# METADATA # { # "name": "openai_draw", # # "display_name": { # "zh": "OpenAI 绘图", # "en": "OpenAI Draw" # }, # "description": { # "zh": "使用 OpenAI 格式的图像生成 API (/v1/images/generations) 根据提示词画图，将图片保存到本地 /sdcard/Download/Operit/plugins/draw/openai_draw/draws/ 目录，并返回 Markdown 图片提示。", # "en": "Generate images via an OpenAI-compatible image generation API (/v1/images/generations) from a prompt, save to /sdcard/Download/Operit/plugins/draw/openai_draw/draws/, and return a Markdown image reference." # }, # "category": "Draw", # "env": [ # { # "name": "OPENAI_API_KEY", # "description": { # "zh": "OpenAI API Key（必填）", # "en": "OpenAI API key (required)" # }, # "required": true # }, # { # "name": "OPENAI_API_BASE_URL", # "description": { # "zh": "OpenAI API Base URL（可选，不填则默认 https://api.openai.com ）", # "en": "OpenAI API base URL (optional; defaults to https://api.openai.com)" # }, # "required": false # }, # { # "name": "OPENAI_IMAGE_MODEL", # "description": { # "zh": "默认绘图模型（可选；当 draw_image 未传 model 时使用）", # "en": "Default image model (optional; used when draw_image doesn't pass model)" # }, # "required": false # } # ], # "tools": [ # { # "name": "draw_image", # "description": { # "zh": "根据提示词调用 OpenAI 格式图像生成接口生成图片，保存到本地并返回 Markdown 图片提示。", # "en": "Generate an image via an OpenAI-compatible image generation endpoint using a prompt, save it locally, and return a Markdown image reference." # }, # "parameters": [ # { "name": "prompt", "description": { "zh": "绘图提示词（英文或中文皆可）", "en": "Prompt for image generation (Chinese or English)" }, "type": "string", "required": true }, # { "name": "model", "description": { "zh": "模型名称（可选；不传则使用环境变量 OPENAI_IMAGE_MODEL，再不行使用默认值）", "en": "Model name (optional; falls back to env OPENAI_IMAGE_MODEL, then default)" }, "type": "string", "required": false }, # { "name": "size", "description": { "zh": "图片尺寸，例如 '1024x1024'，可选", "en": "Image size, e.g. '1024x1024' (optional)" }, "type": "string", "required": false }, # { "name": "file_name", "description": { "zh": "自定义保存到本地的文件名（不含路径和扩展名）", "en": "Custom output file name (without path or extension)" }, "type": "string", "required": false }, # { "name": "api_base_url", "description": { "zh": "OpenAI API Base URL（不传则取环境变量 OPENAI_API_BASE_URL 或默认 https://api.openai.com ）", "en": "OpenAI API base URL (optional; falls back to env OPENAI_API_BASE_URL or https://api.openai.com)" }, "type": "string", "required": false } # ] # } # ] # } import json import re import time import random import traceback HTTP_TIMEOUT_MS = 600000 client = OkHttp.newBuilder() client = client.connectTimeout(HTTP_TIMEOUT_MS) client = client.readTimeout(HTTP_TIMEOUT_MS) client = client.writeTimeout(HTTP_TIMEOUT_MS) client = client.build() DEFAULT_API_BASE_URL = "https://api.openai.com" DEFAULT_MODEL = "gpt-image-1" DRAW_ROOT_DIR = getPluginConfigDir("draw") STORAGE_DIR = f"{DRAW_ROOT_DIR}/openai_draw" DRAWS_DIR = f"{STORAGE_DIR}/draws" def getApiKey(): apiKey = getEnv("OPENAI_API_KEY") if not apiKey: raise Exception("OPENAI_API_KEY 未配置，请在环境变量中设置 OpenAI 的 API Key。") return apiKey def joinUrl(baseUrl, path): normalizedBase = baseUrl if baseUrl.endswith("/") else baseUrl + "/" normalizedPath = path[1:] if path.startswith("/") else path return normalizedBase + normalizedPath def getApiBaseUrl(customBaseUrl=None): fromParam = (customBaseUrl or "").strip() if fromParam: return fromParam fromEnv = (getEnv("OPENAI_API_BASE_URL") or "").strip() if fromEnv: return fromEnv return DEFAULT_API_BASE_URL def getImageEndpoint(baseUrl): trimmed = baseUrl.strip() if not trimmed: return joinUrl(DEFAULT_API_BASE_URL, "v1/images/generations") if "/v1/images/generations" in trimmed: return trimmed # If user passes .../v1, we should append images/generations if trimmed.endswith("/v1") or trimmed.endswith("/v1/"): return joinUrl(trimmed, "images/generations") # Otherwise append v1/images/generations return joinUrl(trimmed, "v1/images/generations") def sanitizeFileName(name): safe = re.sub(r'[\\/:*?"<>|]', '_', name).strip() if not safe: return f"openai_draw_{int(time.time() * 1000)}" return safe[:80] def buildFileName(prompt, customName=None): if customName and len(customName.strip()) > 0: return sanitizeFileName(customName) shortPrompt = (prompt[:40] + "...") if len(prompt) > 40 else prompt base = sanitizeFileName(shortPrompt or "image") timestamp = int(time.time() * 1000) return f"{base}_{timestamp}" async def ensureDirectories(): dirs = [DRAW_ROOT_DIR, STORAGE_DIR, DRAWS_DIR] for dir_ in dirs: try: result = await Tools.Files.mkdir(dir_) if not result.get("successful"): console.warn(f"创建目录失败(可能已存在): {dir_} -> {result.get('details')}") except Exception as e: console.warn(f"创建目录异常: {dir_} -> {getattr(e, 'message', e)}") def normalizeBase64(base64): raw = str(base64 or "").strip() if not raw: return raw # Some providers may return: data:image/png;base64,xxxx prefixIndex = raw.find("base64,") if raw.startswith("data:") and prefixIndex >= 0: return raw[prefixIndex + len("base64,"):].strip() return raw async def callOpenAIImageApi(params): apiKey = getApiKey() apiBaseUrl = getApiBaseUrl(params.get("api_base_url")) endpoint = getImageEndpoint(apiBaseUrl) modelFromParam = (params.get("model") or "").strip() modelFromEnv = (getEnv("OPENAI_IMAGE_MODEL") or "").strip() effectiveModel = modelFromParam or modelFromEnv or DEFAULT_MODEL body = { "model": effectiveModel, "prompt": params["prompt"], "response_format": "b64_json" } if params.get("size") and len(params["size"].strip()) > 0: body["size"] = params["size"].strip() headers = { "accept": "application/json", "content-type": "application/json", "Authorization": f"Bearer {apiKey}" } request = client.newRequest() request = request.url(endpoint) request = request.method("POST") request = request.headers(headers) request = request.body(json.dumps(body), "json") response = await request.build().execute() if not response.isSuccessful(): raise Exception(f"OpenAI 图片 API 调用失败: {response.statusCode} - {response.content}") try: parsed = json.loads(response.content) except Exception as e: msg = str(getattr(e, "message", e)) raise Exception(f"解析 OpenAI 响应失败: {msg}") data = parsed.get("data") if isinstance(parsed, dict) else None item = data[0] if (data and len(data) > 0) else None if not item: raise Exception("OpenAI 响应中未找到 data[0]，请检查接口返回格式。") if item.get("b64_json") and len(str(item["b64_json"]).strip()) > 0: return { "b64_json": str(item["b64_json"]), "revised_prompt": str(item["revised_prompt"]) if item.get("revised_prompt") else None, "effective_model": effectiveModel } if item.get("url") and len(str(item["url"]).strip()) > 0: # Some OpenAI-compatible APIs may return url only. We download it, then read as base64. tmpName = f"openai_tmp_{int(time.time() * 1000)}" tmpPath = f"{DRAWS_DIR}/{tmpName}.png" downloadResult = await Tools.Files.download(str(item["url"]), tmpPath) if not downloadResult.get("successful"): raise Exception(f"下载图片失败: {downloadResult.get('details')}") readBinary = await Tools.Files.readBinary(tmpPath) contentBase64 = readBinary.get("contentBase64") if isinstance(readBinary, dict) else None if not contentBase64 or len(str(contentBase64).strip()) == 0: raise Exception("读取下载图片失败: contentBase64 为空") try: await Tools.Files.deleteFile(tmpPath) except Exception: # ignore pass return { "b64_json": str(contentBase64), "revised_prompt": str(item["revised_prompt"]) if item.get("revised_prompt") else None, "effective_model": effectiveModel } raise Exception("OpenAI 响应中未找到 b64_json 或 url，请检查模型/参数以及接口兼容性。") async def draw_image(params): if not params or not params.get("prompt") or len(params.get("prompt", "").strip()) == 0: raise Exception("参数 prompt 不能为空。") prompt = params["prompt"].strip() await ensureDirectories() apiResult = await callOpenAIImageApi({ "prompt": prompt, "model": params.get("model"), "size": params.get("size"), "api_base_url": params.get("api_base_url") }) baseName = buildFileName(prompt, params.get("file_name")) filePath = f"{DRAWS_DIR}/{baseName}.png" writeResult = await Tools.Files.writeBinary(filePath, normalizeBase64(apiResult["b64_json"])) if not writeResult.get("successful"): raise Exception(f"保存图片失败: {writeResult.get('details')}") fileUri = f"file://{filePath}" markdown = f"![AI生成的图片]({fileUri})" hintLines = [] hintLines.append(f"图片已生成并保存在本地 {DRAWS_DIR}。") hintLines.append(f"本地路径: {filePath}") hintLines.append("") hintLines.append("在后续回答中，请直接输出下面这一行 Markdown 来展示这张图片：") hintLines.append("") hintLines.append(markdown) return { "file_path": filePath, "file_uri": fileUri, "markdown": markdown, "prompt": prompt, "revised_prompt": apiResult.get("revised_prompt") or None, "model": apiResult["effective_model"], "hint": "\n".join(hintLines) } async def draw_image_wrapper(params): try: result = await draw_image(params) complete({ "success": True, "message": f"图片生成成功，已保存到 {DRAWS_DIR}，并返回 Markdown 图片提示。", "data": result }) except Exception as error: console.error("draw_image 执行失败:", error) msg = str(getattr(error, "message", error)) stack = traceback.format_exc() if isinstance(error, Exception) else None complete({ "success": False, "message": f"图片生成失败: {msg}", "error_stack": stack }) exports.draw_image = draw_image_wrapper