image2text.py

import os
from PIL import Image
import gradio as gr
from google import genai

# 初始化 Gemini API
GEMINI_API_KEY = os.environ.get("GEMINI_API_KEY")
client = genai.Client(api_key=GEMINI_API_KEY)

# 定義「圖解釋文」功能
def explain_image(image: Image.Image):
    # 直接把 PIL image 傳進去
    response = client.models.generate_content(
        model="gemini-2.0-flash",
        contents=[image, "使用繁體中文描述這張圖片"],
    )
    # 取出回答
    explanation = response.text
    return explanation

# Gradio 介面
with gr.Blocks() as demo:
    gr.Markdown("## 🧠新興高中 王澄雅 圖片智慧分析（圖 ➜ 文）")
    image_input = gr.Image(type="pil", label="上傳圖片")
    explain_button = gr.Button("解釋圖片")
    output_text = gr.Textbox(label="圖片說明", lines=5)

    explain_button.click(fn=explain_image, inputs=image_input, outputs=output_text)

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