feat: 新增 GPT-1.5 生图节点

__init__.py

@@ -21,7 +21,7 @@ except Exception:
 
 import ssl
 
-from .nodes import NanoBananaPro, BatchNanoBananaPro, GoogleGemini, LoadFile, ImageStitchPro, SaveCleanImage, BatchCleanMetadata, VideoPreview, GoogleVeo, FluxImageEdit, UniversalLLMChat, KlingVideo, KlingFirstLastFrame, KlingMotionControlTest, AspectRatioPreset, MultiResPreview, BatchImagesO1key, Seedance, SeedanceMultiModal, StreamPreview, DoubaoImage
+from .nodes import NanoBananaPro, BatchNanoBananaPro, GoogleGemini, LoadFile, ImageStitchPro, SaveCleanImage, BatchCleanMetadata, VideoPreview, GoogleVeo, FluxImageEdit, UniversalLLMChat, KlingVideo, KlingFirstLastFrame, KlingMotionControlTest, AspectRatioPreset, MultiResPreview, BatchImagesO1key, Seedance, SeedanceMultiModal, StreamPreview, DoubaoImage, O1keyGPTImage
 
 # 报错弹框友好文案（不修改原节点代码，仅在外层统一处理）
 _MSG_TIMEOUT = "API 请求超时，请稍后重试或检查网络。"
@@ -79,6 +79,7 @@ NODE_CLASS_MAPPINGS = {
     "SeedanceMultiModal": SeedanceMultiModal,
     "StreamPreview": StreamPreview,
     "DoubaoImage": DoubaoImage,
+    "O1keyGPTImage": O1keyGPTImage,
 }
 
 NODE_DISPLAY_NAME_MAPPINGS = {
@@ -103,6 +104,7 @@ NODE_DISPLAY_NAME_MAPPINGS = {
     "SeedanceMultiModal": "Seedance 多模态参考生视频",
     "StreamPreview": "流式文本预览",
     "DoubaoImage": "豆包生图",
+    "O1keyGPTImage": "o1key GPT Image",
 }
 
 WEB_DIRECTORY = "./web"

clients/gpt_image_client.py

@@ -0,0 +1,411 @@
+"""
+GPT Image API 客户端
+支持两个接口：
+  - POST /v1/images/generations/  文生图 / 图生图（gpt-image-1 / gpt-image-1.5）
+  - POST /v1/images/edits/        图像编辑（带蒙版 inpainting）
+
+设计原则：
+  - 与 doubao_image_client.py 保持相同的异步 + 同步双入口模式
+  - 图像以 multipart/form-data 方式上传（edits 接口）
+  - generations 接口使用 JSON 请求体，图像以 data URI base64 内联传递
+  - 响应支持 url 和 b64_json 两种格式，优先处理 b64_json（避免二次下载）
+"""
+
+import asyncio
+import base64
+import json
+import time
+from concurrent.futures import ThreadPoolExecutor
+from io import BytesIO
+from typing import List, Optional
+
+import aiohttp
+import numpy as np
+import torch
+from PIL import Image
+
+from ..utils.config import get_api_key_or_raise, get_api_base_url
+from ..utils.image_utils import tensor_to_pil, encode_image_to_base64
+
+# ── 接口端点 ──────────────────────────────────────────────────────────────────
+_ENDPOINT_GENERATIONS = "/v1/images/generations/"
+_ENDPOINT_EDITS       = "/v1/images/edits/"
+
+# ── 超时 ──────────────────────────────────────────────────────────────────────
+_REQUEST_TIMEOUT = 300   # 秒
+
+
+class GptImageClient:
+    """
+    GPT Image API 客户端
+
+    接口说明：
+      generations：JSON body，支持 background / quality / size / n / model
+      edits：multipart/form-data，必须包含 image（PNG），可选 mask（PNG）
+
+    两个接口的响应格式相同：
+      { "data": [ {"url": "..."} | {"b64_json": "..."} ] }
+    """
+
+    def __init__(self):
+        self.api_key  = get_api_key_or_raise("O1KEY_API_KEY")
+        self.base_url = get_api_base_url()
+
+    # ── 认证头 ────────────────────────────────────────────────────────────────
+
+    def _auth_headers(self) -> dict:
+        return {"Authorization": f"Bearer {self.api_key}"}
+
+    def _json_headers(self) -> dict:
+        return {
+            "Authorization": f"Bearer {self.api_key}",
+            "Content-Type": "application/json",
+        }
+
+    # ── 图像转换工具 ──────────────────────────────────────────────────────────
+
+    @staticmethod
+    def _tensor_to_png_bytes(tensor: torch.Tensor) -> bytes:
+        """
+        单张 ComfyUI IMAGE tensor [1, H, W, C] 或 [H, W, C] → PNG bytes
+        """
+        if tensor.dim() == 4:
+            tensor = tensor.squeeze(0)              # [H, W, C]
+        arr = (tensor.cpu().numpy() * 255).clip(0, 255).astype(np.uint8)
+        img = Image.fromarray(arr)
+        buf = BytesIO()
+        img.save(buf, format="PNG")
+        return buf.getvalue()
+
+    @staticmethod
+    def _mask_tensor_to_rgba_png_bytes(mask: torch.Tensor, image_size: tuple) -> bytes:
+        """
+        ComfyUI MASK tensor [1, H, W] 或 [H, W] → RGBA PNG bytes
+        白色区域（mask=1）→ 透明（alpha=0），即 API 将在此处生成新内容。
+        """
+        if mask.dim() == 3:
+            mask = mask.squeeze(0)                  # [H, W]
+
+        h, w = mask.shape
+        ih, iw = image_size
+
+        # 尺寸不一致时给出提示（API 侧也会报错）
+        if (h, w) != (ih, iw):
+            raise ValueError(
+                f"蒙版尺寸 ({h}×{w}) 与图像尺寸 ({ih}×{iw}) 不一致，请保持相同尺寸"
+            )
+
+        alpha = ((1.0 - mask.cpu().numpy()) * 255).clip(0, 255).astype(np.uint8)
+        rgba  = np.zeros((h, w, 4), dtype=np.uint8)
+        rgba[:, :, 3] = alpha                       # 只设 alpha，RGB 全 0
+
+        buf = BytesIO()
+        Image.fromarray(rgba, mode="RGBA").save(buf, format="PNG")
+        return buf.getvalue()
+
+    @staticmethod
+    def _pil_list_to_tensor(images: List[Image.Image]) -> torch.Tensor:
+        """
+        PIL Image 列表 → ComfyUI IMAGE tensor [B, H, W, C]，值域 [0, 1]
+        RGBA 自动转换为 RGBA（保留透明通道）
+        """
+        if not images:
+            placeholder = Image.new("RGBA", (512, 512), (128, 128, 128, 255))
+            images = [placeholder]
+
+        tensors = []
+        for img in images:
+            arr = np.array(img.convert("RGBA")).astype(np.float32) / 255.0
+            tensors.append(torch.from_numpy(arr))
+
+        return torch.stack(tensors, dim=0)          # [B, H, W, 4]
+
+    # ── 响应解析（通用） ─────────────────────────────────────────────────────
+
+    async def _parse_response(
+        self,
+        resp_json: dict,
+        session: aiohttp.ClientSession,
+    ) -> List[Image.Image]:
+        """
+        解析 data 列表，优先取 b64_json，回退到 url 下载
+        """
+        if "error" in resp_json:
+            err = resp_json["error"]
+            msg = (
+                err.get("message") or err.get("msg") or json.dumps(err, ensure_ascii=False)
+                if isinstance(err, dict)
+                else str(err)
+            )
+            raise RuntimeError(f"API 返回错误: {msg}")
+
+        data_list = resp_json.get("data")
+        if not data_list:
+            raise RuntimeError(
+                f"API 响应中未找到 data 字段，完整响应：\n"
+                f"{json.dumps(resp_json, ensure_ascii=False, indent=2)}"
+            )
+
+        images: List[Image.Image] = []
+        for idx, item in enumerate(data_list):
+            b64 = item.get("b64_json", "")
+            url = item.get("url", "")
+
+            if b64:
+                # 优先 base64（无需二次下载）
+                try:
+                    img_bytes = base64.b64decode(b64)
+                    img = Image.open(BytesIO(img_bytes))
+                    images.append(img)
+                    print(f"[o1key GPT Image] 第 {idx + 1} 张 base64 解码完成 "
+                          f"({img.size[0]}×{img.size[1]})")
+                except Exception as e:
+                    raise RuntimeError(f"第 {idx + 1} 张 base64 解码失败: {e}")
+
+            elif url and url.startswith("http"):
+                # 回退：下载 URL
+                async with session.get(url, allow_redirects=True) as r:
+                    if r.status != 200:
+                        raise RuntimeError(
+                            f"图像下载失败 HTTP {r.status}，URL: {url}"
+                        )
+                    img_bytes = await r.read()
+                img = Image.open(BytesIO(img_bytes))
+                images.append(img)
+                print(f"[o1key GPT Image] 第 {idx + 1} 张下载完成 "
+                      f"({img.size[0]}×{img.size[1]})")
+            else:
+                print(f"[o1key GPT Image] 警告：第 {idx + 1} 条数据既无 b64_json 也无 url，已跳过")
+
+        return images
+
+    # ── 文生图 / 图生图（generations 接口）───────────────────────────────────
+
+    async def _generate_async(
+        self,
+        prompt: str,
+        model: str,
+        quality: str,
+        background: str,
+        size: str,
+        n: int,
+        seed: int,
+        image_tensor: Optional[torch.Tensor] = None,
+    ) -> List[Image.Image]:
+        """
+        调用 /v1/images/generations/ 接口。
+        当传入 image_tensor 时，以 data URI 格式内联图像（图生图）。
+        """
+        body: dict = {
+            "model":      model,
+            "prompt":     prompt,
+            "quality":    quality,
+            "background": background,
+            "n":          n,
+            "moderation": "low",
+        }
+
+        # size = "auto" 时不传该字段，让 API 自行决定
+        if size and size != "auto":
+            body["size"] = size
+
+        # seed > 0 时才传递（0 视为不指定）
+        if seed > 0:
+            body["seed"] = seed
+
+        # 图生图：将 tensor 转成 data URI 内联
+        if image_tensor is not None:
+            pil_images = tensor_to_pil(image_tensor)
+            data_urls = []
+            for img in pil_images:
+                b64 = encode_image_to_base64(img, format="PNG")
+                data_urls.append(f"data:image/png;base64,{b64}")
+            body["image"] = data_urls[0] if len(data_urls) == 1 else data_urls
+            mode = f"图生图（参考图 {len(data_urls)} 张）"
+        else:
+            mode = "文生图"
+
+        url = f"{self.base_url}{_ENDPOINT_GENERATIONS}"
+        print(f"[o1key GPT Image] {mode} | 模型={model} | quality={quality} | "
+              f"background={background} | size={size} | n={n}")
+
+        connector = aiohttp.TCPConnector(ssl=False, force_close=True)
+        timeout   = aiohttp.ClientTimeout(total=_REQUEST_TIMEOUT)
+
+        async with aiohttp.ClientSession(connector=connector, timeout=timeout) as session:
+            t0 = time.time()
+            async with session.post(url, json=body, headers=self._json_headers()) as resp:
+                elapsed = time.time() - t0
+                text = await resp.text()
+
+                if resp.status != 200:
+                    try:
+                        err_json = json.loads(text)
+                        err_obj  = err_json.get("error", {})
+                        msg = (
+                            err_obj.get("message") or err_obj.get("msg") or text
+                            if isinstance(err_obj, dict)
+                            else str(err_obj) or text
+                        )
+                    except Exception:
+                        msg = text
+                    raise RuntimeError(f"请求失败 HTTP {resp.status}: {msg}")
+
+                try:
+                    resp_json = json.loads(text)
+                except Exception:
+                    raise RuntimeError(f"响应 JSON 解析失败，原始内容：{text[:500]}")
+
+            print(f"[o1key GPT Image] API 响应耗时 {elapsed:.1f}s")
+            return await self._parse_response(resp_json, session)
+
+    # ── 图像编辑（edits 接口，必须带蒙版）───────────────────────────────────
+
+    async def _edit_async(
+        self,
+        prompt: str,
+        model: str,
+        quality: str,
+        background: str,
+        size: str,
+        n: int,
+        seed: int,
+        image_tensor: torch.Tensor,
+        mask_tensor: Optional[torch.Tensor] = None,
+    ) -> List[Image.Image]:
+        """
+        调用 /v1/images/edits/ 接口（multipart/form-data）。
+        image_tensor 取第一帧（edits 接口仅支持单张输入图）。
+        """
+        # 取第一张图
+        single = image_tensor[:1] if image_tensor.dim() == 4 else image_tensor.unsqueeze(0)
+        image_png = self._tensor_to_png_bytes(single)
+        ih, iw = single.shape[1], single.shape[2]
+
+        # 构建 multipart 表单
+        form = aiohttp.FormData()
+        form.add_field("model",      model)
+        form.add_field("prompt",     prompt)
+        form.add_field("quality",    quality)
+        form.add_field("background", background)
+        form.add_field("n",          str(n))
+        form.add_field("moderation", "low")
+
+        if size and size != "auto":
+            form.add_field("size", size)
+
+        if seed > 0:
+            form.add_field("seed", str(seed))
+
+        # 主图（PNG）
+        form.add_field(
+            "image",
+            image_png,
+            filename="image.png",
+            content_type="image/png",
+        )
+
+        # 蒙版（PNG，RGBA 格式，透明区域为待编辑区）
+        if mask_tensor is not None:
+            mask_png = self._mask_tensor_to_rgba_png_bytes(mask_tensor, (ih, iw))
+            form.add_field(
+                "mask",
+                mask_png,
+                filename="mask.png",
+                content_type="image/png",
+            )
+            mode = "图像编辑（带蒙版）"
+        else:
+            mode = "图像编辑（无蒙版）"
+
+        url = f"{self.base_url}{_ENDPOINT_EDITS}"
+        print(f"[o1key GPT Image] {mode} | 模型={model} | quality={quality} | "
+              f"background={background} | size={size} | n={n}")
+
+        connector = aiohttp.TCPConnector(ssl=False, force_close=True)
+        timeout   = aiohttp.ClientTimeout(total=_REQUEST_TIMEOUT)
+
+        async with aiohttp.ClientSession(connector=connector, timeout=timeout) as session:
+            t0 = time.time()
+            async with session.post(
+                url,
+                data=form,
+                headers=self._auth_headers(),   # Content-Type 由 FormData 自动设置
+            ) as resp:
+                elapsed = time.time() - t0
+                text = await resp.text()
+
+                if resp.status != 200:
+                    try:
+                        err_json = json.loads(text)
+                        err_obj  = err_json.get("error", {})
+                        msg = (
+                            err_obj.get("message") or err_obj.get("msg") or text
+                            if isinstance(err_obj, dict)
+                            else str(err_obj) or text
+                        )
+                    except Exception:
+                        msg = text
+                    raise RuntimeError(f"请求失败 HTTP {resp.status}: {msg}")
+
+                try:
+                    resp_json = json.loads(text)
+                except Exception:
+                    raise RuntimeError(f"响应 JSON 解析失败，原始内容：{text[:500]}")
+
+            print(f"[o1key GPT Image] API 响应耗时 {elapsed:.1f}s")
+            return await self._parse_response(resp_json, session)
+
+    # ── 同步统一入口（供节点调用）────────────────────────────────────────────
+
+    def run_sync(
+        self,
+        prompt: str,
+        model: str,
+        quality: str,
+        background: str,
+        size: str,
+        n: int,
+        seed: int,
+        image_tensor: Optional[torch.Tensor] = None,
+        mask_tensor: Optional[torch.Tensor] = None,
+    ) -> List[Image.Image]:
+        """
+        同步入口，在独立线程中运行事件循环，避免与 ComfyUI 主循环冲突。
+
+        路由逻辑：
+          - 无 image_tensor           → generations 接口（文生图）
+          - 有 image_tensor，无 mask  → generations 接口（图生图，data URI）
+          - 有 image_tensor，有 mask  → edits 接口（图像编辑 + 蒙版）
+        """
+        use_edits = (image_tensor is not None and mask_tensor is not None)
+
+        if use_edits:
+            coro = self._edit_async(
+                prompt=prompt, model=model, quality=quality,
+                background=background, size=size, n=n, seed=seed,
+                image_tensor=image_tensor, mask_tensor=mask_tensor,
+            )
+        else:
+            coro = self._generate_async(
+                prompt=prompt, model=model, quality=quality,
+                background=background, size=size, n=n, seed=seed,
+                image_tensor=image_tensor,
+            )
+
+        def _run():
+            loop = asyncio.new_event_loop()
+            asyncio.set_event_loop(loop)
+            try:
+                return loop.run_until_complete(coro)
+            finally:
+                loop.close()
+
+        with ThreadPoolExecutor(max_workers=1) as executor:
+            future = executor.submit(_run)
+            try:
+                return future.result(timeout=_REQUEST_TIMEOUT + 30)
+            except TimeoutError:
+                raise RuntimeError(
+                    f"o1key GPT Image 请求超时（>{_REQUEST_TIMEOUT}s），请检查网络或稍后重试"
+                )

nodes/__init__.py

@@ -19,5 +19,6 @@ from .multi_res_preview import MultiResPreview
 from .batch_images_o1key import BatchImagesO1key
 from .seedance_video import Seedance, SeedanceMultiModal
 from .doubao_image import DoubaoImage
+from .o1key_gpt_image import O1keyGPTImage
 
-__all__ = ['NanoBananaPro', 'BatchNanoBananaPro', 'GoogleGemini', 'LoadFile', 'ImageStitchPro', 'SaveCleanImage', 'BatchCleanMetadata', 'VideoPreview', 'KlingVideo', 'KlingFirstLastFrame', 'KlingMotionControlTest', 'AspectRatioPreset', 'GoogleVeo', 'FluxImageEdit', 'UniversalLLMChat', 'MultiResPreview', 'BatchImagesO1key', 'Seedance', 'SeedanceMultiModal', 'StreamPreview', 'DoubaoImage']
+__all__ = ['NanoBananaPro', 'BatchNanoBananaPro', 'GoogleGemini', 'LoadFile', 'ImageStitchPro', 'SaveCleanImage', 'BatchCleanMetadata', 'VideoPreview', 'KlingVideo', 'KlingFirstLastFrame', 'KlingMotionControlTest', 'AspectRatioPreset', 'GoogleVeo', 'FluxImageEdit', 'UniversalLLMChat', 'MultiResPreview', 'BatchImagesO1key', 'Seedance', 'SeedanceMultiModal', 'StreamPreview', 'DoubaoImage', 'O1keyGPTImage']

nodes/o1key_gpt_image.py

@@ -0,0 +1,158 @@
+"""
+o1key GPT Image 节点
+支持 gpt-image-1 / gpt-image-1.5 模型的文生图、图生图、图像编辑（带蒙版）
+"""
+
+import time
+import torch
+
+from ..clients.gpt_image_client import GptImageClient
+
+
+class O1keyGPTImage:
+    """
+    o1key GPT Image 节点
+
+    功能：
+      - 文生图：仅提供 prompt
+      - 图生图：提供 prompt + image（无 mask）
+      - 图像编辑：提供 prompt + image + mask（白色区域将被替换）
+
+    参数：
+      - prompt      : 文本提示词（多行）
+      - seed        : 随机种子（0 表示不指定）
+      - quality     : 图像质量 low / medium / high
+      - background  : 背景模式 auto / opaque / transparent
+      - size        : 图像尺寸（auto 让 API 自动决定）
+      - n           : 生成数量 1-8
+      - image       : 可选参考图（用于图生图或编辑）
+      - mask        : 可选蒙版（白色区域将被替换）
+      - model       : 模型选择 gpt-image-1 / gpt-image-1.5
+    """
+
+    @classmethod
+    def INPUT_TYPES(cls):
+        return {
+            "required": {
+                "prompt": ("STRING", {
+                    "default": "",
+                    "multiline": True,
+                    "tooltip": "Text prompt for GPT Image",
+                }),
+            },
+            "optional": {
+                "seed": ("INT", {
+                    "default": 0,
+                    "min": 0,
+                    "max": 2**31 - 1,
+                    "step": 1,
+                    "display": "number",
+                    "control_after_generate": True,
+                    "tooltip": "Random seed (0 = not specified)",
+                }),
+                "quality": (["low", "medium", "high"], {
+                    "default": "low",
+                    "tooltip": "Image quality, affects cost and generation time.",
+                }),
+                "background": (["auto", "opaque", "transparent"], {
+                    "default": "auto",
+                    "tooltip": "Return image with or without background",
+                }),
+                "size": (["auto", "1024x1024", "1024x1536", "1536x1024"], {
+                    "default": "auto",
+                    "tooltip": "Image size (auto = API decides)",
+                }),
+                "n": ("INT", {
+                    "default": 1,
+                    "min": 1,
+                    "max": 8,
+                    "step": 1,
+                    "display": "number",
+                    "tooltip": "How many images to generate",
+                }),
+                "image": ("IMAGE", {
+                    "tooltip": "Optional reference image for image editing.",
+                }),
+                "mask": ("MASK", {
+                    "tooltip": "Optional mask for inpainting (white areas will be replaced)",
+                }),
+                "model": (["gpt-image-1", "gpt-image-1.5"], {
+                    "default": "gpt-image-1.5",
+                }),
+            },
+        }
+
+    RETURN_TYPES = ("IMAGE",)
+    RETURN_NAMES = ("IMAGE",)
+    FUNCTION = "generate"
+    CATEGORY = "o1key/image"
+    OUTPUT_NODE = False
+
+    def generate(
+        self,
+        prompt: str,
+        seed: int = 0,
+        quality: str = "low",
+        background: str = "auto",
+        size: str = "auto",
+        n: int = 1,
+        image=None,
+        mask=None,
+        model: str = "gpt-image-1.5",
+    ):
+        """
+        生成图像（文生图 / 图生图 / 图像编辑）
+
+        路由逻辑：
+          - 无 image           → generations 接口（文生图）
+          - 有 image，无 mask  → generations 接口（图生图）
+          - 有 image，有 mask  → edits 接口（图像编辑 + 蒙版）
+        """
+        start_time = time.time()
+
+        # ── 1. 参数校验 ───────────────────────────────────────────────────────
+        if not prompt or not prompt.strip():
+            raise ValueError("提示词不能为空")
+
+        if mask is not None and image is None:
+            raise ValueError("提供了蒙版但未提供图像，请同时提供 image 和 mask")
+
+        # ── 2. 创建客户端 ─────────────────────────────────────────────────────
+        try:
+            client = GptImageClient()
+        except ValueError as e:
+            if str(e) == "未授权！":
+                print("[o1key GPT Image] 请联系作者授权后方可使用！")
+                raise ValueError("未授权！") from None
+            raise
+
+        # ── 3. 调用 API ───────────────────────────────────────────────────────
+        try:
+            pil_images = client.run_sync(
+                prompt=prompt,
+                model=model,
+                quality=quality,
+                background=background,
+                size=size,
+                n=n,
+                seed=seed,
+                image_tensor=image,
+                mask_tensor=mask,
+            )
+        except Exception as e:
+            error_msg = str(e).split('\n')[0]
+            print(f"[o1key GPT Image] ❌ {error_msg}")
+            raise RuntimeError(error_msg) from None
+
+        # ── 4. PIL → tensor ───────────────────────────────────────────────────
+        output_tensor = GptImageClient._pil_list_to_tensor(pil_images)
+
+        # ── 5. 完成日志 ───────────────────────────────────────────────────────
+        elapsed = time.time() - start_time
+        print(
+            f"[o1key GPT Image] 完成！耗时 {elapsed:.1f}s，"
+            f"输出 {output_tensor.shape[0]} 张 "
+            f"{output_tensor.shape[2]}×{output_tensor.shape[1]}"
+        )
+
+        return (output_tensor,)