feat: 普号增加25s普通视频、增加pro系列高清模型、统一模型名字、过载不再计入错误禁用计数

2026-02-04 02:04:42 +08:00 · 2025-12-24 19:39:28 +08:00
parent 2c2fd44b6a
commit fad83533bf
8 changed files with 226 additions and 69 deletions
--- a/README.md
+++ b/README.md
@@ -110,14 +110,15 @@ python main.py

 | 功能 | 模型 | 说明 |
 |------|------|------|
-| 文生图 | `sora-image*` | 使用 `content` 为字符串 |
-| 图生图 | `sora-image*` | 使用 `content` 数组 + `image_url` |
-| 文生视频 | `sora-video*` | 使用 `content` 为字符串 |
-| 图生视频 | `sora-video*` | 使用 `content` 数组 + `image_url` |
-| 创建角色 | `sora-video*` | 使用 `content` 数组 + `video_url` |
-| 角色生成视频 | `sora-video*` | 使用 `content` 数组 + `video_url` + 文本 |
-| Remix | `sora-video*` | 在 `content` 中包含 Remix ID |
-| 视频分镜 | `sora-video*` | 在 `content` 中使用```[时长s]提示词```格式触发 |
+| 文生图 | `gpt-image*` | 使用 `content` 为字符串 |
+| 图生图 | `gpt-image*` | 使用 `content` 数组 + `image_url` |
+| 文生视频 | `sora2-*` | 使用 `content` 为字符串 |
+| 图生视频 | `sora2-*` | 使用 `content` 数组 + `image_url` |
+| 视频风格 | `sora2-*` | 在提示词中使用 `{风格ID}` 格式,如 `{anime}提示词` |
+| 创建角色 | `sora2-*` | 使用 `content` 数组 + `video_url` |
+| 角色生成视频 | `sora2-*` | 使用 `content` 数组 + `video_url` + 文本 |
+| Remix | `sora2-*` | 在 `content` 中包含 Remix ID |
+| 视频分镜 | `sora2-*` | 在 `content` 中使用```[时长s]提示词```格式触发 |

 ---

@@ -135,20 +136,44 @@ python main.py

 | 模型 | 说明 | 尺寸 |
 |------|------|------|
-| `sora-image` | 文生图（默认） | 360×360 |
-| `sora-image-landscape` | 文生图（横屏） | 540×360 |
-| `sora-image-portrait` | 文生图（竖屏） | 360×540 |
+| `gpt-image` | 文生图（正方形） | 360×360 |
+| `gpt-image-landscape` | 文生图（横屏） | 540×360 |
+| `gpt-image-portrait` | 文生图（竖屏） | 360×540 |

 **视频模型**

+**标准版（Sora2）**
+
 | 模型 | 时长 | 方向 | 说明 |
 |------|------|------|------|
-| `sora-video-10s` | 10秒 | 横屏 | 文生视频/图生视频 |
-| `sora-video-15s` | 15秒 | 横屏 | 文生视频/图生视频 |
-| `sora-video-landscape-10s` | 10秒 | 横屏 | 文生视频/图生视频 |
-| `sora-video-landscape-15s` | 15秒 | 横屏 | 文生视频/图生视频 |
-| `sora-video-portrait-10s` | 10秒 | 竖屏 | 文生视频/图生视频 |
-| `sora-video-portrait-15s` | 15秒 | 竖屏 | 文生视频/图生视频 |
+| `sora2-landscape-10s` | 10秒 | 横屏 | 文生视频/图生视频 |
+| `sora2-landscape-15s` | 15秒 | 横屏 | 文生视频/图生视频 |
+| `sora2-landscape-25s` | 25秒 | 横屏 | 文生视频/图生视频 |
+| `sora2-portrait-10s` | 10秒 | 竖屏 | 文生视频/图生视频 |
+| `sora2-portrait-15s` | 15秒 | 竖屏 | 文生视频/图生视频 |
+| `sora2-portrait-25s` | 25秒 | 竖屏 | 文生视频/图生视频 |
+
+**Pro 版（需要 ChatGPT Pro 订阅）**
+
+| 模型 | 时长 | 方向 | 说明 |
+|------|------|------|------|
+| `sora2pro-landscape-10s` | 10秒 | 横屏 | Pro 质量文生视频/图生视频 |
+| `sora2pro-landscape-15s` | 15秒 | 横屏 | Pro 质量文生视频/图生视频 |
+| `sora2pro-landscape-25s` | 25秒 | 横屏 | Pro 质量文生视频/图生视频 |
+| `sora2pro-portrait-10s` | 10秒 | 竖屏 | Pro 质量文生视频/图生视频 |
+| `sora2pro-portrait-15s` | 15秒 | 竖屏 | Pro 质量文生视频/图生视频 |
+| `sora2pro-portrait-25s` | 25秒 | 竖屏 | Pro 质量文生视频/图生视频 |
+
+**Pro HD 版（需要 ChatGPT Pro 订阅，高清质量）**
+
+| 模型 | 时长 | 方向 | 说明 |
+|------|------|------|------|
+| `sora2pro-hd-landscape-10s` | 10秒 | 横屏 | Pro 高清文生视频/图生视频 |
+| `sora2pro-hd-landscape-15s` | 15秒 | 横屏 | Pro 高清文生视频/图生视频 |
+| `sora2pro-hd-portrait-10s` | 10秒 | 竖屏 | Pro 高清文生视频/图生视频 |
+| `sora2pro-hd-portrait-15s` | 15秒 | 竖屏 | Pro 高清文生视频/图生视频 |
+
+> **注意：** Pro 系列模型需要 ChatGPT Pro 订阅（`plan_type: "chatgpt_pro"`）。如果没有 Pro 账号，请求这些模型会返回错误。

 #### 请求示例

@@ -159,13 +184,14 @@ curl -X POST "http://localhost:8000/v1/chat/completions" \
  -H "Authorization: Bearer han1234" \
  -H "Content-Type: application/json" \
  -d '{
-    "model": "sora-image",
+    "model": "gpt-image",
    "messages": [
      {
        "role": "user",
        "content": "一只可爱的小猫咪"
      }
-    ]
+    ],
+    "stream": true
  }'
 ```

@@ -176,7 +202,7 @@ curl -X POST "http://localhost:8000/v1/chat/completions" \
  -H "Authorization: Bearer han1234" \
  -H "Content-Type: application/json" \
  -d '{
-    "model": "sora-image",
+    "model": "gpt-image",
    "messages": [
      {
        "role": "user",
@@ -205,7 +231,7 @@ curl -X POST "http://localhost:8000/v1/chat/completions" \
  -H "Authorization: Bearer han1234" \
  -H "Content-Type: application/json" \
  -d '{
-    "model": "sora-video-landscape-10s",
+    "model": "sora2-landscape-10s",
    "messages": [
      {
        "role": "user",
@@ -223,7 +249,7 @@ curl -X POST "http://localhost:8000/v1/chat/completions" \
  -H "Authorization: Bearer han1234" \
  -H "Content-Type: application/json" \
  -d '{
-    "model": "sora-video-landscape-10s",
+    "model": "sora2-landscape-10s",
    "messages": [
      {
        "role": "user",
@@ -254,13 +280,14 @@ curl -X POST "http://localhost:8000/v1/chat/completions" \
  -H "Authorization: Bearer han1234" \
  -H "Content-Type: application/json" \
  -d '{
-    "model": "sora-video-landscape-10s",
+    "model": "sora2-landscape-10s",
    "messages": [
      {
        "role": "user",
        "content": "https://sora.chatgpt.com/p/s_68e3a06dcd888191b150971da152c1f5改成水墨画风格"
      }
-    ]
+    ],
+    "stream": true
  }'
 ```

@@ -280,13 +307,14 @@ curl -X POST "http://localhost:8000/v1/chat/completions" \
  -H "Authorization: Bearer han1234" \
  -H "Content-Type: application/json" \
  -d '{
-    "model": "sora-video-landscape-10s",
+    "model": "sora2-landscape-10s",
    "messages": [
      {
        "role": "user",
        "content": "[5.0s]猫猫从飞机上跳伞 [5.0s]猫猫降落 [10.0s]猫猫在田野奔跑"
      }
-    ]
+    ],
+    "stream": true
  }'
 ```

@@ -322,7 +350,7 @@ curl -X POST "http://localhost:8000/v1/chat/completions" \
  -H "Authorization: Bearer han1234" \
  -H "Content-Type: application/json" \
  -d '{
-    "model": "sora-video-landscape-10s",
+    "model": "sora2-landscape-10s",
    "messages": [
      {
        "role": "user",
@@ -340,7 +368,7 @@ curl -X POST "http://localhost:8000/v1/chat/completions" \
  -H "Authorization: Bearer han1234" \
  -H "Content-Type: application/json" \
  -d '{
-    "model": "sora-video-landscape-10s",
+    "model": "sora2-landscape-10s",
    "messages": [
      {
        "role": "user",
@@ -358,7 +386,7 @@ curl -X POST "http://localhost:8000/v1/chat/completions" \
  -H "Authorization: Bearer han1234" \
  -H "Content-Type: application/json" \
  -d '{
-    "model": "sora-video-landscape-10s",
+    "model": "sora2-landscape-10s",
    "messages": [
      {
        "role": "user",
@@ -394,7 +422,7 @@ curl -X POST "http://localhost:8000/v1/chat/completions" \
  -H "Authorization: Bearer han1234" \
  -H "Content-Type: application/json" \
  -d '{
-    "model": "sora-video-landscape-10s",
+    "model": "sora2-landscape-10s",
    "messages": [
      {
        "role": "user",
@@ -421,7 +449,7 @@ curl -X POST "http://localhost:8000/v1/chat/completions" \
  -H "Authorization: Bearer han1234" \
  -H "Content-Type: application/json" \
  -d '{
-    "model": "sora-video-landscape-10s",
+    "model": "sora2-landscape-10s",
    "messages": [
      {
        "role": "user",
@@ -461,7 +489,7 @@ response = requests.post(
        "Content-Type": "application/json"
    },
    json={
-        "model": "sora-video-landscape-10s",
+        "model": "sora2-landscape-10s",
        "messages": [
            {
                "role": "user",
--- a/config/setting.toml
+++ b/config/setting.toml
@@ -27,7 +27,7 @@ base_url = "http://127.0.0.1:8000"

 [generation]
 image_timeout = 300
-video_timeout = 1500
+video_timeout = 3000

 [admin]
 error_ban_threshold = 3
--- a/config/setting_warp.toml
+++ b/config/setting_warp.toml
@@ -27,7 +27,7 @@ base_url = "http://127.0.0.1:8000"

 [generation]
 image_timeout = 300
-video_timeout = 1500
+video_timeout = 3000

 [admin]
 error_ban_threshold = 3
--- a/src/core/config.py
+++ b/src/core/config.py
@@ -163,7 +163,7 @@ class Config:
    @property
    def video_timeout(self) -> int:
        """Get video generation timeout in seconds"""
-        return self._config.get("generation", {}).get("video_timeout", 1500)
+        return self._config.get("generation", {}).get("video_timeout", 3000)

    def set_video_timeout(self, timeout: int):
        """Set video generation timeout in seconds"""
--- a/src/core/database.py
+++ b/src/core/database.py
@@ -144,12 +144,12 @@ class Database:
        if count[0] == 0:
            # Get generation config from config_dict if provided, otherwise use defaults
            image_timeout = 300
-            video_timeout = 1500
+            video_timeout = 3000

            if config_dict:
                generation_config = config_dict.get("generation", {})
                image_timeout = generation_config.get("image_timeout", 300)
-                video_timeout = generation_config.get("video_timeout", 1500)
+                video_timeout = generation_config.get("video_timeout", 3000)

            await db.execute("""
                INSERT INTO generation_config (id, image_timeout, video_timeout)
@@ -401,7 +401,7 @@ class Database:
                CREATE TABLE IF NOT EXISTS generation_config (
                    id INTEGER PRIMARY KEY DEFAULT 1,
                    image_timeout INTEGER DEFAULT 300,
-                    video_timeout INTEGER DEFAULT 1500,
+                    video_timeout INTEGER DEFAULT 3000,
                    created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP,
                    updated_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP
                )
@@ -1013,7 +1013,7 @@ class Database:
                return GenerationConfig(**dict(row))
            # If no row exists, return a default config
            # This should not happen in normal operation as _ensure_config_rows should create it
-            return GenerationConfig(image_timeout=300, video_timeout=1500)
+            return GenerationConfig(image_timeout=300, video_timeout=3000)

    async def update_generation_config(self, image_timeout: int = None, video_timeout: int = None):
        """Update generation configuration"""
@@ -1027,10 +1027,10 @@ class Database:
                current = dict(row)
                # Update only provided fields
                new_image_timeout = image_timeout if image_timeout is not None else current.get("image_timeout", 300)
-                new_video_timeout = video_timeout if video_timeout is not None else current.get("video_timeout", 1500)
+                new_video_timeout = video_timeout if video_timeout is not None else current.get("video_timeout", 3000)
            else:
                new_image_timeout = image_timeout if image_timeout is not None else 300
-                new_video_timeout = video_timeout if video_timeout is not None else 1500
+                new_video_timeout = video_timeout if video_timeout is not None else 3000

            await db.execute("""
                UPDATE generation_config
--- a/src/services/generation_handler.py
+++ b/src/services/generation_handler.py
@@ -19,52 +19,139 @@ from ..core.logger import debug_logger

 # Model configuration
 MODEL_CONFIG = {
-    "sora-image": {
+    "gpt-image": {
        "type": "image",
        "width": 360,
        "height": 360
    },
-    "sora-image-landscape": {
+    "gpt-image-landscape": {
        "type": "image",
        "width": 540,
        "height": 360
    },
-    "sora-image-portrait": {
+    "gpt-image-portrait": {
        "type": "image",
        "width": 360,
        "height": 540
    },
    # Video models with 10s duration (300 frames)
-    "sora-video-10s": {
+    "sora2-landscape-10s": {
        "type": "video",
        "orientation": "landscape",
        "n_frames": 300
    },
-    "sora-video-landscape-10s": {
-        "type": "video",
-        "orientation": "landscape",
-        "n_frames": 300
-    },
-    "sora-video-portrait-10s": {
+    "sora2-portrait-10s": {
        "type": "video",
        "orientation": "portrait",
        "n_frames": 300
    },
    # Video models with 15s duration (450 frames)
-    "sora-video-15s": {
+    "sora2-landscape-15s": {
        "type": "video",
        "orientation": "landscape",
        "n_frames": 450
    },
-    "sora-video-landscape-15s": {
-        "type": "video",
-        "orientation": "landscape",
-        "n_frames": 450
-    },
-    "sora-video-portrait-15s": {
+    "sora2-portrait-15s": {
        "type": "video",
        "orientation": "portrait",
        "n_frames": 450
+    },
+    # Video models with 25s duration (750 frames)
+    "sora2-landscape-25s": {
+        "type": "video",
+        "orientation": "landscape",
+        "n_frames": 750,
+        "model": "sy_8",
+        "size": "small"
+    },
+    "sora2-portrait-25s": {
+        "type": "video",
+        "orientation": "portrait",
+        "n_frames": 750,
+        "model": "sy_8",
+        "size": "small"
+    },
+    # Pro video models (require Pro subscription)
+    "sora2pro-landscape-10s": {
+        "type": "video",
+        "orientation": "landscape",
+        "n_frames": 300,
+        "model": "sy_ore",
+        "size": "small",
+        "require_pro": True
+    },
+    "sora2pro-portrait-10s": {
+        "type": "video",
+        "orientation": "portrait",
+        "n_frames": 300,
+        "model": "sy_ore",
+        "size": "small",
+        "require_pro": True
+    },
+    "sora2pro-landscape-15s": {
+        "type": "video",
+        "orientation": "landscape",
+        "n_frames": 450,
+        "model": "sy_ore",
+        "size": "small",
+        "require_pro": True
+    },
+    "sora2pro-portrait-15s": {
+        "type": "video",
+        "orientation": "portrait",
+        "n_frames": 450,
+        "model": "sy_ore",
+        "size": "small",
+        "require_pro": True
+    },
+    "sora2pro-landscape-25s": {
+        "type": "video",
+        "orientation": "landscape",
+        "n_frames": 750,
+        "model": "sy_ore",
+        "size": "small",
+        "require_pro": True
+    },
+    "sora2pro-portrait-25s": {
+        "type": "video",
+        "orientation": "portrait",
+        "n_frames": 750,
+        "model": "sy_ore",
+        "size": "small",
+        "require_pro": True
+    },
+    # Pro HD video models (require Pro subscription, high quality)
+    "sora2pro-hd-landscape-10s": {
+        "type": "video",
+        "orientation": "landscape",
+        "n_frames": 300,
+        "model": "sy_ore",
+        "size": "large",
+        "require_pro": True
+    },
+    "sora2pro-hd-portrait-10s": {
+        "type": "video",
+        "orientation": "portrait",
+        "n_frames": 300,
+        "model": "sy_ore",
+        "size": "large",
+        "require_pro": True
+    },
+    "sora2pro-hd-landscape-15s": {
+        "type": "video",
+        "orientation": "landscape",
+        "n_frames": 450,
+        "model": "sy_ore",
+        "size": "large",
+        "require_pro": True
+    },
+    "sora2pro-hd-portrait-15s": {
+        "type": "video",
+        "orientation": "portrait",
+        "n_frames": 450,
+        "model": "sy_ore",
+        "size": "large",
+        "require_pro": True
    }
 }

@@ -294,10 +381,20 @@ class GenerationHandler:
                    return

        # Streaming mode: proceed with actual generation
+        # Check if model requires Pro subscription
+        require_pro = model_config.get("require_pro", False)
+
        # Select token (with lock for image generation, Sora2 quota check for video generation)
-        token_obj = await self.load_balancer.select_token(for_image_generation=is_image, for_video_generation=is_video)
+        # If Pro is required, filter for Pro tokens only
+        token_obj = await self.load_balancer.select_token(
+            for_image_generation=is_image,
+            for_video_generation=is_video,
+            require_pro=require_pro
+        )
        if not token_obj:
-            if is_image:
+            if require_pro:
+                raise Exception("No available Pro tokens. Pro models require a ChatGPT Pro subscription.")
+            elif is_image:
                raise Exception("No available tokens for image generation. All tokens are either disabled, cooling down, locked, or expired.")
            else:
                raise Exception("No available tokens for video generation. All tokens are either disabled, cooling down, Sora2 quota exhausted, don't support Sora2, or expired.")
@@ -383,12 +480,18 @@ class GenerationHandler:
                    )
                else:
                    # Normal video generation
+                    # Get model and size from config (default to sy_8 and small for backward compatibility)
+                    sora_model = model_config.get("model", "sy_8")
+                    video_size = model_config.get("size", "small")
+
                    task_id = await self.sora_client.generate_video(
                        clean_prompt, token_obj.token,
                        orientation=model_config["orientation"],
                        media_id=media_id,
                        n_frames=n_frames,
-                        style_id=style_id
+                        style_id=style_id,
+                        model=sora_model,
+                        size=video_size
                    )
            else:
                task_id = await self.sora_client.generate_image(
@@ -1271,10 +1374,16 @@ class GenerationHandler:
            # Get n_frames from model configuration
            n_frames = model_config.get("n_frames", 300)  # Default to 300 frames (10s)

+            # Get model and size from config (default to sy_8 and small for backward compatibility)
+            sora_model = model_config.get("model", "sy_8")
+            video_size = model_config.get("size", "small")
+
            task_id = await self.sora_client.generate_video(
                full_prompt, token_obj.token,
                orientation=model_config["orientation"],
-                n_frames=n_frames
+                n_frames=n_frames,
+                model=sora_model,
+                size=video_size
            )
            debug_logger.log_info(f"Video generation started, task_id: {task_id}")

@@ -1282,7 +1391,7 @@ class GenerationHandler:
            task = Task(
                task_id=task_id,
                token_id=token_obj.id,
-                model=f"sora-video-{model_config['orientation']}",
+                model=f"sora2-video-{model_config['orientation']}",
                prompt=full_prompt,
                status="processing",
                progress=0.0
@@ -1375,7 +1484,7 @@ class GenerationHandler:
            task = Task(
                task_id=task_id,
                token_id=token_obj.id,
-                model=f"sora-video-{model_config['orientation']}",
+                model=f"sora2-video-{model_config['orientation']}",
                prompt=f"remix:{remix_target_id} {clean_prompt}",
                status="processing",
                progress=0.0
--- a/src/services/load_balancer.py
+++ b/src/services/load_balancer.py
@@ -17,13 +17,14 @@ class LoadBalancer:
        # Use image timeout from config as lock timeout
        self.token_lock = TokenLock(lock_timeout=config.image_timeout)

-    async def select_token(self, for_image_generation: bool = False, for_video_generation: bool = False) -> Optional[Token]:
+    async def select_token(self, for_image_generation: bool = False, for_video_generation: bool = False, require_pro: bool = False) -> Optional[Token]:
        """
        Select a token using random load balancing

        Args:
            for_image_generation: If True, only select tokens that are not locked for image generation and have image_enabled=True
            for_video_generation: If True, filter out tokens with Sora2 quota exhausted (sora2_cooldown_until not expired), tokens that don't support Sora2, and tokens with video_enabled=False
+            require_pro: If True, only select tokens with ChatGPT Pro subscription (plan_type="chatgpt_pro")

        Returns:
            Selected token or None if no available tokens
@@ -56,6 +57,13 @@ class LoadBalancer:
        if not active_tokens:
            return None

+        # Filter for Pro tokens if required
+        if require_pro:
+            pro_tokens = [token for token in active_tokens if token.plan_type == "chatgpt_pro"]
+            if not pro_tokens:
+                return None
+            active_tokens = pro_tokens
+
        # If for video generation, filter out tokens with Sora2 quota exhausted and tokens without Sora2 support
        if for_video_generation:
            from datetime import datetime
--- a/src/services/sora_client.py
+++ b/src/services/sora_client.py
@@ -254,8 +254,20 @@ class SoraClient:
        return result["id"]
    
    async def generate_video(self, prompt: str, token: str, orientation: str = "landscape",
-                            media_id: Optional[str] = None, n_frames: int = 450, style_id: Optional[str] = None) -> str:
-        """Generate video (text-to-video or image-to-video)"""
+                            media_id: Optional[str] = None, n_frames: int = 450, style_id: Optional[str] = None,
+                            model: str = "sy_8", size: str = "small") -> str:
+        """Generate video (text-to-video or image-to-video)
+
+        Args:
+            prompt: Video generation prompt
+            token: Access token
+            orientation: Video orientation (landscape/portrait)
+            media_id: Optional image media_id for image-to-video
+            n_frames: Number of frames (300/450/750)
+            style_id: Optional style ID
+            model: Model to use (sy_8 for standard, sy_ore for pro)
+            size: Video size (small for standard, large for HD)
+        """
        inpaint_items = []
        if media_id:
            inpaint_items = [{
@@ -267,9 +279,9 @@ class SoraClient:
            "kind": "video",
            "prompt": prompt,
            "orientation": orientation,
-            "size": "small",
+            "size": size,
            "n_frames": n_frames,
-            "model": "sy_8",
+            "model": model,
            "inpaint_items": inpaint_items,
            "style_id": style_id
        }