first commit

.gitattributes

@@ -33,3 +33,6 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+*.png filter=lfs diff=lfs merge=lfs -text
+*.jpeg filter=lfs diff=lfs merge=lfs -text
+*.jpg filter=lfs diff=lfs merge=lfs -text

README.md

@@ -1,7 +1,7 @@
 ---
 title: WeShopAI Pose Generator
-emoji: 🏆
-colorFrom: indigo
+emoji: 🙆‍♂️
+colorFrom: pink
 colorTo: blue
 sdk: gradio
 sdk_version: 5.25.2

app.py

@@ -0,0 +1,437 @@
+import gradio as gr
+import hmac
+import hashlib
+import time
+import os
+import requests
+from io import BytesIO
+from PIL import Image
+import uuid
+
+# from dotenv import load_dotenv
+
+# load_dotenv()
+
+example_path = os.path.join(os.path.dirname(__file__), 'assets')
+human_list = os.listdir(os.path.join(example_path, "human"))
+human_list_path = [os.path.join(example_path, "human", human) for human in human_list]
+
+base_url = os.getenv('base_url')
+upload_image_url = os.getenv('upload_image_url')
+create_save_task_url = os.getenv('create_save_task_url')
+execute_task_url = os.getenv('execute_task_url')
+query_task_url = os.getenv('query_task_url')
+secret_key = os.getenv('secret_key')
+app_id = os.getenv('app_id')
+agent_version = os.getenv('agent_version')
+agent_name = os.getenv('agent_name')
+
+def parse_response(response):
+    data = {}
+    msg = ''
+    if response.status_code == 200:
+        try:
+            datas = response.json()
+            if datas:
+                data = datas.get("data")
+                if not data:
+                    msg = datas.get("msg")
+                    if not msg:
+                        msg = "Field error."
+            else:
+                msg = "The parsing result is empty."
+        except Exception as e:
+            msg = f"parse error: {repr(e)}."
+    else:
+        msg = f'request error.'
+    return data, msg
+
+
+def generate_signature(key, did, timestamp):
+    data = f"{did}:{timestamp}{app_id}"
+    h = hmac.new(key.encode(), data.encode(), hashlib.sha256)
+    return h.hexdigest()
+
+
+def url_to_image(url, ip):
+    headers = {
+        "user-agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36",
+        'X-Forwarded-For': ip
+    }
+    try:
+        response = requests.get(url, headers=headers, timeout=30)
+    except:
+        return None
+    if response.status_code == 200:
+        img = Image.open(BytesIO(response.content))
+        return img
+    return None
+
+
+def start_task(task_id, did, ip):
+    timestamp = str(int(time.time()))
+    signature = generate_signature(
+        key=secret_key,
+        did=did,
+        timestamp=timestamp,
+    )
+    headers = {
+        'Did': did,
+        'X-Timestamp': timestamp,
+        'X-Signature': signature,
+        'X-Forwarded-For': ip,
+        'X-AppId': app_id,
+    }
+    data = {
+        "agentVersion": agent_version,
+        "agentName": agent_name,
+        "taskId": task_id,
+        "runFreeAsFallback": False
+    }
+    response = requests.post(base_url + execute_task_url, json=data, headers=headers)
+    data, msg = parse_response(response)
+    return data, msg
+
+
+def create_task(image_url, did, ip):
+    timestamp = str(int(time.time()))
+    signature = generate_signature(
+        key=secret_key,
+        did=did,
+        timestamp=timestamp,
+    )
+    headers = {
+        'Did': did,
+        'X-Timestamp': timestamp,
+        'X-Signature': signature,
+        'X-Forwarded-For': ip,
+        'X-AppId': app_id,
+    }
+    data = {
+        "agentVersion": agent_version,
+        "agentName": agent_name,
+        "image": image_url
+    }
+    response = requests.post(base_url + create_save_task_url, json=data, headers=headers)
+    data, msg = parse_response(response)
+    return data, msg
+
+
+def save_task(image_url, show_image, task_id, text_description, did, ip):
+    timestamp = str(int(time.time()))
+    signature = generate_signature(
+        key=secret_key,
+        did=did,
+        timestamp=timestamp,
+    )
+    headers = {
+        'Did': did,
+        'X-Timestamp': timestamp,
+        'X-Signature': signature,
+        'X-Forwarded-For': ip,
+        'X-AppId': app_id,
+    }
+    data = {
+        "agentVersion": agent_version,
+        "agentName": agent_name,
+        "image": image_url,
+        "showImage": show_image,
+        "taskId": task_id,
+        "textDescription": text_description,
+    }
+    response = requests.post(base_url + create_save_task_url, json=data, headers=headers)
+    data, msg = parse_response(response)
+    return data, msg
+
+
+def query_task(task_id, execution_id, did, ip):
+    timestamp = str(int(time.time()))
+    signature = generate_signature(
+        key=secret_key,
+        did=did,
+        timestamp=timestamp,
+    )
+    headers = {
+        'Did': did,
+        'X-Timestamp': timestamp,
+        'X-Signature': signature,
+        'X-Forwarded-For': ip,
+        'X-AppId': app_id,
+    }
+    data = {
+        "agentVersion": agent_version,
+        "agentName": agent_name,
+        "taskId": task_id,
+        "executionId": execution_id,
+    }
+    response = requests.post(base_url + query_task_url, json=data, headers=headers)
+    data, msg = parse_response(response)
+    return data, msg
+
+
+def upload_image(image, did, ip):
+    if image is None:
+        return None
+    timestamp = str(int(time.time()))
+    signature = generate_signature(
+        key=secret_key,
+        did=did,
+        timestamp=timestamp,
+    )
+    upload_url = upload_image_url
+    image_format = image.format if image.format else "PNG"
+    mime_type = f"image/{image_format.lower()}"
+    with BytesIO() as m_img:
+        image.save(m_img, format=image_format)
+        m_img.seek(0)
+        files = {'image': (f"main_image.{image_format.lower()}", m_img, mime_type)}
+        headers = {
+            'Did': did,
+            'X-Timestamp': timestamp,
+            'X-Signature': signature,
+            'X-Forwarded-For': ip,
+            'X-AppId': app_id,
+        }
+        response = requests.post(base_url + upload_url, files=files, headers=headers)
+        data, msg = parse_response(response)
+        return data, msg
+
+
+def load_description(file_path):
+    with open(file_path, 'r', encoding='utf-8') as file:
+        content = file.read()
+    return content
+
+
+def generate_image(main_image, text_description, did, request: gr.Request):
+    if not did:
+        did = str(uuid.uuid4())
+    if main_image is None:
+        m = "Please upload both the main image and the background reference image before generating."
+        return gr.Warning(m), did
+
+    client_ip = request.client.host
+    x_forwarded_for = request.headers.get('x-forwarded-for')
+    if x_forwarded_for:
+        client_ip = x_forwarded_for
+   
+    upload_image_data, upload_image_msg = upload_image(
+        image=main_image,
+        did=did,
+        ip=client_ip,
+    )
+    if not upload_image_data:
+        return gr.Warning(upload_image_msg), did
+    image_url = upload_image_data.get("image")
+    if not image_url:
+        m = 'Upload image failed.'
+        return gr.Warning(m), did
+    create_task_data, create_task_msg = create_task(
+        image_url=image_url,
+        did=did,
+        ip=client_ip,
+    )
+    if not create_task_data:
+        return gr.Warning(create_task_msg), did
+    task_id = create_task_data.get("taskId")
+    show_image = create_task_data.get("showImage")
+    if not task_id or not show_image:
+        m = 'Create task failed.'
+        return gr.Warning(m), did
+    save_task_data, save_task_msg = save_task(
+        image_url=image_url,
+        show_image=show_image,
+        task_id=task_id,
+        text_description=text_description,
+        did=did,
+        ip=client_ip,
+    )
+    if not save_task_data:
+        return gr.Warning(save_task_msg), did
+    save_task_id = save_task_data.get("taskId")
+    save_show_image = save_task_data.get("showImage")
+    if not save_task_id or not save_show_image:
+        return gr.Warning('Save task failed'), did
+    start_task_data, start_task_msg = start_task(
+        task_id=save_task_id,
+        did=did,
+        ip=client_ip,
+    )
+    if not start_task_data:
+        return gr.Warning(start_task_msg), did
+    execution_id = start_task_data.get("executionId")
+    if not execution_id:
+        m = "The task failed to start."
+        return gr.Warning(m), did
+    start_time = int(time.time())
+    while True:
+        m = "Query task failed."
+        query_task_data, query_task_msg = query_task(
+            task_id=save_task_id,
+            execution_id=execution_id,
+            did=did,
+            ip=client_ip,
+        )
+        print("Query task data:", query_task_data)
+        print("Query task msg:", query_task_msg)
+        print("Time:", int(time.time()) - start_time)
+        if not query_task_data:
+            return gr.Warning(query_task_msg), did
+        executions = query_task_data.get("executions")
+        if not executions:
+            return gr.Warning(m), did
+        results = executions[0].get("result")
+        if not results:
+            return gr.Warning(m), did
+        status = results[0].get("status")
+        if status == "Failed":
+            return gr.Warning(m), did
+        elif status == "Success" or status == "Blocked":
+            img = results[0].get("image")
+            if img and str(img).strip() != "":
+                return url_to_image(img, ip=client_ip), did
+        end_time = int(time.time())
+        if end_time - start_time > 3600:
+            m = 'Query task timeout.'
+            return gr.Warning(m), did
+        time.sleep(4)
+
+
+def preprocess_image(main_image):
+    return main_image
+
+
+css = """  
+    .text-box {
+        min-height: 260px !important;
+    }
+    .image-container {
+        min-height: 260px !important;
+    }
+    .image-container img {  
+        max-height: 500px;  
+        width: auto;  
+    }  
+    .hide-buttons .source-selection {  
+        display: none;  
+    }  
+    #example-images .gallery {
+        display: flex;  
+        flex-wrap: wrap;  
+    }
+    #example-images .gallery-item .container{
+        width: 100%;
+        max-width: 100%;
+    }
+    #example-images .gallery-item {
+        flex: 0 0 30%; 
+        max-width: 30%; 
+        box-sizing: border-box; 
+        display: flex;  
+        text-align: center;
+        justify-content: center;
+
+    }
+    @media (max-width: 767px) {
+        #example-res-images th {
+            font-size: 12px;
+            word-wrap: break-word; 
+            word-break: break-word;
+            white-space: normal; 
+            overflow-wrap: break-word; 
+        }
+    }
+"""
+
+with gr.Blocks(css=css) as WeShop:
+    current_did = gr.State(value='')
+    gr.HTML(load_description("assets/title.html"))
+    with gr.Row():
+        with gr.Column():
+            gr.Markdown("#### Step 1: Upload a person image")
+            main_image_input = gr.Image(
+                type="pil",
+                label="Main Image",
+                elem_classes=["image-container", "hide-buttons"]
+            )
+            clothing_example = gr.Examples(
+                inputs=main_image_input,
+                examples_per_page=12,
+                examples=human_list_path,
+                elem_id="example-images",
+                outputs=main_image_input
+            )
+
+        with gr.Column():
+            gr.Markdown("#### Step 2: Input pose description")
+            text_description_input = gr.Textbox(
+                label="Text Description",
+                placeholder="Enter pose description here...",
+                lines=9,
+                elem_classes=["text-box"]
+            )
+            text_description_example = gr.Examples(
+                inputs=text_description_input,
+                examples_per_page=12,
+                examples=[
+                    "Keep the background unchanged and only change the model's pose",
+                    "Follow-shot view",
+                    "Switch to a back view",
+                    "Change pose",
+                    "The scene stays the same, while the outfit is shown from multiple angles",
+                    "Dynamic walking pose",
+                    "Simulating a model walking motion",
+                    "Low-angle upward shoot",
+                ],
+                outputs=text_description_input
+           )
+        with gr.Column():
+            with gr.Row():
+                with gr.Column():
+                    gr.Markdown("#### Step 3: Press 'Generate' to get the result")
+                    output = gr.Image(
+                        label="Result",
+                        elem_classes=["image-container", "hide-buttons"],
+                        interactive=False
+                    )
+            with gr.Row():
+                submit_button = gr.Button("Generate")
+                submit_button.click(
+                    fn=generate_image,
+                    inputs=[main_image_input, text_description_input, current_did],
+                    outputs=[output, current_did],
+                    concurrency_limit=None
+                )
+            with gr.Column():
+                show_case = gr.Examples(
+                    examples=[
+                        [
+                            "assets/human/07.png", 
+                            "Dynamic walking pose", 
+                            "assets/examples/result_07_01.jpeg",
+                        ],
+                        [
+                            "assets/human/01.png", 
+                            "Follow-shot view",
+                            "assets/examples/result_01_01.jpeg"
+                        ],
+                        [
+                            "assets/human/06.png", 
+                            "Change pose",
+                            "assets/examples/result_06_01.jpeg"
+                        ],
+                        [
+                            "assets/human/04.png", 
+                            "Keep the background unchanged and only change the model's pose",
+                            "assets/examples/result_04_01.jpeg"
+                        ],
+                    ],
+                    inputs=[main_image_input, text_description_input, output],
+                    elem_id="example-res-images"
+                )
+    main_image_input.upload(
+        fn=preprocess_image,
+        inputs=[main_image_input],
+        outputs=main_image_input
+    )
+WeShop.queue(api_open=False).launch(show_api=False)

assets/title.html

@@ -0,0 +1,100 @@
+<!DOCTYPE html>
+<html lang="en">
+  <head>
+    <meta charset="UTF-8" />
+    <title>WeShopAI Pose Generator</title>
+    <style>
+      .title-text {
+        display: flex;
+        justify-content: center;
+        align-items: center;
+        text-align: center;
+        font-size: 40px;
+      }
+
+      .center-content {
+        display: flex;
+        justify-content: center;
+        align-items: center;
+        text-align: center;
+      }
+
+      .content-text {
+        text-align: center;
+        font-size: 14px;
+        padding-right: 10%;
+        padding-left: 10%;
+      }
+
+      @media (max-width: 767px) {
+        .title-text {
+          font-size: 30px;
+        }
+
+        .content-text {
+          font-size: 12px;
+          padding-right: 2%;
+          padding-left: 2%;
+        }
+      }
+
+      @media (max-width: 450px) {
+        .title-text {
+          font-size: 24px;
+        }
+
+        .content-text {
+          font-size: 12px;
+          padding-right: 0;
+          padding-left: 0;
+        }
+      }
+
+      @media (max-width: 360px) {
+        .title-text {
+          font-size: 22px;
+        }
+      }
+    </style>
+  </head>
+  <body>
+    <div>
+      <div>
+        <div class="title-text">
+          <b>WeShopAI Pose Generator</b>
+        </div>
+        <br />
+        <div class="center-content">
+          <a href="https://www.weshop.com/"
+            ><img
+              src="https://img.shields.io/static/v1?label=Official Website&message=CN&color=red"
+          /></a>
+          &ensp;
+          <a href="https://www.weshop.ai/"
+            ><img
+              src="https://img.shields.io/static/v1?label=Official Website&message=EN&color=blue"
+          /></a>
+        </div>
+        <br />
+        <div class="content-text">
+          Commercial clients are encouraged to evaluate the API capabilities
+          offered at weshop.ai.
+          <a
+            href="https://open.weshop.ai/authorization/apikey"
+            style="font-weight: bold"
+            target="_blank"
+            >Get an API Key</a
+          >
+          to explore further.
+        </div>
+        <br />
+        <div class="content-text">
+          <strong>Disclaimer: </strong>The WeShopAI-Pose-Generator Demo on
+          HuggingFace space is free for trials. Any solicitation for payment
+          based on the free features we provide on HuggingFace Space is a
+          fraudulent act. Beware of scams and do not fall victim to deceit.
+        </div>
+      </div>
+    </div>
+  </body>
+</html>