import gradio as gr
import os

def process_inputs(image, audio, emotion=None, cfg_scale=1.2):
    # 处理逻辑
    current_dir = os.path.dirname(os.path.abspath(__file__))
    video_path = os.path.join(current_dir, "src", "prevideo", "obama.mp4")
    return video_path
with gr.Blocks(theme=gr.themes.Soft(), css=".gradio-container {max-width: 960px !important; margin: 0 auto !important}") as demo:
    gr.HTML(
        """
        <div align='center'>
            <h1>AI数字人视频生成</h1>
            <p style="display:flex; justify-content: center; gap: 10px;">
                <a href='#'><img src='https://img.shields.io/badge/AI-Digital_Human-blue'></a>
                <a href='#'><img src='https://img.shields.io/badge/Video-Generation-red'></a>
                <a href='#'><img src='https://img.shields.io/badge/Emotion-Control-green'></a>
            </p>
        </div>
        """
    )

    with gr.Row(variant="panel"):
        with gr.Column(scale=1):
            with gr.Row():
                source_image = gr.Image(
                    label="上传参考图片",
                    type="filepath"
                )

            with gr.Row():
                driving_audio = gr.Audio(
                    label="上传音频",
                    type="filepath"
                )

            with gr.Row():
                emotion_dropdown = gr.Dropdown(
                    label="选择情感",
                    choices=["开心", "悲伤", "愤怒", "平静"],
                    value="平静"
                )

            with gr.Row():
                cfg_slider = gr.Slider(
                    label="生成强度",
                    minimum=1.0,
                    maximum=3.0,
                    step=0.05,
                    value=1.2
                )

            submit_button = gr.Button("生成视频", variant="primary")

        with gr.Column(scale=1):
            output_video = gr.Video(
                label="生成的视频",
                height=480,
                width=640,
                autoplay=True
            )

    gr.Markdown(
        """
        ---
        ### **使用说明**
        1. 上传一张清晰的正面照片
        2. 上传想要数字人说的音频
        3. 选择期望的情感表现
        4. 调整生成强度（数值越大，表情越夸张）
        5. 点击"生成视频"按钮

        ### **免责声明**
        本项目仅用于学术研究，不对用户生成的内容负责。用户在使用此生成模型时需要对自己的行为负责。
        """
    )
    
    submit_button.click(
        fn=process_inputs,
        inputs=[
            source_image,
            driving_audio,
            emotion_dropdown,
            cfg_slider
        ],
        outputs=output_video
    )
demo.launch()