更新两个小功能，即梦端点无法改成v3和增加通义万相接入功能，增加摩搭社区Z-image模型并使用每日免费500次请求

.gitignore

@@ -58,6 +58,10 @@ history/*
 .gemini/
 /image_providers.yaml
 /text_providers.yaml
+*_providers.yaml
+*_providers.yml
+!*_providers.yaml.example
+!*_providers.yml.example
 
 # Logs
 *.log

README.md

@@ -263,6 +263,13 @@ providers:
 
 ## 更新日志
 
+### v1.4.2 (2025-12-18)
+- 增加摩搭社区Z-image模型并使用每日免费500次请求
+
+### v1.4.1 (2025-12-17)
+- 🔧 修复即梦AI端点无法从 `/v1/images/generations` 修改为 `/v3/images/generations` 的问题，支持即梦AI的 `doubao-seedream-4-5-251128` 模型
+- ✨ 新增通义万相 `wan2.6-t2i`（类型选择：通义万相 Wan2.6（文生图V2））
+
 ### v1.4.0 (2025-11-30)
 - 🏗️ 后端架构重构：拆分单体路由为模块化蓝图（history、images、generation、outline、config）
 - 🏗️ 前端组件重构：提取可复用组件（ImageGalleryModal、OutlineModal、ShowcaseBackground等）

backend/app.py

@@ -1,7 +1,7 @@
 import logging
 import sys
 from pathlib import Path
-from flask import Flask, send_from_directory
+from flask import Flask, send_from_directory, request
 from flask_cors import CORS
 from backend.config import Config
 from backend.routes import register_routes
@@ -78,6 +78,8 @@ def serve_index():
         # 处理 Vue Router 的 HTML5 History 模式
         @app.errorhandler(404)
         def fallback(e):
+            if request.path.startswith('/api/'):
+                return {"success": False, "error": "Not Found"}, 404
             return send_from_directory(app.static_folder, 'index.html')
     else:
         @app.route('/')

backend/generators/factory.py

@@ -4,6 +4,8 @@
 from .google_genai import GoogleGenAIGenerator
 from .openai_compatible import OpenAICompatibleGenerator
 from .image_api import ImageApiGenerator
+from .wan26_t2i import Wan26T2IGenerator
+from .modelscope_z_image import ModelScopeZImageGenerator
 
 
 class ImageGeneratorFactory:
@@ -15,6 +17,8 @@ class ImageGeneratorFactory:
         'openai': OpenAICompatibleGenerator,
         'openai_compatible': OpenAICompatibleGenerator,
         'image_api': ImageApiGenerator,
+        'wan2.6-t2i': Wan26T2IGenerator,
+        'modelscope_z_image': ModelScopeZImageGenerator,
     }
 
     @classmethod

backend/generators/image_api.py

@@ -36,13 +36,12 @@ class ImageApiGenerator(ImageGeneratorBase):
     def __init__(self, config: Dict[str, Any]):
         super().__init__(config)
         logger.debug("初始化 ImageApiGenerator...")
-        self.base_url = config.get('base_url', 'https://api.example.com').rstrip('/').rstrip('/v1')
-        self.model = config.get('model', 'default-model')
-        self.default_aspect_ratio = config.get('default_aspect_ratio', '3:4')
-        self.image_size = config.get('image_size', '4K')
-
+
+        # 原始 Base URL
+        base_url = (config.get('base_url', 'https://api.example.com') or 'https://api.example.com').strip().rstrip('/')
+
         # 支持自定义端点路径
-        endpoint_type = config.get('endpoint_type', '/v1/images/generations')
+        endpoint_type = (config.get('endpoint_type', '/v1/images/generations') or '/v1/images/generations').strip()
         # 兼容旧的简写格式
         if endpoint_type == 'images':
             endpoint_type = '/v1/images/generations'
@@ -52,6 +51,28 @@ def __init__(self, config: Dict[str, Any]):
         if not endpoint_type.startswith('/'):
             endpoint_type = '/' + endpoint_type
         self.endpoint_type = endpoint_type
+
+        # 智能处理 Base URL: 
+        # 如果 endpoint_type 是 /v3/xxx，且 base_url 结尾也是 /v3，则去掉 base_url 的 /v3
+        # 避免拼接成 /v3/v3/xxx
+
+        # 提取版本前缀 (e.g. /v1, /v3)
+        import re
+        version_match = re.search(r'^/(v\d+)', self.endpoint_type)
+        if version_match:
+            version_prefix = version_match.group(1) # e.g. /v3
+            if base_url.endswith(version_prefix):
+                 base_url = base_url[:-len(version_prefix)].rstrip('/')
+
+        # 兼容旧逻辑：如果 base_url 结尾是 /v1 但 endpoint 也是 /v1 开头，去掉 base_url 的
+        elif base_url.endswith('/v1') and self.endpoint_type.startswith('/v1'):
+             base_url = base_url[:-3].rstrip('/')
+
+        self.base_url = base_url
+
+        self.model = config.get('model', 'default-model')
+        self.default_aspect_ratio = config.get('default_aspect_ratio', '3:4')
+        self.image_size = config.get('image_size', '4K')
 
         logger.info(f"ImageApiGenerator 初始化完成: base_url={self.base_url}, model={self.model}, endpoint={self.endpoint_type}")
 

backend/generators/modelscope_z_image.py

@@ -0,0 +1,175 @@
+import logging
+import re
+import time
+from typing import Dict, Any, Optional, List
+
+import requests
+
+from .base import ImageGeneratorBase
+
+logger = logging.getLogger(__name__)
+
+
+class ModelScopeZImageGenerator(ImageGeneratorBase):
+    def __init__(self, config: Dict[str, Any]):
+        super().__init__(config)
+
+        base_url = (config.get('base_url') or 'https://api-inference.modelscope.cn').strip().rstrip('/')
+        endpoint_type = (config.get('endpoint_type') or '/v1/images/generations').strip()
+        if not endpoint_type.startswith('/'):
+            endpoint_type = '/' + endpoint_type
+
+        version_match = re.search(r'^/(v\d+)', endpoint_type)
+        if version_match:
+            version_prefix = '/' + version_match.group(1)
+            if base_url.endswith(version_prefix):
+                base_url = base_url[:-len(version_prefix)].rstrip('/')
+        elif base_url.endswith('/v1') and endpoint_type.startswith('/v1'):
+            base_url = base_url[:-3].rstrip('/')
+
+        self.base_url = base_url
+        self.endpoint_type = endpoint_type
+
+        self.model = config.get('model') or 'Tongyi-MAI/Z-Image-Turbo'
+        self.task_endpoint = (config.get('task_endpoint') or '/v1/tasks').strip()
+        if not self.task_endpoint.startswith('/'):
+            self.task_endpoint = '/' + self.task_endpoint
+
+        self.poll_interval_seconds = float(config.get('poll_interval_seconds') or 3)
+        self.max_wait_seconds = float(config.get('max_wait_seconds') or 300)
+        self.max_prompt_chars = int(config.get('max_prompt_chars') or 1900)
+
+        logger.info(
+            f"ModelScopeZImageGenerator 初始化完成: base_url={self.base_url}, model={self.model}, endpoint={self.endpoint_type}"
+        )
+
+    def _normalize_prompt(self, prompt: str) -> str:
+        text = (prompt or "").strip()
+        max_chars = self.max_prompt_chars
+        if max_chars < 100:
+            max_chars = 100
+        if len(text) <= max_chars:
+            return text
+        return text[:max_chars].rstrip()
+
+    def validate_config(self) -> bool:
+        if not self.api_key:
+            raise ValueError(
+                "ModelScope API Key 未配置。\n"
+                "解决方案：在系统设置页面编辑该服务商，填写 API Key"
+            )
+        if not self.base_url:
+            raise ValueError(
+                "ModelScope Base URL 未配置。\n"
+                "解决方案：在系统设置页面编辑该服务商，填写 Base URL（例如 https://api-inference.modelscope.cn）"
+            )
+        return True
+
+    def generate_image(
+        self,
+        prompt: str,
+        model: Optional[str] = None,
+        **kwargs
+    ) -> bytes:
+        self.validate_config()
+
+        model_id = (model or self.model).strip()
+        if not model_id:
+            raise ValueError(
+                "ModelScope 模型未配置。\n"
+                "解决方案：在系统设置页面编辑该服务商，填写模型（例如 Tongyi-MAI/Z-Image-Turbo）"
+            )
+
+        normalized_prompt = self._normalize_prompt(prompt)
+        create_url = f"{self.base_url}{self.endpoint_type}"
+        headers = {
+            "Authorization": f"Bearer {self.api_key}",
+            "Content-Type": "application/json",
+            "X-ModelScope-Async-Mode": "true",
+        }
+        payload: Dict[str, Any] = {
+            "model": model_id,
+            "prompt": normalized_prompt,
+            "n": 1,
+            "size": kwargs.get("size") or self.config.get("size") or "1024x1024",
+        }
+
+        logger.info(f"ModelScope Z-Image 提交任务: model={model_id}, url={create_url}")
+        response = requests.post(create_url, headers=headers, json=payload, timeout=60)
+
+        if response.status_code != 200:
+            detail = response.text[:800]
+            raise Exception(
+                f"ModelScope 图片生成请求失败 (状态码: {response.status_code})\n"
+                f"请求地址: {create_url}\n"
+                f"错误详情: {detail}"
+            )
+
+        data = response.json() or {}
+        task_id = data.get("task_id") or data.get("id")
+        if not task_id:
+            raise Exception(
+                "ModelScope 响应中未找到 task_id。\n"
+                f"响应片段: {str(data)[:800]}"
+            )
+
+        task_url = f"{self.base_url}{self.task_endpoint}/{task_id}"
+        status_headers = {
+            "Authorization": f"Bearer {self.api_key}",
+            "Content-Type": "application/json",
+            "X-ModelScope-Task-Type": "image_generation",
+        }
+
+        deadline = time.time() + self.max_wait_seconds
+        last_status: Optional[str] = None
+        while True:
+            if time.time() > deadline:
+                raise Exception(
+                    f"ModelScope 任务超时（{self.max_wait_seconds}s）。task_id={task_id}, last_status={last_status}"
+                )
+
+            status_resp = requests.get(task_url, headers=status_headers, timeout=60)
+            if status_resp.status_code != 200:
+                detail = status_resp.text[:800]
+                raise Exception(
+                    f"ModelScope 任务查询失败 (状态码: {status_resp.status_code})\n"
+                    f"请求地址: {task_url}\n"
+                    f"错误详情: {detail}"
+                )
+
+            task_data = status_resp.json() or {}
+            task_status = (task_data.get("task_status") or task_data.get("status") or "").upper()
+            last_status = task_status or last_status
+
+            if task_status == "SUCCEED":
+                output_images = task_data.get("output_images") or []
+                if not isinstance(output_images, list) or not output_images:
+                    raise Exception(
+                        "ModelScope 任务成功但未返回图片地址。\n"
+                        f"响应片段: {str(task_data)[:800]}"
+                    )
+                image_url = output_images[0]
+                if not isinstance(image_url, str) or not image_url.strip():
+                    raise Exception(
+                        "ModelScope 返回的图片地址无效。\n"
+                        f"响应片段: {str(task_data)[:800]}"
+                    )
+                img_resp = requests.get(image_url, timeout=120)
+                if img_resp.status_code != 200:
+                    raise Exception(
+                        f"下载图片失败 (状态码: {img_resp.status_code})\n"
+                        f"图片地址: {image_url}\n"
+                        f"错误详情: {img_resp.text[:200]}"
+                    )
+                return img_resp.content
+
+            if task_status == "FAILED":
+                error_msg = (
+                    task_data.get("message")
+                    or task_data.get("error")
+                    or task_data.get("output")
+                    or "未知错误"
+                )
+                raise Exception(f"ModelScope 图片生成失败: {error_msg}")
+
+            time.sleep(self.poll_interval_seconds)