Upload 2 files

app.py

@@ -0,0 +1,171 @@
+import gradio as gr
+from suggestion import generate_outfit_advice
+from clothGen import cloth_gen
+from user_dress import user_cloths
+import requests
+import os
+from io import BytesIO
+from PIL import Image
+from cal_compatibility import cal_compatibility
+
+
+gen_pic_num = 6
+
+
+def get_select_index(evt: gr.SelectData, gallery):
+    print(gallery[evt.index][0])
+    # response = requests.get(gallery[evt.index][0])
+    # img = Image.open(BytesIO(response.content))
+    return gallery[evt.index][0]
+
+
+def update_choices(dropout1, dropout2,):
+    if dropout1 == "男":
+        option = ['倒三角形', '矩形', '苹果形', '沙漏型', '胖型']
+    else:
+        option = ["梨形", "草莓形", "沙漏形", "标准", "苹果形"]
+    dropout2 = gr.Dropdown(choices=option)
+    return dropout2
+
+
+with gr.Blocks(css="styles.css", theme=gr.themes.Base()) as demo:
+    with gr.Row():
+        # 左侧模块
+        with gr.Column(scale=1):
+            with gr.Row():
+                gr.Markdown(""
+                            "# 用户信息"
+                            "")
+            with gr.Row():
+                text_input1 = gr.Textbox(label="用户姓名", min_width=100)
+                text_input2 = gr.Textbox(label="身高/cm", min_width=100)
+                text_input3 = gr.Textbox(label="体重/kg", min_width=100)
+                text_input4 = gr.Textbox(label="腰围/cm", min_width=100)
+                text_input5 = gr.Textbox(label="胸围/cm", min_width=100)
+                text_input6 = gr.Textbox(label="臀围/cm", min_width=100)
+                text_input7 = gr.Textbox(label="肩宽/cm", min_width=100)
+                text_input8 = gr.Textbox(label="腿长/cm", min_width=100)
+                text_input9 = gr.Textbox(label="臂长/cm", min_width=100)
+                dropdown_options1 = ["男", "女"]
+                dropdown_input1 = gr.Dropdown(choices=dropdown_options1, label="性别", min_width=100)
+                dropdown_input2 = gr.Dropdown(choices=[], label="体型分类", min_width=100)
+                dropdown_input1.change(fn=update_choices, inputs=[dropdown_input1, dropdown_input2], outputs=dropdown_input2)
+                dropdown_options3 = ["浅色", "中等偏黄色", "中等偏褐色", "深色"]
+                dropdown_input3 = gr.Dropdown(choices=dropdown_options3, label="肤色", min_width=100)
+                text_input10 = gr.Textbox(label="穿衣风格偏好", min_width=1000)
+                text_input11 = gr.Textbox(label="生话方式和场景需求", min_width=1000)
+                text_input12 = gr.Textbox(label="其他特殊需求", min_width=1000)
+            with gr.Row():
+                user_pic = gr.Image(label="用户照片", value="model.jpg", height=550, width=300)
+
+        # 右侧模块
+        with gr.Column(scale=2):
+            with gr.Row():
+                gr.Markdown(""
+                            "# 穿搭建议"
+                            "")
+            with gr.Row():
+                text_output1 = gr.Textbox(label="穿搭建议", lines=12, max_lines=12, interactive=False, show_label=False,
+                                          min_width=1000)
+                submit_button_1 = gr.Button("AI智能分析，生成穿搭建议", min_width=1000)
+            image_output_1 = gr.Image(label="显示图像", value="image 209.png")
+
+            gallery_1 = gr.Gallery(
+                label="服装", elem_id="gallery",
+                value=[
+                    # os.path.join(example_path, '上衣/_WEB_2016_09_26__2016092617451357e8ee2957aa1_TD.jpg'),
+                    # os.path.join(example_path, '上衣/_WEB_2016_09_27__2016092717211057ea3a069c749_TD.jpg'),
+                    # os.path.join(example_path, '上衣/_WEB_2016_09_27__2016092717391657ea3e446ce3f_TD.jpg'),
+                    # os.path.join(example_path, '上衣/_WEB_2016_09_27__2016092717573057ea428a305bc_TD.jpg'),
+                    # os.path.join(example_path, '上衣/_WEB_2016_09_28__2016092810150157eb27a56a631_TD.jpg'),
+                    # os.path.join(example_path, '上衣/_WEB_2016_09_28__2016092810464557eb2f15e1df3_TD.jpg'),
+                ],
+                columns=[4], rows=[2], object_fit="contain", height=250, min_width=450)
+
+            gallery_3 = gr.Gallery(
+                label="配饰", elem_id="gallery",
+                value=[
+                    'downloads/access_1.jpg',
+                    'downloads/access_2.jpg',
+                    'downloads/access_3.jpg',
+                    'downloads/access_4.jpg',
+                    'downloads/access_5.jpg',
+                    'downloads/access_6.jpg',
+                ],
+                columns=[4], rows=[2], object_fit="contain", height=250, min_width=450)
+            submit_button_2 = gr.Button("AI智能分析，生成民族服饰")
+
+        with gr.Column(scale=2):
+            with gr.Row():
+                gr.Markdown(""
+                            "# 搭配生成"
+                            "")
+            with gr.Row():
+                gallery_4 = gr.Gallery(
+                    label="套装", elem_id="gallery",
+                    value=[],
+                    columns=[3], rows=[1], object_fit="contain", height=180, min_width=450)
+            with gr.Row():
+                submit_button_3 = gr.Button("服饰及搭配兼容性排序")
+            with gr.Row():
+                image_output_5 = gr.Image(label="显示图像", show_label=False, min_width=200, height=350)
+                intro = gr.Textbox(label="服饰介绍", lines=14, max_lines=14)
+            with gr.Row():
+                submit_button_4 = gr.Button("虚拟试穿")
+            with gr.Row():
+                with gr.Column(scale=1):
+                    gallery_user = gr.Gallery(
+                        label="试穿结果",
+                        elem_id="gallery",
+                        value=[],
+                        columns=[3], rows=[2],
+                        object_fit="contain",
+                        min_width=200,
+                        height=350,
+                    )
+                with gr.Column(scale=1):
+                    feedback = gr.Textbox(label="反馈", placeholder="可以从款式、颜色、图案、风格倾向、文化偏好角度进行反馈", lines=1,
+                                          max_lines=1, elem_id="feedback")
+                    submit_button_5 = gr.Button("反馈")
+    with gr.Row():
+        gr.Markdown("""
+                    女性体型备注：
+                    1. **梨形身材**：臀围比胸围**至少大5.08厘米**
+                    2. **草莓形身材**：臀围比胸围**至少小5.08厘米**
+                    3. **沙漏形身材**：胸围比腰围**至少大3.81厘米**且腰部线条明显
+                    4. **标准身材**：胸围比腰围**至少大3.81厘米**且腰部线条不明显
+                    5. **苹果形身材**：胸围比腰围**至少小3.81厘米**
+                    
+                    男性体型备注：
+                    1. **倒三角形身材**：肩宽比腰围**至少大10厘米**  
+                    2. **矩形身材**：肩宽与腰围的差异**小于5厘米**  
+                    3. **苹果形身材**：腰围比肩宽**至少大7.5厘米**  
+                    4. **沙漏型身材**：腰围比肩宽或臀围**至少小10厘米**，且肩宽和臀围的差异**小于5厘米**  
+                    5. **胖型身材**：腰围比胸围或肩宽**至少大10厘米**
+                    """)
+
+    submit_button_1.click(fn=generate_outfit_advice,
+                          inputs=[text_input1, text_input2, text_input3, text_input4, text_input5,
+                                  text_input6, text_input7, text_input8, text_input9, dropdown_input1,
+                                  dropdown_input2, dropdown_input3, text_input10, text_input11, text_input12,
+                                  feedback, user_pic],
+                          outputs=text_output1)
+    submit_button_2.click(fn=cloth_gen,
+                          inputs=[text_output1, dropdown_input1],
+                          outputs=[gallery_1, image_output_5, intro])
+
+    gallery_1.select(get_select_index, gallery_1, image_output_5)
+    submit_button_3.click(fn=cal_compatibility,
+                          inputs=[],
+                          outputs=[gallery_4])
+    submit_button_4.click(fn=user_cloths,
+                          inputs=[user_pic, image_output_5],
+                          outputs=gallery_user)
+    submit_button_5.click(fn=generate_outfit_advice,
+                          inputs=[text_input1, text_input2, text_input3, text_input4, text_input5,
+                                  text_input6, text_input7, text_input8, text_input9, dropdown_input1,
+                                  dropdown_input2, dropdown_input3, text_input10, text_input11, text_input12,
+                                  feedback, user_pic],
+                          outputs=text_output1)
+
+demo.launch(server_port=7860, share=True)

user_dress.py

@@ -0,0 +1,106 @@
+from io import BytesIO
+import fal_client
+import os
+from PIL import Image
+import requests
+import time
+
+UPLOAD_FOLDER = 'uploads'
+DOWNLOAD_FOLDER = 'downloads'
+
+
+def user_dress_cat(user_pic, cloth_gen, index):
+    time_1 = time.time()
+    filename_1 = 'user_image.jpg'
+    filename_2 = 'cloth_image.jpg'
+    file_path_1 = os.path.join(UPLOAD_FOLDER, filename_1)
+    file_path_2 = os.path.join(UPLOAD_FOLDER, filename_2)
+    Image.fromarray(user_pic).save(file_path_1)
+    Image.fromarray(cloth_gen).save(file_path_2)
+    time_2 = time.time()
+    save_time = time_2 - time_1
+    print(f"save_time:{save_time}")
+    time_1 = time.time()
+    handler = fal_client.submit(
+        "fal-ai/cat-vton",
+        arguments={
+            "human_image_url": fal_client.upload_file(file_path_1),
+            "garment_image_url": fal_client.upload_file(file_path_2),
+            "cloth_type": "overall"
+        },
+    )
+    request_id = handler.request_id
+    result = fal_client.result("fal-ai/cat-vton", request_id)
+    time_2 = time.time()
+    cat_time = time_2 - time_1
+    print(f"cat_time:{cat_time}")
+    time_1 = time.time()
+    response = requests.get(result['image']['url'])
+    time_2 = time.time()
+    url_time = time_2 - time_1
+    print(f"url_time:{url_time}")
+    time_1 = time.time()
+    save_directory = "downloads"
+    if response.status_code == 200:
+        filename = os.path.join(save_directory, f"cat-vton_{index}.png")
+        with open(filename, 'wb') as f:
+            f.write(response.content)
+    else:
+        print(f"Failed to download image from {result['image']['url']}")
+    time_2 = time.time()
+    downloads_time = time_2 - time_1
+    print(f"downloads_time:{downloads_time}")
+    return os.path.join(save_directory, f"cat-vton_{index}.png")
+
+
+def user_dress_idm(user_pic, cloth_gen, index):
+    time_1 = time.time()
+    filename_1 = 'user_image.jpg'
+    filename_2 = 'cloth_image.jpg'
+    file_path_1 = os.path.join(UPLOAD_FOLDER, filename_1)
+    file_path_2 = os.path.join(UPLOAD_FOLDER, filename_2)
+    Image.fromarray(user_pic).save(file_path_1)
+    Image.fromarray(cloth_gen).save(file_path_2)
+    time_2 = time.time()
+    save_time = time_2 - time_1
+    print(f"save_time:{save_time}")
+    time_1 = time.time()
+    handler = fal_client.submit(
+        "fal-ai/idm-vton",
+        arguments={
+            "human_image_url": fal_client.upload_file(file_path_1),
+            "garment_image_url": fal_client.upload_file(file_path_2),
+            "description": "long-sleeved long coat"
+        },
+    )
+    request_id = handler.request_id
+    result = fal_client.result("fal-ai/idm-vton", request_id)
+    time_2 = time.time()
+    cat_time = time_2 - time_1
+    print(f"idm_time:{cat_time}")
+    time_1 = time.time()
+    response = requests.get(result['image']['url'])
+    time_2 = time.time()
+    url_time = time_2 - time_1
+    print(f"url_time:{url_time}")
+    time_1 = time.time()
+    save_directory = "downloads"
+    if response.status_code == 200:
+        filename = os.path.join(save_directory, f"idm-vton_{index}.png")
+        with open(filename, 'wb') as f:
+            f.write(response.content)
+    else:
+        print(f"Failed to download image from {result['image']['url']}")
+    time_2 = time.time()
+    downloads_time = time_2 - time_1
+    print(f"downloads_time:{downloads_time}")
+    return os.path.join(save_directory, f"idm-vton_{index}.png")
+
+
+def user_cloths(user_pic, cloth_gen):
+    user_cloth = []
+    for i in range(1, 4):
+        user_cloth.append(user_dress_cat(user_pic, cloth_gen, i))
+    for i in range(4, 6):
+        user_cloth.append(user_dress_idm(user_pic, cloth_gen, i))
+    return user_cloth