app.py

import os
import time
from typing import List, Tuple, Optional

import google.generativeai as genai
import gradio as gr
from PIL import Image

print("google-generativeai:", genai.__version__)

GOOGLE_API_KEY = os.environ.get("GOOGLE_API_KEY")

TITLE = """<h1 align="center">👗 Gemini Personal Stylist Chatbot 🛍️</h1>"""
#SUBTITLE = """<h2 align="center">Chat with Your Personal Stylist using Gemini Vision Pro 🖼️</h2>"""

IMAGE_WIDTH = 512

def preprocess_image(image: Image.Image) -> Image.Image:
    image_height = int(image.height * IMAGE_WIDTH / image.width)
    return image.resize((IMAGE_WIDTH, image_height))

def user(text_prompt: str, chatbot: List[Tuple[str, str]]):
    return "", chatbot + [[text_prompt, None]]

def bot(
    google_key: str,
    image_prompt: Optional[Image.Image],
    text_prompt: str,
    chatbot: List[Tuple[str, str]]
):
    google_key = google_key or GOOGLE_API_KEY
    if not google_key:
        raise ValueError("GOOGLE_API_KEY is not set. Please set it up.")

    genai.configure(api_key=google_key)
    instructions = "You are an expert stylist. You can assist anyone regardless gender, race or any other demographic diversity, look stylish and confident. "
    images_with_prompt = [text_prompt, instructions] + [preprocess_image(image_prompt)]

    model = genai.GenerativeModel('gemini-pro-vision')
    response = model.generate_content(images_with_prompt, stream=True)
    response.resolve()

    chatbot[-1][1] = ""
    for chunk in response:
        for i in range(0, len(chunk.text), 10):
            chatbot[-1][1] += chunk.text[i:i + 10]
            time.sleep(0.01)
            yield chatbot

google_key_component = gr.Textbox(
    label="GOOGLE API KEY",
    type="password",
    placeholder="Enter your Google API Key",
    visible=GOOGLE_API_KEY is None
)

image_prompt_component = gr.Image(type="pil", label="Upload Image")
chatbot_component = gr.Chatbot(label='Gemini Personal Stylist')
text_prompt_component = gr.Textbox(
    placeholder="Describe your style needs",
    label="Enter your prompt and press Enter"
)
run_button_component = gr.Button("Ask Stylist")

user_inputs = [text_prompt_component, chatbot_component]
bot_inputs = [google_key_component, image_prompt_component, text_prompt_component, chatbot_component]

with gr.Blocks() as demo:
    gr.HTML(TITLE)
    gr.HTML(SUBTITLE)
    with gr.Column():
        google_key_component.render()
        image_prompt_component.render()
        chatbot_component.render()
        text_prompt_component.render()
        run_button_component.render()

    run_button_component.click(
        fn=user,
        inputs=user_inputs,
        outputs=[text_prompt_component, chatbot_component],
        queue=False
    ).then(
        fn=bot, inputs=bot_inputs, outputs=[chatbot_component]
    )

    text_prompt_component.submit(
        fn=user,
        inputs=user_inputs,
        outputs=[text_prompt_component, chatbot_component],
        queue=False
    ).then(
        fn=bot, inputs=bot_inputs, outputs=[chatbot_component]
    )

demo.launch()