Update use_template_app.py

use_template_app.py

@@ -1,463 +1,464 @@
-import gradio as gr
-import requests
-import pathlib
-import zipfile
-import io
-import urllib.parse # For URL sanitization
-import xml.etree.ElementTree as ET
-import base64
-from PIL import Image # For handling image types if needed by Gradio gallery
-import time
-import shutil
-import uuid # For unique filenames
-
-# --- Configuration ---
-TEMPLATE_API_BASE_URL = "http://127.0.0.1:8001"  # Explicitly use 127.0.0.1
-SVG2PNG_API_URL = 'https://innoai-svg2png-api.hf.space/convert' # From svg2png_guiapp.py
-TEMP_DIR = pathlib.Path("temp_gradio_app_files") # Renamed for clarity
-
-# Ensure temp directory exists
-TEMP_DIR.mkdir(parents=True, exist_ok=True) # No need to clean it aggressively on each start
-
-# --- Helper Functions ---
-
-def sanitize_url(url_string: str) -> str:
-    """
-    Safely percent-encodes the path and query components of an HTTP/HTTPS URL.
-    Attempts to be idempotent if parts of the URL are already percent-encoded.
-    Parentheses in the path are preserved.
-    """
-    if not url_string or not isinstance(url_string, str) or url_string.startswith('data:'):
-        return url_string
-    try:
-        scheme, netloc, path, query, fragment = urllib.parse.urlsplit(url_string)
-
-        if scheme not in ['http', 'https']:
-            # For other schemes, or if splitting fails to identify a scheme, return original.
-            # This could happen for URNs, mailto, etc., which we don't want to modify.
-            return url_string
-
-        # Sanitize path: unquote to normalize, then quote.
-        # `safe` keeps slashes, existing valid percent-encodings (%), and specified chars like parentheses.
-        sanitized_path = urllib.parse.quote(urllib.parse.unquote(path), safe='/%;:@&+$,=?#()~')
-
-        sanitized_query = ""
-        if query:
-            # Unquote the original query to handle cases where it might be partially encoded
-            unquoted_query = urllib.parse.unquote(query)
-            # Parse the unquoted query string into a dictionary of parameters
-            parsed_query_params = urllib.parse.parse_qs(unquoted_query, keep_blank_values=True)
-            # urlencode will quote keys and values appropriately for the query string.
-            sanitized_query = urllib.parse.urlencode(parsed_query_params, doseq=True, quote_via=urllib.parse.quote)
-            
-        final_url = urllib.parse.urlunsplit((scheme, netloc, sanitized_path, sanitized_query, fragment))
-        # print(f"Sanitize URL Input: {url_string}, Output: {final_url}") # Debugging line
-        return final_url
-    except Exception as e:
-        print(f"Warning: Could not sanitize URL '{url_string}': {e}")
-        return url_string
-
-def fetch_template_details(m_param, u_param, category_param, name_param):
-    """
-    Fetches template details from the template_server.py API.
-    Expected to return a dict with 'name', 'thumbnail_url', 'svg_download_url'.
-    """
-    username = urllib.parse.quote(u_param) 
-    category = urllib.parse.quote(category_param)
-    template_name = urllib.parse.quote(name_param)
-
-    api_url = f"{TEMPLATE_API_BASE_URL}/gradio_template_details/{username}/{category}/{template_name}"
-    print(f"Fetching template details from: {api_url}") 
-    try:
-        response = requests.get(api_url, timeout=10)
-        response.raise_for_status()
-        return response.json() 
-    except requests.exceptions.RequestException as e:
-        print(f"Error fetching template details from {api_url}: {e}")
-        return None
-
-def download_svg_content(svg_url):
-    """Downloads the raw SVG content from a given URL."""
-    try:
-        response = requests.get(svg_url, timeout=10)
-        response.raise_for_status()
-        return response.text # SVG is XML text
-    except requests.exceptions.RequestException as e:
-        print(f"Error downloading SVG content from {svg_url}: {e}")
-        return None
-
-def replace_background_in_svg(svg_content_str: str, new_image_path: str):
-    """
-    Replaces a designated image in the SVG string with a base64 data URI.
-    Also converts all other external http/https image URLs in the SVG to base64 data URIs.
-    """
-    try:
-        namespaces = {'xlink': 'http://www.w3.org/1999/xlink', 'svg': 'http://www.w3.org/2000/svg'}
-        for prefix, uri in namespaces.items():
-            ET.register_namespace(prefix, uri if prefix else '')
-        root = ET.fromstring(svg_content_str)
-
-        def _image_bytes_to_data_uri(image_bytes: bytes, resource_identifier: str) -> str:
-            """
-            Converts image bytes to a data URI.
-            mime_type_source can be a full MIME type string, a file path, or a URL.
-            """
-            mime_type = "image/png" # Default
-            if "/" in resource_identifier: # Check if it looks like a MIME type string or a URL/path with extension
-                # Try to get from Content-Type header if it's a requests.Response object (not applicable here directly)
-                # For now, infer from path/URL extension or if it's already a MIME type string
-                if resource_identifier.startswith("image/"): # Already a MIME type
-                    mime_type = resource_identifier
-                else: # Infer from extension
-                    ext = pathlib.Path(resource_identifier).suffix.lower()
-                    if ext == ".png": mime_type = "image/png"
-                    elif ext in [".jpg", ".jpeg"]: mime_type = "image/jpeg"
-                    elif ext == ".gif": mime_type = "image/gif"
-                    elif ext == ".webp": mime_type = "image/webp"
-                    elif ext == ".svg": mime_type = "image/svg+xml"
-            
-            encoded_image = base64.b64encode(image_bytes).decode('utf-8')
-            return f"data:{mime_type};base64,{encoded_image}"
-
-        # 1. Handle the user-uploaded image (new_image_path)
-        background_image_element = None
-        ids_to_check = ["background_image", "placeholder_image", "product_image", "main_image", "image_to_replace", "底图"]
-        for img_id in ids_to_check:
-            found_element = root.find(f".//svg:image[@id='{img_id}']", namespaces)
-            if found_element is None:
-                 found_element = root.find(f".//{{*}}image[@id='{img_id}']")
-            if found_element is not None:
-                background_image_element = found_element
-                break
-        
-        if background_image_element is None:
-            first_image_svg_ns = root.find('.//svg:image', namespaces)
-            if first_image_svg_ns is not None:
-                background_image_element = first_image_svg_ns
-            else:
-                first_image_any_ns = root.find('.//image')
-                if first_image_any_ns is not None:
-                    background_image_element = first_image_any_ns
-                else:
-                    print("Warning: No suitable <image> tag found in SVG to replace with new_image_path.")
-        
-        if background_image_element is not None:
-            try:
-                new_image_path_obj = pathlib.Path(new_image_path)
-                if not new_image_path_obj.is_file():
-                    print(f"Error: New image file not found or is not a file: {new_image_path}")
-                    # Potentially return None or raise error, for now, it will skip this replacement
-                else:
-                    with open(new_image_path_obj, "rb") as f:
-                        new_image_bytes = f.read()
-                    data_uri = _image_bytes_to_data_uri(new_image_bytes, new_image_path) # Pass path for MIME
-                    
-                    xlink_href_attr_namespaced = f"{{{namespaces['xlink']}}}href"
-                    attr_to_set_on_target = None
-                    if xlink_href_attr_namespaced in background_image_element.attrib:
-                        attr_to_set_on_target = xlink_href_attr_namespaced
-                    elif 'href' in background_image_element.attrib:
-                        attr_to_set_on_target = 'href'
-                    else:
-                        attr_to_set_on_target = xlink_href_attr_namespaced if 'xlink' in namespaces else 'href'
-                    background_image_element.set(attr_to_set_on_target, data_uri)
-            except Exception as e:
-                print(f"Error processing and embedding new_image_path '{new_image_path}': {e}")
-
-
-        # 2. Process all other http/https images in the SVG
-        if root is not None:
-            # Create a list of all image elements to iterate over.
-            # This avoids issues with modifying the tree while iterating over findall() results directly.
-            all_image_elements_in_tree = list(root.findall('.//svg:image', namespaces)) + \
-                                         [el for el in root.findall('.//image') if el.tag != f"{{{namespaces['svg']}}}image"]
-
-            for image_el in all_image_elements_in_tree:
-                if image_el == background_image_element: # Already processed (or attempted)
-                    continue
-
-                href_val = None
-                attr_name_to_set = None
-                xlink_href_qname = f"{{{namespaces['xlink']}}}href"
-
-                if image_el.get(xlink_href_qname) is not None:
-                    href_val = image_el.get(xlink_href_qname)
-                    attr_name_to_set = xlink_href_qname
-                elif image_el.get('href') is not None:
-                    href_val = image_el.get('href')
-                    attr_name_to_set = 'href'
-                
-                if href_val and href_val.startswith(('http://', 'https://')):
-                    try:
-                        safe_download_url = sanitize_url(href_val) # Sanitize before download
-                        # print(f"DEBUG: Downloading remote image for embedding: {safe_download_url}")
-                        response = requests.get(safe_download_url, timeout=20, stream=True)
-                        response.raise_for_status()
-                        
-                        image_bytes = response.content
-                        
-                        content_type_header = response.headers.get('Content-Type')
-                        mime_type_source = content_type_header.split(';')[0].strip() if content_type_header and '/' in content_type_header else href_val
-                        
-                        data_uri = _image_bytes_to_data_uri(image_bytes, mime_type_source)
-                        image_el.set(attr_name_to_set, data_uri)
-                        # print(f"DEBUG: Replaced remote URL {href_val} with data URI.")
-                    except requests.exceptions.RequestException as e_req:
-                        print(f"Failed to download or process remote image {href_val}: {e_req}")
-                    except Exception as e_proc:
-                        print(f"Error processing remote image {href_val} into data URI: {e_proc}")
-            
-        if root is None: 
-            print("Error: SVG root is None after parsing, cannot proceed.")
-            return None
-
-        modified_svg_bytes = ET.tostring(root, encoding='UTF-8', method='xml', xml_declaration=True)
-        
-        try:
-            debug_svg_filename = f"debug_embedded_svg_{uuid.uuid4().hex[:8]}.svg"
-            debug_svg_path = TEMP_DIR / debug_svg_filename
-            with open(debug_svg_path, "wb") as f_debug:
-                f_debug.write(modified_svg_bytes)
-            print(f"Saved SVG with all images embedded for debugging to: {debug_svg_path}")
-        except Exception as e_save:
-            print(f"Error saving debug embedded SVG: {e_save}")
-            
-        return modified_svg_bytes
-
-    except ET.ParseError as e:
-        print(f"Error parsing SVG XML: {e}")
-        return None
-    except Exception as e:
-        print(f"An unexpected error occurred during SVG manipulation: {e}")
-        return None
-
-def convert_svg_bytes_to_png_api(svg_content_bytes: bytes, original_template_name: str, index: int):
-    """
-    Converts SVG bytes to PNG bytes using the external API.
-    Saves the PNG to a temporary file and returns the path.
-    """
-    if not svg_content_bytes:
-        return None
-    try:
-        # The API expects a file upload
-        files = {'svg_file': ('generated_svg.svg', svg_content_bytes, 'image/svg+xml')}
-        response = requests.post(SVG2PNG_API_URL, files=files, timeout=60)
-        response.raise_for_status()
-        
-        # Use UUID for more robust unique filenames
-        png_filename = f"{original_template_name.replace(' ','_')}_output_{index}_{uuid.uuid4().hex[:8]}.png"
-        temp_png_path = TEMP_DIR / png_filename
-        with open(temp_png_path, "wb") as f:
-            f.write(response.content)
-        return str(temp_png_path)
-        
-    except requests.exceptions.RequestException as e:
-        print(f"Error converting SVG to PNG via API: {e}")
-        if hasattr(e, 'response') and e.response is not None:
-            print(f"API Response: {e.response.text}")
-        return None
-    except Exception as e:
-        print(f"Error saving PNG: {e}")
-        return None
-
-# --- Gradio App Logic ---
-
-def generate_images_from_template(original_svg_download_url: str, template_name_for_file: str, uploaded_image_files: list, request: gr.Request):
-    """
-    Main processing function for Gradio.
-    Takes original SVG URL, template name, and list of uploaded image file objects.
-    """
-    if not original_svg_download_url:
-        gr.Warning("模板信息未加载，请确保URL参数正确。")
-        return [], None
-    if not uploaded_image_files:
-        gr.Warning("请上传至少一张图片。")
-        return [], None
-
-    # 1. Download original SVG content
-    original_svg_content = download_svg_content(original_svg_download_url)
-    if not original_svg_content:
-        gr.Error("无法下载原始SVG模板。")
-        return [], None
-
-    generated_png_paths = []
-    processed_count = 0
-    
-    for i, uploaded_file_obj in enumerate(uploaded_image_files):
-        # uploaded_file_obj.name is the temporary path to the uploaded file
-        new_image_path = uploaded_file_obj.name 
-        
-        # 2. Replace background in SVG for each uploaded image
-        modified_svg_bytes = replace_background_in_svg(original_svg_content, new_image_path)
-        if not modified_svg_bytes:
-            gr.Warning(f"处理图片 {i+1} 失败：无法修改SVG。")
-            continue # Skip this image
-
-        # 3. Convert modified SVG to PNG
-        # Use template_name_for_file for unique output filenames
-        png_path = convert_svg_bytes_to_png_api(modified_svg_bytes, template_name_for_file, i + 1)
-        if png_path:
-            generated_png_paths.append(png_path)
-            processed_count += 1
-        else:
-            gr.Warning(f"处理图片 {i+1} 失败：无法转换为PNG。")
-            
-    if not generated_png_paths:
-        gr.Info("未能成功生成任何图片。")
-        return [], None
-
-    zip_buffer = io.BytesIO()
-    with zipfile.ZipFile(zip_buffer, "w", zipfile.ZIP_DEFLATED) as zf:
-        for png_path_str in generated_png_paths:
-            png_file = pathlib.Path(png_path_str)
-            zf.write(png_file, arcname=png_file.name) 
-    
-    zip_buffer.seek(0)
-    
-    zip_filename = f"{template_name_for_file.replace(' ','_')}_batch_{uuid.uuid4().hex[:8]}.zip" # Added UUID
-    temp_zip_path = TEMP_DIR / zip_filename
-    with open(temp_zip_path, "wb") as f:
-        f.write(zip_buffer.getvalue())
-
-    gr.Info(f"成功生成 {processed_count} 张图片!")
-    return generated_png_paths, str(temp_zip_path)
-
-
-def initial_load_template_info(request: gr.Request):
-    """
-    Loads initial template information based on URL query parameters.
-    Downloads the thumbnail and returns its local path.
-    """
-    query_params = request.query_params
-    m = query_params.get("m")
-    u = query_params.get("u")
-    category = query_params.get("category")
-    name = query_params.get("name") # This is the template name
-
-    if not all([m, u, category, name]):
-        print("Initial load: URL parameters (m, u, category, name) are incomplete or missing.")
-        return None, "无模板信息 (请检查URL参数)", None, "无模板"
-
-    details = fetch_template_details(m, u, category, name)
-    
-    local_thumbnail_path = None
-    template_display_name = "错误: 无法加载模板信息" # Default error message
-    svg_download_url = None
-    
-    if details and "name" in details: 
-        template_display_name = details["name"] # Use name from details for display
-        svg_download_url = details.get("svg_download_url")
-
-        if details.get("thumbnail_url"):
-            try:
-                thumb_url = details["thumbnail_url"]
-                print(f"Fetching thumbnail from: {thumb_url}")
-                thumb_response = requests.get(thumb_url, timeout=10)
-                thumb_response.raise_for_status()
-                
-                # Create a unique filename for the thumbnail in TEMP_DIR
-                # Use name from URL params for file naming consistency if needed, or details['name']
-                safe_template_name_for_file = name.replace(' ','_').replace('/','_').replace('\\\\','_') # Basic sanitization
-                thumb_filename = f"thumb_{safe_template_name_for_file}_{uuid.uuid4().hex[:8]}.png"
-                local_thumbnail_path = TEMP_DIR / thumb_filename
-                with open(local_thumbnail_path, "wb") as f:
-                    f.write(thumb_response.content)
-                print(f"Thumbnail saved to: {local_thumbnail_path}")
-            except requests.exceptions.RequestException as e:
-                print(f"Error downloading thumbnail from {details.get('thumbnail_url')}: {e}")
-                local_thumbnail_path = None 
-            except Exception as e:
-                print(f"Error saving thumbnail: {e}")
-                local_thumbnail_path = None
-    else:
-        print(f"Failed to load template details for m={m}, u={u}, category={category}, name={name}")
-        # template_display_name is already set to the default error message
-
-    return str(local_thumbnail_path) if local_thumbnail_path else None, \
-           template_display_name, \
-           svg_download_url, \
-           name # 'name' from URL param for template_name_for_file_state
-
-# --- Gradio Interface ---
-with gr.Blocks(theme=gr.themes.Soft(), title="SVG模板批量图片生成器") as demo:
-    gr.Markdown("## 使用SVG模板批量生成图片")
-    gr.Markdown("从URL加载模板，上传您自己的图片替换模板中的底图，然后批量生成并下载。")
-
-    # Hidden state to store original SVG download URL and template name for file ops
-    original_svg_download_url_state = gr.State()
-    template_name_for_file_state = gr.State() # To preserve the name from URL param for consistent file naming
-
-    with gr.Row():
-        with gr.Column(scale=1, min_width=200):
-            template_thumbnail_display = gr.Image(label="当前模板缩略图", interactive=False, height=200, type="filepath") # Ensure type is filepath
-            template_name_display = gr.Textbox(label="当前模板名称", interactive=False)
-        with gr.Column(scale=3):
-            uploaded_images_input = gr.Files(
-                label="上传您的图片 (可多选)", 
-                file_count="multiple", 
-                file_types=["image"] # Accepts .png, .jpg, .jpeg, .gif, .webp etc.
-            )
-    
-    generate_button = gr.Button("🚀 立即生成图片", variant="primary", scale=1)
-    
-    with gr.Accordion("生成结果预览与下载", open=True):
-        output_gallery = gr.Gallery(
-            label="生成图片预览", 
-            show_label=True, 
-            elem_id="output_gallery", 
-            columns=[4], 
-            object_fit="contain", 
-            height="auto"
-            # type="filepath" is default for Gallery if fed filepaths
-        )
-        output_zip_file = gr.File(label="下载所有生成图片的ZIP包", interactive=False, type="filepath") # Ensure type is filepath
-
-    # Load initial template info based on URL parameters when the interface loads.
-    # The `initial_load_template_info` function will parse request.query_params.
-    # It's crucial that `gr.Request` is passed to it.
-    # `inputs=None` with `request: gr.Request` in function signature works.
-    demo.load(
-        initial_load_template_info, 
-        inputs=None, # gr.Request is implicitly passed if type-hinted in the function
-        outputs=[
-            template_thumbnail_display, 
-            template_name_display, 
-            original_svg_download_url_state,
-            template_name_for_file_state
-        ]
-    )
-    
-    generate_button.click(
-        generate_images_from_template, 
-        inputs=[
-            original_svg_download_url_state, 
-            template_name_for_file_state,
-            uploaded_images_input
-            # gr.Request is also implicitly passed here
-        ],
-        outputs=[output_gallery, output_zip_file]
-    )
-
-if __name__ == "__main__":
-    # To run this app:
-    # 1. Ensure your template_server.py is running (e.g., on http://localhost:8001)
-    #    and has the /template_details/{username}/{category}/{template_name} endpoint.
-    # 2. Run this script: python use_template_app.py
-    # 3. Open your browser to the Gradio link, appending parameters, e.g.:
-    #    http://127.0.0.1:7860/?m=pub&u=all&category=%E5%AE%B6%E7%94%B5&name=%E6%A8%A1%E6%9D%BF_%E5%86%B0%E7%82%B9%E4%BB%B7
-    
-    # For easy testing, you might want to create a dummy template_server.py endpoint
-    # or hardcode some details if the server isn't ready.
-    
-    demo.launch()
-
-    # Optional: Cleanup TEMP_DIR logic can be added here if needed for long-running servers
-    # For development, manual cleanup or OS temp cleaning is often sufficient.
-    # Example:
-    # importate atexit
-    # def cleanup_temp_dir():
-    #     if TEMP_DIR.exists():
-    #         print(f"Cleaning up temp directory: {TEMP_DIR}")
-    #         shutil.rmtree(TEMP_DIR)
-    # atexit.register(cleanup_temp_dir) # This might be too aggressive for dev
-
+import gradio as gr
+import requests
+import pathlib
+import zipfile
+import io
+import urllib.parse # For URL sanitization
+import xml.etree.ElementTree as ET
+import base64
+from PIL import Image # For handling image types if needed by Gradio gallery
+import time
+import shutil
+import uuid # For unique filenames
+import os
+
+# --- Configuration ---
+TEMPLATE_API_BASE_URL = os.getenv("mubanapi")
+SVG2PNG_API_URL = os.getenv("convertapi")
+TEMP_DIR = pathlib.Path("temp_gradio_app_files") # Renamed for clarity
+
+# Ensure temp directory exists
+TEMP_DIR.mkdir(parents=True, exist_ok=True) # No need to clean it aggressively on each start
+
+# --- Helper Functions ---
+
+def sanitize_url(url_string: str) -> str:
+    """
+    Safely percent-encodes the path and query components of an HTTP/HTTPS URL.
+    Attempts to be idempotent if parts of the URL are already percent-encoded.
+    Parentheses in the path are preserved.
+    """
+    if not url_string or not isinstance(url_string, str) or url_string.startswith('data:'):
+        return url_string
+    try:
+        scheme, netloc, path, query, fragment = urllib.parse.urlsplit(url_string)
+
+        if scheme not in ['http', 'https']:
+            # For other schemes, or if splitting fails to identify a scheme, return original.
+            # This could happen for URNs, mailto, etc., which we don't want to modify.
+            return url_string
+
+        # Sanitize path: unquote to normalize, then quote.
+        # `safe` keeps slashes, existing valid percent-encodings (%), and specified chars like parentheses.
+        sanitized_path = urllib.parse.quote(urllib.parse.unquote(path), safe='/%;:@&+$,=?#()~')
+
+        sanitized_query = ""
+        if query:
+            # Unquote the original query to handle cases where it might be partially encoded
+            unquoted_query = urllib.parse.unquote(query)
+            # Parse the unquoted query string into a dictionary of parameters
+            parsed_query_params = urllib.parse.parse_qs(unquoted_query, keep_blank_values=True)
+            # urlencode will quote keys and values appropriately for the query string.
+            sanitized_query = urllib.parse.urlencode(parsed_query_params, doseq=True, quote_via=urllib.parse.quote)
+            
+        final_url = urllib.parse.urlunsplit((scheme, netloc, sanitized_path, sanitized_query, fragment))
+        # print(f"Sanitize URL Input: {url_string}, Output: {final_url}") # Debugging line
+        return final_url
+    except Exception as e:
+        print(f"Warning: Could not sanitize URL '{url_string}': {e}")
+        return url_string
+
+def fetch_template_details(m_param, u_param, category_param, name_param):
+    """
+    Fetches template details from the template_server.py API.
+    Expected to return a dict with 'name', 'thumbnail_url', 'svg_download_url'.
+    """
+    username = urllib.parse.quote(u_param) 
+    category = urllib.parse.quote(category_param)
+    template_name = urllib.parse.quote(name_param)
+
+    api_url = f"{TEMPLATE_API_BASE_URL}/gradio_template_details/{username}/{category}/{template_name}"
+    print(f"Fetching template details from: {api_url}") 
+    try:
+        response = requests.get(api_url, timeout=10)
+        response.raise_for_status()
+        return response.json() 
+    except requests.exceptions.RequestException as e:
+        print(f"Error fetching template details from {api_url}: {e}")
+        return None
+
+def download_svg_content(svg_url):
+    """Downloads the raw SVG content from a given URL."""
+    try:
+        response = requests.get(svg_url, timeout=10)
+        response.raise_for_status()
+        return response.text # SVG is XML text
+    except requests.exceptions.RequestException as e:
+        print(f"Error downloading SVG content from {svg_url}: {e}")
+        return None
+
+def replace_background_in_svg(svg_content_str: str, new_image_path: str):
+    """
+    Replaces a designated image in the SVG string with a base64 data URI.
+    Also converts all other external http/https image URLs in the SVG to base64 data URIs.
+    """
+    try:
+        namespaces = {'xlink': 'http://www.w3.org/1999/xlink', 'svg': 'http://www.w3.org/2000/svg'}
+        for prefix, uri in namespaces.items():
+            ET.register_namespace(prefix, uri if prefix else '')
+        root = ET.fromstring(svg_content_str)
+
+        def _image_bytes_to_data_uri(image_bytes: bytes, resource_identifier: str) -> str:
+            """
+            Converts image bytes to a data URI.
+            mime_type_source can be a full MIME type string, a file path, or a URL.
+            """
+            mime_type = "image/png" # Default
+            if "/" in resource_identifier: # Check if it looks like a MIME type string or a URL/path with extension
+                # Try to get from Content-Type header if it's a requests.Response object (not applicable here directly)
+                # For now, infer from path/URL extension or if it's already a MIME type string
+                if resource_identifier.startswith("image/"): # Already a MIME type
+                    mime_type = resource_identifier
+                else: # Infer from extension
+                    ext = pathlib.Path(resource_identifier).suffix.lower()
+                    if ext == ".png": mime_type = "image/png"
+                    elif ext in [".jpg", ".jpeg"]: mime_type = "image/jpeg"
+                    elif ext == ".gif": mime_type = "image/gif"
+                    elif ext == ".webp": mime_type = "image/webp"
+                    elif ext == ".svg": mime_type = "image/svg+xml"
+            
+            encoded_image = base64.b64encode(image_bytes).decode('utf-8')
+            return f"data:{mime_type};base64,{encoded_image}"
+
+        # 1. Handle the user-uploaded image (new_image_path)
+        background_image_element = None
+        ids_to_check = ["background_image", "placeholder_image", "product_image", "main_image", "image_to_replace", "底图"]
+        for img_id in ids_to_check:
+            found_element = root.find(f".//svg:image[@id='{img_id}']", namespaces)
+            if found_element is None:
+                 found_element = root.find(f".//{{*}}image[@id='{img_id}']")
+            if found_element is not None:
+                background_image_element = found_element
+                break
+        
+        if background_image_element is None:
+            first_image_svg_ns = root.find('.//svg:image', namespaces)
+            if first_image_svg_ns is not None:
+                background_image_element = first_image_svg_ns
+            else:
+                first_image_any_ns = root.find('.//image')
+                if first_image_any_ns is not None:
+                    background_image_element = first_image_any_ns
+                else:
+                    print("Warning: No suitable <image> tag found in SVG to replace with new_image_path.")
+        
+        if background_image_element is not None:
+            try:
+                new_image_path_obj = pathlib.Path(new_image_path)
+                if not new_image_path_obj.is_file():
+                    print(f"Error: New image file not found or is not a file: {new_image_path}")
+                    # Potentially return None or raise error, for now, it will skip this replacement
+                else:
+                    with open(new_image_path_obj, "rb") as f:
+                        new_image_bytes = f.read()
+                    data_uri = _image_bytes_to_data_uri(new_image_bytes, new_image_path) # Pass path for MIME
+                    
+                    xlink_href_attr_namespaced = f"{{{namespaces['xlink']}}}href"
+                    attr_to_set_on_target = None
+                    if xlink_href_attr_namespaced in background_image_element.attrib:
+                        attr_to_set_on_target = xlink_href_attr_namespaced
+                    elif 'href' in background_image_element.attrib:
+                        attr_to_set_on_target = 'href'
+                    else:
+                        attr_to_set_on_target = xlink_href_attr_namespaced if 'xlink' in namespaces else 'href'
+                    background_image_element.set(attr_to_set_on_target, data_uri)
+            except Exception as e:
+                print(f"Error processing and embedding new_image_path '{new_image_path}': {e}")
+
+
+        # 2. Process all other http/https images in the SVG
+        if root is not None:
+            # Create a list of all image elements to iterate over.
+            # This avoids issues with modifying the tree while iterating over findall() results directly.
+            all_image_elements_in_tree = list(root.findall('.//svg:image', namespaces)) + \
+                                         [el for el in root.findall('.//image') if el.tag != f"{{{namespaces['svg']}}}image"]
+
+            for image_el in all_image_elements_in_tree:
+                if image_el == background_image_element: # Already processed (or attempted)
+                    continue
+
+                href_val = None
+                attr_name_to_set = None
+                xlink_href_qname = f"{{{namespaces['xlink']}}}href"
+
+                if image_el.get(xlink_href_qname) is not None:
+                    href_val = image_el.get(xlink_href_qname)
+                    attr_name_to_set = xlink_href_qname
+                elif image_el.get('href') is not None:
+                    href_val = image_el.get('href')
+                    attr_name_to_set = 'href'
+                
+                if href_val and href_val.startswith(('http://', 'https://')):
+                    try:
+                        safe_download_url = sanitize_url(href_val) # Sanitize before download
+                        # print(f"DEBUG: Downloading remote image for embedding: {safe_download_url}")
+                        response = requests.get(safe_download_url, timeout=20, stream=True)
+                        response.raise_for_status()
+                        
+                        image_bytes = response.content
+                        
+                        content_type_header = response.headers.get('Content-Type')
+                        mime_type_source = content_type_header.split(';')[0].strip() if content_type_header and '/' in content_type_header else href_val
+                        
+                        data_uri = _image_bytes_to_data_uri(image_bytes, mime_type_source)
+                        image_el.set(attr_name_to_set, data_uri)
+                        # print(f"DEBUG: Replaced remote URL {href_val} with data URI.")
+                    except requests.exceptions.RequestException as e_req:
+                        print(f"Failed to download or process remote image {href_val}: {e_req}")
+                    except Exception as e_proc:
+                        print(f"Error processing remote image {href_val} into data URI: {e_proc}")
+            
+        if root is None: 
+            print("Error: SVG root is None after parsing, cannot proceed.")
+            return None
+
+        modified_svg_bytes = ET.tostring(root, encoding='UTF-8', method='xml', xml_declaration=True)
+        
+        try:
+            debug_svg_filename = f"debug_embedded_svg_{uuid.uuid4().hex[:8]}.svg"
+            debug_svg_path = TEMP_DIR / debug_svg_filename
+            with open(debug_svg_path, "wb") as f_debug:
+                f_debug.write(modified_svg_bytes)
+            print(f"Saved SVG with all images embedded for debugging to: {debug_svg_path}")
+        except Exception as e_save:
+            print(f"Error saving debug embedded SVG: {e_save}")
+            
+        return modified_svg_bytes
+
+    except ET.ParseError as e:
+        print(f"Error parsing SVG XML: {e}")
+        return None
+    except Exception as e:
+        print(f"An unexpected error occurred during SVG manipulation: {e}")
+        return None
+
+def convert_svg_bytes_to_png_api(svg_content_bytes: bytes, original_template_name: str, index: int):
+    """
+    Converts SVG bytes to PNG bytes using the external API.
+    Saves the PNG to a temporary file and returns the path.
+    """
+    if not svg_content_bytes:
+        return None
+    try:
+        # The API expects a file upload
+        files = {'svg_file': ('generated_svg.svg', svg_content_bytes, 'image/svg+xml')}
+        response = requests.post(SVG2PNG_API_URL, files=files, timeout=60)
+        response.raise_for_status()
+        
+        # Use UUID for more robust unique filenames
+        png_filename = f"{original_template_name.replace(' ','_')}_output_{index}_{uuid.uuid4().hex[:8]}.png"
+        temp_png_path = TEMP_DIR / png_filename
+        with open(temp_png_path, "wb") as f:
+            f.write(response.content)
+        return str(temp_png_path)
+        
+    except requests.exceptions.RequestException as e:
+        print(f"Error converting SVG to PNG via API: {e}")
+        if hasattr(e, 'response') and e.response is not None:
+            print(f"API Response: {e.response.text}")
+        return None
+    except Exception as e:
+        print(f"Error saving PNG: {e}")
+        return None
+
+# --- Gradio App Logic ---
+
+def generate_images_from_template(original_svg_download_url: str, template_name_for_file: str, uploaded_image_files: list, request: gr.Request):
+    """
+    Main processing function for Gradio.
+    Takes original SVG URL, template name, and list of uploaded image file objects.
+    """
+    if not original_svg_download_url:
+        gr.Warning("模板信息未加载，请确保URL参数正确。")
+        return [], None
+    if not uploaded_image_files:
+        gr.Warning("请上传至少一张图片。")
+        return [], None
+
+    # 1. Download original SVG content
+    original_svg_content = download_svg_content(original_svg_download_url)
+    if not original_svg_content:
+        gr.Error("无法下载原始SVG模板。")
+        return [], None
+
+    generated_png_paths = []
+    processed_count = 0
+    
+    for i, uploaded_file_obj in enumerate(uploaded_image_files):
+        # uploaded_file_obj.name is the temporary path to the uploaded file
+        new_image_path = uploaded_file_obj.name 
+        
+        # 2. Replace background in SVG for each uploaded image
+        modified_svg_bytes = replace_background_in_svg(original_svg_content, new_image_path)
+        if not modified_svg_bytes:
+            gr.Warning(f"处理图片 {i+1} 失败：无法修改SVG。")
+            continue # Skip this image
+
+        # 3. Convert modified SVG to PNG
+        # Use template_name_for_file for unique output filenames
+        png_path = convert_svg_bytes_to_png_api(modified_svg_bytes, template_name_for_file, i + 1)
+        if png_path:
+            generated_png_paths.append(png_path)
+            processed_count += 1
+        else:
+            gr.Warning(f"处理图片 {i+1} 失败：无法转换为PNG。")
+            
+    if not generated_png_paths:
+        gr.Info("未能成功生成任何图片。")
+        return [], None
+
+    zip_buffer = io.BytesIO()
+    with zipfile.ZipFile(zip_buffer, "w", zipfile.ZIP_DEFLATED) as zf:
+        for png_path_str in generated_png_paths:
+            png_file = pathlib.Path(png_path_str)
+            zf.write(png_file, arcname=png_file.name) 
+    
+    zip_buffer.seek(0)
+    
+    zip_filename = f"{template_name_for_file.replace(' ','_')}_batch_{uuid.uuid4().hex[:8]}.zip" # Added UUID
+    temp_zip_path = TEMP_DIR / zip_filename
+    with open(temp_zip_path, "wb") as f:
+        f.write(zip_buffer.getvalue())
+
+    gr.Info(f"成功生成 {processed_count} 张图片!")
+    return generated_png_paths, str(temp_zip_path)
+
+
+def initial_load_template_info(request: gr.Request):
+    """
+    Loads initial template information based on URL query parameters.
+    Downloads the thumbnail and returns its local path.
+    """
+    query_params = request.query_params
+    m = query_params.get("m")
+    u = query_params.get("u")
+    category = query_params.get("category")
+    name = query_params.get("name") # This is the template name
+
+    if not all([m, u, category, name]):
+        print("Initial load: URL parameters (m, u, category, name) are incomplete or missing.")
+        return None, "无模板信息 (请检查URL参数)", None, "无模板"
+
+    details = fetch_template_details(m, u, category, name)
+    
+    local_thumbnail_path = None
+    template_display_name = "错误: 无法加载模板信息" # Default error message
+    svg_download_url = None
+    
+    if details and "name" in details: 
+        template_display_name = details["name"] # Use name from details for display
+        svg_download_url = details.get("svg_download_url")
+
+        if details.get("thumbnail_url"):
+            try:
+                thumb_url = details["thumbnail_url"]
+                print(f"Fetching thumbnail from: {thumb_url}")
+                thumb_response = requests.get(thumb_url, timeout=10)
+                thumb_response.raise_for_status()
+                
+                # Create a unique filename for the thumbnail in TEMP_DIR
+                # Use name from URL params for file naming consistency if needed, or details['name']
+                safe_template_name_for_file = name.replace(' ','_').replace('/','_').replace('\\\\','_') # Basic sanitization
+                thumb_filename = f"thumb_{safe_template_name_for_file}_{uuid.uuid4().hex[:8]}.png"
+                local_thumbnail_path = TEMP_DIR / thumb_filename
+                with open(local_thumbnail_path, "wb") as f:
+                    f.write(thumb_response.content)
+                print(f"Thumbnail saved to: {local_thumbnail_path}")
+            except requests.exceptions.RequestException as e:
+                print(f"Error downloading thumbnail from {details.get('thumbnail_url')}: {e}")
+                local_thumbnail_path = None 
+            except Exception as e:
+                print(f"Error saving thumbnail: {e}")
+                local_thumbnail_path = None
+    else:
+        print(f"Failed to load template details for m={m}, u={u}, category={category}, name={name}")
+        # template_display_name is already set to the default error message
+
+    return str(local_thumbnail_path) if local_thumbnail_path else None, \
+           template_display_name, \
+           svg_download_url, \
+           name # 'name' from URL param for template_name_for_file_state
+
+# --- Gradio Interface ---
+with gr.Blocks(theme=gr.themes.Soft(), title="SVG模板批量图片生成器") as demo:
+    gr.Markdown("## 使用SVG模板批量生成图片")
+    gr.Markdown("从URL加载模板，上传您自己的图片替换模板中的底图，然后批量生成并下载。")
+
+    # Hidden state to store original SVG download URL and template name for file ops
+    original_svg_download_url_state = gr.State()
+    template_name_for_file_state = gr.State() # To preserve the name from URL param for consistent file naming
+
+    with gr.Row():
+        with gr.Column(scale=1, min_width=200):
+            template_thumbnail_display = gr.Image(label="当前模板缩略图", interactive=False, height=200, type="filepath") # Ensure type is filepath
+            template_name_display = gr.Textbox(label="当前模板名称", interactive=False)
+        with gr.Column(scale=3):
+            uploaded_images_input = gr.Files(
+                label="上传您的图片 (可多选)", 
+                file_count="multiple", 
+                file_types=["image"] # Accepts .png, .jpg, .jpeg, .gif, .webp etc.
+            )
+    
+    generate_button = gr.Button("🚀 立即生成图片", variant="primary", scale=1)
+    
+    with gr.Accordion("生成结果预览与下载", open=True):
+        output_gallery = gr.Gallery(
+            label="生成图片预览", 
+            show_label=True, 
+            elem_id="output_gallery", 
+            columns=[4], 
+            object_fit="contain", 
+            height="auto"
+            # type="filepath" is default for Gallery if fed filepaths
+        )
+        output_zip_file = gr.File(label="下载所有生成图片的ZIP包", interactive=False, type="filepath") # Ensure type is filepath
+
+    # Load initial template info based on URL parameters when the interface loads.
+    # The `initial_load_template_info` function will parse request.query_params.
+    # It's crucial that `gr.Request` is passed to it.
+    # `inputs=None` with `request: gr.Request` in function signature works.
+    demo.load(
+        initial_load_template_info, 
+        inputs=None, # gr.Request is implicitly passed if type-hinted in the function
+        outputs=[
+            template_thumbnail_display, 
+            template_name_display, 
+            original_svg_download_url_state,
+            template_name_for_file_state
+        ]
+    )
+    
+    generate_button.click(
+        generate_images_from_template, 
+        inputs=[
+            original_svg_download_url_state, 
+            template_name_for_file_state,
+            uploaded_images_input
+            # gr.Request is also implicitly passed here
+        ],
+        outputs=[output_gallery, output_zip_file]
+    )
+
+if __name__ == "__main__":
+    # To run this app:
+    # 1. Ensure your template_server.py is running (e.g., on http://localhost:8001)
+    #    and has the /template_details/{username}/{category}/{template_name} endpoint.
+    # 2. Run this script: python use_template_app.py
+    # 3. Open your browser to the Gradio link, appending parameters, e.g.:
+    #    http://127.0.0.1:7860/?m=pub&u=all&category=%E5%AE%B6%E7%94%B5&name=%E6%A8%A1%E6%9D%BF_%E5%86%B0%E7%82%B9%E4%BB%B7
+    
+    # For easy testing, you might want to create a dummy template_server.py endpoint
+    # or hardcode some details if the server isn't ready.
+    
+    demo.launch()
+
+    # Optional: Cleanup TEMP_DIR logic can be added here if needed for long-running servers
+    # For development, manual cleanup or OS temp cleaning is often sufficient.
+    # Example:
+    # importate atexit
+    # def cleanup_temp_dir():
+    #     if TEMP_DIR.exists():
+    #         print(f"Cleaning up temp directory: {TEMP_DIR}")
+    #         shutil.rmtree(TEMP_DIR)
+    # atexit.register(cleanup_temp_dir) # This might be too aggressive for dev
+