diff --git a/.trae/skills/image-tools/SKILL.md b/.trae/skills/image-tools/SKILL.md index d5d6de4..1c8944b 100644 --- a/.trae/skills/image-tools/SKILL.md +++ b/.trae/skills/image-tools/SKILL.md @@ -1,7 +1,6 @@ - --- -## 引擎1：豆包万相（Doubao Seedream） +## 豆包 Seedream（字节跳动/火山引擎） 使用火山引擎 Ark API 的 Doubao Seedream 模型，支持文生图、图生图、多图融合、图片编辑。 @@ -16,12 +15,20 @@ | 模式 | `image`参数 | 用途 | |---|---|---| | 文生图 | 省略或`[]` | 从文字描述生成图片 | -| 图生图 | `["url1"]` | 风格转换、精修、变换 | -| 多图融合 | `["url1","url2",...]` | 最多14张参考图融合 | +| 图生图 | `["url1"]` | 基于参考图生成新图，可通过prompt控制保留原图产品外观 | +| 多图融合 | `["url1","url2",...]` | 最多10张参考图融合，prompt中用"图1/图2"指代 | | 图片编辑 | `["url1"]` + 编辑提示词 | 移除/替换元素、换背景等 | 默认模型: `doubao-seedream-5-0-260128` +### 图生图说明 + +Seedream 的 `image` 参数传入的是**参考图**，模型会根据 prompt 重新生成图片，**不能100%像素级保留原图**。但通过精确的 prompt 可在一定程度上保持产品形态： +- 好的做法：`"保留图1中U盘产品的外观和形态不变，仅更换背景为白色，添加科技光效"` +- 不好的做法：`"U盘插在电脑上"`（描述太笼统，产品外观会偏离原图） + + + ### 文生图（Windows PowerShell） ```powershell @@ -51,12 +58,18 @@ $imgBytes = [System.IO.File]::ReadAllBytes((Resolve-Path "path\to\input.png")) $imgBase64 = [System.Convert]::ToBase64String($imgBytes) $tmpFile = "$env:TEMP\request.json" -$prompt = "YOUR_EDIT_INSTRUCTIONS_HERE" -$json = @" -{"model":"doubao-seedream-5-0-260128","prompt":"$prompt","image":["data:image/png;base64,$imgBase64"],"size":"2k","sequential_image_generation":"disabled","response_format":"url","stream":false,"watermark":false} -"@ $sw = [System.IO.StreamWriter]::new($tmpFile, $false, (New-Object System.Text.UTF8Encoding $false)) -$sw.Write($json); $sw.Close() +$sw.WriteLine('{') +$sw.WriteLine(' "model": "doubao-seedream-5-0-260128",') +$sw.WriteLine(' "prompt": "保留图1中产品的外观和形态不变，更换背景为白色",') +$sw.WriteLine(' "image": ["data:image/png;base64,' + $imgBase64 + '"],') +$sw.WriteLine(' "size": "2048x2048",') +$sw.WriteLine(' "sequential_image_generation": "disabled",') +$sw.WriteLine(' "response_format": "url",') +$sw.WriteLine(' "stream": false,') +$sw.WriteLine(' "watermark": false') +$sw.WriteLine('}') +$sw.Close() $response = curl.exe -s -X POST https://ark.cn-beijing.volces.com/api/v3/images/generations ` -H "Content-Type: application/json" ` @@ -69,6 +82,8 @@ curl.exe -s -o $outputPath $imageUrl Write-Output $outputPath ``` +> **注意**：当 prompt 中包含中文引号等特殊字符时，不要用 PowerShell 的 `@"..."@` here-string 拼接 JSON，应改用 `StreamWriter.WriteLine()` 逐行写入，避免解析错误。 + ### 多图融合 ```json @@ -99,71 +114,70 @@ Write-Output $outputPath } ``` -### 豆包参数参考 +### Seedream 参数参考 | 参数 | 类型 | 必填 | 说明 | |---|---|---|---| -| `model` | string | 是 | `doubao-seedream-5-0-260128`（默认）或 `doubao-seededit-3-0-i2i`（精确编辑） | +| `model` | string | 是 | `doubao-seedream-5-0-260128`（默认，5.0）、`doubao-seedream-4-5-251128`（4.5）、`doubao-seedream-4-0-250828`（4.0） | | `prompt` | string | 是 | 文本提示词，最多~300中文字或~600英文词 | -| `image` | string[] | 否 | 参考图片URL或Base64（`data:image/png;base64,...`），最多14张 | -| `size` | string | 否 | `2k`（默认）、`3k`、`4k` 或 `WIDTHxHEIGHT`（如`2048x1024`），必须小写 | +| `image` | string[] | 否 | 参考图片URL或Base64（`data:image/jpeg;base64,...`），最多10张；prompt中用"图1/图2"指代顺序 | +| `size` | string | 否 | `2k`（默认）、`3k`（仅5.0）、`4k`（4.5/4.0） 或 `WIDTHxHEIGHT`（如`2048x2048`） | | `response_format` | string | 否 | `url`（下载链接，24h有效）或 `b64_json` | +| `output_format` | string | 否 | `jpeg`（默认）或 `png`（仅5.0支持） | | `sequential_image_generation` | string | 否 | `disabled`（默认，单图）或 `auto`（模型决定批量数） | | `sequential_image_generation_options` | object | 否 | `{ "max_images": N }` N为1-15，仅auto模式 | | `stream` | boolean | 否 | `false`（等待全部）或 `true`（逐张返回） | -| `watermark` | boolean | 否 | 添加"AI Generated"水印，默认`false` | -| `guidance_scale` | float | 否 | 文本遵循度(CFG)，范围[1,10]，仅seededit-3-0-i2i | -| `seed` | int32 | 否 | 可复现种子，仅seededit-3-0-i2i | +| `watermark` | boolean | 否 | 添加"AI Generated"水印，默认`true`（建议设`false`） | +| `tools` | array | 否 | `[{ "type": "web_search" }]` 联网搜索（仅5.0） | -### 豆包尺寸指南 +### Seedream 尺寸指南 | 别名 | 近似分辨率 | 适用场景 | |---|---|---| | `2k` | ~2048px | **默认** — 质量与速度平衡 | -| `3k` | ~3072px | 高质量，产品图 | -| `4k` | ~4096px | 最终生产质量 | -| `WIDTHxHEIGHT` | 自定义 | 如`2048x1024`宽幅横幅 | +| `3k` | ~3072px | 高质量，产品图（仅5.0） | +| `4k` | ~4096px | 最终生产质量（4.5/4.0） | +| `WIDTHxHEIGHT` | 自定义 | 如`2048x2048`、`2048x1024` | -像素约束：两个维度必须在[1280, 4096]，最大3:1宽高比。 +**像素约束**： +- 方式1（别名）：用`2k`/`3k`等，prompt中描述宽高比 +- 方式2（精确像素）：总像素范围 [3686400, 16777216]，宽高比 [1/16, 16]，每个维度 [1280, 4096] +- 两种方式不可混用 -### 豆包提示词技巧 +### Seedream 提示词技巧 - **描述场景**: 主体、动作、环境、氛围 - **指定风格**: "电影大片感"、"水彩画风"、"赛博朋克"、"oc渲染"、"超现实主义" - **质量关键词**: "光线追踪"、"景深"、"极致的光影"、"动态模糊" -- **编辑指令**: "把背景换成海滩"、"移除画面中的人物"、"将风格改为水彩画" +- **编辑指令**: "保留图1中产品外观不变，把背景换成海滩"、"移除画面中的人物" +- **多图指代**: 多图输入时用"图1""图2"明确指代，如"保留图1的U盘，放到图2的笔记本上" - **中文提示词原生支持** — 无需翻译为英文 +### 文字渲染技巧（重要） + +Seedream 对文字渲染有特殊优化，但需要正确的提示词写法： + +- **必须用双引号包裹文字内容**：Seedream 对双引号内的文字有特殊渲染优先级，准确率远高于不加引号 + - ✅ 正确：`标题文字"NAS U盘"`，`标签显示"自动内网穿透"` + - ❌ 错误：`标题文字NAS U盘`，`标签显示自动内网穿透` +- **强制清晰度约束**：加"笔画清晰醒目、无乱码无模糊"等约束条件 +- **指定字体样式**：明确"黑色加粗印刷体"、"白色大号粗体"等，避免模糊描述 +- **避免过多小字**：小字号文字像素占比太低，极易模糊出错；尽量只生成少量大号标题文字 +- **商用场景建议**：对文字精度要求高时，采用"AI生成无文字底图 + 手动/程序叠加文字"的工作流 + ### Windows注意事项 | 问题 | 解决方案 | |---|---| | PowerShell 5.1给JSON加UTF-8 BOM | **不要**用`Set-Content`或`[System.Text.Encoding]::UTF8`，用`[System.IO.StreamWriter]::new($path, $false, (New-Object System.Text.UTF8Encoding $false))` | | PowerShell破坏curl中的`"` | **始终**写JSON到临时文件，用`-d "@$tmpFile"` | +| prompt含中文引号/特殊字符 | 用`StreamWriter.WriteLine()`逐行写JSON，不要用`@"..."@` here-string | | curl别名冲突 | 用`curl.exe`而非`curl` | | 续行符 | 用反引号`` ` `` | -### 豆包错误处理 +### Seedream 错误处理 - **401**: API Key无效 — 检查Authorization头 -- **400**: 请求错误 — 检查提示词长度、图片格式、尺寸值 +- **400**: 请求错误 — 检查提示词长度、图片格式、尺寸值（像素不低于3686400） - **429**: 限流 — 等待后重试 - **500**: 服务器错误 — 稍后重试 - -### 精确编辑模型 - -需要精细编辑时，`doubao-seededit-3-0-i2i`提供额外控制： - -```json -{ - "model": "doubao-seededit-3-0-i2i", - "prompt": "Change the dress color to red", - "image": ["data:image/png;base64,..."], - "guidance_scale": 5.5, - "seed": 42, - "size": "adaptive", - "response_format": "url", - "watermark": false -} -``` - diff --git a/.trae/skills/image-tools/references/doubao-api-reference.md b/.trae/skills/image-tools/references/doubao-api-reference.md index f03f7d5..47e7e41 100644 --- a/.trae/skills/image-tools/references/doubao-api-reference.md +++ b/.trae/skills/image-tools/references/doubao-api-reference.md @@ -1,18 +1,21 @@ -# Volcengine Ark Image API Reference +# 火山引擎 Ark 图像生成 API 参考 -## Endpoint +> 本文档描述字节跳动/火山引擎的豆包 Seedream 模型 API。 +> 阿里云万相（Wan）是不同厂商的产品，参见万相相关文档。 + +## 端点 `POST https://ark.cn-beijing.volces.com/api/v3/images/generations` -## Authentication +## 认证 ``` Authorization: Bearer ark-119a37ed-f123-42fc-aea6-b38d47d47fc0-f7ddf ``` -## Response Format +## 响应格式 -### Success Response (response_format: "url") +### 成功响应 (response_format: "url") ```json { @@ -26,7 +29,7 @@ Authorization: Bearer ark-119a37ed-f123-42fc-aea6-b38d47d47fc0-f7ddf } ``` -### Success Response (response_format: "b64_json") +### 成功响应 (response_format: "b64_json") ```json { @@ -40,7 +43,7 @@ Authorization: Bearer ark-119a37ed-f123-42fc-aea6-b38d47d47fc0-f7ddf } ``` -### Error Response +### 错误响应 ```json { @@ -51,26 +54,46 @@ Authorization: Bearer ark-119a37ed-f123-42fc-aea6-b38d47d47fc0-f7ddf } ``` -## Available Models +## 可用模型 -| Model ID | Capabilities | Max Size | -|---|---|---| -| `doubao-seedream-5-0-260128` | Text-to-image, image-to-image, multi-image fusion, editing, batch | 2K | -| `doubao-seedream-4-5-t2i` | Text-to-image, image-to-image, multi-image fusion | 4K | -| `doubao-seedream-4-0-t2i` | Text-to-image, image-to-image, multi-image fusion | 4K | -| `doubao-seedream-5-0-lite-t2i` | Text-to-image, batch, web search | 3K | -| `doubao-seedream-3-0-t2i` | Text-to-image (with seed/guidance_scale) | 2K | -| `doubao-seededit-3-0-i2i` | Dedicated image editing (with seed/guidance_scale) | adaptive | +| Model ID | 能力 | 最大尺寸 | 备注 | +|---|---|---|---| +| `doubao-seedream-5-0-260128` | 文生图、图生图、多图融合、编辑、批量、联网搜索 | 3K | **默认推荐** | +| `doubao-seedream-5-0-lite-260128` | 文生图、批量、联网搜索 | 3K | 轻量版，速度更快 | +| `doubao-seedream-4-5-251128` | 文生图、图生图、多图融合 | 4K | 高质量 | +| `doubao-seedream-4-0-250828` | 文生图、图生图、多图融合 | 4K | 稳定版 | +| `doubao-seededit-3-0-i2i` | 精确图片编辑 (seed/guidance_scale) | adaptive | **当前API Key无权限** | -## Image Input Formats +> ⚠️ `doubao-seededit-3-0-i2i` 精确编辑模型在当前 API Key 下返回 404，暂不可用。 -- **URL**: Publicly accessible HTTP/HTTPS URL +## 图片输入格式 + +- **URL**: 公网可访问的 HTTP/HTTPS URL - **Base64**: `data:image/;base64,` - - Format must be lowercase: `data:image/png;base64,...` (not `PNG`) - - Supported formats: png, jpeg, webp + - 格式必须小写: `data:image/png;base64,...`（不是 `PNG`） + - 支持格式: png, jpeg, jpg, webp, bmp, tiff, gif + - 文件大小: 不超过 10MB + - 总像素: ≤ 6000×6000 + - 宽高比: [1/16, 16] -## Size Constraints +## 尺寸约束 -- Pixel range: `[1280, 4096]` for both width and height -- Max aspect ratio: 3:1 (long edge to short edge) -- Custom size format: `"WIDTHxHEIGHT"` (e.g., `"2048x1024"`) +### 方式1：分辨率别名 +- `2K`（默认）、`3K`（仅5.0）、`4K`（4.5/4.0） +- 需在 prompt 中描述宽高比/形状/用途，由模型判断具体像素 + +### 方式2：精确像素 +- 格式: `"WIDTHxHEIGHT"`（如 `"2048x2048"`） +- 总像素范围: [3686400, 16777216]（即最小约 2560×1440） +- 宽高比范围: [1/16, 16] +- 每个维度: [1280, 4096] + +> ⚠️ 两种方式不可混用 + +## image 参数说明 + +- 类型: `string[]`（URL或Base64数组） +- 最多 **10** 张参考图 +- 输入图 + 输出图总数 ≤ 15 +- prompt 中用"图1""图2"等指代图片顺序 +- **重要**: image 传入的是参考图，模型会根据 prompt 重新生成，不能100%像素级保留原图产品外观 diff --git a/.trae/skills/image-tools/scripts/__pycache__/wan_image_edit.cpython-313.pyc b/.trae/skills/image-tools/scripts/__pycache__/wan_image_edit.cpython-313.pyc deleted file mode 100644 index bb845fa..0000000 Binary files a/.trae/skills/image-tools/scripts/__pycache__/wan_image_edit.cpython-313.pyc and /dev/null differ diff --git a/.trae/skills/image-tools/scripts/wan_image_edit.py b/.trae/skills/image-tools/scripts/wan_image_edit.py deleted file mode 100644 index 612b893..0000000 --- a/.trae/skills/image-tools/scripts/wan_image_edit.py +++ /dev/null @@ -1,123 +0,0 @@ -import os -import sys -import base64 -import mimetypes -import requests -from urllib.request import urlretrieve - -API_KEY = os.getenv("DASHSCOPE_API_KEY", "") -DEFAULT_API_KEY = "sk-ef591b52e48840c6a63899c8901a9428" -if not API_KEY: - API_KEY = DEFAULT_API_KEY - print(f"使用默认API Key (前8位): {DEFAULT_API_KEY[:8]}...") - -def encode_image(file_path): - mime_type, _ = mimetypes.guess_type(file_path) - if not mime_type: - mime_type = "image/png" - with open(file_path, "rb") as f: - return f"data:{mime_type};base64,{base64.b64encode(f.read()).decode('utf-8')}" - -def wan_image_edit( - prompt, - image_path, - output_path="wan_output.png", - size="2K", - negative_prompt="", - n=1, - prompt_extend=True -): - """ - 万相2.7图生图 - 基于输入图片和提示词生成新图 - - 参数: - prompt: 提示词，描述想要的效果 - image_path: 输入图片路径 - output_path: 输出图片路径 - size: 输出尺寸 ("2K" 或 "1K") - negative_prompt: 反向提示词 - n: 生成数量 (1-4) - prompt_extend: 是否启用提示词扩展 - - 示例: - python wan_image_edit.py "keep the product, white background" "product.png" "output.png" - """ - print(f"提示词: {prompt}") - print(f"输入图片: {image_path}") - print(f"输出路径: {output_path}") - print(f"尺寸: {size}") - - url = "https://dashscope.aliyuncs.com/api/v1/services/aigc/multimodal-generation/generation" - headers = { - "Content-Type": "application/json", - "Authorization": f"Bearer {API_KEY}" - } - - content = [{"text": prompt}] - if image_path: - content.append({"image": encode_image(image_path)}) - - payload = { - "model": "wan2.7-image-pro", - "input": { - "messages": [{ - "role": "user", - "content": content - }] - }, - "parameters": { - "size": size, - "n": n, - "prompt_extend": prompt_extend, - "watermark": False - } - } - - if negative_prompt: - payload["parameters"]["negative_prompt"] = negative_prompt - - try: - resp = requests.post(url, headers=headers, json=payload, timeout=120) - result = resp.json() - - if resp.status_code == 200 and "output" in result: - choices = result["output"].get("choices", []) - if choices: - for idx, choice in enumerate(choices): - content_list = choice.get("message", {}).get("content", []) - for item in content_list: - if item.get("type") == "image": - image_url = item["image"] - out_path = output_path if n == 1 else output_path.replace(".png", f"_{idx}.png").replace(".jpg", f"_{idx}.jpg") - urlretrieve(image_url, out_path) - print(f"成功: {out_path}") - return True - else: - msg = result.get("message", str(result)) - print(f"失败: {msg}") - return False - - except requests.exceptions.Timeout: - print("请求超时，请重试") - return False - except Exception as e: - print(f"错误: {e}") - return False - -if __name__ == "__main__": - if len(sys.argv) < 3: - print("用法: python wan_image_edit.py <提示词> <输入图片> [输出路径] [尺寸]") - print("") - print("示例:") - print(' python wan_image_edit.py "keep the product, white background" "product.png" "output.png"') - print(' python wan_image_edit.py "cyberpunk style" "product.png" "output.png" "2K"') - print("") - print("尺寸选项: 1K (1024x1024), 2K (2048x2048)") - sys.exit(1) - - prompt = sys.argv[1] - image_path = sys.argv[2] - output_path = sys.argv[3] if len(sys.argv) > 3 else "wan_output.png" - size = sys.argv[4] if len(sys.argv) > 4 else "2K" - - wan_image_edit(prompt, image_path, output_path, size) \ No newline at end of file diff --git a/产品/NAS U盘/image/实拍图.png b/产品/NAS U盘/image/实拍图.png new file mode 100644 index 0000000..7b372fa Binary files /dev/null and b/产品/NAS U盘/image/实拍图.png differ diff --git a/产品/nascab/产品名称.md b/产品/NAS U盘/产品名称.md similarity index 100% rename from 产品/nascab/产品名称.md rename to 产品/NAS U盘/产品名称.md diff --git a/产品/NAS U盘/详情/NAS U盘主图.png b/产品/NAS U盘/详情/NAS U盘主图.png new file mode 100644 index 0000000..ddce082 Binary files /dev/null and b/产品/NAS U盘/详情/NAS U盘主图.png differ diff --git a/产品/nascab/image/214357667.jpg b/产品/nascab/image/214357667.jpg deleted file mode 100644 index 82de519..0000000 Binary files a/产品/nascab/image/214357667.jpg and /dev/null differ diff --git a/产品/nascab/image/214357698.jpg b/产品/nascab/image/214357698.jpg deleted file mode 100644 index 2b02f2f..0000000 Binary files a/产品/nascab/image/214357698.jpg and /dev/null differ diff --git a/产品/nascab/image/scene_01_主图白底.png b/产品/nascab/image/scene_01_主图白底.png deleted file mode 100644 index a97247a..0000000 Binary files a/产品/nascab/image/scene_01_主图白底.png and /dev/null differ diff --git a/产品/nascab/image/scene_02_相册备份.png b/产品/nascab/image/scene_02_相册备份.png deleted file mode 100644 index ac52513..0000000 Binary files a/产品/nascab/image/scene_02_相册备份.png and /dev/null differ diff --git a/产品/nascab/image/scene_02_相册备份_v2.png b/产品/nascab/image/scene_02_相册备份_v2.png deleted file mode 100644 index c182db5..0000000 Binary files a/产品/nascab/image/scene_02_相册备份_v2.png and /dev/null differ diff --git a/产品/nascab/image/scene_02_相册备份_v3.png b/产品/nascab/image/scene_02_相册备份_v3.png deleted file mode 100644 index 5540240..0000000 Binary files a/产品/nascab/image/scene_02_相册备份_v3.png and /dev/null differ diff --git a/产品/nascab/image/scene_02_相册备份_v4.png b/产品/nascab/image/scene_02_相册备份_v4.png deleted file mode 100644 index 3ecf032..0000000 Binary files a/产品/nascab/image/scene_02_相册备份_v4.png and /dev/null differ diff --git a/产品/nascab/image/scene_03_影视播放.png b/产品/nascab/image/scene_03_影视播放.png deleted file mode 100644 index 129f22c..0000000 Binary files a/产品/nascab/image/scene_03_影视播放.png and /dev/null differ diff --git a/产品/nascab/image/scene_04_多设备互联.png b/产品/nascab/image/scene_04_多设备互联.png deleted file mode 100644 index fc6d713..0000000 Binary files a/产品/nascab/image/scene_04_多设备互联.png and /dev/null differ diff --git a/产品/nascab/image/scene_05_文件共享.png b/产品/nascab/image/scene_05_文件共享.png deleted file mode 100644 index f21f2aa..0000000 Binary files a/产品/nascab/image/scene_05_文件共享.png and /dev/null differ diff --git a/产品/nascab/image/scene_06_加密安全.png b/产品/nascab/image/scene_06_加密安全.png deleted file mode 100644 index fdc447d..0000000 Binary files a/产品/nascab/image/scene_06_加密安全.png and /dev/null differ diff --git a/产品/nascab/详情/__pycache__/doubao_image.cpython-313.pyc b/产品/nascab/详情/__pycache__/doubao_image.cpython-313.pyc deleted file mode 100644 index 863996f..0000000 Binary files a/产品/nascab/详情/__pycache__/doubao_image.cpython-313.pyc and /dev/null differ diff --git a/产品/nascab/详情/add_text.py b/产品/nascab/详情/add_text.py deleted file mode 100644 index 61d6b37..0000000 --- a/产品/nascab/详情/add_text.py +++ /dev/null @@ -1,70 +0,0 @@ -# -*- coding: utf-8 -*- -"""在图片上叠加文字标签：标题 + 副标题，带半透明背景条""" -from PIL import Image, ImageDraw, ImageFont -import sys -import os - -def add_text_overlay(image_path, title, subtitle, output_path): - img = Image.open(image_path).convert("RGBA") - w, h = img.size - - # 创建文字图层 - txt_layer = Image.new("RGBA", (w, h), (0, 0, 0, 0)) - draw = ImageDraw.Draw(txt_layer) - - # 字体大小根据图片宽度自适应 - title_size = max(48, w // 18) - sub_size = max(28, w // 30) - - # 尝试加载中文字体 - font_paths = [ - r"C:\Windows\Fonts\msyhbd.ttc", # 微软雅黑粗体 - r"C:\Windows\Fonts\msyh.ttc", # 微软雅黑 - r"C:\Windows\Fonts\simhei.ttf", # 黑体 - ] - title_font = None - sub_font = None - for fp in font_paths: - if os.path.exists(fp): - title_font = ImageFont.truetype(fp, title_size) - sub_font = ImageFont.truetype(fp, sub_size) - break - - if title_font is None: - title_font = ImageFont.load_default() - sub_font = ImageFont.load_default() - - # 文字位置：左上角，带边距 - margin_x = w // 15 - margin_y = h // 10 - - # 绘制半透明背景条 - title_bbox = draw.textbbox((margin_x, margin_y), title, font=title_font) - sub_bbox = draw.textbbox((margin_x, title_bbox[3] + 12), subtitle, font=sub_font) - - bar_top = margin_y - 16 - bar_bottom = sub_bbox[3] + 16 - bar_right = max(title_bbox[2], sub_bbox[2]) + 30 - - draw.rectangle( - [(margin_x - 16, bar_top), (bar_right, bar_bottom)], - fill=(0, 0, 0, 140) - ) - - # 绘制标题（白色粗体） - draw.text((margin_x, margin_y), title, font=title_font, fill=(255, 255, 255, 255)) - - # 绘制副标题（浅白色） - draw.text((margin_x, title_bbox[3] + 12), subtitle, font=sub_font, fill=(220, 230, 255, 230)) - - # 合成 - result = Image.alpha_composite(img, txt_layer).convert("RGB") - result.save(output_path, "PNG") - print(f"Saved: {output_path}") - - -if __name__ == "__main__": - if len(sys.argv) < 5: - print("Usage: python add_text.py <subtitle> <output_path>") - sys.exit(1) - add_text_overlay(sys.argv[1], sys.argv[2], sys.argv[3], sys.argv[4]) diff --git a/产品/nascab/详情/doubao_image.py b/产品/nascab/详情/doubao_image.py deleted file mode 100644 index 07b88ab..0000000 --- a/产品/nascab/详情/doubao_image.py +++ /dev/null @@ -1,70 +0,0 @@ -# -*- coding: utf-8 -*- -"""豆包万相 (Doubao Seedream) 图片生成工具 - 支持文生图/图生图""" -import sys -import os -import json -import base64 -import urllib.request -import urllib.error - -API_URL = "https://ark.cn-beijing.volces.com/api/v3/images/generations" -API_KEY = "ark-119a37ed-f123-42fc-aea6-b38d47d47fc0-f7ddf" -MODEL = "doubao-seedream-5-0-260128" - - -def generate(prompt, input_image=None, output_path="output.png", size="2k"): - """生成图片。input_image 为本地路径时走图生图，否则文生图。""" - body = { - "model": MODEL, - "prompt": prompt, - "size": size, - "sequential_image_generation": "disabled", - "response_format": "url", - "stream": False, - "watermark": False, - } - - if input_image and os.path.exists(input_image): - with open(input_image, "rb") as f: - img_b64 = base64.b64encode(f.read()).decode("ascii") - ext = os.path.splitext(input_image)[1].lower().lstrip(".") - mime = "jpeg" if ext in ("jpg", "jpeg") else ("png" if ext == "png" else "octet-stream") - body["image"] = [f"data:image/{mime};base64,{img_b64}"] - - data = json.dumps(body).encode("utf-8") - req = urllib.request.Request( - API_URL, - data=data, - headers={ - "Content-Type": "application/json", - "Authorization": f"Bearer {API_KEY}", - }, - method="POST", - ) - - try: - with urllib.request.urlopen(req, timeout=180) as resp: - result = json.loads(resp.read().decode("utf-8")) - except urllib.error.HTTPError as e: - err_body = e.read().decode("utf-8", errors="replace") - print(f"HTTP {e.code}: {err_body}", file=sys.stderr) - sys.exit(1) - - image_url = result["data"][0]["url"] - print(f"Generated URL: {image_url}") - - # 下载图片 - urllib.request.urlretrieve(image_url, output_path) - print(f"Saved to: {output_path}") - - -if __name__ == "__main__": - # 用法: python doubao_image.py <prompt> <output_path> [input_image] [size] - if len(sys.argv) < 3: - print("Usage: python doubao_image.py <prompt> <output_path> [input_image] [size]") - sys.exit(1) - prompt = sys.argv[1] - output_path = sys.argv[2] - input_image = sys.argv[3] if len(sys.argv) > 3 else None - size = sys.argv[4] if len(sys.argv) > 4 else "2k" - generate(prompt, input_image, output_path, size) diff --git a/产品/nascab/详情/详情页文案.md b/产品/nascab/详情/详情页文案.md deleted file mode 100644 index 9909a8b..0000000 --- a/产品/nascab/详情/详情页文案.md +++ /dev/null @@ -1,74 +0,0 @@ -# NasCabOS 商品详情页文案 - -> 产品全称：智能存储NAS相册备份无线传输共享视频影视播放手机电脑互联 -> 软件名称：NasCabOS -> 官网：https://nas.cab/ - ---- - -## 屏1｜产品主图 - -![产品主图](../image/scene_01_主图白底.png) - -**NasCabOS 智能私有云** -手机电脑互联 · 相册备份 · 影视播放 · 无需专用硬件 - ---- - -## 屏2｜相册备份 - -![相册备份场景](../image/scene_02_相册备份.png) - -**手机照片自动备份 永不丢失** -- 插上U盘，手机照片自动无线备份到私有云 -- 时间线/地图视图，智能相册自动分类 -- 人脸识别 · AI分类，一键整理海量照片 -- 不改变原有文件夹结构，只读取和组织 - ---- - -## 屏3｜影视播放 - -![影视播放场景](../image/scene_03_影视播放.png) - -**私人影院 随时随地看** -- 手机/iPad/电视多端播放，实时转码 -- 几乎支持所有视频格式，蓝光原盘也能播 -- 自动匹配海报，影视库式管理 -- 出门在外也能远程看家里的电影 - ---- - -## 屏4｜多设备互联 - -![多设备互联场景](../image/scene_04_多设备互联.png) - -**手机 电脑 平板 一盘全通** -- Windows / macOS / Android / iOS / TV / Web 全平台 -- 局域网内所有设备同时访问，互不干扰 -- 无需专用硬件，电脑装上就能用 -- 插上U盘，旧电脑秒变NAS服务器 - ---- - -## 屏5｜文件共享 - -![文件共享场景](../image/scene_05_文件共享.png) - -**全家共享 一人管文件全家都能用** -- FTP / SFTP / WebDAV 三协议共享 -- 每个成员独立账号，独立权限 -- 在线上传下载，远程也能访问 -- 挂载云盘，聚合多个网盘到一个入口 - ---- - -## 屏6｜加密安全 - -![加密安全场景](../image/scene_06_加密安全.png) - -**AES256加密 隐私绝不泄露** -- 加密空间，隐藏敏感文件 -- 2FA双因素认证，Google/Microsoft验证器 -- IP黑白名单，控制谁能访问 -- WebRTC加密传输，数据不走第三方服务器