Update app.py

app.py

@@ -1,112 +1,3 @@
-import gradio as gr
-import torch
+import streamlit as st
 from transformers import pipeline
-from datasets import load_dataset
-from PIL import Image
-import numpy as np
-from collections import Counter
-import functools
-'''
-# 使用标准库的缓存装饰器替代Gradio缓存
-@functools.lru_cache(maxsize=None)
-def load_models():
-    return {
-        "detector": pipeline(
-            "object-detection", 
-            model="facebook/detr-resnet-50",
-            device=0 if torch.cuda.is_available() else -1
-        ),
-        "generator": pipeline(
-            "text2text-generation",
-            model="google/flan-t5-base",  # 改用基础版降低资源需求
-            device=0 if torch.cuda.is_available() else -1
-        )
-    }
-
-# 数据集加载函数（移除Gradio缓存）
-def load_dataset_data():
-    ds = load_dataset("AntZet/home_decoration_objects_images")
-    return ds['train'].to_pandas()
-
-# 颜色分析函数保持不变
-def get_dominant_colors(img, n_colors=3):
-    arr = np.array(img.resize((100,100)))
-    pixels = arr.reshape(-1,3)
-    from sklearn.cluster import KMeans
-    kmeans = KMeans(n_clusters=n_colors)
-    kmeans.fit(pixels)
-    return [f"#{int(c[0]):02x}{int(c[1]):02x}{int(c[2]):02x}" for c in kmeans.cluster_centers_]
-
-# 核心处理函数
-def generate_recommendation(target_style):
-    try:
-        models = load_models()
-        df = load_dataset_data()
-        
-        style_df = df[df['style'] == target_style.lower()]
-        if len(style_df) < 3:
-            return f"⚠️ Not enough samples for {target_style} style"
-        
-        sample_images = style_df.sample(5)['image']
-        
-        all_objects = []
-        color_palette = []
-        
-        for img in sample_images:
-            detected = models["detector"](img)
-            all_objects += [obj['label'] for obj in detected if obj['score'] > 0.9]
-            color_palette += get_dominant_colors(img)
-        
-        top_objects = Counter(all_objects).most_common(3)
-        top_colors = Counter(color_palette).most_common(3)
-        
-        prompt = f"""Create interior design recommendations for {target_style} style:
-        Key objects: {[o[0] for o in top_objects]}
-        Color palette: {[c[0] for c in top_colors]}
-        Include: 3 essentials, 2 budget tips, common mistakes"""
-        
-        advice = models["generator"](prompt, max_length=300)[0]['generated_text']
-        
-        output = f"## 🎨 {target_style.title()} Style Guide\n\n"
-        output += "### 🪑 Key Objects\n" + "\n".join(
-            [f"- {o[0]} ({o[1]}x)" for o in top_objects]) + "\n\n"
-        output += "### 🎨 Colors\n" + "\n".join(
-            [f"<span style='color:{c[0]};'>■</span> {c[0]}" for c in top_colors]) + "\n\n"
-        output += "### 💡 Advice\n" + advice.replace(". ", ".\n")
-        
-        return output
-    
-    except Exception as e:
-        return f"❌ Error: {str(e)}"
-
-# Gradio界面保持不变
-with gr.Blocks(title="Design Assistant") as demo:
-    gr.Markdown("# 🏡 AI Design Advisor")
-    
-    with gr.Row():
-        style_input = gr.Dropdown(
-            label="Select Style",
-            choices=["Industrial", "Scandinavian", "Bohemian", "Modern"],
-            value="Industrial"
-        )
-    
-    submit_btn = gr.Button("Generate Plan", variant="primary")
-    
-    with gr.Row():
-        output = gr.Markdown()
-        gallery = gr.Gallery(
-            label="Examples",
-            object_fit="contain",
-            height="300px"
-        )
-    
-    def update_gallery(style):
-        df = load_dataset_data()
-        return df[df['style'] == style.lower()].sample(3)['image'].tolist()
-    
-    style_input.change(update_gallery, inputs=style_input, outputs=gallery)
-    submit_btn.click(generate_recommendation, inputs=style_input, outputs=output)
-
-if __name__ == "__main__":
-    demo.launch()
-    '''
+import torch