app.py

import base64
import os

import gradio as gr
from openai import OpenAI

client = OpenAI(
    api_key=os.getenv('HUNYUAN_API_KEY'),
    base_url="https://api.hunyuan.cloud.tencent.com/v1"
)

def generate_caption(image_path, question):
    # 将图片转换为Base64
    with open(image_path, "rb") as image_file:
        base64_image = base64.b64encode(image_file.read()).decode('utf-8')
    
    # 构建消息结构
    messages = [{
        "role": "user",
        "content": [
            {"type": "text", "text": question},
            {
                "type": "image_url",
                "image_url": {
                    "url": f"data:image/jpeg;base64,{base64_image}"
                }
            }
        ]
    }]

    # 调用混元视觉模型
    response = client.chat.completions.create(
        model="hunyuan-vision",
        messages=messages,
        stream=True,
        extra_body={
            "stream_moderation": True,
            "enable_enhancement": False
        }
    )

    # 流式处理响应
    full_response = ""
    for chunk in response:
        token = chunk.choices[0].delta.content
        if token:
            full_response += token
            yield full_response



# 创建Gradio界面
title = "Hunyuan-Vision图生文Demo"
theme = gr.themes.Soft(
    primary_hue="teal",
    secondary_hue="blue",
    font=[gr.themes.GoogleFont("Noto Sans SC"), "Arial", "sans-serif"]
)

with gr.Blocks(title=title, theme=theme) as demo:
    # ================= 头部区域 =================
    gr.Markdown(f"""
    <div style="text-align: center;">
        <h1 style="color: #2E86C1; border-bottom: 3px solid #AED6F1; padding-bottom: 10px;">🖼️ {title}</h1>
        <p style="color: #616A6B;">上传图片并输入问题，体验腾讯混元视觉大模型的图像理解能力</p>
    </div>
    """)
    
    # ================= 主体区域 =================
    with gr.Row(variant="panel"):
        # 左侧输入列
        with gr.Column(scale=3):
            with gr.Group(label="输入区域"):
                image_input = gr.Image(
                    type="filepath",
                    label="上传图片",
                    height=400,
                    show_download_button=False,
                    elem_classes="preview-box"
                )
                question_input = gr.Textbox(
                    label="问题描述",
                    placeholder="请输入关于图片的问题...",
                    value="请详细描述图片中的场景、人物和细节",
                    lines=2
                )
                with gr.Row():
                    clear_btn = gr.Button("清空", variant="secondary")
                    submit_btn = gr.Button("生成描述", variant="primary")
        
        # 右侧输出列
        with gr.Column(scale=4):
            with gr.Group(label="生成结果"):
                output = gr.Textbox(
                    label="描述内容",
                    interactive=False,
                    show_copy_button=True,
                    lines=12,
                    max_lines=20,
                    autoscroll=True
                )
    
    # ================= 示例区域 =================
    with gr.Accordion("🖼️ 点击查看示例", open=False):
        with gr.Row():
            gr.Examples(
                examples=[
                    ["tencent.png", "图片中的天气状况如何？"],
                    ["tencent.png", "描述参会人员的衣着特征"]
                ],
                inputs=[image_input, question_input],
                label="快速示例"
            )
    
    # ================= 交互逻辑 =================
    submit_btn.click(
        fn=generate_caption,
        inputs=[image_input, question_input],
        outputs=output,
        api_name="generate"
    )
    
    clear_btn.click(
        fn=lambda: [None, "", ""],
        outputs=[image_input, question_input, output],
        queue=False
    )

# ================= 自定义样式 =================
css = """
.preview-box img {border-radius: 10px; box-shadow: 0 4px 6px rgba(0,0,0,0.1);}
.preview-box:hover img {transform: scale(1.02);}
button#generate {transition: all 0.3s ease;}
"""
demo.css = css

if __name__ == "__main__":
    demo.queue(default_concurrency_limit=100)
    demo.launch(
        server_port=7860,
        show_error=True,
        favicon_path="favicon.ico",
        max_threads=100
    )

# # 创建Gradio界面
# title="Hunyuan-Vision图生文Demo"
# with gr.Blocks(title=title) as demo:
#     gr.Markdown(f"# 🖼️ {title}")
#     with gr.Row():
#         with gr.Column():
#             image_input = gr.Image(type="filepath", label="上传图片")
#             question_input = gr.Textbox(label="输入问题", value="请描述图片内容")
#             submit_btn = gr.Button("生成描述")
#         output = gr.Textbox(label="描述结果", interactive=False)

#     submit_btn.click(
#         fn=generate_caption,
#         inputs=[image_input, question_input],
#         outputs=output
#     )

# if __name__ == "__main__":
#     demo.queue(default_concurrency_limit=100)
#     demo.launch(max_threads=100)