED = 999999

example_path = os.path.join(os.path.dirname(__file__), 'assets')

human_list = os.listdir(os.path.join(example_path,"human"))
human_list_path = [os.path.join(example_path,"human",human) for human in human_list]

css="""
#col-left {
    margin: 0 auto;
    max-width: 430px;
}
#col-mid {
    margin: 0 auto;
    max-width: 430px;
}
#col-right {
    margin: 0 auto;
    max-width: 430px;
}
#col-showcase {
    margin: 0 auto;
    max-width: 1100px;
}
#button {
    color: blue;
}
"""

def load_description(fp):
    with open(fp, 'r', encoding='utf-8') as f:
        content = f.read()
    return content

with gr.Blocks(css=css) as Tryon:
    gr.HTML(load_description("assets/title.md"))
    with gr.Row():
        with gr.Column(elem_id = "col-left"):
            gr.HTML("""
            <div style="display: flex; justify-content: center; align-items: center; text-align: center; font-size: 20px;">
                <div>
                Step 1.  Upload a person image ⬇️
                </div>
            </div>
            """)
        with gr.Column(elem_id = "col-mid"):
            gr.HTML("""
            <div style="display: flex; justify-content: center; align-items: center; text-align: center; font-size: 20px;">
                <div>
                Step 2. Enter a garment description ⬇️
                </div>
            </div>
            """)
        with gr.Column(elem_id = "col-right"):
            gr.HTML("""
            <div style="display: flex; justify-content: center; align-items: center; text-align: center; font-size: 20px;">
                <div>
                Step 3. Press "Run" to get try-on results
                </div>
            </div>
            """)
    with gr.Row():
        with gr.Column(elem_id = "col-left"):
            imgs = gr.Image(label="Person image", sources='upload', type="numpy")
            example = gr.Examples(
                inputs=imgs,
                examples_per_page=12,
                examples=human_list_path
            )
        with gr.Column(elem_id = "col-mid"):
            garm_prompt = gr.Textbox(label="Garment description", placeholder="Enter a description of the garment...")
            example_prompts = gr.Examples(
                inputs=garm_prompt,
                examples=["A red t-shirt", "Blue jeans", "A floral summer dress", "A black leather jacket"]
            )
        with gr.Column(elem_id = "col-right"):
            image_out = gr.Image(label="Result", show_share_button=False)
            with gr.Row():
                seed = gr.Slider(
                    label="Seed",
                    minimum=0,
                    maximum=MAX_SEED,
                    step=1,
                    value=0,
                )
                randomize_seed = gr.Checkbox(label="Random seed", value=True)
            with gr.Row():
                seed_used = gr.Number(label="Seed used")
                result_info = gr.Text(label="Response")
            test_button = gr.Button(value="Run", elem_id="button")

    test_button.click(fn=tryon, inputs=[imgs, garm_prompt, seed, randomize_seed], outputs=[image_out, seed_used, result_info], api_name='tryon', concurrency_limit=40)

    with gr.Column(elem_id = "col-showcase"):
        gr.HTML("""
        <div style="display: flex; justify-content: center; align-items: center; text-align: center; font-size: 20px;">
            <div> </div>
            <br>
            <div>
            Virtual try-on examples in pairs of person images and garment descriptions
            </div>
        </div>
        """)
        show_case = gr.Examples(
            examples=[
                ["assets/examples/model2.png", "A blue t-shirt", "assets/examples/result2.png"],
                ["assets/examples/model3.png", "A red dress", "assets/examples/result3.png"],
                ["assets/examples/model1.png", "A black suit", "assets/examples/result1.png"],
            ],
            inputs=[imgs, garm_prompt, image_out],
            label=None
        )

Tryon.launch()