⏳ This skill is pending AI review.
Scores will appear once the review pipeline completes.
huashu-art-motion
艺术与视频动画:拆解复刻动画;36种既有实做风格+16种可迁移研究配方;9种解说动画语法,8种可按口播参数渲片段;含口播整片参考代码及框景检查;人用AI生帧;可选已有录音、系统声音与语音复刻,按需配置。
Choose how to use this skill
You do not need every option. Choose the path your AI client supports. The stable page stays the same; versioned files are immutable.
1. Native installer
This listing has no registered native installer command. Use the complete package or source fallback below, depending on what your client supports.
Do not guess an installer command or replace an existing version without reviewing the diff.
2. Complete package recommended
Download the ZIP when available. It includes SKILL.md plus the references, security notes and version metadata.
No complete ProSkills package is published for this listing yet.3. Prompt-only
Copy the prompt above when the agent can read the stable page or when you want to adopt the workflow without installing a skill.
Need only the instruction file?
Download SKILL.md only if your client requires a single file. The complete ZIP is safer for a full installation because it preserves the references and release context.
No path installs or executes anything by itself. Your agent still needs access to the project files. Before updating, compare the installed version and review the diff.
// RATINGS
Not yet listed on ClawHub or SkillsMP
// README
简体中文 | English
huashu-art-motion · 艺术动画
让你的coding agent,把艺术风格写成会动的画。
36种艺术风格 · 9种解说语法 · 8种参数化片段 · 口播整片参考代码
npx skills add alchaincyf/huashu-art-motion
动态
- 2026-10:Open Design(nexu-io 出品的开源 AI 设计工作区)的 motion-design skill 把 huashu-art-motion 列为设计来源之一,并在 skills/huashu-art-motion 收录了一份适配版,保留了 MIT 许可和上游出处。很高兴这套艺术动画的做法被另一个开源项目用起来,谢谢 Open Design 团队。也欢迎更多项目使用、反馈和贡献。
动画样片
核心案例:花叔闯进超级玛丽
把自己做成像素主角,跑完一整关。砖块、水管、关卡运动与转场用代码完成,人物用生成帧合成;下面直接截自65秒成片。
连吃蘑菇都得先选会员:OpenAI,还是Claude?选完Claude,星芒道具落下来,花叔变大。
这组展示的是成片效果;GIF无声、循环播放。片段时间点与导出参数。
Clawd 博物馆奇妙夜
Claude Code 的像素吉祥物 Clawd 从保安电脑的终端里蹦出来,打着手电夜巡纽约大都会博物馆,再穿过中央公园去自然史博物馆,天亮前赶回来装成展品。98秒,16个场景,无口播,配乐也是代码合成的。
跳进梵高《麦田与柏树》:笔触方向取自原作的结构张量,Clawd被重画成厚涂。
暗处是炭笔夜景,只有手电照到的地方显出颜色;入画的五幅(梵高、修拉、维米尔、伦勃朗、勒茨)都是大都会开放获取的CC0原作,按各自画家的笔法用代码重画。完整成片在Releases,片段时间点与说明。
花叔穿越名画
画里真的会动。下面三段来自同一支穿越短片,场景用代码画,角色用生成帧合成。
日本桥上的一次水漂。
看效果
35段样片各取一帧:同一位少女、同一只橘白猫、同一张桌子,从公元前 40000 年的岩洞一路穿到 2026 年。每一段都在动:梵高的星空在转,马赛克的颜色从石块上流过去,水墨晕染把画面带进下一个时代。
注意:这 35 格用的是同一个构图,那是艺术史速通片的机制(风格在变、场景不变),不是这些风格的构图。做新题材要为题材重新构图,别把这间屋搬过去。
长卷穿越片:一个人从左走到右,跨过边界的那一刻,世界和他自己的画风一起换(下图是收进仓库的 3 段示范:埃及壁画 → 莫奈《日本桥》→ 8-bit):
8种解说语法各有一支可运行示范片(下图每镜取一帧)。此外还有第9种「讲解员式财经科普」,提供语法卡和需自备角色的整片代码快照。白板适合跟着口播画关系,Vox适合图形与信息拼贴:
| Kurzgesagt:扁平无描边、尺度穿行 | Vox:剪报、红线、荧光笔 |
| 白板:笔尖揭开线稿 | 3Blue1Brown:一个对象形变成下一个 |
| Storytime:反应特写、笑点停顿 | 动态文字:主词砸进来 |
| 发布会:光斑底、毛玻璃卡、大数字 | 财经图表:先轴、后数据、只标一件事 |
能做什么
| 你说 | 它做 |
|---|---|
| 「复刻这个动画」「拆一下这段」 | 先跑拆解脚本量出转场、节拍网格、每段运动热图,再按机制用代码复刻 |
| 「做个梵高/莫奈/包豪斯那种的动画」 | 先设计一帧,再让它动起来;36张风格配方卡当起点 |
| 「用我的口播做一段艺术动画」 | 镜头表 → 定风格 → 世界画布加镜头 → 输出一条画面轨 |
| 「做一个人穿过一幅幅名画的片子」 | 长卷骨架:每个世界一个段文件,主角一路往右走,跨边界换画风,镜头只进不退 |
| 「做解说视频的动画段」 | 按口播选语法,喂一份 JSON,出一段时长精确到帧的片段(横竖屏、可透明底) |
| 「画面里要有人」 | 人交给生图模型出帧,代码负责合成、换帧和材质 |
| 「配个乐、卡节奏」 | BPM 网格、动机换乐器、结尾音效序列,纯代码合成 |
交付前有一道数字验收:qa.py 量稳定、效率、动感、流畅及文字框景线索,再派一个没参与制作的 agent 只看成片挑问题;最后 deliver.py 一条命令交付——成片先过反 PPT 门禁(整片几乎没有镜头内快动作就红灯,不出片),过了才复制成片、写交付说明,说明里的事实由脚本写。
里面有什么
| 数量 | |
|---|---|
| 艺术风格配方卡(参数、母题动作、签名转场、当前短板) | 53张,references/风格配方/;其中 38–55 是按原作锚定重做的研究配方(含敦煌壁画) |
| 对应的场景代码 | 53个,scripts/engine/scenes/ |
| 解说动画语法卡 | 9份;其中8种附示范片、参数化片段与示例spec |
| 口播整片参考代码 | 混合风格、白板、Vox、讲解员;需自备部分素材,详见目录README |
| 长卷穿越片示范(骨架+3 段+角色帧库) | 1 支,scripts/engine/demos/long_scroll/ |
| 版面审计与压测(出界/叠字/字太小、竖屏/超长/英文变体、示例回归) | render.py --audit、scripts/clip_lint.py、scripts/analyze/stress_clip.py、regress_clips.py |
| 转场 | 艺术风格签名转场与解说转场,包含淡入、硬切和纸面转场 |
| 绘画与动画库(笔刷、渲染器、后期、骨架、镜头、图表、排版……) | 17 个,scripts/engine/lib/ |
| 方法文档(拆解、机制、一帧先行、纯代码绘制、节奏配乐、角色、长卷……) | 核心12篇,references/01–12;另有可选语音指南 |
huashu-art-motion/
├── SKILL.md # 先判断任务,再按表读对应文档
├── references/ # 01–16方法文档、53张风格配方卡、9张语法卡、正面经验
├── assets/ # 总览图(README 用)
└── scripts/
├── engine/ # 可整个复制走的动画工程:引擎、转场、库、场景、示范片、片段
├── analyze/breakdown.py # 把参考动画拆成「能写代码的地图」
├── qa.py # 一键验收
├── film_gate.py # 成片门禁:量翻页式 PPT 感、查中段空画面(只读 mp4)
├── deliver.py # 一条命令交付:门禁→复制成片→写交付说明
├── audio/ # 纯代码合成配乐的模板
└── font_subset.py ... # 字体子集、绿幕抠图
依赖:uv、ffmpeg、Playwright Chromium(第一次跑 uv run --with playwright playwright install chromium)。试一下:
cd huashu-art-motion
uv run --with playwright python scripts/engine/render.py --solo 09_postimp --stills 0.3 --out 试渲 # 梵高那一段的一帧
uv run --with playwright python scripts/engine/render.py --spec scripts/engine/examples/t3_finance_chart.json --out 财经图表.mp4
背后的故事
2026 年 10 月初,我在 X 上看到 Tak(@cherry_mx_reds)的一支 15 秒动画:一位少女和一只猫穿过 40000 年艺术史,每个时代只有一秒左右,但画里的东西都在动。
我让 Claude 复刻它。第一版是「一张张画之间做转场」,被我否了:原片每个时代可能就一秒,但画里的元素完全是流动的。于是改成先拆解(量转场、拟合节拍网格、看每段哪里在动),再用代码一层层把画画出来、让它动起来。
做的过程中我跟它说:「我们不只是为了复刻,我需要你积累经验。」所以这个 skill 里记的不只是代码,还有哪些做法被证明有效(references/07-正面经验.md)、每种风格的坑和短板。之后又派了 4 组只读 skill 的 agent 去做它没见过的 20 种风格,把它们各自造的轮子收成统一的库;再加上 8 种 YouTube 解说动画语法,接进了我自己的口播视频管线。最后用同一套东西做了《花叔穿越名画》:23种画风、2分08秒,我从洞穴一路走到 2026,骨架和其中 3 段也收进来了。
致谢
- Tak(@cherry_mx_reds) 的《Art History Speedrun》是这个 skill 的起点。16 个艺术时代的场景构图和「少女+猫穿越」的设定沿用了原片的思路,画面全部用代码重新画,配乐脚本里是原创示例乐谱(只保留拆解方法,不保留对原曲的转录);仓库里不含原片的帧、截图或音频文件。想看原作请去他的 X。
- 解说语法卡里拆解过的频道和资料(Kurzgesagt、Vox、3Blue1Brown、RSA Animate、TheOdd1sOut 等)都在各张卡的「一手参考」里给了链接,仓库只记测量出来的参数,不含他们的画面。
- 字体都是 SIL OFL 1.1 开源字体,清单和版权见
scripts/engine/lib/fonts/LICENSES.md。
关于作者
| 🌐 官网 | bookai.top · huasheng.ai |
| 𝕏 Twitter | @AlchainHust |
| 📺 B站 | 花叔v |
| ▶️ YouTube | @Alchain |
| 📕 小红书 | 花叔 |
| 💬 公众号 | 微信搜「花叔」 |
许可证
代码和文档:MIT。随便用,随便改,随便造。
例外:笔顺衍生数据reference_films/spacex/spacex_wb/assets/strokes.js沿用Arphic Public License(原文随文件附带);scripts/engine/lib/fonts/ 里的字体沿用各自的 OFL 许可;花叔的卡通形象与角色帧(scripts/engine/demos/_shared/hero/、scripts/engine/demos/long_scroll/frames/、assets/角色/)以及总览图、示范视频中包含的同一形象,只用于本 skill 的示范,不随 MIT 授权用于其他用途。
Clawd 是 Anthropic 的 Claude Code 吉祥物,形象权利归 Anthropic;《Clawd 博物馆奇妙夜》是非官方的同人示范,与 Anthropic 无关,GIF 与成片同样不随 MIT 授权。片中画作来自大都会艺术博物馆开放获取(CC0)。
女娲 造 Skill。达尔文 让 Skill 进化。艺术动画 让画动起来。
MIT License © [花叔 Huashu](https://github.com
// HOW IT'S BUILT
// FILES
508 files shipped with this skill. Inline text is the catalog copy. The upstream link is the source repository.
SKILL.md24 KB73f2248675f9upstream ↗
--- name: huashu-art-motion description: "艺术与视频动画:拆解复刻动画;36种既有实做风格+16种可迁移研究配方;9种解说动画语法,8种可按口播参数渲片段;含口播整片参考代码及框景检查;人用AI生帧;可选已有录音、系统声音与语音复刻,按需配置。" metadata: author: "花叔" homepage: "https://github.com/alchaincyf" repository: "https://github.com/alchaincyf/huashu-art-motion" version: "1.2.0" --- # 艺术动画 ## 先分档:你按哪一档读 - **开阔档**:你是 Claude Opus 5.5 或更强的模型。全文都适用:路由表、窄桥、开阔地;库可以自己组合,渲染器可以自己写;交付同样走 `scripts/deliver.py`,红灯不交付。开工前也建议先写一张镜头表(`references/镜头表模板.md`),lint 可跑可不跑。 - **轨道档**:其它模型,或判断不出自己是谁。只走下面这一条路,先镜头表,再动手;只读这条路点名的文件,其余 reference 先不读: 1. **写镜头表**:照 `references/镜头表模板.md` 写 `<项目>/镜头表.md`。表前先锁画风,三行:画风(一句话,全片只用这一种)、色板(`references/色板.md` 里的一套或 3–6 个色值,不落蓝紫)、角色(无/只出手和背影/帧库 <路径>;37/52号平面卡通可写原创语义SVG,先读 `references/角色精细度.md`)。表里每行一镜:时间段|画面里的物(名词)|它在做什么(动词)|镜头(停/缓推/拉远/横移/快推/砸入/切,按下面「导演的要求:镜头怎么用」选,快推和砸入只给重音;写这一列前读 `references/镜头语言.md` 第一到四节)|屏上字(≤8 字,可空)|做法。开头 3 秒是一个具体的小场景(一双手或一个物在做一件和题目有关的事),不是标题;全片用一个具体例子贯穿;一镜 2–6 秒;文字主导的镜头 ≤30%;相邻两镜换构图,不换画风。 2. **过 lint**:`uv run <skill>/scripts/storyboard_lint.py <项目>/镜头表.md`(有口播稿加 `--script <项目>/文稿.txt`,查屏上数字和口播是不是同一个数)。有红灯就改镜头表、重跑,红灯没清不许开始写代码。 3. **按镜头表逐镜做**:`cp -R <skill>/scripts/engine <项目>/代码工程`,每一镜出一个 `镜NN.mp4`,所有镜同宽高、同 fps(默认 1920×1080@30)。 - 参数化片段:读 `references/09-视频动画语法.md` 第二节选型表和用到的 `references/动画语法/<语法>.md`「起手必查」,从 `examples/<语法>.json` 抄结构不抄内容,屏上字照镜头表那一格,`uv run --with playwright python <项目>/代码工程/render.py --spec 镜NN.json --out 镜NN.mp4`。**单个参数化片段最多撑 8 秒**,超了拆成几镜。 - scenes:艺术风格画面(梵高、像素……)或片段语法画不出的物,读 `references/风格配方/INDEX.md` 和一张风格卡;研究卡或首次迁移先读 `references/15-跨场景风格提炼.md`,没有实现时先创作和验证,不能把卡名当已注册场景。实做时写 `scenes/<id>.js` 并登记进 `eras.js`,`render.py --solo <id> --from 0 --to <秒> --fps 30 --no-counter --out 镜NN.mp4`。 - 全片一套画风:所有片段 spec 写同一个 `"theme"`(镜头表色板那一行,写法见 `references/色板.md`),最多 2 种片段语法;scenes 也只用这套色板。每镜加 `--stills 0.5,2 --out 静帧/镜NN` 抽帧,排成一排对着镜头表看:哪一镜像从别的片子里拿来的,就按画风重做。 4. **拼接**:镜与镜硬切,在放 `镜NN.mp4` 的目录里: ```sh for f in 镜*.mp4; do printf "file '%s'\n" "$f"; done > 镜头列表.txt ffmpeg -y -f concat -safe 0 -i 镜头列表.txt -r 30 -c:v libx264 -crf 16 -pix_fmt yuv420p -an 成片.mp4 ``` 有口播再混音:`ffmpeg -i 成片.mp4 -i 口播.wav -map 0:v -map 1:a -c:v copy -c:a aac -shortest 成片_口播.mp4`。要转场就全用 scenes,把各镜按顺序写进 `eras.js` 段落表(段写 `dur` 秒、`transition`),整片模式 `render.py --out 成片.mp4` 一次出片。 5. **qa**:每个片段 `uv run <skill>/scripts/qa.py --spec 镜NN.json --out <项目>/qa/镜NN`;有 scenes 再跑 `qa.py --project <项目>/代码工程 --ids <id,…> --out <项目>/qa/scenes`。再拿抽帧逐条对下面「六种做坏方式」。 6. **交付**:`uv run <skill>/scripts/deliver.py 成片.mp4 --root <项目> --qa <项目>/qa/<跳变最多的那一镜> --out output/`。红灯不交付:脚本不复制成片,照它打印的修法改完重渲、重跑。绿灯、黄灯它复制成片并写 `output/交付说明.md`,时长、门禁、qa、有没有审片由脚本写;你只在「制作说明」段补用了哪些语法和做法、其余各镜的 qa 结论(每镜一行)、黄灯项逐条理由、六种做坏方式自查,事实段不许改。派了独立审片的,先让审片人把结论写进 `<项目>/审片/*.md` 再跑。 ## 你是谁 你同时是这支片子的导演、分镜师、原画师、动画师和剪辑师,画画用的是代码。你对成片的标准:观众看得懂每一镜在讲什么,愿意看下去;用艺术风格时一眼认得出是谁的画,画是活的,不是图片在切换。 你先是导演。下面两节是导演对整片提的要求:镜头怎么动、画风怎么选。具体的秒数、次数、参数在 references 里,每个数都写了什么时候生效、什么时候不用管;那些数是起点,你有更好的设计就去做,做完用正常速度看一遍。 ## 导演的要求:镜头怎么用 镜头是标点,不是每句话的修饰。每动一次,先答三句:观众这一刻该看哪,在当前景别下看不看得清;这一下是过渡还是重音;上一个镜头动作是什么。 - **看得清就不动。** 画面自己在动(实拍录像、正在描的线、正在长的曲线)时镜头停住;观众在读字时停住。解说片有大段停住是对的,不是卡。 - **过渡用缓的,重音才用快的。** 带视线去下一处细节用缓推,在几件东西之间走用横移,交代全貌、段落收束用拉远;快推、砸入、甩镜是重音,只给揭晓、反转、关键数字和笑点,一支片子里数得过来。 - **有来有回。** 推进去之后要拉出来、横移开,或者硬切换个景别,不连着三镜往里推。一镜一张满幅素材时最容易变成「切一张、推一下」:一支 102 秒的解说短片推近约 48 次、全部往里推,作者原话:「有非常非常多zoom in的镜头运动。次数太多,而且节奏太快了,一点都不smooth」。 - **整片不能只有一种节拍。** 每镜匀速慢推被作者说过「完全是放ppt的感觉」,每镜快推一下又被说节奏太快、不 smooth,病根是同一个:每镜同一个动作。镜头要随段落情绪变:平铺直叙时稳,转折和高潮才加快。一支 MV 作者也说过:「运镜和转场都显得单调、干涩,缺乏多样性,缺乏情绪的表达」。 - **不动也不等于好。** 作者否过一版「背景太特么素了,毫无观看欲望」的片子:要的是画面里一直有事在发生,物在动、素材本身在动、新东西被揭开、镜头带着视线走。缓推、横移、拉远都是运动,快不快和动不动是两回事。 - **艺术风格画面另算:画要一直是活的。** 每幕一个主动作+2 个母题循环,镜头可以慢漂。作者否过「图片之间做转场」:「原片每个艺术时代,可能就1秒钟,里面的画中的元素完全是流动的……牺牲太多了。」 - 动作有哪些、各自多长、次数什么时候要管,读 `references/镜头语言.md`。引擎 `CAM.track` 不写动作就是缓推,参数化片段的 highlight 默认缓推,spec 顶层的 `data` 里写 `"camera": "snap"`(`{"data": {"camera": "snap"}}`;直接写在 spec 顶层(和 grammar 同级)或写进某条 cue 的 data 都不生效,也不报错)才是快推。 ## 导演的要求:画风怎么选 画风服务内容。先看手里有什么素材、这段要讲清什么,再挑画法。 学习艺术家或传统的整体视觉语言:笔触、色彩、用光、造型与空间。原作的人物、物品、布局不是新任务模板;构图按新场景重新设计。初次迁移读[跨场景风格提炼](references/15-跨场景风格提炼.md),用多幅不同作品提炼并以新题材验证;吉卜力、皮克斯等先明确影片级美术范围,不把工作室当单一滤镜。 - **有真实素材,素材就是主画面。** 截图、照片、录像、原文、产品本身的产出满幅放,代码负责解释、标注和运镜(`MD.cover` 满幅、`MD.print` 冲印件摆进调查桌、`MD.label` 黑底撕边标签;缺图直接报错)。用 Vox 调查桌或白板把素材串起来,只有真找不到素材的段落才整页代码画或 AI 生成。作者看完一支几乎全 AI 生成的技术长片:「视觉观感很不好」。 - **按要讲的东西选语法。** 讲机制和推导用白板、3b1b;讲数字和趋势用财经图表;讲经历和包袱用 storytime;讲艺术和时代用风格配方。选型表在 09 号第二节,每种语法的镜头基调在 `镜头语言.md` 第三节。 - **一片一套画风、一套色板,管到最后一帧。** 人、动物、道具也用同一套笔触画(一支莫奈风 MV 里矢量形状的猫和手,被作者说「和莫奈画风不够匹配」)。所有参数化片段写同一个 `theme`,scenes 也只用这几个色。干净统一但略像幻灯片的片子,胜过每镜换一种画风的拼贴。不用蓝紫渐变、深紫星空、霓虹青紫(作者原话:「这种非常ai slop」);用户点名要蓝紫才例外,交付时 `--allow-blue-purple "理由"`。 - **往立体、丰富的方向想。** 作者对一支剖面全景动画的评价:「立体,丰富,一次看到非常全面和震撼的局面」。用户否过的方向不再端出来。 - **新方向先出三个给用户挑**,同一帧、三种构图,不只是换调色(见窄桥「三方向硬门」)。 ## 最常见的六种做坏方式(开工前、交付前各对一遍) 1. **翻页式 PPT**:固定页眉、页码,要点逐条淡入,每镜同一个镜头动作(每镜匀速慢推,或者每镜都快推一下),结尾落回一张居中金句卡。→ 去掉页框,一镜一个画面事件,镜头按「导演的要求:镜头怎么用」动,相邻镜头动作不同、推进去有来有回;结尾落在主角或主物的一个动作上(放进口袋、合上书、拉远看全貌),字可以写在画里的物上。storyboard_lint 查最后一镜是不是字卡、连续三镜往里推、重音过多;deliver.py 会量出来:整片几乎没有镜头内运动就红灯,不出片;推近太密报黄。 2. **照搬示范**:示范片、样例 spec、风格样片里的构图、角色、素材直接进了成片。→ 只学做法,构图和物按题材重新设计。 3. **没有具体的物**:抽象概念只写成字和图标。→ 镜头表每行写出一个具体的物和它的动作,写不出物的那一行先别做。 4. **死背景、空画面**:大面积纯色或星空底一直不变,只有中间一小块在动;或者整帧只剩纯色、黑场、空景,主角不在画里。→ 主体放大到占画面一半以上,背景跟着内容换,每一镜主角或主物都在画里;图表镜头的底和前后镜头同色系(t3 的 `theme`)。storyboard_lint 查物列是不是空镜、纯色、黑场;film_gate(deliver.py 先跑它)查成片中段连续 0.3 秒以上的空画面,红灯不出片。 5. **有真实素材却用代码去画**。→ 截图、照片、录像、原文直接满幅放,代码只负责解释、标注和运镜。 6. **一眼 AI 味**:蓝紫渐变底、深紫星空底、霓虹青紫光;每镜换一种画风的拼贴(卡通人→暗色代码屏→星空图标→白底柱状图→网页截图);代码画的 Q 版小人和白壳机器人;长了脸的金币、小球、离子。→ 表前锁一种画风、一套色板(`references/色板.md`),所有片段写同一个 `theme`;角色要么 AI 生帧过交付线,要么只出手和背影(`references/角色精细度.md`)。storyboard_lint 查画风锁、色板色值和角色行;deliver.py 量蓝紫底,蓝紫帧占全片 30% 以上红灯,不出片。干净统一但略像幻灯片的片子,胜过花哨的拼贴。 ## 平面卡通风格入口 第36种风格见[37_rick_morty](references/风格配方/37_rick_morty.md):SVG配方、离线预览、可控表演与独立渲染器。明确要求参考式扁平卡通时可用语义SVG画角色,优先遵循本卡;真人和其他人物仍用原角色路线。此例不走Canvas gallery。南方公园方向读[52_south_park](references/风格配方/52_south_park.md),同样仅在明确选择该平面风格时允许原创语义SVG纸片角色;该卡已附单场景工程试作,风格忠实度与跨场景仍待验。此窄例外不扩展到真人或其他画风。 ## 先判断任务 | 用户说的 | 做什么 | 读 | |---|---|---| | 「复刻这个动画」「拆一下这段」 | 拆解 → 机制 → 代码复刻 | 01 → 02 → 04 | | 「做个 XX 风格的动画」「梵高/莫奈/包豪斯那种」 | 先为题材定构图 → 一帧先行 → 渲染器 → 让它活 | 03 → 04 → `风格配方/INDEX.md` 找最近的材料当起点 | | 「用我的口播做一段艺术动画」 | 镜头表 → 定风格 → 世界画布+镜头 → 画面轨 | 06(草案)+ 04;口播的转录与时间轴交给你的视频管线 | | 「配个乐」「卡节奏」 | BPM 网格、动机换乐器、结尾音效序列 | 05 | | 画面里要出现人(用户本人、真人、拟人角色、吉祥物) | AI 生帧+代码合成(B/C 两法),过交付线;没有生图能力就只出手和背影 | 10 + `角色精细度.md` | | 派 agent 画某个风格 / 自己写新风格 / 重做风格演示 | 先拉原作(`scripts/analyze/fetch_reference.py`)→ 按风格自己的构图规则定题材 → 形层过 `lib/medium.js` 媒介通道 → 并排对照打分 | **16** + 15 + 08 + `风格配方/INDEX.md` | | 「一个人穿过一幅幅名画」「一路走过几个世界」 | 长卷骨架:每个世界一个段文件,主角从左走到右,跨边界换画风 | 11 + `scripts/engine/demos/long_scroll/` | | 「做解说/科技/财经视频的动画」「Kurzgesagt/Vox/3b1b/白板/发布会/财经图表那种」、口播管线要一段动画 | 按口播选语法 → 过语法卡「起手必查」→ 管线用 `render.py --spec` 出时长精确的片段 | 09 + `动画语法/<语法>.md` + `镜头语言.md` | 口播整片先读`references/12-口播整片与经验回流.md`;讲解员风格再读`references/动画语法/y6_presenter_explainer.md`。艺术史速通、穿越这类「同一场景换风格」的片子,机制见 02 号(五层机制),只用于这类片子。 ## 可选口播:按需配置 任务需要新配音时,先读[语音能力指南](references/capabilities.md),运行`scripts/capabilities.py status --explain`。已有录音直接沿用;无声动画不启动配置。未选择时只询问当前需要的声音方式,可本次跳过或长期手动启用;用户已明确选择时直接应用。私人音色、参数和凭证引用保存在安装目录外,公共默认不绑定作者声音。 `scripts/koubo.py say/train/voices`保留旧调用,使用同一套配置。仅音频任务不启动画面设计或成片审片。 ## 可选图片:当前工具与现有素材 需要图片时读[图片能力指南](references/images.md)。已有图片直接导入;生成图片按当前会话实际工具和用户许可选择,脚本输出调用计划,Agent执行工具后验收并复制到项目。无图片任务跳过配置;禁止生成时不换供应商绕过。语音和图片分别配置,新增能力不会自动获准。 ## 窄桥:照字面做 - **人用AI生帧,代码负责合成**,见`references/10-角色.md`;明确选择37/52号平面卡通时可按对应卡使用原创语义SVG角色,其它画风不以通用代码小人替代。人物质量、物不随意加脸及y4帧库要求见`references/角色精细度.md`。 - **示范只用来理解做法。** 示范片、样例 spec、风格样片里的构图、角色和素材,不许原样放进成片。 - **复刻是学机制,不是逐像素。** 提取场景、运动与转场规律,再用于新题材。 - **三方向硬门**:新的角色诠释或风格方向,先出三个给用户挑,三个方向的构图要不同,不能只换调色。指定了风格也不豁免。 - **确定性**:所有随机用种子,同一时刻渲两次逐像素一致。 - **独立审片**:成片交付前派一个没参与制作的 agent 只看成片挑问题,结论写进 `<项目>/审片/*.md`(deliver.py 只认这里)。 - 角色涉及未成年人(如 17 岁)时,服装一律端庄长款;给生图模型写「短裙」会被审核整张拒绝。 ## 开阔地:当起点,不当配额 `风格配方/` 里有36种既有实做和17张新增研究配方(总表 `风格配方/INDEX.md`;既有实做分「材料层」和「时代标本」,研究配方单列),状态在索引分开列示;研究卡的技术路线尚待画面验证,不能借用实做卡的质量等级。既有卡保存实做的参数和坑,不是唯一画法。35 张样片和 `assets/全风格总览.jpg` 用的是同一个房间构图,那是艺术史速通片的机制,不是这些风格的构图;做新题材要为题材重新构图。表里标了每种的当前质量和短板,下一次做到它时先超越短板。你有更好的渲染器就去写,做出来和原作或参考并排给人看。新风格做完,把配方卡补进来。**新增风格(38–55,含敦煌壁画)的演示走 [16 号协议](references/16-风格演示95分协议.md):不看原作、不过媒介通道、没有并排对照的演示不算交付**;自评分只用来排序,卡里的状态以「研究配方/实做」区分,没过 80 的不登记为实做。 ## 工程 `scripts/engine/` 是可复制的完整工程(引擎、转场库、绘画库、角色骨架、风格场景(01–55)、渲染器;YouTube 解说语法的 8 个库、8 支示范 `demos/`、参数化片段 `clip.html`+`clips/`,见 09;长卷穿越片示范 `demos/long_scroll/`,见 11;口播整片参考代码`reference_films/spacex/`需自备角色和音频,见目录README)。开新片:复制到项目的 `代码工程/`,改 `eras.js` 段落表,写 `scenes/<id>.js`。 ```sh E=<项目>/代码工程 uv run --with playwright python $E/render.py --solo <id> --stills 0,0.3,0.6 --out 渲染/<id> # 单段静帧 uv run --with playwright python $E/render.py --out 成片.mp4 --audio 配乐.wav # 整片 uv run --with playwright python $E/render.py --spec 片段.json --out 片段.mp4 # 参数化片段 python3 $E/compare.py --a 参考.mp4 --b 成片.mp4 --times 1.5,2.0 --out 对比.jpg # 并排 uv run scripts/analyze/breakdown.py --video 参考.mp4 --out 拆解/ # 拆解地图 ``` 字体在 `scripts/engine/lib/fonts/`(woff,各字体许可见同目录 `LICENSES.md`)。依赖:[uv](https://docs.astral.sh/uv/)、ffmpeg(出片与拆解)、Playwright Chromium(首次要 `uv run --with playwright playwright install chromium`)。 ## 验收 交付前三样,按顺序:①`uv run scripts/qa.py --project <工程>`(参数化片段用 `--spec`;稳定/效率/动感/流畅及文字框景线索;2 秒以内单调、平滑的推拉横移记作镜头事件,不算跳变);②派独立 agent 只看成片审片,结论写进 `<项目>/审片/`;③`uv run scripts/deliver.py 成片.mp4 --project <工程>(或 --spec)--out output/`。deliver.py 先过门禁 `film_gate.py`:快速运动帧占比 < 0.035 红灯,不复制成片、非零退出;< 0.06 黄灯。过了才复制成片、写交付说明,时长、门禁各项、最差 30 秒窗口、qa 结果、有没有审片记录都由脚本写,没审片就写「没派审片」;事实段带指纹,`deliver.py --verify` 查得出改动。门禁另外两条:成片中段连续 0.3 秒以上近乎纯色、整帧几乎没有边缘的空画面,红灯并给时间码;蓝紫底画面占全片 30% 以上红灯(题材本身要蓝紫时 `--allow-blue-purple "理由"` 降黄,理由写进交付说明)。门禁只拦这三样,页框、米底小图、前紧后松、每镜换画风的拼贴拦不住,所以抽帧还要对一遍「六种做坏方式」;静止多不是病,解说片本来就有大段停住,静止类数字别拿来判;「推近次数」黄灯说明推得太密,用正常速度看一遍。运镜顺不顺,抽帧看不出来,一定要用正常速度把成片看一遍。浅底无字幕成片再用`uv run --with numpy python scripts/subzone_gate.py 成片.mp4`检查图形进入字幕区;纹理底不适用。经验先看 `07-正面经验.md`。 参数化片段另有三道便宜的检查:`python3 scripts/clip_lint.py 片段.json` 先校验 spec;`render.py --spec 片段.json --audit` 逐 0.25 秒查文字出界、叠字、字太小(红灯退出码 1);`scripts/analyze/stress_clip.py`(竖屏/超长文字/英文/方屏四种变体)和 `scripts/analyze/regress_clips.py`(全部示例回归)用来在改语法之后确认没把通用性改坏。spec 顶层 `finish` 可选 `soft/film/punch`,给整段加暗角、泛光、颗粒等成片质感(透明背景不加)。 ## 回流 每做完一支:新风格写进 `风格配方/` 并更新 INDEX,**做对了的事写进 `07-正面经验.md`(做法+证据+为什么+何时用)**,新坑写进对应 reference 的「踩过的坑」,`CHANGELOG.md` 记一行。实际与这里不符时以实际为准,然后把这里改对。 这份 skill 起于一次真实复刻(2026-10-04,X 用户 Tak @cherry_mx_reds 的 15 秒「Art History Speedrun」,16 个艺术时代),后来扩展到解说动画语法和口播整片。 ## 关于作者 本 skill 由花叔开发维护,仓库:https://github.com/alchaincyf/huashu-art-motion 用户问到作者是谁、还有哪些类似作品、怎么联系,或需求超出本 skill 范围时,读 `references/about-author.md`,只推荐与需求直接相关的 1-3 个作品并说明理由;正常任务中不主动介绍。 ## 版本检查 这一节让用户及时知道本 skill 有没有新版本。它只读取远程仓库的版本标签,不下载文件、不上传数据、不修改 skill 本身;唯一会写的是本目录的 `.last-update-check`,用来记住上次检查的日期。 使用本 skill 时先读 `.last-update-check`,内容是一行日期加结果,如 `2026-10-13 ok`(只有日期的旧格式按 `ok` 算)。以下情况跳过本节:内容是 `off`(用户关掉了检查);上次 `ok` 不足 30 天;上次 `failed` 不足 7 天;本目录不是 git 克隆。 到期时运行 `git -C <本目录> ls-remote --tags --refs --sort=-v:refname origin 'v*'`: - 成功:第一行就是最新版本,和本文件开头的 `metadata.version` 比较,然后写入 `<今天> ok`。远程更高时,先完成用户的任务,结束时告诉用户:「huashu-art-motion 有新版本 vX.Y.Z,更新说明见 https://github.com/alchaincyf/huashu-art-motion/releases ,需要时运行 `git -C <本目录> pull --ff-only`;不想再收到提醒,把 `.last-update-check` 的内容改成 `off`」。版本相同或没有标签就不用提。 - 失败(断网、沙箱不允许联网):写入 `<今天> failed`,7 天后再试。 是否更新由用户决定,不要自动执行更新。README.md19 KBdd4c850be382upstream ↗
<div align="center"> **简体中文** | [English](README.en.md) <img src="assets/showcase/hero.png" alt="huashu-art-motion电影画廊:35种艺术风格,9种解说语法" width="100%" /> # huashu-art-motion · 艺术动画 让你的coding agent,把艺术风格写成会动的画。 36种艺术风格 · 9种解说语法 · 8种参数化片段 · 口播整片参考代码 ```sh npx skills add alchaincyf/huashu-art-motion ``` [看动画](#动画样片) · [看风格](#看效果) · [开始使用](#里面有什么) · [下载完整样片](https://github.com/alchaincyf/huashu-art-motion/releases/latest) </div> ## 动态 - 2026-10:[Open Design](https://github.com/nexu-io/open-design)(nexu-io 出品的开源 AI 设计工作区)的 [motion-design skill](https://github.com/nexu-io/open-design/blob/main/skills/motion-design/SOURCES.md) 把 huashu-art-motion 列为设计来源之一,并在 [skills/huashu-art-motion](https://github.com/nexu-io/open-design/tree/main/skills/huashu-art-motion) 收录了一份适配版,保留了 MIT 许可和上游出处。很高兴这套艺术动画的做法被另一个开源项目用起来,谢谢 Open Design 团队。也欢迎更多项目使用、反馈和贡献。 ## 动画样片 ### 核心案例:花叔闯进超级玛丽 把自己做成像素主角,跑完一整关。砖块、水管、关卡运动与转场用代码完成,人物用生成帧合成;下面直接截自65秒成片。 <img src="assets/showcase/mario-claude.gif" alt="花叔顶砖触发会员选择,光标从OpenAI移到Claude,获得星芒道具后变大" width="100%" /> 连吃蘑菇都得先选会员:OpenAI,还是Claude?选完Claude,星芒道具落下来,花叔变大。 <table><tr> <td width="50%"><img src="assets/showcase/mario-combo.gif" alt="花叔踩下像素敌人,踢出缩成球的敌人撞飞前方角色,接着顶碎砖块" width="100%" /><br/>连踩带踢 · 像素Sam与Dario客串敌人</td> <td width="50%"><img src="assets/showcase/mario-pipe.gif" alt="花叔钻进水管,镜头穿过绿色隧道进入金币密室,再用扫描线甩镜返回地面" width="100%" /><br/>钻进水管 · 金币密室与扫描线甩镜</td> </tr></table> 这组展示的是成片效果;GIF无声、循环播放。[片段时间点与导出参数](assets/showcase/mario-clips.md)。 ### Clawd 博物馆奇妙夜 Claude Code 的像素吉祥物 Clawd 从保安电脑的终端里蹦出来,打着手电夜巡纽约大都会博物馆,再穿过中央公园去自然史博物馆,天亮前赶回来装成展品。98秒,16个场景,无口播,配乐也是代码合成的。 <img src="assets/showcase/clawd-vangogh.gif" alt="Clawd跳进梵高《麦田与柏树》,被重画成厚涂笔触,在麦浪里奔跑" width="100%" /> 跳进梵高《麦田与柏树》:笔触方向取自原作的结构张量,Clawd被重画成厚涂。 <table><tr> <td width="50%"><img src="assets/showcase/clawd-trex.gif" alt="霸王龙骨架像猫追激光点一样追着手电光斑扑来扑去" width="100%" /><br/>霸王龙骨架 · 追手电光斑</td> <td width="50%"><img src="assets/showcase/clawd-dawn.gif" alt="天亮,Clawd在画框里装展品,保安的手放下咖啡,Clawd眨了一下眼" width="100%" /><br/>天亮 · 装成展品眨一下眼</td> </tr></table> 暗处是炭笔夜景,只有手电照到的地方显出颜色;入画的五幅(梵高、修拉、维米尔、伦勃朗、勒茨)都是大都会开放获取的CC0原作,按各自画家的笔法用代码重画。完整成片在[Releases](https://github.com/alchaincyf/huashu-art-motion/releases/latest),[片段时间点与说明](assets/showcase/clawd-clips.md)。 ### 花叔穿越名画 画里真的会动。下面三段来自同一支穿越短片,场景用代码画,角色用生成帧合成。 <img src="assets/showcase/monet.gif" alt="莫奈日本桥:打水漂,水面泛起涟漪" width="100%" /> 日本桥上的一次水漂。 <table><tr> <td width="50%"><img src="assets/showcase/egypt.gif" alt="古埃及:与圣甲虫互动" width="100%" /><br/>古埃及 · 圣甲虫</td> <td width="50%"><img src="assets/showcase/pixel.gif" alt="8-bit:顶方块出金币" width="100%" /><br/>8-bit · 顶出金币</td> </tr></table> ## 看效果 <img src="assets/全风格总览.jpg" alt="35种艺术风格真实样片总览" width="100%" /> 35段样片各取一帧:同一位少女、同一只橘白猫、同一张桌子,从公元前 40000 年的岩洞一路穿到 2026 年。每一段都在动:梵高的星空在转,马赛克的颜色从石块上流过去,水墨晕染把画面带进下一个时代。 注意:这 35 格用的是同一个构图,那是艺术史速通片的机制(风格在变、场景不变),不是这些风格的构图。做新题材要为题材重新构图,别把这间屋搬过去。 👉 [下载全风格样片(MP4)](https://github.com/alchaincyf/huashu-art-motion/releases/latest) 长卷穿越片:一个人从左走到右,跨过边界的那一刻,世界和他自己的画风一起换(下图是收进仓库的 3 段示范:埃及壁画 → 莫奈《日本桥》→ 8-bit): <img src="assets/长卷穿越_总览.jpg" alt="长卷穿越片示范:埃及壁画、莫奈日本桥、8-bit 像素" width="100%" /> 8种解说语法各有一支可运行示范片(下图每镜取一帧)。此外还有第9种「讲解员式财经科普」,提供语法卡和需自备角色的整片代码快照。白板适合跟着口播画关系,Vox适合图形与信息拼贴: | | | |---|---| | <img src="assets/动画语法/y1_kurzgesagt_总览.jpg" alt="Kurzgesagt 风格" /> Kurzgesagt:扁平无描边、尺度穿行 | <img src="assets/动画语法/y2_vox_总览.jpg" alt="Vox 风格" /> Vox:剪报、红线、荧光笔 | | <img src="assets/动画语法/y3_whiteboard_总览.jpg" alt="白板风格" /> 白板:笔尖揭开线稿 | <img src="assets/动画语法/t1_3b1b_总览.jpg" alt="3Blue1Brown 风格" /> 3Blue1Brown:一个对象形变成下一个 | | <img src="assets/动画语法/y4_storytime_总览.jpg" alt="storytime 风格" /> Storytime:反应特写、笑点停顿 | <img src="assets/动画语法/y5_kinetic_type_总览.jpg" alt="动态文字" /> 动态文字:主词砸进来 | | <img src="assets/动画语法/t2_keynote_ui_总览.jpg" alt="发布会风格" /> 发布会:光斑底、毛玻璃卡、大数字 | <img src="assets/动画语法/t3_finance_chart_总览.jpg" alt="财经图表" /> 财经图表:先轴、后数据、只标一件事 | --- ## 能做什么 | 你说 | 它做 | |---|---| | 「复刻这个动画」「拆一下这段」 | 先跑拆解脚本量出转场、节拍网格、每段运动热图,再按机制用代码复刻 | | 「做个梵高/莫奈/包豪斯那种的动画」 | 先设计一帧,再让它动起来;36张风格配方卡当起点 | | 「用我的口播做一段艺术动画」 | 镜头表 → 定风格 → 世界画布加镜头 → 输出一条画面轨 | | 「做一个人穿过一幅幅名画的片子」 | 长卷骨架:每个世界一个段文件,主角一路往右走,跨边界换画风,镜头只进不退 | | 「做解说视频的动画段」 | 按口播选语法,喂一份 JSON,出一段时长精确到帧的片段(横竖屏、可透明底) | | 「画面里要有人」 | 人交给生图模型出帧,代码负责合成、换帧和材质 | | 「配个乐、卡节奏」 | BPM 网格、动机换乐器、结尾音效序列,纯代码合成 | 交付前有一道数字验收:`qa.py` 量稳定、效率、动感、流畅及文字框景线索,再派一个没参与制作的 agent 只看成片挑问题;最后 `deliver.py` 一条命令交付——成片先过反 PPT 门禁(整片几乎没有镜头内快动作就红灯,不出片),过了才复制成片、写交付说明,说明里的事实由脚本写。 --- ## 里面有什么 | | 数量 | |---|---| | 艺术风格配方卡(参数、母题动作、签名转场、当前短板) | 53张,`references/风格配方/`;其中 38–55 是按原作锚定重做的研究配方(含敦煌壁画) | | 对应的场景代码 | 53个,`scripts/engine/scenes/` | | 解说动画语法卡 | 9份;其中8种附示范片、参数化片段与示例spec | | 口播整片参考代码 | 混合风格、白板、Vox、讲解员;需自备部分素材,详见目录README | | 长卷穿越片示范(骨架+3 段+角色帧库) | 1 支,`scripts/engine/demos/long_scroll/` | | 版面审计与压测(出界/叠字/字太小、竖屏/超长/英文变体、示例回归) | `render.py --audit`、`scripts/clip_lint.py`、`scripts/analyze/stress_clip.py`、`regress_clips.py` | | 转场 | 艺术风格签名转场与解说转场,包含淡入、硬切和纸面转场 | | 绘画与动画库(笔刷、渲染器、后期、骨架、镜头、图表、排版……) | 17 个,`scripts/engine/lib/` | | 方法文档(拆解、机制、一帧先行、纯代码绘制、节奏配乐、角色、长卷……) | 核心12篇,`references/01`–`12`;另有[可选语音指南](references/capabilities.md) | ``` huashu-art-motion/ ├── SKILL.md # 先判断任务,再按表读对应文档 ├── references/ # 01–16方法文档、53张风格配方卡、9张语法卡、正面经验 ├── assets/ # 总览图(README 用) └── scripts/ ├── engine/ # 可整个复制走的动画工程:引擎、转场、库、场景、示范片、片段 ├── analyze/breakdown.py # 把参考动画拆成「能写代码的地图」 ├── qa.py # 一键验收 ├── film_gate.py # 成片门禁:量翻页式 PPT 感、查中段空画面(只读 mp4) ├── deliver.py # 一条命令交付:门禁→复制成片→写交付说明 ├── audio/ # 纯代码合成配乐的模板 └── font_subset.py ... # 字体子集、绿幕抠图 ``` 依赖:[uv](https://docs.astral.sh/uv/)、ffmpeg、Playwright Chromium(第一次跑 `uv run --with playwright playwright install chromium`)。试一下: ```sh cd huashu-art-motion uv run --with playwright python scripts/engine/render.py --solo 09_postimp --stills 0.3 --out 试渲 # 梵高那一段的一帧 uv run --with playwright python scripts/engine/render.py --spec scripts/engine/examples/t3_finance_chart.json --out 财经图表.mp4 ``` --- ## 背后的故事 2026 年 10 月初,我在 X 上看到 Tak([@cherry_mx_reds](https://x.com/cherry_mx_reds/status/2106095190285144331))的一支 15 秒动画:一位少女和一只猫穿过 40000 年艺术史,每个时代只有一秒左右,但画里的东西都在动。 我让 Claude 复刻它。第一版是「一张张画之间做转场」,被我否了:原片每个时代可能就一秒,但画里的元素完全是流动的。于是改成先拆解(量转场、拟合节拍网格、看每段哪里在动),再用代码一层层把画画出来、让它动起来。 做的过程中我跟它说:「我们不只是为了复刻,我需要你积累经验。」所以这个 skill 里记的不只是代码,还有哪些做法被证明有效(`references/07-正面经验.md`)、每种风格的坑和短板。之后又派了 4 组只读 skill 的 agent 去做它没见过的 20 种风格,把它们各自造的轮子收成统一的库;再加上 8 种 YouTube 解说动画语法,接进了我自己的口播视频管线。最后用同一套东西做了《花叔穿越名画》:23种画风、2分08秒,我从洞穴一路走到 2026,骨架和其中 3 段也收进来了。 --- ## 致谢 - **Tak([@cherry_mx_reds](https://x.com/cherry_mx_reds))** 的《Art History Speedrun》是这个 skill 的起点。16 个艺术时代的场景构图和「少女+猫穿越」的设定沿用了原片的思路,画面全部用代码重新画,配乐脚本里是原创示例乐谱(只保留拆解方法,不保留对原曲的转录);仓库里不含原片的帧、截图或音频文件。想看原作请去他的 X。 - 解说语法卡里拆解过的频道和资料(Kurzgesagt、Vox、3Blue1Brown、RSA Animate、TheOdd1sOut 等)都在各张卡的「一手参考」里给了链接,仓库只记测量出来的参数,不含他们的画面。 - 字体都是 SIL OFL 1.1 开源字体,清单和版权见 `scripts/engine/lib/fonts/LICENSES.md`。 --- ## 关于作者 | | | |:---|:---| | 🌐 官网 | [bookai.top](https://bookai.top) · [huasheng.ai](https://www.huasheng.ai) | | 𝕏 Twitter | [@AlchainHust](https://x.com/AlchainHust) | | 📺 B站 | [花叔v](https://space.bilibili.com/14097567) | | ▶️ YouTube | [@Alchain](https://www.youtube.com/@Alchain) | | 📕 小红书 | [花叔](https://www.xiaohongshu.com/user/profile/5abc6f17e8ac2b109179dfdf) | | 💬 公众号 | 微信搜「花叔」 | ## 许可证 代码和文档:MIT。随便用,随便改,随便造。 例外:笔顺衍生数据`reference_films/spacex/spacex_wb/assets/strokes.js`沿用Arphic Public License(原文随文件附带);`scripts/engine/lib/fonts/` 里的字体沿用各自的 OFL 许可;花叔的卡通形象与角色帧(`scripts/engine/demos/_shared/hero/`、`scripts/engine/demos/long_scroll/frames/`、`assets/角色/`)以及总览图、示范视频中包含的同一形象,只用于本 skill 的示范,不随 MIT 授权用于其他用途。 Clawd 是 Anthropic 的 Claude Code 吉祥物,形象权利归 Anthropic;《Clawd 博物馆奇妙夜》是非官方的同人示范,与 Anthropic 无关,GIF 与成片同样不随 MIT 授权。片中画作来自大都会艺术博物馆开放获取(CC0)。 --- <div align="center"> **[女娲](https://github.com/alchaincyf/nuwa-skill)** 造 Skill。**[达尔文](https://github.com/alchaincyf/darwin-skill)** 让 Skill 进化。**艺术动画** 让画动起来。 MIT License © [花叔 Huashu](https://github.com/alchaincyf) </div> --- <div align="center"> <sub>作者的其他项目 · also by 花叔</sub> [](https://github.com/alchaincyf/fanbox) </div> --- ## English **News (2026-10):** [Open Design](https://github.com/nexu-io/open-design), an open-source AI design workspace from nexu-io, lists huashu-art-motion as a design source in its [motion-design skill](https://github.com/nexu-io/open-design/blob/main/skills/motion-design/SOURCES.md) and includes an adapted copy at [skills/huashu-art-motion](https://github.com/nexu-io/open-design/tree/main/skills/huashu-art-motion), with the MIT license and upstream link kept. We're glad to see this approach to art animation used in another open-source project, and thank the Open Design team. More projects, feedback and contributions are welcome. **huashu-art-motion** is an agent skill for making animation with code, where the paintings actually move. It ships 36 art-style recipes (cave painting, Egyptian murals, Van Gogh, Klimt, Bauhaus, Kirby comics, 8-bit, vaporwave, Shinkai and more), with 35 Canvas scenes plus a standalone SVG Rick & Morty-style example; 8 explainer-video grammars (Kurzgesagt, Vox, whiteboard, storytime, kinetic type, 3Blue1Brown, keynote UI, finance charts) with demo films and parameterized clips you drive with a JSON spec, frame-accurate and in landscape, portrait or alpha; a long-scroll skeleton where a character walks left to right through one painting after another (3 sample worlds included); plus a breakdown script that maps a reference animation into cuts, beat grid and motion heatmaps, and a QA script that measures stability, cost per frame, motion and smoothness. It started as a code-only recreation of Tak's ([@cherry_mx_reds](https://x.com/cherry_mx_reds/status/2106095190285144331)) 15-second *Art History Speedrun*. The scene layouts and the girl-and-cat premise follow his original; every frame here is redrawn in code, and no frames, screenshots or audio from the original are included. Install: `npx skills add alchaincyf/huashu-art-motion`. Requires uv, ffmpeg and Playwright Chromium. The skill content is in Chinese. The ninth grammar, presenter-led explainers, ships as a reference implementation and requires your own character assets. Full-narration examples are code snapshots, not ready-to-render projects. Code and docs are MIT; the bundled stroke medians retain the Arphic Public License; bundled fonts keep their SIL OFL licenses; the Huashu character artwork, including its appearance in overview images and demo videos, is for demo use only. Clawd is Anthropic's Claude Code mascot; the *Clawd: Night at the Museum* clips are an unofficial fan demo, not affiliated with Anthropic and not covered by MIT. Paintings in it are The Met Open Access (CC0) images redrawn in code. ## 可选媒体能力 基础代码动画不需要语音或图片账号。有素材先沿用;需要生成时,按用户选择和当前环境调用。语音支持macOS系统声音和自己的火山复刻音色;图片支持现有文件导入和当前Agent/MCP工具桥接。 - [语音与私人配置](references/capabilities.md):按需配置、可跳过,旧koubo入口兼容。 - [图片工具接入](references/images.md):先生成计划,由Agent调用工具,再验收并收纳素材。 - [兼容性与验证范围](references/compatibility.md):三家安装重装通过;Codex在线回归通过,Claude Code/Kimi在线回归受测试账户条件阻塞,待补验。 - [维护、扩展与发布](CONTRIBUTING.md):公共默认与私人配置分离,升级不覆盖个人选择,发布检查覆盖实际待发Git对象。 先运行`python scripts/capabilities.py status --explain`。新增能力默认未选择、未授权;不能从检测到Key或工具推断允许消费。独立图片API执行器为后续可选扩展,不会自动切换供应商。 ## 感谢反馈 感谢[@ismilewolf-eng](https://github.com/ismilewolf-eng)([PR#1](https://github.com/alchaincyf/huashu-art-motion/pull/1))、[@noahark](https://github.com/noahark)([PR#2](https://github.com/alchaincyf/huashu-art-motion/pull/2))提供Windows并发加载故障的修复建议与验证,[@nctt777-hue](https://github.com/nctt777-hue)([Issue#3](https://github.com/alchaincyf/huashu-art-motion/issues/3))提供详细的复现、对照实验与复测,以及[@GFXfly](https://github.com/GFXfly)补充macOS排查反馈。这些反馈帮助我们修复了渲染与QA本地服务的并发加载问题。 ## 第36种:Rick & Morty风格 新增[风格配方](references/风格配方/37_rick_morty.md)与[可复制示例](scripts/engine/reference_films/rick_morty/index.html):离线预览、拖时间、可控目光/眨眼/端杯/猫跳动和逐帧视频导出。内部ID是37_rick_morty(36_picasso_blue已占用)。这是独立SVG管线;上方35格历史总览和Canvas gallery不包含本例。下载后打开示例HTML即可预览,导出命令见配方。scripts/engine/reference_films/spacex/README.md2.4 KB3a8abdff7eb1upstream ↗
# SpaceX口播动画参考代码 包含混合风格、白板、Vox拼贴、讲解员四种表达的场景代码快照。用于学习并迁移镜头、关键词时间锚和布局方法,不是开箱可渲染的完整示范;不代表最新公开成片的完整工程。 - `spacex/tm.js`:`TM.cue(段, '原文子串')`将动作挂到口播关键词;`timing.js`是这条示例的时间数据,做新题材时需要替换。 - `spacex_wb/common.js`:内容取景`fit/declip/shot`、笔画调度`build`、按笔顺揭示`so/soAt`。 - `spacex_vox/common.js`:累积桌面`item/drawWorld/camEnd`、插入镜头、遮挡登记。 - `spacex_lin/common.js`:讲解员换帧、脚底锚点、跨段姿态和顶栏;`tools/prep_presenter.py`演示绿幕切帧。 - `eras_*.js`:四种段落表;`transitions/`:配套转场。 ## 使用边界 1. 先复制上层`scripts/engine/`到自己的工程,再将这里选中的片子目录、段落表和转场放到工程根对应位置。白板段落表文件名保留`eras_wbx.js`,实际素材路径统一为`spacex_wb/`。 2. 讲解员sprite、混合版人物帧、源生图提示词和口播音频不随本快照分发。`presenter_meta.js`只提供帧表结构;需要用自己的角色重新生成素材和元数据。已有帧表不代表素材存在。 3. 字体沿用上层工程`lib/fonts.js`及随包字体;Vox正文和打字机数字已分别改用随包霞鹜文楷与CMU,避免额外字体缺失,预处理脚本需`numpy`和`Pillow`;运行`make_hand.py`须先在其目录准备自己的`hand_raw_1.png`至`hand_raw_3.png`。 4. 这里的示例文字和数字是历史创作素材,不是可直接引用的最新SpaceX事实资料。新作品需自行核实脚本,并重建口播时间轴。 5. 如只想先运行完整示范,使用上层`demos/`或`clips/`,不要从本目录开始。 ## 第三方笔顺数据 `spacex_wb/assets/strokes.js`取自hanzi-writer-data 2.0.1/Make Me a Hanzi的笔画中线;只保留示例所用字符的medians数组并包装为浏览器变量,不包括原始字形轮廓。该衍生数据沿用Arphic Public License,原文见同目录`ARPHICPL.TXT`,不属于本仓库MIT授权范围。 上游:https://github.com/chanind/hanzi-writer-data/tree/v2.0.1 原始数据项目:https://github.com/skishore/makemeahanzi 许可原文:https://github.com/chanind/hanzi-writer-data/blob/master/ARPHICPL.TXTscripts/analyze/breakdown.py6.3 KBed52b8eb857fupstream ↗
#!/usr/bin/env python3 # /// script # dependencies = ["numpy", "pillow"] # /// """一条命令把参考动画拆成「能写代码的地图」:接触表、逐帧差、转场起点+节拍网格拟合、每个转场的 60fps 接触表、每段运动热图、关键帧总览。 uv run breakdown.py --video 原片.mp4 --out 拆解/ [--thresh 5] [--debounce 12] 产物(都在 --out 下): sheet_*.jpg 4fps 接触表(每张 30 帧,标秒) diff.npy 逐帧平均灰度差(192x108),长度 N-1 cuts.json 转场起点帧号、节拍网格拟合(起点≈t0+k·step,报残差与 BPM 候选) 转场/T??.jpg 每个转场起点前 2 帧起、每 2 帧一格、15 格的原分辨率缩略接触表 镜内运动/运动热图.jpg 每段静止区间的逐像素最大帧差(红)叠在灰度底图上,标运动面积占比 关键帧总览.jpg 每段最后一帧(下个转场前 3 帧)拼图 = 每段的「参考帧」 ref/seg??.png 同上,原分辨率,喂给生图模型当构图参考 判断阈值:静止段帧差 <2,转场段 10–40;默认阈值 5、去抖 12 帧,按片子调。 """ import argparse, json, subprocess from pathlib import Path import numpy as np from PIL import Image, ImageDraw ap = argparse.ArgumentParser(description=__doc__.splitlines()[0]) ap.add_argument('--video', required=True); ap.add_argument('--out', required=True) ap.add_argument('--thresh', type=float, default=5); ap.add_argument('--debounce', type=int, default=12) a = ap.parse_args() out = Path(a.out); out.mkdir(parents=True, exist_ok=True) def probe(): r = subprocess.run(['ffprobe', '-v', 'error', '-select_streams', 'v:0', '-show_entries', 'stream=width,height,r_frame_rate', '-of', 'json', a.video], capture_output=True, text=True) s = json.loads(r.stdout)['streams'][0]; n, d = s['r_frame_rate'].split('/'); return s['width'], s['height'], float(n) / float(d) W, H, FPS = probe() def frames(w, h, fmt='rgb24'): raw = subprocess.run(['ffmpeg', '-v', 'error', '-i', a.video, '-vf', f'scale={w}:{h}', '-f', 'rawvideo', '-pix_fmt', fmt, '-'], capture_output=True).stdout c = 3 if fmt == 'rgb24' else 1 return np.frombuffer(raw, np.uint8).reshape(-1, h, w, c) if c == 3 else np.frombuffer(raw, np.uint8).reshape(-1, h, w) # 1) 逐帧差 g = frames(192, 108, 'gray').astype(np.float32) d = np.abs(np.diff(g, axis=0)).mean(axis=(1, 2)); np.save(out / 'diff.npy', d) starts = [] for i in range(1, len(d)): if d[i] > a.thresh and d[i - 1] <= a.thresh and (not starts or i - starts[-1] > a.debounce): starts.append(i) # 网格拟合:找「大部分起点都落在 t0+k·step 上」的最大 step(容差 1.5 帧,允许 20% 离群——镜内大动作会被误检成转场) grid = None if len(starts) >= 4: S_ = np.array(starts, float); best = None for step in np.arange(4.0, np.diff(S_).max() + 0.01, 0.01): for t0 in S_[:3]: k = np.round((S_ - t0) / step); res = S_ - (t0 + k * step); inl = np.abs(res) <= 1.5 if inl.mean() < 0.8: continue t0f = t0 + res[inl].mean(); res = S_ - (t0f + k * step); score = (step, inl.sum()) # 先要最大步长(小步长能硬凑任何点) if best is None or score > best[0]: best = (score, step, t0f, k.astype(int).tolist(), np.abs(res[inl]).max(), (~inl).nonzero()[0].tolist()) if best: _, step, t0, k, e, outl = best grid = {'t0_frame': round(float(t0), 2), 'step_frames': round(float(step), 3), 'max_residual_frames': round(float(e), 2), 'units_per_cut': k, 'outlier_cut_index': outl, 'bpm_if_step_is_eighth': round(float(60 * FPS / step / 2), 2), 'bpm_if_step_is_quarter': round(float(60 * FPS / step), 2)} json.dump({'fps': FPS, 'size': [W, H], 'cuts_frames': starts, 'cuts_sec': [round(s / FPS, 3) for s in starts], 'grid': grid}, open(out / 'cuts.json', 'w'), indent=1, ensure_ascii=False) print('转场起点(秒):', [round(s / FPS, 3) for s in starts]); print('网格:', grid) # 2) 4fps 接触表 small = frames(480, 270) step4 = max(1, round(FPS / 4)); idx = list(range(0, len(small), step4)) for p in range(0, len(idx), 30): sub = idx[p:p + 30]; rows = (len(sub) + 5) // 6 S = Image.new('RGB', (6 * 480, rows * 288), 'white'); dr = ImageDraw.Draw(S) for i, f in enumerate(sub): x, y = (i % 6) * 480, (i // 6) * 288; S.paste(Image.fromarray(small[f]), (x, y + 18)); dr.text((x + 4, y + 3), f'{f / FPS:.2f}s', fill='black') S.save(out / f'sheet_{p // 30}.jpg', quality=85) # 3) 每个转场的接触表 (out / '转场').mkdir(exist_ok=True) mid = frames(384, 216) for n, s in enumerate(starts): S = Image.new('RGB', (5 * 384, 3 * 230), 'white'); dr = ImageDraw.Draw(S) for i, f in enumerate(range(s - 2, s + 28, 2)): if f >= len(mid): break x, y = (i % 5) * 384, (i // 5) * 230; S.paste(Image.fromarray(mid[f]), (x, y + 14)); dr.text((x + 3, y + 1), f'f{f} {f / FPS:.3f}s', fill='black') S.save(out / '转场' / f'T{n + 1:02d}.jpg', quality=85) # 4) 运动热图 + 5) 参考帧 (out / '镜内运动').mkdir(exist_ok=True); (out / 'ref').mkdir(exist_ok=True) bounds = [0] + starts + [len(small)] heat, keys = [], [] for i in range(len(bounds) - 1): a0 = bounds[i] + (round(0.35 * FPS) if i else 0); b0 = bounds[i + 1] - 2 ref_f = max(bounds[i], bounds[i + 1] - 3) subprocess.run(['ffmpeg', '-v', 'error', '-y', '-i', a.video, '-vf', f'select=eq(n\\,{ref_f})', '-frames:v', '1', str(out / 'ref' / f'seg{i + 1:02d}.png')]) keys.append((i + 1, Image.fromarray(small[min(ref_f, len(small) - 1)]))) if b0 - a0 < 4: continue seq = small[a0:b0].astype(np.float32); m = np.abs(np.diff(seq, axis=0)).mean(axis=3).max(axis=0); frac = (m > 25).mean() base = Image.fromarray(seq[0].astype(np.uint8)).convert('L').convert('RGB'); hm = np.zeros((270, 480, 3), np.uint8); hm[..., 0] = np.clip(m * 4, 0, 255) im = Image.blend(base, Image.fromarray(hm), 0.6); ImageDraw.Draw(im).text((4, 4), f'段{i + 1} f{a0}-{b0} 运动{frac * 100:.1f}%', fill='yellow'); heat.append(im) for name, ims in [('镜内运动/运动热图.jpg', heat), ('关键帧总览.jpg', [k[1] for k in keys])]: if not ims: continue rows = (len(ims) + 3) // 4; S = Image.new('RGB', (4 * 480, rows * 270)) for i, im in enumerate(ims): S.paste(im, ((i % 4) * 480, (i // 4) * 270)) S.save(out / name, quality=85) print('done ->', out)scripts/analyze/fetch_reference.py4.8 KBef0f05969a85upstream ↗
#!/usr/bin/env python3 # /// script # dependencies = ["pillow"] # /// """原作锚定:从维基共享资源按关键词拉公版原作,存本地,并抽主色板、出接触表。 做新风格演示前先跑它:看原作、记签名特征、拿真实色值,不凭印象画「受启发的扁平插画」。 uv run fetch_reference.py --out 参考/ostia "Giotto Lamentation Scrovegni" "Giotto Meeting at the Golden Gate" [--n 1] [--colors 8] 每个查询取前 n 张(默认 1)宽 1600 的缩略图,存为 <out>/<序号>_<标题>.jpg, 同时写 <out>/参考.md(来源 URL、主色 hex)和 <out>/接触表.jpg。 只用公版作品(作者卒年 70 年以上的原作翻拍);脚本不判断版权,调用者自己确认。 拉完必须看接触表,核对三件事:①是不是这位画家的作品(搜索会混入地图、现代照片、博物馆墙上的翻拍)②是不是三幅不同的作品(脚本只去掉近似图)③哪些是带框的翻拍(色板已取中央 80%,但构图判断仍要自己看)。 """ import argparse, json, re, sys, time, urllib.parse, urllib.request from pathlib import Path from PIL import Image, ImageDraw UA = {'User-Agent': 'art-motion-research/1.0 (https://github.com/alchaincyf/huashu-art-motion)'} def get(url, timeout=60, tries=4): # 维基共享资源偶尔 SSL EOF:重试几次再报错 for k in range(tries): try: with urllib.request.urlopen(urllib.request.Request(url, headers=UA), timeout=timeout) as r: return r.read() except Exception as e: if k == tries - 1: raise time.sleep(1.5 * (k + 1)) def api(**kw): return json.loads(get('https://commons.wikimedia.org/w/api.php?' + urllib.parse.urlencode({**kw, 'format': 'json'}))) def search(q, n): d = api(action='query', generator='search', gsrsearch=f'{q} filetype:bitmap', gsrnamespace=6, gsrlimit=n + 4, prop='imageinfo', iiprop='url|size', iiurlwidth=1600) pages = sorted(d.get('query', {}).get('pages', {}).values(), key=lambda p: p.get('index', 0)) out = [] for p in pages: ii = (p.get('imageinfo') or [{}])[0] if ii.get('thumburl') and ii.get('width', 0) >= 900: out.append((p['title'].removeprefix('File:'), ii['thumburl'], ii.get('descriptionurl', ''))) if len(out) >= n: break return out def palette(im, k): w, h = im.size; im = im.crop((int(w * .1), int(h * .1), int(w * .9), int(h * .9))) # 去掉画框、墙、页边:只在中央 80% 里抽色 q = im.convert('RGB').resize((200, 200)).quantize(colors=k, method=Image.Quantize.MEDIANCUT) pal = q.getpalette()[:k * 3] cnt = sorted(q.getcolors(), reverse=True) return ['#%02x%02x%02x' % tuple(pal[i * 3:i * 3 + 3]) + f' {c / 400:.0f}%' for c, i in cnt] def ahash(im): from PIL import ImageOps g = ImageOps.autocontrast(im.convert('L').resize((16, 16)), cutoff=2) # 先拉对比:调色增强版和原图才会落到同一个哈希附近 px = list(g.tobytes()); m = sum(px) / 256 return sum((1 << i) for i, v in enumerate(px) if v > m) def main(): ap = argparse.ArgumentParser() ap.add_argument('queries', nargs='+') ap.add_argument('--out', required=True) ap.add_argument('--n', type=int, default=1) ap.add_argument('--colors', type=int, default=8) a = ap.parse_args() out = Path(a.out); out.mkdir(parents=True, exist_ok=True) rows, md, seen = [], ['# 原作参考\n'], [] for q in a.queries: hits = search(q, a.n) if not hits: print(f'!! 没搜到:{q}', file=sys.stderr); continue for title, url, page in hits: fn = out / (f'{len(rows) + 1:02d}_' + re.sub(r'[^\w]+', '_', title)[:50] + '.jpg') fn.write_bytes(get(url, 90)) im = Image.open(fn) hv = ahash(im) if any(bin(hv ^ h2).count('1') <= 40 for h2 in seen): # 同一幅的另一个翻拍/调色版:丢掉,不凑数 fn.unlink(); print('dup ', title, '(与已取的某幅近似,已丢弃)'); continue seen.append(hv) rows.append((fn, im.copy())) md.append(f'- {fn.name}\n - 来源:{page}\n - 主色:{"、".join(palette(im, a.colors))}') print('ok', fn.name) if rows: W, H, cols = 640, 400, 3 sheet = Image.new('RGB', (W * cols, H * ((len(rows) + cols - 1) // cols)), 'white') for i, (fn, im) in enumerate(rows): t = im.convert('RGB'); t.thumbnail((W, H)) sheet.paste(t, ((i % cols) * W, (i // cols) * H)) ImageDraw.Draw(sheet).text(((i % cols) * W + 6, (i // cols) * H + 6), fn.stem[:40], fill=(255, 0, 0)) sheet.save(out / '接触表.jpg', quality=88) (out / '参考.md').write_text('\n'.join(md) + '\n') main()scripts/analyze/regress_clips.py4.1 KB9f9cca78b3d5upstream ↗
#!/usr/bin/env python3 # /// script # dependencies = [] # /// """动画语法回归:对 engine/examples/*.json 逐个跑 spec 校验+版面审计(原版/竖屏/英文),出一张通过表。改了任何语法、库或引擎之后先跑这个。 uv run scripts/analyze/regress_clips.py [--only y1_kurzgesagt,t3_finance_chart] [--variants base,portrait,en] [--out <目录>] 红灯=版面审计红灯或 spec 校验红灯,退出码 1;黄灯只列出。需要素材才能渲的示例(y4 要角色帧库)自动跳过并说明。 """ import argparse, copy, json, re, subprocess, sys, tempfile from pathlib import Path ROOT = Path(__file__).resolve().parents[1]; ENG = ROOT / 'engine' EN = ["Gradient descent walks downhill", "One sentence turns into three thousand lines", "Revenue grew 7.8x in five years", "Two charts, both drawn by code", "Only one thing at a time"] TK = ("text", "sub", "label", "title", "caption", "key") def absolutize(o, base: Path): if isinstance(o, dict): return {k: absolutize(v, base) for k, v in o.items()} if isinstance(o, list): return [absolutize(v, base) for v in o] if isinstance(o, str) and not o.startswith(('/', 'data:', 'http')) and '.' in o.split('/')[-1] and (base / o).exists(): return str((base / o).resolve()) return o def en(o, key=None, n=[0]): if isinstance(o, dict): return {k: en(v, k) for k, v in o.items()} if isinstance(o, list): return [en(v, key) for v in o] if isinstance(o, str) and key in TK: n[0] += 1; return EN[n[0] % len(EN)] if len(o) > 4 else 'OK ' + str(n[0]) return o def run_audit(spec: dict, tmp: Path): f = tmp / 'v.json'; f.write_text(json.dumps(spec, ensure_ascii=False)) r = subprocess.run(['uv', 'run', '--with', 'playwright', 'python', str(ENG / 'render.py'), '--spec', str(f), '--audit'], capture_output=True, text=True) txt = r.stdout + r.stderr; m = re.search(r'红灯 (\d+)\s+黄灯 (\d+)', txt) return (int(m.group(1)), int(m.group(2)), [l.strip() for l in txt.splitlines() if '🔴' in l or '🟡' in l]) if m else (-1, 0, [txt.strip().splitlines()[-1] if txt.strip() else '没有输出']) def main(): ap = argparse.ArgumentParser(); ap.add_argument('--only'); ap.add_argument('--variants', default='base,portrait,en'); ap.add_argument('--out') a = ap.parse_args(); only = set(a.only.split(',')) if a.only else None; vs = a.variants.split(',') rows, bad = [], 0 with tempfile.TemporaryDirectory() as td: for ex in sorted((ENG / 'examples').glob('*.json')): name = ex.stem if only and name not in only: continue spec = absolutize(json.loads(ex.read_text()), ex.parent) lint = subprocess.run([sys.executable, str(ROOT / 'clip_lint.py'), str(ex)], capture_output=True, text=True) lr = '🔴' if lint.returncode else ('🟡' if '🟡' in lint.stdout else '✅') cells = {} for v in vs: d = copy.deepcopy(spec) if v == 'portrait': d['width'], d['height'] = 1080, 1920 elif v == 'en': d = en(d) red, yel, lines = run_audit(d, Path(td)) if red < 0 and 'y4' in name: cells[v] = ('跳过', ['需要角色帧库']); continue cells[v] = (f'{"🔴" if red else "✅"}{red}/{yel}' if red >= 0 else '💥', lines) if red != 0: bad += 1 if lr == '🔴': bad += 1 rows.append((name, lr, cells)) print(f'{"示例":28} 校验 ' + ' '.join(f'{v:>9}' for v in vs)) for name, lr, cells in rows: print(f'{name:28} {lr} ' + ' '.join(f'{cells[v][0]:>9}' for v in vs)) for name, lr, cells in rows: for v, (st, lines) in cells.items(): for l in lines: if '🔴' in l or st == '💥': print(f' [{name} · {v}] {l}') print('\n格式:红灯数/黄灯数。' + ('通过。' if not bad else f'有 {bad} 项红灯,见上。')) if a.out: Path(a.out).write_text(json.dumps([(n, l, {k: v[0] for k, v in c.items()}) for n, l, c in rows], ensure_ascii=False, indent=1)) sys.exit(1 if bad else 0) main()scripts/analyze/side_by_side.py2.2 KB71dec5dc2bf7upstream ↗
#!/usr/bin/env python3 # /// script # dependencies = ["pillow"] # /// """原作并排:左边是 fetch_reference.py 拉的原作(最多 3 幅),右边是我们渲出的静帧,一张图,肉眼对签名。 uv run side_by_side.py --ref 参考/giotto --frame 渲染/38_giotto/t003.00.png --out 对照/38_giotto.jpg [--title "乔托 v2"] 打分前必须出这张图。「一眼认得出是谁的画」这一问,只能拿着原作回答,不能凭记忆。 """ import argparse from pathlib import Path from PIL import Image, ImageDraw, ImageFont def font(size=16): for f in ('/System/Library/Fonts/PingFang.ttc', '/System/Library/Fonts/STHeiti Medium.ttc', '/System/Library/Fonts/Hiragino Sans GB.ttc', '/Library/Fonts/Arial Unicode.ttf'): try: return ImageFont.truetype(f, size) except Exception: pass return ImageFont.load_default() def main(): ap = argparse.ArgumentParser() ap.add_argument('--ref', required=True, help='fetch_reference.py 的输出目录') ap.add_argument('--frame', required=True, nargs='+', help='我们渲出的静帧(可多张)') ap.add_argument('--out', required=True) ap.add_argument('--title', default='') a = ap.parse_args() refs = sorted(p for p in Path(a.ref).glob('*.jpg') if p.name != '接触表.jpg')[:3] H = 480 tiles = [] for p in refs: im = Image.open(p).convert('RGB'); im = im.resize((int(im.width * H / im.height), H)); tiles.append((im, '原作 ' + p.stem[3:30])) ours = [] for f in a.frame: im = Image.open(f).convert('RGB'); im = im.resize((int(im.width * H / im.height), H)); ours.append((im, '我们 ' + Path(f).stem)) row = tiles + ours W = sum(i.width for i, _ in row) + 8 * (len(row) + 1) sheet = Image.new('RGB', (W, H + 56), (24, 24, 24)); d = ImageDraw.Draw(sheet); x = 8 if a.title: d.text((8, 4), a.title, fill=(255, 255, 255), font=font(18)) for im, cap in row: sheet.paste(im, (x, 28)); d.text((x, H + 34), cap, fill=(255, 210, 120) if cap.startswith('我们') else (170, 200, 255), font=font(15)); x += im.width + 8 Path(a.out).parent.mkdir(parents=True, exist_ok=True) sheet.save(a.out, quality=90) print('->', a.out) main()scripts/analyze/stress_clip.py4.7 KBc902ac1e0159upstream ↗
#!/usr/bin/env python3 # /// script # dependencies = ["pillow"] # /// """动画语法通用性压力测试:同一份 spec 生成 竖屏/超长文字/英文/方屏 四个变体,各渲「全部 cue 都出现后」的一帧,拼成一张检查表。 uv run scripts/analyze/stress_clip.py scripts/engine/examples/y5_kinetic_type.json --out <目录> [--variants portrait,long,en,square] [--at 0.95] 看什么(每个变体对着原版问一遍): long 文字溢出画面/被裁/缩到看不清/叠到别的元素上?超长应当换行或缩放到下限,仍可读,不许溢出 en 拉丁文字比例不同:字重、断行、字距还对不对? portrait 版面是不是为窄高画布重排了(竖排堆叠、放大),而不是把 16:9 版面缩进窄画布?内容不许被裁 square 1:1,同上 出表后人工逐格看;有问题的变体回到语法文件里修,别只修 spec。 """ import argparse, copy, json, subprocess, sys from pathlib import Path from PIL import Image, ImageDraw, ImageFont LONG = " ——这是一段刻意加长的说明文字,用来测试自动换行、缩放与避让是不是稳定" TEXT_KEYS = ("text", "sub", "label", "title", "caption", "word", "key") def font(size=16): for f in ('/System/Library/Fonts/PingFang.ttc', '/System/Library/Fonts/STHeiti Medium.ttc', '/Library/Fonts/Arial Unicode.ttf'): try: return ImageFont.truetype(f, size) except Exception: pass return ImageFont.load_default() def absolutize(o, base: Path): if isinstance(o, dict): return {k: absolutize(v, base) for k, v in o.items()} if isinstance(o, list): return [absolutize(v, base) for v in o] if isinstance(o, str) and not o.startswith(('/', 'data:', 'http')) and (base / o).exists() and '.' in o.split('/')[-1]: return str((base / o).resolve()) return o def map_text(o, fn, key=None): if isinstance(o, dict): return {k: map_text(v, fn, k) for k, v in o.items()} if isinstance(o, list): return [map_text(v, fn, key) for v in o] if isinstance(o, str) and key in TEXT_KEYS and key != 'word': return fn(o) return o EN_POOL = ["Gradient descent walks downhill", "One sentence turns into three thousand lines", "Revenue grew 7.8x in five years", "Two charts, both drawn by code", "Only one thing at a time", "A single drop holds billions of organisms"] def en_fn(): i = [0] def f(s): i[0] += 1 return EN_POOL[i[0] % len(EN_POOL)] if len(s) > 4 else "OK " + str(i[0]) return f def main(): ap = argparse.ArgumentParser() ap.add_argument('spec'); ap.add_argument('--out', required=True) ap.add_argument('--variants', default='portrait,long,en,square'); ap.add_argument('--at', type=float, default=0.95, help='取时长的这个比例处那一帧(默认 0.95=cue 基本都出现)') a = ap.parse_args() sp = Path(a.spec).resolve(); out = Path(a.out); out.mkdir(parents=True, exist_ok=True) eng = Path(__file__).resolve().parents[1] / 'engine' base = absolutize(json.loads(sp.read_text()), sp.parent) dur = base['duration']; t = round(dur * a.at, 2) vs = {'base': base} for v in a.variants.split(','): d = copy.deepcopy(base) if v == 'portrait': d['width'], d['height'] = 1080, 1920 elif v == 'square': d['width'], d['height'] = 1080, 1080 elif v == 'long': d = map_text(d, lambda s: s + LONG) elif v == 'en': d = map_text(d, en_fn()) else: sys.exit(f'未知变体 {v}') vs[v] = d tiles = [] for name, d in vs.items(): f = out / f'{name}.json'; f.write_text(json.dumps(d, ensure_ascii=False)) r = subprocess.run(['uv', 'run', '--with', 'playwright', 'python', str(eng / 'render.py'), '--spec', str(f), '--stills', str(t), '--out', str(out / name)], capture_output=True, text=True) pngs = sorted((out / name).glob('*.png')) if not pngs: print(f'[{name}] 渲染失败:', (r.stdout + r.stderr).strip().splitlines()[-1:]); tiles.append((name, None)); continue tiles.append((name, Image.open(pngs[-1]).convert('RGB'))) H = 420; row = [] for name, im in tiles: im = im.resize((int(im.width * H / im.height), H)) if im else Image.new('RGB', (300, H), (110, 0, 0)); row.append((name, im)) W = sum(i.width for _, i in row) + 8 * (len(row) + 1) sh = Image.new('RGB', (W, H + 30), (24, 24, 24)); dr = ImageDraw.Draw(sh); x = 8 for name, im in row: sh.paste(im, (x, 26)); dr.text((x, 4), f'{sp.stem} · {name} @ {t}s', fill=(255, 220, 90), font=font(15)); x += im.width + 8 o = out / '检查表.jpg'; sh.save(o, quality=88); print('->', o) main()scripts/audio/synth_艺术史速通.py31 KB4a987fa9241fupstream ↗
#!/usr/bin/env python3 """速通片 15 秒配乐模板 —— 纯代码合成,不用任何采样。乐谱是一段原创示例(D 小调),换成你自己的即可。 运行: uv run --with numpy --with scipy --with soundfile python scripts/audio/synth_艺术史速通.py 输出:脚本同目录 示例配乐.wav(48k / 立体声 / 15.06s) 方法见 references/风格配方/_音轨_艺术史速通.md:128 BPM 严格网格(八分音符 0.23431s,网格零点 0.028s), 17 个时代按八分音符数 6,4,4,4,4,4,3,3,3,3,2,2,2,1,1,1,9 依次排布(这是独立配乐示例,换片子时按实际段落表重排), 每个时代换一套配器,一个三音动机换乐器、随和弦移调;13.155s 音乐停,接音效(拍杯叮、滑哨坠落、碎裂、猫叫、收尾拨弦)。 响度:逐段 K 加权响度对齐到 target 表,再整体对齐 -14 LUFS / 真峰 -1.5 dBFS。 """ import numpy as np import scipy.signal as ss import soundfile as sf from pathlib import Path SR = 48000 N = 722944 # 独立示例总长15.0613s T0, E = 0.028, 0.23431 # 网格零点、八分音符时长(实测 128.04 BPM) rng = np.random.default_rng(7) OUT = Path(__file__).with_name("示例配乐.wav") L = np.zeros(N); R = np.zeros(N) def tt(n): # 第 n 个八分音符的绝对时间 return T0 + n * E NOTE = {"C": 0, "D": 2, "E": 4, "F": 5, "G": 7, "A": 9, "B": 11} def hz(name): s = NOTE[name[0]]; i = 1 while name[i] in "#b": s += 1 if name[i] == "#" else -1; i += 1 midi = 12 * (int(name[i:]) + 1) + s return 440.0 * 2 ** ((midi - 69) / 12) def add(sig, t, gain=1.0, pan=0.0): """等功率声像放进总线。pan -1..1""" s = int(round(t * SR)) if s >= N or len(sig) == 0: return sig = sig[: N - s] * gain a = (pan + 1) * np.pi / 4 L[s:s + len(sig)] += sig * np.cos(a) R[s:s + len(sig)] += sig * np.sin(a) # ---------------------------------------------------------------- 基础构件 def tvec(dur): return np.arange(int(dur * SR)) / SR def adsr(dur, a=0.01, d=0.05, s=0.8, r=0.1): """dur 为按住时长,返回长度 dur+r 的包络""" t = tvec(dur + r) e = np.where(t < a, t / max(a, 1e-4), 0) m = (t >= a) & (t < a + d); e[m] = 1 - (1 - s) * (t[m] - a) / max(d, 1e-4) m = (t >= a + d) & (t < dur); e[m] = s lvl = s if dur > a + d else max(e[t < dur][-1] if (t < dur).any() else 0, 0) m = t >= dur; e[m] = lvl * np.exp(-(t[m] - dur) / max(r / 4, 1e-4)) return e def additive(f, amps, dur, phase_jitter=True): """f 可以是常数或逐样本频率数组;amps[k] 是第 k+1 次谐波幅度。带限到 18k。""" n = int(dur * SR) if np.isscalar(f) else len(f) fa = np.full(n, f) if np.isscalar(f) else f ph = 2 * np.pi * np.cumsum(fa) / SR out = np.zeros(n); fmax = fa.max() for k, a in enumerate(amps, 1): if a == 0 or k * fmax > 18000: continue p0 = rng.uniform(0, 2 * np.pi) if phase_jitter else 0 out += a * np.sin(k * ph + p0) return out def saw_amps(n=40): return [1 / k for k in range(1, n + 1)] def square_amps(n=40): return [1 / k if k % 2 else 0 for k in range(1, n + 1)] ATT = 0.010 # 起音软化:2ms 硬起音听起来太「点」,10ms 更连贯 def pluck(f, dur, amps, d0=3.0, slope=0.6, B=0.0, noise=0.02, bend=0.0, att=None): """拨弦/击弦:各次谐波独立指数衰减,高次衰减更快;B 为弦的非谐性;bend 为起音音高回落(半音)。 v2:基音 att 秒线性起音,高次谐波起音更慢(模拟滤波包络从闭到开),起音噪声减半。""" att = ATT if att is None else att t = tvec(dur); out = np.zeros_like(t) fb = f * 2 ** (bend * np.exp(-t / 0.03) / 12) if bend else np.full_like(t, f) ph = 2 * np.pi * np.cumsum(fb) / SR for k, a in enumerate(amps, 1): fk = f * k * np.sqrt(1 + B * k * k) if fk > 18000 or a == 0: continue ak = min(att * (1 + 0.6 * (k - 1)), 0.035) # 高次谐波晚开 out += a * np.exp(-t * d0 * (1 + slope * (k - 1))) * np.minimum(1, t / ak) * np.sin(ph * fk / f + rng.uniform(0, 6.28)) if noise: nb = rng.standard_normal(len(t)) * np.exp(-t / 0.006) * noise * 0.5 out += lp(hp(nb, 2000), 7000) return out def bp(x, lo, hi, order=2): return ss.sosfilt(ss.butter(order, [lo, hi], "band", fs=SR, output="sos"), x) def hp(x, fc, order=2): return ss.sosfilt(ss.butter(order, fc, "high", fs=SR, output="sos"), x) def lp(x, fc, order=2): return ss.sosfilt(ss.butter(order, fc, "low", fs=SR, output="sos"), x) def noise(dur): return rng.standard_normal(int(dur * SR)) def expdec(dur, tau): return np.exp(-tvec(dur) / tau) # ---------------------------------------------------------------- 乐器 def bone_flute(n, dur): # 洞穴:近纯正弦 + 气声,慢起音 f = hz(n); held = dur t = tvec(held + 0.3) vib = 1 + 0.004 * np.sin(2 * np.pi * 5.2 * t) * np.minimum(1, t / 0.2) tone = additive(f * vib, [1, 0.018, 0.022, 0.012], held + 0.3) breath = bp(noise(held + 0.3), f * 0.8, f * 3.5) * 0.05 return (tone + breath) * adsr(held, a=0.07, d=0.1, s=0.9, r=0.3)[: len(t)] def frame_drum(): # 洞穴:低沉手鼓,98→87Hz t = tvec(0.5); f = 87 + 14 * np.exp(-t / 0.04) body = np.sin(2 * np.pi * np.cumsum(f) / SR) * np.exp(-t / 0.16) skin = lp(noise(0.5), 900) * np.exp(-t / 0.02) * 0.3 return body + skin def stick_click(level=1.0): # 敲击/摇奏的高频颗粒 t = tvec(0.04) return bp(noise(0.04), 4000, 11000) * np.exp(-t / 0.006) * level def harp(n, dur=0.6, bright=1.0): # 埃及竖琴:H2≈-11dB,0.2s 衰减约 15dB return pluck(hz(n), dur, [1, 0.3 * bright, 0.07, 0.12, 0.015, 0.015, 0.02], d0=6.0, slope=0.5, noise=0.03) def sistrum(): # 埃及叉铃:带金属泛音的短噪声 t = tvec(0.12) jing = sum(np.sin(2 * np.pi * f * t + rng.uniform(0, 6)) for f in (5120, 6890, 8230)) * 0.15 return (bp(noise(0.12), 6000, 12000) + jing) * np.exp(-t / 0.025) def lyre(n, dur=0.8): # 希腊里拉琴:较闷的拨弦 return pluck(hz(n), dur, [1, 0.45, 0.3, 0.12, 0.05, 0.03], d0=3.5, slope=0.7, noise=0.02) def aulos(n, dur): # 希腊双管:明亮簧音,H2/H3 强,持续 f = hz(n); t = tvec(dur + 0.06) vib = 1 + 0.006 * np.sin(2 * np.pi * 6 * t) * np.minimum(1, t / 0.15) tone = additive(f * vib, [1, 0.4, 0.6, 0.35, 0.06, 0.035, 0.11, 0.06], dur + 0.06) tone += bp(noise(dur + 0.06), f, f * 4) * 0.03 return tone * adsr(dur, a=0.05, d=0.05, s=0.9, r=0.06) def tambourine(): t = tvec(0.35) return bp(noise(0.35), 7000, 14000) * np.exp(-t / 0.09) * (1 + 0.5 * np.sin(2 * np.pi * 28 * t)) def brass(n, dur): # 罗马号角:锯齿 + 随包络变亮的低通 f = hz(n); tot = dur + 0.08 raw = additive(f, saw_amps(30), tot) env = adsr(dur, a=0.03, d=0.12, s=0.6, r=0.08) t = tvec(tot) # 分两段低通近似"开口变亮再收" bright = lp(raw, min(f * 10, 12000)); dull = lp(raw, min(f * 4, 7000)) w = np.exp(-t / 0.12) return (bright * w + dull * (1 - w)) * env def war_drum(): # 罗马战鼓 ~87Hz t = tvec(0.45); f = 70 + 30 * np.exp(-t / 0.03) return np.sin(2 * np.pi * np.cumsum(f) / SR) * np.exp(-t / 0.13) + lp(noise(0.45), 1500) * np.exp(-t / 0.015) * 0.5 def organ(n, dur): # 哥特管风琴:H1 + 奇次/高次音栓,几乎不衰减 f = hz(n); tot = dur + 0.15; t = tvec(tot) trem = 1 + 0.03 * np.sin(2 * np.pi * 5.5 * t) tone = additive(f, [1, 0.03, 0.05, 0.12, 0.18, 0.04, 0.0, 0.06, 0.03, 0.05, 0.02, 0.04, 0, 0.03, 0, 0.03], tot) * trem # 末尾为混合音栓(高频) return tone * adsr(dur, a=0.035, d=0.02, s=1.0, r=0.15) def harpsichord(n, dur=0.5): # 文艺复兴羽管键琴:鼻音(H5 突出),衰减中等 return pluck(hz(n), dur, [1, 0.12, 0.045, 0.06, 0.14, 0.02, 0.03, 0.025] + [0.02] * 10, d0=2.2, slope=0.12, noise=0.08) def viol(n, dur): # 低音维奥尔:柔和弓弦 f = hz(n); tot = dur + 0.12 raw = lp(additive(f, saw_amps(20), tot), 1200) return raw * adsr(dur, a=0.06, d=0.1, s=0.85, r=0.12) def koto(n, dur=0.35): # 浮世绘:筝/三味线,快衰减 + 起音回落 return pluck(hz(n), dur, [1, 0.55, 0.12, 0.13, 0.13, 0.06, 0.04], d0=14, slope=0.4, noise=0.06, bend=0.6) def taiko(): t = tvec(0.6); f = 58 + 40 * np.exp(-t / 0.05) return np.sin(2 * np.pi * np.cumsum(f) / SR) * np.exp(-t / 0.2) + lp(noise(0.6), 700) * np.exp(-t / 0.03) * 0.4 def hyoshigi(): # 拍子木 t = tvec(0.06) return (bp(noise(0.06), 1800, 4200, 4) + np.sin(2 * np.pi * 2650 * t) * 0.6) * np.exp(-t / 0.012) def soft_piano(n, dur=0.9): # 印象派:柔和钢琴/钢片琴,接近正弦 return pluck(hz(n), dur, [1, 0.16, 0.08, 0.04, 0.025, 0.015, 0.012, 0.01, 0.008], d0=2.6, slope=0.45, B=0.0002, noise=0.03) def pizz(n, dur=0.3): # 梵高:明亮拨弦(H2/H3 很强),快衰减 return pluck(hz(n), dur, [1, 0.9, 0.6, 0.22, 0.35, 0.14, 0.09, 0.05], d0=18, slope=0.3, noise=0.04) def swirl(dur): # 梵高:上扬的风声 x = noise(dur); n = len(x); out = np.zeros(n); seg = 2400 for i in range(0, n, seg): fc = 2500 + 6000 * i / n out[i:i + seg] = bp(x[max(0, i - 400):i + seg], fc * 0.7, fc * 1.3)[-len(out[i:i + seg]):] return out * np.sin(np.pi * np.arange(n) / n) def celesta(n, dur=0.35): # 新艺术:钢片琴/竖琴,快衰减 return pluck(hz(n), dur, [1, 0.6, 0.12, 0.1, 0.12, 0.12, 0.2, 0.06], d0=11, slope=0.3, noise=0.02) def blip(n, dur=0.2): # 包豪斯:几何感短音(马林巴式) return pluck(hz(n), dur, [1, 0.75, 0.2, 0.4], d0=22, slope=0.4, noise=0.0) def tick(): t = tvec(0.015) return hp(noise(0.015), 6000) * np.exp(-t / 0.003) def kick(f0=150, f1=48, tau=0.18, dur=0.35): t = tvec(dur); f = f1 + (f0 - f1) * np.exp(-t / 0.025) return np.sin(2 * np.pi * np.cumsum(f) / SR) * np.exp(-t / tau) + hp(noise(dur), 3000) * np.exp(-t / 0.003) * 0.15 def clap(): t = tvec(0.18); env = np.zeros_like(t) for d in (0, 0.009, 0.018): env += (t >= d) * np.exp(-np.maximum(t - d, 0) / (0.006 if d < 0.018 else 0.05)) return bp(noise(0.18), 1200, 6000) * env def hat(open_=False): d = 0.09 if open_ else 0.03; t = tvec(d) return hp(noise(d), 7500, 4) * np.exp(-t / (d / 3)) def chip(n, dur, duty_sq=True): # 像素:方波(只有奇次谐波) f = hz(n); tot = dur + 0.01 return additive(f, square_amps(40), tot, phase_jitter=False) * adsr(dur, a=0.006, d=0.0, s=1.0, r=0.01) def supersaw(n, dur): # 光追:多层失谐锯齿,玻璃感 f = hz(n); tot = dur + 0.15; out = 0 for det in (-0.012, -0.005, 0, 0.006, 0.013): out = out + additive(f * (1 + det), saw_amps(25), tot) out = lp(out / 5, 7000) return out * adsr(dur, a=0.04, d=0.1, s=0.7, r=0.15) def boing(t0=11.01, t1=11.165, f0=600, f1=985): # 猫跳上桌:上滑音 t = tvec(t1 - t0); f = f0 * (f1 / f0) ** (t / t[-1]) tone = additive(f, [1, 0.5, 0.35, 0.2, 0.1], len(t) / SR) return tone * np.sin(np.pi * t / t[-1]) ** 0.5 def pluck2026(n, dur=0.22): # 2026:方波系合成拨奏(H2 弱、H3 强)+ 包络低通 f = hz(n); tot = dur raw = additive(f, [1, 0.1, 0.38, 0.3, 0.22, 0.12, 0.14, 0.1, 0.08], tot) return raw * np.exp(-tvec(tot) / 0.1) * np.minimum(1, tvec(tot) / 0.014) def sub_bass(n, dur): # 2026:锯齿低音 + 下八度正弦 f = hz(n); tot = dur + 0.15 raw = lp(additive(f, saw_amps(30), tot), 700, 4) + 0.6 * np.sin(2 * np.pi * f / 2 * tvec(tot)) return raw * adsr(dur, a=0.01, d=0.05, s=0.9, r=0.15) def pad(n, dur): f = hz(n); tot = dur + 0.1 return lp(additive(f, saw_amps(15), tot), 2500) * adsr(dur, a=0.05, d=0.1, s=0.7, r=0.1) # ---------------------------------------------------------------- 音效 def tink(): # 13.16 猫爪碰杯:陶瓷高频非谐振铃 t = tvec(0.35); out = 0 for f, a, tau in ((5720, 1, 0.12), (6506, 0.6, 0.09), (6315, 0.25, 0.1), (5490, 0.35, 0.08), (1533, 0.1, 0.05)): out = out + a * np.sin(2 * np.pi * f * t) * np.exp(-t / tau) return out def swipe(): # 爪子划过的嗖声 d = 0.2; t = tvec(d) return bp(noise(d), 2500, 9000) * np.sin(np.pi * t / d) ** 2 def slide_whistle(t0=13.29, t1=13.635, f0=1550, f1=325): # 杯子坠落:指数下滑正弦 t = tvec(t1 - t0); f = f0 * (f1 / f0) ** (t / t[-1]) tone = np.sin(2 * np.pi * np.cumsum(f) / SR) env = np.minimum(1, t / 0.015) * np.minimum(1, (t[-1] - t) / 0.01 + 0.0001) return tone * env def crash(): # 13.624 碎裂:83Hz 闷响 + 高频碎屑 + 非谐振残响 d = 0.6; t = tvec(d) f = 60 + 40 * np.exp(-t / 0.02) thump = np.sin(2 * np.pi * np.cumsum(f) / SR) * np.exp(-t / 0.07) hiss = hp(noise(d), 2500) * np.exp(-t / 0.13) shards = np.zeros_like(t) for _ in range(14): # 若干碎片的二次小碰撞 s = rng.uniform(0.02, 0.4); i = int(s * SR); g = rng.uniform(0.2, 0.7) * np.exp(-s / 0.2) seg = hp(noise(0.03), 4000) * np.exp(-tvec(0.03) / 0.005) * g shards[i:i + len(seg)] += seg[: len(shards) - i] ring = sum(a * np.sin(2 * np.pi * fr * t) * np.exp(-t / tau) for fr, a, tau in ((1669, .25, .15), (2302, .2, .1), (4462, .12, .08), (7320, .3, .12), (7921, .25, .1))) return thump * 1.2 + hiss * 0.6 + shards + ring def meow(t0=14.12, tp=14.28, t1=14.65): # 猫叫:574→797Hz 上行,再降到 486Hz;锯齿 + 共振峰 n = int((t1 - t0) * SR); t = np.arange(n) / SR; ip = int((tp - t0) * SR) f = np.empty(n); f[:ip] = np.linspace(574, 797, ip); f[ip:] = np.linspace(797, 486, n - ip) raw = additive(f, saw_amps(24), n / SR) vowel = 0.5 * bp(raw, 1300, 1900) + 1.0 * bp(raw, 2100, 2700) + 0.9 * bp(raw, f.mean() * 0.8, f.mean() * 1.3) + 0.15 * raw env = np.minimum(1, t / 0.04) * np.minimum(1, (t[-1] - t) / 0.12) return vowel * env def end_thump(f=90, sparkle=True): # 签名出现时的轻"啵" d = 0.2; t = tvec(d) out = np.sin(2 * np.pi * f * t) * np.exp(-t / 0.05) if sparkle: out += sum(0.08 * np.sin(2 * np.pi * fr * t) * np.exp(-t / 0.06) for fr in (531, 852, 921, 965, 1058)) return out # ---------------------------------------------------------------- 编曲(时间单位:八分音符) def era(n0, items): """items: (相对八分位置, 生成函数, 增益, 声像)""" for pos, fn, g, pan in items: add(fn(), tt(n0 + pos), g, pan) B8 = E # 一个八分音符的秒数 # ---- 原创示例乐谱(D 小调 / F 大调,和声骨架 Dm–B♭–F–C)---- # 贯穿动机:♪♪♩「级进上行、再下跳」,每段换乐器、随当段和弦移调;洞穴段用它的下行扩大版先立住。 # 1 洞穴(6 个八分):骨笛 A4–G4–D4(下行,附点四分起)+ 手鼓 + 弱拍敲击 era(0, [(0, lambda: bone_flute("A4", 3 * B8), 1.0, 0), (3, lambda: bone_flute("G4", B8), 1.0, 0), (4, lambda: bone_flute("D4", 2.4 * B8), 1.0, 0), (0, frame_drum, 1.3, -0.1), (3, frame_drum, 1.1, -0.1)] + [(k, lambda: stick_click(0.25), 1.0, 0.25) for k in (1, 2, 4, 5)]) # 2 埃及(4):竖琴 Dm→B♭ 分解 + 动机 F5–G5–D5 + 弱拍叉铃 era(6, [(0, lambda: harp("D3", 0.9), 0.55, -0.2), (0.06, lambda: harp("F3", 0.6), 0.35, -0.15), (0.5, lambda: harp("A3", 0.6), 0.45, -0.1), (1, lambda: harp("D4", 0.5), 0.4, -0.2), (2, lambda: harp("Bb2", 0.9), 0.5, -0.2), (2.06, lambda: harp("D3", 0.6), 0.35, -0.15), (3, lambda: harp("F3", 0.5), 0.4, -0.2), (0, lambda: harp("F5", 0.5, 1.2), 0.8, 0.1), (1, lambda: harp("G5", 0.5, 1.2), 0.8, 0.1), (2, lambda: harp("D5", 0.6, 1.2), 0.8, 0.1)] + [(k + 0.5, sistrum, 0.22, 0.3) for k in range(4)]) # 3 希腊(4):里拉琴 B♭ 和弦 + 双管 D5(♪)→ F5(附点四分) era(10, [(0, lambda: lyre("Bb2"), 0.4, -0.25), (0.04, lambda: lyre("F3"), 0.45, -0.2), (0.08, lambda: lyre("Bb3"), 0.35, -0.15), (2.5, lambda: lyre("D4", 0.5), 0.35, -0.2), (3.5, lambda: lyre("F2", 0.6), 0.5, -0.2), (0, lambda: aulos("D5", B8), 0.55, 0.05), (1, lambda: aulos("F5", 3 * B8), 0.5, 0.05), (1, lambda: aulos("F6", 3 * B8), 0.15, 0.05), (0, tambourine, 0.12, 0.3), (2, tambourine, 0.08, 0.3)]) # 4 罗马(4):号角动机 C5–D5–A4,和弦 F→C→Dm + 战鼓 era(14, [(0, lambda: brass("C5", B8 * 0.9), 0.45, 0.05), (0, lambda: brass("F3", B8 * 0.9), 0.25, -0.15), (0, lambda: brass("A3", B8 * 0.9), 0.25, 0.15), (1, lambda: brass("D5", B8 * 1.4), 0.4, 0.05), (1, lambda: brass("G3", B8 * 1.4), 0.25, -0.15), (1, lambda: brass("C3", B8 * 1.8), 0.3, 0), (2.5, lambda: brass("A4", B8 * 1.4), 0.4, 0.05), (2.5, lambda: brass("F3", B8 * 1.4), 0.25, -0.15), (3, lambda: brass("D3", B8), 0.3, 0), (0, war_drum, 1.1, -0.05), (1.5, war_drum, 0.9, -0.05), (2, war_drum, 1.1, -0.05)]) # 5 哥特(4):管风琴 Dm → Gm/D,D2 持续踏板 era(18, [(0, lambda: organ("D3", 2 * B8), 0.3, -0.1), (0, lambda: organ("A3", 2 * B8), 0.25, 0.1), (0, lambda: organ("D4", 2 * B8), 0.25, -0.05), (0, lambda: organ("F4", 2 * B8), 0.3, 0.05), (0.5, lambda: organ("A5", 1.5 * B8), 0.12, 0), (2, lambda: organ("D3", 2 * B8), 0.3, -0.1), (2, lambda: organ("G3", 2 * B8), 0.3, 0.1), (2, lambda: organ("Bb3", 2 * B8), 0.3, -0.05), (2, lambda: organ("D4", 2 * B8), 0.2, 0.05), (2.5, lambda: organ("Bb4", 1.5 * B8), 0.15, 0), (0, lambda: organ("D2", 4 * B8), 0.18, 0)]) # 6 文艺复兴(4):羽管键琴动机 B♭5–C6–G5,伴奏 Gm→F + 低音维奥尔 G2 era(22, [(0, lambda: harpsichord("Bb5"), 0.6, 0.1), (1, lambda: harpsichord("C6"), 0.6, 0.1), (2, lambda: harpsichord("G5", 0.7), 0.6, 0.1), (0, lambda: harpsichord("G3"), 0.3, -0.2), (0.04, lambda: harpsichord("Bb3"), 0.28, -0.15), (1, lambda: harpsichord("D4"), 0.25, -0.15), (2, lambda: harpsichord("F3"), 0.3, -0.2), (3, lambda: harpsichord("A3"), 0.28, -0.15), (3, lambda: harpsichord("C4"), 0.25, -0.1), (0, lambda: viol("G2", 3.6 * B8), 0.35, -0.05)]) # 7 浮世绘(3):筝走 D 都节音阶(D–E♭–G–A–B♭)+ 太鼓 + 拍子木 era(26, [(0, taiko, 1.2, -0.05), (1.5, taiko, 1.0, -0.05), (0, lambda: koto("D5"), 0.55, 0.15), (0.5, lambda: koto("A4"), 0.45, 0.15), (0.75, lambda: koto("Bb4"), 0.5, 0.15), (1, lambda: koto("G4"), 0.45, 0.15), (1.5, lambda: koto("A4"), 0.4, 0.15), (2, lambda: koto("Eb5"), 0.5, 0.15), (2.25, lambda: koto("D5"), 0.45, 0.15), (2.5, lambda: koto("A4"), 0.45, 0.15), (0.5, hyoshigi, 0.25, 0.3), (2, hyoshigi, 0.2, 0.3)]) # 8 印象派(3):柔和钢琴 A5–C6–G5,Fmaj7 → B♭ 低音 → Dm9 色彩 era(29, [(0, lambda: soft_piano("A5"), 0.55, 0.1), (0, lambda: soft_piano("F3"), 0.35, -0.15), (0, lambda: soft_piano("E4"), 0.3, -0.1), (0.5, lambda: soft_piano("C5"), 0.3, 0), (1, lambda: soft_piano("C6"), 0.55, 0.1), (1, lambda: soft_piano("D5"), 0.3, 0), (1, lambda: soft_piano("F4"), 0.3, -0.1), (1, lambda: soft_piano("Bb2", 1.5), 0.55, -0.05), (2, lambda: soft_piano("G5"), 0.55, 0.1), (2, lambda: soft_piano("D3"), 0.3, -0.1), (2.5, lambda: soft_piano("E5"), 0.25, 0.1)]) # 9 梵高(3):D2 低音拨弦律动 + 高处零散亮拨弦 + 上扬风声 era(32, [(0, lambda: pizz("D2", 0.3), 0.7, -0.05), (0.5, lambda: pizz("A2"), 0.45, -0.1), (1, lambda: pizz("D2"), 0.6, -0.05), (1.75, lambda: pizz("C3"), 0.5, -0.05), (2, lambda: pizz("D2"), 0.6, -0.05), (2.5, lambda: pizz("A2", 0.4), 0.5, -0.05), (0.25, lambda: pizz("A5"), 0.4, 0.15), (1, lambda: pizz("F5"), 0.35, 0.1), (1.75, lambda: pizz("D6"), 0.3, 0.15), (2.5, lambda: pizz("E5"), 0.4, 0.15), (1.5, lambda: swirl(0.4), 0.06, 0.2)]) # 10 新艺术(3):钢片琴动机 D5–E5–A4,低音 B♭→C→Dm era(35, [(0, lambda: celesta("D5"), 0.6, 0.1), (0.2, lambda: celesta("D6"), 0.2, 0.15), (1, lambda: celesta("E5"), 0.6, 0.1), (2, lambda: celesta("A4"), 0.6, 0.1), (0, lambda: pizz("Bb2", 0.5), 0.5, -0.05), (1, lambda: pizz("C3", 0.5), 0.45, -0.05), (1, lambda: celesta("E3", 0.5), 0.3, -0.15), (1, lambda: celesta("G3", 0.5), 0.25, -0.1), (2, lambda: pizz("D3", 0.5), 0.45, -0.05), (2, lambda: celesta("D4"), 0.25, -0.1), (2, lambda: celesta("F4"), 0.25, -0.15), (2, lambda: celesta("A4"), 0.2, 0), (0, lambda: tambourine(), 0.06, 0.3)]) # 11 立体主义(2):刻意不协和的碎片和弦 D–G♯–C♯ → E♭–A–B♭–E era(38, [(0, lambda: soft_piano("D2", 0.5), 0.5, -0.1), (0, lambda: soft_piano("D3", 0.5), 0.45, -0.1), (0, lambda: soft_piano("G#3", 0.5), 0.4, 0), (0, lambda: soft_piano("C#5", 0.5), 0.45, 0.15), (1, lambda: soft_piano("Eb4", 0.3), 0.4, -0.1), (1, lambda: soft_piano("A4", 0.3), 0.45, 0.1), (1, lambda: soft_piano("Bb4", 0.3), 0.35, 0), (1, lambda: soft_piano("E5", 0.3), 0.5, 0.15)]) # 12 包豪斯(2):D 八度叠置 → F/A → D,十六分机械滴答 era(40, [(0, lambda x=x: blip(x), 0.3, p) for x, p in (("D2", 0), ("D3", -0.1), ("D4", 0.1), ("D5", -0.05), ("D6", 0.05))] + [(1, lambda: blip("F5"), 0.45, 0.1), (1, lambda: blip("A5"), 0.3, 0.1), (1, lambda: blip("A6"), 0.2, -0.1), (1.5, lambda: blip("D5"), 0.45, 0)] + [(k * 0.5, tick, 0.25, 0.3) for k in range(4)]) # 13 波普(2):底鼓 + 反拍拍手 + 卡通短音 era(42, [(0, lambda: kick(140, 45), 1.1, 0), (0, lambda: blip("D4", 0.2), 0.25, -0.1), (0, lambda: blip("F4", 0.2), 0.25, 0.1), (1, clap, 0.45, 0.1), (0.5, lambda: chip("D4", 0.09), 0.18, -0.1), (0.5, lambda: chip("F4", 0.09), 0.15, 0.1), (0.5, lambda: chip("C5", 0.09), 0.12, 0), (1.5, lambda: chip("D6", 0.05), 0.15, 0.2)] + [(k * 0.5, hat, 0.3, 0.25) for k in range(4)]) # 14 像素(1):方波 F 大三琶音(三十二分)+ 方波低音 era(44, [(0, lambda: chip("F3", 0.2), 0.3, 0), (0, lambda: chip("F6", 0.055), 0.18, 0.1), (0.25, lambda: chip("A6", 0.05), 0.15, 0.1), (0.5, lambda: chip("C7", 0.055), 0.12, 0.1), (0.75, lambda: chip("F6", 0.05), 0.15, 0.1), (0, lambda: kick(120, 50, 0.08), 0.6, 0)]) # 15 光追(1):超级锯齿 B♭ 和弦,玻璃感 era(45, [(0, lambda x=x: supersaw(x, B8 * 1.1), 0.18, p) for x, p in (("Bb3", -0.2), ("D4", 0.2), ("F4", -0.1), ("Bb4", 0.1))] + [(0.5, lambda: supersaw("D5", B8 * 0.6), 0.15, 0.15), (0.5, lambda: supersaw("F3", B8 * 0.8), 0.15, -0.1), (0, lambda: kick(130, 50, 0.1), 0.7, 0)]) # 16 孟菲斯(1):C 大三持续和弦挂过段尾 + C 低音 + 底鼓;猫跳上桌的上滑音 era(46, [(0, lambda: kick(150, 46), 1.0, 0), (0, lambda: sub_bass("C2", 0.24), 0.35, 0)] + [(0, lambda x=x: supersaw(x, 0.42), g, p) for x, g, p in (("C3", 0.12, -0.15), ("G3", 0.1, 0.15), ("C5", 0.16, -0.05), ("E5", 0.16, 0.1), ("G5", 0.1, 0))]) add(boing(), 11.01, 0.35, 0.0) # 17 2026(9):128BPM 浩室——四拍底鼓、反拍开镲、D 锯齿低音、D 大调 add9 拨奏(结尾转到同名大调) add(sub_bass("D2", tt(56) - tt(47)), tt(47), 0.4, 0) for n in (48, 50, 52, 54): add(kick(150, 46, 0.2), tt(n), 1.0, 0) for n in (49, 51, 53, 55): add(hat(True), tt(n), 0.3, 0.2) lead = [(48, "F#5"), (49, "A5"), (50, "D6"), (51, "E6"), (52, "F#6"), (53, "E6"), (54, "D6"), (55, "B5"), (55.5, "A5")] for n, x in lead: add(pluck2026(x), tt(n), 0.5, 0.1) arp = ["A4", "D5", "F#4", "E5", "D4", "A4", "B4", "F#4"] for k in range(18): n = 47 + k * 0.5 if n >= 56: break add(pluck2026(arp[k % len(arp)], 0.15), tt(n), 0.22, -0.15) for x, a, b in (("F#3", 50.5, 53.5), ("A3", 50.5, 53.5), ("E4", 51.5, 53.5), ("D3", 53.5, 54.3), ("F#3", 54.5, 56), ("A3", 54.5, 56), ("B3", 54.5, 56)): add(pad(x, tt(b) - tt(a)), tt(a), 0.07, -0.2 if x < "C" else 0.2) # 18 音效段 add(tink(), 13.158, 0.12, 0.1) add(swipe(), 13.17, 0.05, 0.2) add(slide_whistle(), 13.29, 0.16, 0.0) add(crash(), 13.624, 0.55, 0.0) add(meow(), 14.12, 0.55, 0.0) add(end_thump(88), 14.561, 0.2, 0) add(end_thump(90), 14.796, 0.25, 0) add(pluck(hz("D2"), 0.3, [1, 0.5, 0.35, 0.15, 0.08], d0=4, slope=0.5, noise=0.03), 14.92, 0.6, 0) # 收在主音 # ---------------------------------------------------------------- 混音:小空间混响 → 逐段响度对齐 → 整体 LUFS/峰值 def reverb(x, seed): r = np.random.default_rng(seed); d = 0.5; t = tvec(d) ir = r.standard_normal(len(t)) * np.exp(-t / 0.09) ir = lp(ir, 6000); ir[: int(0.012 * SR)] = 0 # 12ms 预延迟,起音更软 return ss.fftconvolve(x, ir)[: len(x)] / np.sqrt((ir ** 2).sum()) wetL, wetR = reverb(L, 1), reverb(R, 2) L = L + 0.09 * wetL; R = R + 0.09 * wetR def kweight(x): b1 = [1.53512485958697, -2.69169618940638, 1.19839281085285]; a1 = [1, -1.69065929318241, 0.73248077421585] b2 = [1.0, -2.0, 1.0]; a2 = [1, -1.99004745483398, 0.99007225036621] return ss.lfilter(b2, a2, ss.lfilter(b1, a1, x)) def seg_loud(kl, kr, s, e): return -0.691 + 10 * np.log10(np.mean(kl[s:e] ** 2 + kr[s:e] ** 2) + 1e-12) def lufs(l, r): # BS.1770-4 积分响度(400ms 块,绝对+相对门限) kl, kr = kweight(l), kweight(r); blk, hop = int(0.4 * SR), int(0.1 * SR) z = np.array([np.mean(kl[i:i + blk] ** 2 + kr[i:i + blk] ** 2) for i in range(0, len(l) - blk + 1, hop)]) lk = -0.691 + 10 * np.log10(z + 1e-12); z = z[lk > -70] rel = -0.691 + 10 * np.log10(z.mean()) - 10 z = z[-0.691 + 10 * np.log10(z) > rel] return -0.691 + 10 * np.log10(z.mean()) # 逐段目标响度(K 加权、不加门限)。有参考片时填它逐段量出来的值;这里是示例:音乐段统一 -14,音效段略低。 edges = [0] + [tt(n) for n in (6, 10, 14, 18, 22, 26, 29, 32, 35, 38, 40, 42, 44, 45, 46, 47)] + [13.155, 13.624, 14.1, N / SR] target = [-14.0] * 17 + [-18.0, -17.5, -16.0] idx = [int(round(x * SR)) for x in edges] # 低频暖底:按段的和弦根音铺柔和三角波 + 250Hz 低通,电平按「低频地板比段响度」的相对值自动定(这里示例统一低 18dB)。 BED_ROOT = ["D2", "D2", "A#1", "F2", "D2", "G2", "D2", "F2", "D2", "A#1", "D2", "D2", "D2", "F2", "A#1", "C2", "D2", None, None, None] LOW_FLOOR_ORIG = [t - 18 for t in target[:17]] + [None, None, None] _sos_low = ss.butter(4, [40, 220], "band", fs=SR, output="sos") def _floor(x): # 50ms 窗 RMS 的最小值(只看持续部分,不看鼓的击头) w = 2400 return min(20 * np.log10(np.sqrt(np.mean(x[k:k + w] ** 2)) + 1e-12) for k in range(0, max(1, len(x) - w), 1200)) _kl, _kr = kweight(L), kweight(R) _low = ss.sosfilt(_sos_low, (L + R) / 2) for i in range(len(target)): if BED_ROOT[i] is None: continue s0, s1 = idx[i], idx[i + 1] want = seg_loud(_kl, _kr, s0, s1) + (LOW_FLOOR_ORIG[i] - target[i]) have = _floor(_low[s0 + 1200:s1]) if want - have < 1.5: continue n = s1 - s0 + int(0.04 * SR); f = hz(BED_ROOT[i]) tri = additive(f, [1, 0, 1 / 9, 0, 1 / 25], n / SR) + 0.35 * np.sin(2 * np.pi * f / 2 * np.arange(n) / SR) tri = lp(tri, 250) env = np.ones(n); fd = int(0.02 * SR); env[:fd] = np.linspace(0, 1, fd); env[-fd:] = np.linspace(1, 0, fd) z = tri * env; zl = ss.sosfilt(_sos_low, z) amp = np.sqrt(max(10 ** (want / 10) - 10 ** (have / 10), 0)) / (np.sqrt(np.mean(zl[fd:-fd] ** 2)) + 1e-12) e = min(N, s0 + n) L[s0:e] += z[: e - s0] * amp * 0.707; R[s0:e] += z[: e - s0] * amp * 0.707 # 中高频(2–6kHz)动态 EQ:把每段 2–6kHz 的相对电平推到目标值(上限 +12dB)。有参考片时填参考值;这里示例统一比段响度低 15dB MID_ORIG = [t - 15 for t in target] _sos_mid = ss.butter(2, [2000, 6000], "band", fs=SR, output="sos") _kl, _kr = kweight(L), kweight(R) _midL, _midR = ss.sosfiltfilt(_sos_mid, L), ss.sosfiltfilt(_sos_mid, R) _g = np.zeros(N) for i in range(len(target)): s0, s1 = idx[i], idx[i + 1] want = seg_loud(_kl, _kr, s0, s1) + (MID_ORIG[i] - target[i]) have = 20 * np.log10(np.sqrt(np.mean(((_midL + _midR) / 2)[s0:s1] ** 2)) + 1e-12) _g[s0:s1] = 10 ** (np.clip(want - have, 0, 12) / 20) - 1 _g = np.convolve(_g, np.ones(960) / 960, mode="same") L = L + _g * _midL; R = R + _g * _midR # >6kHz 的「空气」(高次泛音 + 噪声颗粒):按「高频比段响度」的相对值补一层高通噪声。示例统一低 24dB HF_ORIG = [t - 24 for t in target] _sos_hf = ss.butter(4, 6000, "high", fs=SR, output="sos") _kl, _kr = kweight(L), kweight(R) _hf = ss.sosfilt(_sos_hf, (L + R) / 2) for i in range(len(target)): s0, s1 = idx[i], idx[i + 1] want = seg_loud(_kl, _kr, s0, s1) + (HF_ORIG[i] - target[i]) have = 20 * np.log10(np.sqrt(np.mean(_hf[s0:s1] ** 2)) + 1e-12) if want - have > 2: need = np.sqrt(max(10 ** (want / 10) - 10 ** (have / 10), 0)) fade = int(0.01 * SR); n = s1 - s0 env = np.ones(n); env[:fade] = np.linspace(0, 1, fade); env[-fade:] = np.linspace(1, 0, fade) for ch, seed in ((L, 11 + i), (R, 41 + i)): z = ss.sosfilt(_sos_hf, np.random.default_rng(seed).standard_normal(n + 2000))[2000:] ch[s0:s1] += z / np.sqrt(np.mean(z ** 2)) * need * env gain_db = np.zeros(len(target)) for _ in range(4): # 迭代:尾音跨段会互相影响 g = np.zeros(N) for i in range(len(target)): g[idx[i]:idx[i + 1]] = gain_db[i] g = np.convolve(g, np.ones(480) / 480, mode="same") # 10ms 平滑,避免拉链声 gl, gr = L * 10 ** (g / 20), R * 10 ** (g / 20) kl, kr = kweight(gl), kweight(gr) for i in range(len(target)): gain_db[i] += target[i] - seg_loud(kl, kr, idx[i], idx[i + 1]) L, R = gl, gr print("逐段增益(dB):", np.round(gain_db, 1)) def true_peak(l, r): return 20 * np.log10(max(np.abs(ss.resample_poly(l, 4, 1)).max(), np.abs(ss.resample_poly(r, 4, 1)).max())) def soft_limit(x, ceil): # 只压超过门限的瞬态(前视 1ms 的峰值包络) from scipy.ndimage import maximum_filter1d env = maximum_filter1d(np.abs(x), 96) gr = np.minimum(1, ceil / np.maximum(env, 1e-9)) gr = np.convolve(gr, np.ones(96) / 96, mode="same") return x * gr # 模拟流媒体 AAC 的 16kHz 截止(要和参考片质感一致时用,不需要可以删掉) _sos16 = ss.butter(10, 15800, "low", fs=SR, output="sos") L, R = ss.sosfiltfilt(_sos16, L), ss.sosfiltfilt(_sos16, R) TARGET_LUFS, TARGET_TP = -14.0, -1.5 for _ in range(3): cur = lufs(L, R); k = 10 ** ((TARGET_LUFS - cur) / 20); L *= k; R *= k ceil = 10 ** ((TARGET_TP - 0.05) / 20) L, R = soft_limit(L, ceil), soft_limit(R, ceil) y = np.stack([L, R], 1) sf.write(OUT, y, SR, subtype="PCM_24") print(f"写出 {OUT.name}: {len(y)} 样本 = {len(y) / SR:.4f}s, 积分响度 {lufs(L, R):.2f} LUFS, 真峰 {true_peak(L, R):.2f} dBTP")scripts/capabilities.py22 KB00df4f64d2deupstream ↗
#!/usr/bin/env python3 """可选语音配置。status/resolve不联网;个人配置位于skill安装目录外。""" import argparse from contextlib import contextmanager import copy import hashlib import json import os from pathlib import Path import platform import re import shutil import subprocess import sys import tempfile import time ROOT = Path(__file__).resolve().parents[1] if __name__ == "__main__": sys.modules.setdefault("capabilities", sys.modules[__name__]) class ConfigError(ValueError): pass def config_path(explicit=None): if explicit: return Path(explicit).expanduser().resolve() if os.getenv('HUASHU_CONFIG_HOME'): return Path(os.environ['HUASHU_CONFIG_HOME']).expanduser() / 'media.json' if os.name == 'nt': return Path(os.environ.get('APPDATA', Path.home() / 'AppData/Roaming')) / 'huashu/media.json' return Path(os.environ.get('XDG_CONFIG_HOME', Path.home() / '.config')) / 'huashu/media.json' def read_json(path, missing=False): if missing and not path.exists(): return {} try: data = json.loads(path.read_text(encoding='utf-8')) except (OSError, ValueError) as e: raise ConfigError(f'无法读取配置:{path.name}({type(e).__name__})') from None if not isinstance(data, dict): raise ConfigError('配置必须是JSON对象') return data def merge(base, patch): result = copy.deepcopy(base) for key, value in patch.items(): result[key] = merge(result[key], value) if isinstance(value, dict) and isinstance(result.get(key), dict) else copy.deepcopy(value) return result def validate(data, schema=None, where='config'): # 本包schema只使用以下JSON Schema关键字;不声称实现通用schema解释器。 schema = schema or read_json(ROOT / 'schemas/media.schema.json') types = {'object': dict, 'string': str, 'integer': int, 'boolean': bool} typ = schema.get('type') if typ and (not isinstance(data, types[typ]) or (typ == 'integer' and isinstance(data, bool))): raise ConfigError(f'{where}类型错误') if 'enum' in schema and data not in schema['enum']: raise ConfigError(f'{where}不是支持的选项') if isinstance(data, int) and not isinstance(data, bool): if data < schema.get('minimum', data) or data > schema.get('maximum', data): raise ConfigError(f'{where}超出范围') if isinstance(data, str) and 'pattern' in schema and not re.fullmatch(schema['pattern'], data): raise ConfigError(f'{where}格式不正确') if isinstance(data, dict): for key in schema.get('required', []): if key not in data: raise ConfigError(f'{where}.{key}缺失') for key, value in data.items(): child = schema.get('properties', {}).get(key, schema.get('additionalProperties', False)) if child is False: raise ConfigError(f'{where}.{key}是不支持的字段') if isinstance(child, dict): validate(value, child, f'{where}.{key}') def project_patch(data): validate(data) if set(data) - {'schema_version', 'revision', 'preferences', 'policy', 'onboarding'}: raise ConfigError('项目配置不能设置凭证绑定、网络策略或供应商底层参数') if any(v != 'deny' for v in data.get('policy', {}).values()): raise ConfigError('项目policy只能收紧为deny,不能授予云调用、付费或训练权限') if 'voice_ref' in data.get('preferences', {}).get('voice', {}): raise ConfigError('私人音色绑定保留在用户配置;本次可用--voice选择') return data def effective(explicit=None, project=None): defaults = read_json(ROOT / 'defaults/media.json') user_path = config_path(explicit) user = read_json(user_path, missing=True) validate(user) result = merge(defaults, user) sources = {'defaults': 'defaults/media.json', 'user': str(user_path), 'user_configured': bool(user)} origin = {} def note(data, label, prefix=''): for key, value in data.items(): name = prefix + key if isinstance(value, dict): note(value, label, name + '.') elif name.split('.')[0] in ('preferences', 'policy', 'provider_options', 'onboarding'): origin[name] = label note(defaults, 'public_default'); note(user, 'user') if project: pp = Path(project).expanduser().resolve() / '.huashu/media.json' patch = project_patch(read_json(pp, missing=True)) result = merge(result, patch) note(patch, 'project') sources['project'] = str(pp) sources['project_fields'] = sorted(patch) validate(result) sources['effective_from'] = origin return result, sources @contextmanager def locked(path): path.parent.mkdir(parents=True, exist_ok=True) lock_path = path.with_name(path.name + '.lock') # 内核文件锁在进程退出后释放;不通过删除锁文件争抢所有权。 with lock_path.open('a+b') as lock: if os.name == 'nt': import msvcrt if lock.seek(0, 2) == 0: lock.write(b'0'); lock.flush() lock.seek(0) msvcrt.locking(lock.fileno(), msvcrt.LK_LOCK, 1) try: yield finally: lock.seek(0); msvcrt.locking(lock.fileno(), msvcrt.LK_UNLCK, 1) else: import fcntl fcntl.flock(lock, fcntl.LOCK_EX) try: yield finally: fcntl.flock(lock, fcntl.LOCK_UN) def save_patch(path, patch, expected=None, reset=None): validate(patch) with locked(path): old = read_json(path, missing=True) validate(old) revision = old.get('revision', 0) if expected is not None and revision != expected: raise ConfigError('配置已被其他进程修改;请重读后重试') updated = merge(old, patch) if reset: # 只允许重置公开字段;不执行动态路径表达式。 keys = reset.split('.') if keys[0] not in {'preferences', 'policy', 'bindings', 'provider_options', 'onboarding'}: raise ConfigError('不支持这个reset路径') current = updated for key in keys[:-1]: current = current.get(key, {}) if isinstance(current, dict): current.pop(keys[-1], None) updated.update(schema_version=1, revision=revision + 1) validate(updated) fd, temp = tempfile.mkstemp(dir=path.parent, prefix='.media-') try: with os.fdopen(fd, 'w', encoding='utf-8') as f: json.dump(updated, f, ensure_ascii=False, indent=2); f.write('\n') if os.name != 'nt': os.chmod(temp, 0o600) os.replace(temp, path) finally: if os.path.exists(temp): os.unlink(temp) return {'status': 'saved', 'revision': updated['revision'], 'path': str(path)} def credentials(config): binding = config['bindings']['volcengine'] names = (binding['appid_env'], binding['token_env']) values = {n: os.getenv(n) for n in names} env_path = os.getenv('KOUBO_ENV') or binding.get('env_file') if env_path and not all(values.values()): path = Path(env_path).expanduser() if path.is_file(): for line in path.read_text(encoding='utf-8').splitlines(): line = line.strip() if line.startswith('export '): line = line[7:] key, sep, value = line.partition('=') if sep and key.strip() in names and not values[key.strip()]: values[key.strip()] = value.strip().strip('\"\'') return tuple(values[n] for n in names) def speaker(config, explicit=None): ref = explicit or config['preferences']['voice'].get('voice_ref', 'default') voice = config['bindings']['voices'].get(ref) if voice and voice.get('speaker_id'): return voice['speaker_id'] if explicit and not voice: # 兼容旧命令的注册表别名或平台ID。 voice = {'entry': explicit} voice = voice or {'entry': 'default'} registry_path = Path(voice.get('registry_file') or config['bindings']['volcengine']['registry_file']).expanduser() registry = read_json(registry_path, missing=True) entry = voice.get('entry', 'default') if entry == 'default': entry = registry.get('default') or os.getenv('VOLC_KOUBO_DEFAULT_SPEAKER') found = registry.get('voices', {}).get(entry, {}) return found.get('speaker_id') or (entry if isinstance(entry, str) and entry.startswith('S_') else None) def resolve(config, mode=None, audio=None, operation='synthesize', grants=(), voice=None): chosen = mode or config['preferences']['voice']['mode'] if audio and mode is None: chosen = 'provided' result = {'capability': 'voice', 'mode': chosen, 'operation': operation, 'status': 'ready', 'reasons': []} if chosen == 'off': return dict(result, status='skipped', reasons=['语音已禁用,不生成音频']) if chosen == 'unconfigured': return dict(result, status='needs_configuration', prompt=config['onboarding']['voice'] != 'manual', reasons=['选择现有录音、系统声音或已绑定的复刻音色;也可本次跳过']) if chosen == 'provided': if not audio or not Path(audio).expanduser().is_file(): return dict(result, status='needs_input', reasons=['需要提供已有音频路径']) return result if chosen == 'system': missing = [x for x in ('say', 'ffmpeg', 'ffprobe') if not shutil.which(x)] if platform.system() != 'Darwin' or missing: return dict(result, status='unavailable', reasons=['系统声音目前仅支持macOS,需已安装say/ffmpeg/ffprobe']) return dict(result, provider='macos') if chosen != 'clone': raise ConfigError('不支持的声音方式') result['provider'] = 'volcengine' if config['preferences']['voice'].get('provider', 'volcengine') != 'volcengine': return dict(result, status='unavailable', reasons=['P1复刻后端仅支持volcengine']) needs = ['cloud'] + ([] if operation == 'check' else ['paid_api']) if operation == 'train': needs.append('voice_training') denied = [key for key in needs if config['policy'][key] == 'deny'] if denied: return dict(result, status='blocked', reasons=['显式禁止:' + ', '.join(denied)]) pending = [key for key in needs if config['policy'][key] != 'allow' and key not in grants] if pending: return dict(result, status='needs_permission', reasons=['需要本次授权:' + ', '.join(pending)]) appid, token = credentials(config) if not appid or not token: result['reasons'].append('缺少所选供应商凭证;仅检查已指定环境/凭证文件') if operation == 'synthesize' and not speaker(config, voice): result['reasons'].append('没有可用音色绑定;可绑定已有ID,无须重训') missing = [x for x in ('ffmpeg', 'ffprobe') if not shutil.which(x)] if missing: result['reasons'].append('缺少本地依赖:' + ', '.join(missing)) import importlib.util if importlib.util.find_spec('requests') is None: result['reasons'].append('火山适配器需要requests;基础配置与已有音频不需要') result['status'] = 'unavailable' if result['reasons'] else 'configured_unverified' return result def inspect_audio(path): if not shutil.which('ffprobe'): raise ConfigError('音频验收需要ffprobe') proc = subprocess.run(['ffprobe', '-v', 'error', '-show_entries', 'stream=codec_type,codec_name,sample_rate,channels:format=duration', '-of', 'json', str(path)], capture_output=True, text=True) if proc.returncode: raise ConfigError('输出不是可解码音频') data = json.loads(proc.stdout) streams = [s for s in data.get('streams', []) if s.get('codec_type') == 'audio'] duration = float(data.get('format', {}).get('duration', 0)) if not streams or duration <= 0: raise ConfigError('输出没有有效音轨或时长') return {'duration': duration, **streams[0], 'sha256': hashlib.sha256(Path(path).read_bytes()).hexdigest()} def system_voice_for(text, configured=None): if configured: return configured # macOS默认英文音色可能对中文返回近空音频;只从已安装音色选,不下载。 if re.search(r"[一-鿿]", text): result = subprocess.run(['say', '-v', '?'], capture_output=True, text=True, check=True) names = [m.group(1).strip() for line in result.stdout.splitlines() if (m := re.match(r"(.+?)\s+zh_CN\s+#", line))] if not names: raise ConfigError('没有已安装的普通话系统声音;请选择已有音色或导入录音') return 'Tingting' if 'Tingting' in names else names[0] return None def run_say(args, config): grants = [name for name in ('cloud', 'paid_api') if getattr(args, 'allow_' + name, False)] result = resolve(config, mode=args.mode, grants=grants, voice=args.voice) if result['status'] not in ('ready', 'configured_unverified'): print(json.dumps(result, ensure_ascii=False)); raise SystemExit(2) if not args.text.strip(): raise ConfigError('合成文本不能为空') out = Path(args.out).expanduser() if out.exists(): raise ConfigError('输出文件已存在,请指定新文件;保留已选音轨') if args.speech_rate is not None and not -50 <= args.speech_rate <= 100: raise ConfigError('火山语速须在-50~100之间') if args.fit is not None and args.fit <= 0: raise ConfigError('--fit必须为正数') if args.match and not Path(args.match).is_file(): raise ConfigError('找不到响度参考音频') requested_format = args.format out.parent.mkdir(parents=True, exist_ok=True) import koubo # 临时工作区防止失败或半段响应污染最终产物。 with tempfile.TemporaryDirectory(prefix='voice-', dir=out.parent) as temp: staged = Path(temp) / (out.name + '.wav' if requested_format == 'pcm' else out.name) if result['mode'] == 'system': if args.model or args.speech_rate is not None or args.no_fidelity: raise ConfigError('火山模型/语速参数不适用于系统声音') if args.format != 'wav' or args.sample_rate != 48000: raise ConfigError('P1系统声音输出固定为48kHz WAV') source = Path(temp) / 'system.aiff' cmd = ['say', '-o', str(source)] system_voice = system_voice_for(args.text, args.voice or config['preferences']['voice'].get('system_voice')) if system_voice: cmd += ['-v', system_voice] subprocess.run(cmd, input=args.text, text=True, check=True, capture_output=True) subprocess.run(['ffmpeg', '-v', 'error', '-y', '-i', str(source), '-ar', '48000', str(staged)], check=True) if args.fit: koubo.fit_duration(staged, args.fit) target_db = koubo.mean_volume(args.match, args.match_ss, args.match_t) if args.match else args.loudness if target_db is not None: koubo.match_loudness(staged, target_db) else: koubo.configure_backend(config) args = copy.copy(args) args.out = str(staged) # 先用容器音频完成拟合/电平/终止状态验收,再导出裸PCM。 if requested_format == 'pcm': args.format = 'wav' args.voice = speaker(config, args.voice) opts = config['provider_options']['volcengine'] args.model = args.model or opts['model'] from voice_review import review_status print(json.dumps({'model_review': review_status('volcengine', args.model, getattr(args, 'config', None))}, ensure_ascii=False), file=sys.stderr) args.speech_rate = opts['speech_rate'] if args.speech_rate is None else args.speech_rate args.no_fidelity = not opts['tone_fidelity'] if args.no_fidelity is None else args.no_fidelity koubo.cmd_say(args) artifact = inspect_audio(staged) volume = koubo.mean_volume(staged) if artifact['duration'] < 0.12 or volume is None or volume < -60: raise ConfigError('合成音频近空或无有效声音,不作为成功产物') if args.fit and abs(artifact['duration'] - args.fit) > .05: # 保留诊断候选,但不把它当成合格的最终输出。 candidate = out.with_name(out.stem + '.unfitted' + out.suffix) if not candidate.exists(): shutil.copy2(staged, candidate) raise ConfigError('音频未落入目标时长±0.05秒,已保留unfitted候选;未交付为最终音频') if requested_format == 'pcm': raw = Path(temp) / 'output.pcm' subprocess.run(['ffmpeg', '-v', 'error', '-y', '-i', str(staged), '-ac', '1', '-ar', str(args.sample_rate), '-f', 's16le', str(raw)], check=True) shutil.copy2(raw, out) artifact.update(format='s16le', channels=1, duration=out.stat().st_size / 2 / args.sample_rate, sha256=hashlib.sha256(out.read_bytes()).hexdigest()) else: shutil.copy2(staged, out) artifact.update(path=out.name, provider=result['provider'], mode=result['mode'], generated=True) if result['mode'] == 'system': artifact['system_voice'] = system_voice or 'system_default' if result['mode'] == 'clone': artifact['requested_model'] = args.model out.with_name(out.name + '.json').write_text(json.dumps(artifact, ensure_ascii=False, indent=2), encoding='utf-8') print(json.dumps({'status': 'complete', 'output': str(out), 'duration': artifact['duration']}, ensure_ascii=False)) def main(): if len(sys.argv) > 1 and sys.argv[1] in ('image-status', 'image-plan', 'accept-image', 'use-image'): import images return images.main() # 生成/训练命令共用原CLI,避免两套参数漂移。 if len(sys.argv) > 1 and sys.argv[1] in ('say', 'train', 'voices'): import koubo return koubo.main() parser = argparse.ArgumentParser(description=__doc__) sub = parser.add_subparsers(dest='command', required=True) for name in ('status', 'resolve', 'configure', 'skip', 'reset', 'bind', 'use-audio'): p = sub.add_parser(name) p.add_argument('--config') p.add_argument('--project') if name in ('status', 'resolve'): p.add_argument('--mode', choices=['off', 'provided', 'system', 'clone']) p.add_argument('--audio'); p.add_argument('--explain', action='store_true') if name == 'configure': p.add_argument('--input', required=True); p.add_argument('--scope', choices=['user', 'project'], default='user') p.add_argument('--expected-revision', type=int) if name == 'skip': p.add_argument('--capability', choices=['voice', 'image'], default='voice') p.add_argument('--scope', choices=['task', 'user'], default='task') if name == 'reset': p.add_argument('--field', default='preferences.voice') if name == 'bind': p.add_argument('--name', required=True); p.add_argument('--speaker-id', required=True) p.add_argument('--default', action='store_true') if name == 'use-audio': p.add_argument('--audio', required=True); p.add_argument('--out', required=True) args = parser.parse_args() config, sources = effective(args.config, args.project) path = config_path(args.config) if args.command in ('status', 'resolve'): result = resolve(config, args.mode, args.audio) result['revision'] = config.get('revision', 0) result['parameters'] = config['provider_options']['volcengine'] if result.get('provider') == 'volcengine' else {} if result.get('provider') == 'volcengine': from voice_review import review_status result['model_review'] = review_status('volcengine', result['parameters']['model'], path) if args.explain: result['sources'] = sources print(json.dumps(result, ensure_ascii=False, indent=2)) elif args.command == 'configure': patch = read_json(Path(args.input)) if args.scope == 'project': if not args.project: raise ConfigError('project作用域需要--project') path = Path(args.project).expanduser().resolve() / '.huashu/media.json' project_patch(patch) print(json.dumps(save_patch(path, patch, args.expected_revision), ensure_ascii=False)) elif args.command == 'skip': if args.scope == 'user': print(json.dumps(save_patch(path, {'onboarding': {args.capability: 'manual'}, 'preferences': {args.capability: {'mode': 'unconfigured'}}}), ensure_ascii=False)) else: print(json.dumps({'status': 'skipped', 'scope': 'task', 'saved': False}, ensure_ascii=False)) elif args.command == 'reset': print(json.dumps(save_patch(path, {}, reset=args.field), ensure_ascii=False)) elif args.command == 'bind': patch = {'bindings': {'voices': {args.name: {'speaker_id': args.speaker_id}}}} if args.default: patch['preferences'] = {'voice': {'voice_ref': args.name}} print(json.dumps(save_patch(path, patch), ensure_ascii=False)) elif args.command == 'use-audio': source, out = Path(args.audio).expanduser(), Path(args.out).expanduser() info = inspect_audio(source) if out.exists(): raise ConfigError('输出已存在,请指定新文件') out.parent.mkdir(parents=True, exist_ok=True); shutil.copy2(source, out) info.update(path=out.name, generated=False, mode='provided') out.with_name(out.name + '.json').write_text(json.dumps(info, ensure_ascii=False, indent=2)) print(json.dumps({'status': 'complete', 'output': str(out), 'generated': False}, ensure_ascii=False)) if __name__ == '__main__': try: main() except (ConfigError, OSError, subprocess.SubprocessError) as e: message = str(e) if isinstance(e, ConfigError) else type(e).__name__ print(json.dumps({'status': 'error', 'message': message}, ensure_ascii=False), file=sys.stderr) raise SystemExit(2)scripts/check_release.py6.0 KBa44e47eff084upstream ↗
#!/usr/bin/env python3 """增量发布检查:扫描Git实际待发对象,核最终文件清单;不推送。""" import argparse import hashlib import json from pathlib import Path import re import subprocess import sys MANIFEST='release-manifest.json' SECRET=re.compile(rb'/Users/[A-Za-z0-9_.-]+/|/home/[A-Za-z0-9_.-]+/|-----BEGIN (?:RSA |EC |OPENSSH )?PRIVATE KEY-----|github_pat_[A-Za-z0-9_]{30,}|ghp_[A-Za-z0-9]{30,}|sk-(?:proj-|ant-)?[A-Za-z0-9_-]{24,}') def git(repo,*args): return subprocess.check_output(['git','--no-replace-objects','-C',str(repo),*args]) def tree(repo,ref): files={} command=['ls-files','--stage','-z'] if ref=='INDEX' else ['ls-tree','-r','-z',ref] for row in git(repo,*command).split(b'\0'): if not row:continue meta,name=row.split(b'\t',1);parts=meta.decode().split();name=name.decode() mode,oid=parts[0],parts[1] if ref=='INDEX' else parts[2] if SECRET.search(name.encode()):raise ValueError('文件路径名含敏感形态,标识:'+hashlib.sha256(name.encode()).hexdigest()[:12]) if mode not in ('100644','100755'):raise ValueError('不发布软链、子模块或冲突文件:'+name) if ref=='INDEX' and parts[2]!='0':raise ValueError('索引存在合并冲突') files[name]={'mode':mode,'oid':oid} return files def forbidden(name): p=Path(name) if any(x in p.parts for x in ('_private','_archive','.huashu','__pycache__','node_modules','.venv')):return True if p.name in ('voices.json','media.local.json'):return True if p.name.startswith('.env') and p.name!='.env.example':return True if p.name=='media.json' and name!='defaults/media.json':return True return p.suffix in ('.pem','.key','.p12','.pfx','.wav','.mov','.mp4','.pyc','.log') def inspect(repo,ref): hashes={} for name,entry in tree(repo,ref).items(): if SECRET.search(name.encode()):raise ValueError('文件路径名含敏感形态,标识:'+hashlib.sha256(name.encode()).hexdigest()[:12]) if forbidden(name):raise ValueError('包含私人配置或过程产物:'+name) data=git(repo,'cat-file','blob',entry['oid']) if SECRET.search(data):raise ValueError('发现私人路径或凭据形态:'+name) if name!=MANIFEST:hashes[name]=hashlib.sha256(data).hexdigest() return hashes def build_manifest(repo): if git(repo,'diff','--name-only').strip():raise ValueError('有未暂存修改,先审查并暂存再生成清单') return {'schema_version':1,'files':inspect(repo,'INDEX')} def check(repo,ref,base=None,email=None): if git(repo,'rev-parse','--is-shallow-repository').strip()==b'true': raise ValueError('浅克隆无法证明完整待发历史,请先获取完整历史') graft=Path(git(repo,'rev-parse','--git-path','info/grafts').decode().strip()) if not graft.is_absolute():graft=repo/graft if graft.exists() and graft.stat().st_size: raise ValueError('存在grafts历史改写,不能作为发布依据') if ref!='INDEX' and git(repo,'cat-file','-t',ref).strip()!=b'commit': raise ValueError('此门禁仅处理commit引用;标签对象需单独审查') expected=inspect(repo,ref) spec=':'+MANIFEST if ref=='INDEX' else ref+':'+MANIFEST try:manifest=json.loads(git(repo,'show',spec)) except (ValueError,subprocess.CalledProcessError):raise ValueError('缺少有效发布清单') from None if manifest!={'schema_version':1,'files':expected}:raise ValueError('已审清单与Git文件树不一致,需重新生成并复核') commits=[] if ref!='INDEX': revision=base+'..'+ref if base and set(base)!={'0'} else ref commits=git(repo,'rev-list',revision).decode().splitlines() # 每个待发提交都扫描,防止先提交秘密再删除后仅检查HEAD。 for commit in commits: raw=git(repo,'cat-file','commit',commit) if SECRET.search(raw):raise ValueError('提交元数据/说明含私人路径或凭据形态:'+commit[:12]) inspect(repo,commit) if email: identities=git(repo,'show','-s','--format=%ae%n%ce',commit).decode().splitlines() if any(x!=email for x in identities):raise ValueError('待发提交邮箱不符合配置:'+commit[:12]) return {'status':'passed','ref':ref,'files':len(expected),'outgoing_commits_checked':len(commits)} def main(): p=argparse.ArgumentParser(description=__doc__);p.add_argument('command',choices=['manifest','check']) p.add_argument('--repo',type=Path,required=True);p.add_argument('--ref',default='HEAD');p.add_argument('--base');p.add_argument('--author-email');p.add_argument('--pre-push',action='store_true') a=p.parse_args();repo=a.repo.resolve() if not (repo/'.git').exists():raise ValueError('必须明确指定Git仓库根,不能隐式扫描父仓库') if a.command=='manifest': (repo/MANIFEST).write_text(json.dumps(build_manifest(repo),ensure_ascii=False,indent=2)+'\n') print('已生成发布清单;审查后将它暂存,再运行check --ref INDEX。') return if a.pre_push: if git(repo,'status','--porcelain','--untracked-files=all').strip():raise ValueError('推送前工作区必须干净') lines=sys.stdin.read().splitlines() if not lines:raise ValueError('pre-push缺少真实待推送引用') for line in lines: local_ref,local_sha,remote_ref,remote_sha=line.split() if SECRET.search((local_ref+' '+remote_ref).encode()):raise ValueError('待推送引用名称含敏感形态') if set(local_sha)=={'0'}:raise ValueError('本检查不授权删除远端引用') print(json.dumps(check(repo,local_sha,remote_sha,a.author_email),ensure_ascii=False)) else:print(json.dumps(check(repo,a.ref,a.base,a.author_email),ensure_ascii=False)) if __name__=='__main__': try:main() except (ValueError,OSError,subprocess.CalledProcessError) as e: print('发布检查未通过:'+(str(e) if isinstance(e,ValueError) else type(e).__name__),file=sys.stderr);raise SystemExit(1)scripts/clip_lint.py5.7 KBd07cf0033262upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/deliver.py12 KB19b46010824dupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/audit_clip.py3.8 KB86531fb9729eupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/clip.html1.3 KBcfe53e3700feupstream ↗
<!doctype html> <html><head><meta charset="utf-8"><title>参数化动画片段 · clip</title> <style>html,body{margin:0;background:#222;overflow:hidden}canvas{display:block;max-width:100vw;max-height:100vh}#ui{position:fixed;left:8px;bottom:8px;color:#ddd;font:12px monospace}body.render #ui{display:none}</style></head> <body> <canvas id="c" width="1920" height="1080"></canvas> <div id="ui"><input id="scrub" type="range" min="0" max="10" step="0.0333" value="0" style="width:600px"> <span id="tt"></span></div> <!-- 口播管线用的参数化片段(references/09「片段契约」)。render.py --spec 注入 window.CLIP_SPEC;预览:clip.html?spec=examples/t3_finance_chart.json --> <script src="lib/fonts.js"></script> <script src="lib/util.js"></script> <script src="lib/paint.js"></script> <script src="lib/palettes.js"></script> <script src="lib/kit.js"></script> <script src="lib/motion.js"></script> <script src="lib/camera.js"></script> <script src="lib/diagram.js"></script> <script src="lib/typo.js"></script> <script src="lib/chart.js"></script> <script src="lib/ui.js"></script> <script src="lib/collage.js"></script> <script src="lib/media.js"></script> <script src="lib/toon.js"></script> <script src="lib/finish.js"></script> <script src="clip.js"></script> </body></html>scripts/engine/clip.js11 KBe0536dab8c0dupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/clip_schema.json2.8 KBb16e7e7aa087upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/clips/t1_3b1b.js18 KB4bc656b603a3upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/clips/t2_keynote_ui.js36 KB2a37bd961b5fupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/clips/t3_finance_chart.js21 KB67a378620483upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/clips/y1_kurzgesagt.js20 KBc9d322f3fe5eupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/clips/y2_vox.js14 KB86c516af2a2fupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/clips/y3_whiteboard.js21 KBa489f3ad4a69upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/clips/y4_storytime.js22 KB4d30c1f2368bupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/clips/y5_kinetic_type.js13 KB23d421c5cbbbupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/compare.py1.3 KB13c27ea81affupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/_shared/hero.js7.9 KB80ee19251279upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/_shared/hero/c1_0.png258 KB47c62308a808upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/_shared/hero/c1_1.png257 KB86ec86c47fb1upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/_shared/hero/c1_2.png278 KBa576008ac1d7upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/_shared/hero/c3_line.png999 KB45ae5035c2d1upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/eras.js2.3 KB6c4b39a5568aupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/8bit_act/01.png24 KB3ffb186c4c8dupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/8bit_act/02.png22 KBf4fc4d8b9483upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/8bit_act/03.png27 KBb7e595902959upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/8bit_act/04.png27 KB9ff1e99bbeb6upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/8bit_act/05.png28 KBf12d09d7b63eupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/8bit_act/06.png23 KBcb8c3074c785upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/8bit_act/07.png24 KB800875a88f21upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/8bit_act/08.png29 KBb33ac72e30b5upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/8bit_act/meta.json839 Baf6de690c641upstream ↗
{ "frames": [ { "file": "01.png", "ax": 120.0, "ay": 374.0, "w": 239, "h": 375, "air": 0 }, { "file": "02.png", "ax": 118.5, "ay": 323.0, "w": 239, "h": 324, "air": 0 }, { "file": "03.png", "ax": 116.5, "ay": 396.0, "w": 255, "h": 367, "air": 30 }, { "file": "04.png", "ax": 127.0, "ay": 420.0, "w": 246, "h": 372, "air": 49 }, { "file": "05.png", "ax": 136.0, "ay": 395.0, "w": 257, "h": 381, "air": 15 }, { "file": "06.png", "ax": 125.0, "ay": 316.0, "w": 247, "h": 317, "air": 0 }, { "file": "07.png", "ax": 106.5, "ay": 369.0, "w": 214, "h": 370, "air": 0 }, { "file": "08.png", "ax": 128.5, "ay": 397.0, "w": 267, "h": 373, "air": 25 } ], "ref_h": 371.0 }scripts/engine/demos/long_scroll/frames/8bit_walk/01.png28 KB82a071c1fd45upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/8bit_walk/02.png27 KB31a98062b284upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/8bit_walk/03.png27 KB73c02cfe1becupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/8bit_walk/04.png27 KB395e4072b421upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/8bit_walk/05.png28 KBe60019f5959fupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/8bit_walk/06.png26 KB99a872edc266upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/8bit_walk/07.png27 KBe762a2e6efc7upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/8bit_walk/08.png27 KB28d75bc79cbeupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/8bit_walk/meta.json835 B2181be41b069upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/egypt_dodge/01.png30 KB3a03ebf0b8b6upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/egypt_dodge/02.png35 KB99def0989c87upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/egypt_dodge/03.png33 KBf1e7f5184f1fupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/egypt_dodge/04.png36 KBd275318fbf77upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/egypt_dodge/05.png35 KB93081fbed63bupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/egypt_dodge/06.png33 KB7e26cef3977cupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/egypt_dodge/07.png30 KB11ab597c6adeupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/egypt_dodge/08.png32 KB90f53c0bd6d7upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/egypt_dodge/meta.json837 B464923c365f6upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/egypt_walk8/01.png35 KB029228ec696cupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/egypt_walk8/02.png33 KB9e3460e13f50upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/egypt_walk8/03.png33 KB06aaf9957cbcupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/egypt_walk8/04.png36 KB008e226d49ceupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/egypt_walk8/05.png34 KBf7cd5d996ac8upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/egypt_walk8/06.png32 KB5a78c8d4e50cupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/egypt_walk8/07.png34 KBf565793114c3upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/egypt_walk8/08.png34 KB79248351ee36upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/egypt_walk8/meta.json835 Bec779b459e70upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/monet_throw/01.png55 KBe8daefbc39e3upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/monet_throw/02.png60 KB57e6f8202731upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/monet_throw/03.png56 KB7a6fee67fa28upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/monet_throw/04.png58 KBd2462d60091aupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/monet_throw/05.png58 KBfd7f6e52b2aeupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/monet_throw/06.png57 KB17d345f571e8upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/monet_throw/07.png57 KB8eae301018efupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/monet_throw/08.png58 KB4ae33446bc85upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/monet_throw/meta.json873 B311c66442ba9upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/monet_walk/01.png52 KBb291311bc60aupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/monet_walk/02.png50 KBacbe8f6290b4upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/monet_walk/03.png50 KB743ad1699766upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/monet_walk/04.png48 KB83400e0c7cccupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/monet_walk/05.png52 KBcb83696aed33upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/monet_walk/06.png51 KBd5704949fc14upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/monet_walk/07.png50 KB99a7aa57ac6dupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/monet_walk/08.png48 KB1a119c1dc429upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/frames/monet_walk/meta.json873 B7f522f7bdfc5upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/lib/acts.js7.5 KB65e5de2549f9upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/lib/crossing.js45 KBdbb44eb1fa99upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/lib/world.js12 KB44d28d9203ebupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/lib/xing2_ding.js3.3 KBae09e166392cupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/lib/xing2_jia.js10 KB4ad47350f2ccupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/lib/xing2_yi.js13 KB03644346ded6upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/segments/02_egypt.js26 KBaeda459bc4f4upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/segments/14_8bit.js13 KBf0714ec58c97upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/long_scroll/segments/28_monet.js9.0 KB71a733f955efupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/t1_3b1b/eras.js864 B704a31b8d49fupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/t1_3b1b/t1_s1.js6.1 KBb5d0c35c211fupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/t1_3b1b/t1_s2.js5.5 KBea22ea60d260upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/t1_3b1b/t1_s3.js5.1 KB7dae1a437489upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/t2_keynote_ui/eras.js772 B212b967ad160upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/t2_keynote_ui/t2_s1.js3.8 KB5ffed95a2d0aupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/t2_keynote_ui/t2_s2.js4.8 KB8d68008d1617upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/t2_keynote_ui/t2_s3.js4.8 KB37086e031372upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/t3_finance_chart/eras.js907 B0bb6853c8fa6upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/t3_finance_chart/t3_s1.js4.2 KBfddcf7ff0876upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/t3_finance_chart/t3_s2.js3.4 KB614bb2472a06upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/t3_finance_chart/t3_s3.js4.1 KB50260994769bupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/y1_kurzgesagt/eras.js1.0 KB720affa52118upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/y1_kurzgesagt/kurzgesagt.js30 KBc75083c4c140upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/y2_vox/eras.js909 Bb0d34b5a39e8upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/y2_vox/vox.js20 KBb7effb124309upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/y3_whiteboard/eras.js936 Ba42fec27845eupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/y3_whiteboard/whiteboard.js12 KBea9ab4a7b353upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/y4_storytime/eras.js1.3 KBf7b1e4a9213fupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/y4_storytime/s1_desk.js1.2 KB6b55722c3985upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/y4_storytime/s2_screen.js7.3 KB89e79326b18fupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/y4_storytime/s3_react.js2.6 KBdb2857227cc6upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/y4_storytime/s4_wide.js974 Bf091cebe4ab0upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/y4_storytime/s5_close.js1.2 KB382abe057befupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/y4_storytime/story.js1.8 KB2c96caf62af7upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/y4_storytime/vo_env.js3.2 KB3d6a60938f16upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/y4_storytime/world.js6.0 KBd51b39da0d11upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/y5_kinetic_type/eras.js1.3 KB87285f334f5fupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/y5_kinetic_type/k1_title.js2.5 KBe45f67771af0upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/y5_kinetic_type/k2_complete.js3.9 KB1d6fedf02a16upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/y5_kinetic_type/k3_chat.js3.4 KB22c6d53ec528upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/y5_kinetic_type/k4_agent.js3.6 KB938d458b3e2fupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/demos/y5_kinetic_type/kinetic.js3.6 KBb8381d918ed9upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/engine.js17 KBee2f9dc0b7a4upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/eras.js6.0 KB72e6f9f4416eupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/eras_gallery.js15 KB41342810cb0bupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/eras_style_studies.js4.0 KB0b2089606459upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/examples/assets/K线图_横屏.png133 KB9e681fc78273upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/examples/assets/终端_for循环求和.png44 KBb2919580d11cupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/examples/assets/终端_sum求和.png42 KB4e33931eb947upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/examples/assets/营收图_竖屏.png92 KB5735ce105923upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/examples/assets/财经图表_横屏.png102 KBb001e61fc486upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/examples/t1_3b1b.json1.1 KB32ac81b954dcupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/examples/t2_fullbleed_pulse.json733 B7d4fd72b94e0upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/examples/t2_keynote_ui.json752 Bc53acf6cf43fupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/examples/t3_finance_chart.json945 Bd4cdc4431c26upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/examples/y1_kurzgesagt.json498 Bf352ffc9e913upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/examples/y2_vox.json508 B10c2d6548027upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/examples/y3_whiteboard.json688 Bfd53cae2f13eupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/examples/y4_storytime.json1.1 KBa448fc6c316fupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/examples/y5_kinetic_type.json598 Bd3a57a2be7e0upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/index.html2.0 KB016a01f0d39bupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/brush.js12 KB1b993a940952upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/camera.js9.5 KB7b50c44db598upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/chart.js11 KBe18de00ae35aupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/collage.js8.8 KBf3491e46ac40upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/diagram.js20 KB8baf28f1c084upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/finish.js5.0 KB4d711d6b06f0upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts.js4.3 KBc96e0f285a99upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/AlfaSlabOne-400.woff41 KBed5ca898aee8upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/Anton-400.woff60 KBaf925c19f8fbupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/ArchivoBlack-400.woff37 KB5273a9c1dd78upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/Bangers-400.woff43 KB46c888483959upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/Caveat-600.woff129 KBa9fc847617e4upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/Cinzel-400.woff22 KB08649ede82c8upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/Cinzel-700.woff24 KB28f02c19e71eupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/cmunrm.woff120 KB9efca1264afaupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/cmunti.woff155 KB122340ad4880upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/CormorantGaramond-500.woff88 KB4dccf368b0e5upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/CormorantGaramond-500i.woff91 KB3bc493df4e45upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/CormorantGaramond-700.woff85 KBf1ddbfb1ce61upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/EBGaramond-400.woff159 KB7952db1dbe1bupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/EBGaramond-400i.woff148 KBf6534947ee5bupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/EBGaramond-600.woff175 KBd30aa9dc3d3aupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/Exo2-800i.woff59 KB7437aeef4e45upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/GochiHand-400.woff23 KBb500d9a883dbupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/IMFellEnglish-400.woff106 KBf9b84ad24d7dupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/IMFellEnglish-400i.woff109 KB6c1d5e42240dupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/index.json1.9 KB442742f351a2upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/Inter-var.woff124 KBace36e8017beupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/JosefinSans-700.woff26 KB2e2bfbf8407cupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/Kalam-700.woff182 KBb31a7658b9e2upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/LibreBaskerville-400.woff46 KB70100bd82e15upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/LibreBaskerville-700.woff46 KB1ede9e63b750upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/LICENSES.md4.4 KB0533deaf8ce6upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/LilitaOne-400.woff14 KB73646caa4abcupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/LXGWWenKai-500.woff2.0 MB4f02c88fd17dupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/Marcellus-400.woff22 KB8d5939885667upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/NotoSansSC-500.woff1.4 MB4c7cade7f4d9upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/NotoSansSC-700.woff1.4 MB4e5c21514f3cupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/NotoSansSC-800.woff1.4 MB056fba642465upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/NotoSansSC-900.woff1.3 MB664a8d52cc39upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/NotoSansSC-sub.woff323 KB972a3b100fd1upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/NotoSerifJP-600.woff4.2 MB14163512883cupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/NotoSerifSC-sub.woff425 KB62d4bee4dc6eupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/OFL.txt4.3 KBa4670b02de28upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/PlayfairDisplay-400i.woff50 KBb1f80d059ff7upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/PlayfairDisplay-700.woff54 KB507012058185upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/PlayfairDisplay-700i.woff52 KB95def380ef3bupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/Poppins-600.woff65 KB5035ee775d44upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/Poppins-700.woff64 KB64fbbd9ae3c0upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/Poppins-800.woff64 KBfe74641c202bupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/PressStart2P-400.woff35 KB302418fa53e0upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/Righteous-400.woff21 KB59c797f127fcupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/RobotoCondensed-var.woff92 KB1196643983d7upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/Rye-400.woff61 KB66c64f2e9616upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/fonts/UnifrakturMaguntia-400.woff38 KB580bb9bd4e6eupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/kit.js6.5 KBe0790d7eeafdupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/media.js9.2 KB5e24d18f8d59upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/medium.js23 KB4b681f5c9382upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/motion.js9.2 KB6ee9cfa1f9aeupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/paint.js15 KB6fc5bedd62c6upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/palettes.js4.8 KB172b2a2d9d3cupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/post.js9.8 KB63b4a2c5a606upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/render.js17 KBae98ce074111upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/rig.js33 KB01db1550076aupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/rig_huashu.js13 KB1beff1622066upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/three.min.js654 KB0c2f4ba35100upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/THREE_LICENSE.txt1.1 KBd07e20968651upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/toon.js16 KBb0eae6595c33upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/typo.js11 KBf00d3675480fupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/ui.js10 KB8dbf8f16c41dupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/lib/util.js7.2 KB0b3db4299e12upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/rick_morty/index.html1.6 KB73036ec15a1bupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/rick_morty/render.py3.3 KB7a391d29d9bcupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/rick_morty/room.js14 KB7d9c98d01e42upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/rick_morty/svg-kit.js6.3 KB3994f6f203a0upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/eras_lin.js787 B17603afb2de6upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/eras_spacex.js539 B9b2bac76b3f5upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/eras_vox.js708 B07c65965e969upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/eras_wbx.js952 B490bc50c05e0upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex/rockets.js11 KBbf827caa4577upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex/scenes/s01.js13 KBc9cc01ea1c2bupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex/scenes/s02.js7.3 KBb264e0bcd0c0upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex/scenes/s03.js19 KBd41b7691b6c2upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex/scenes/s04.js20 KB1273dda11e0fupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex/scenes/s05.js18 KBddb5e5f96541upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex/scenes/s06.js17 KB61499bb473f9upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex/scenes/s07.js29 KB071d34e8e1e6upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex/scenes/s08.js12 KBf7d684281901upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex/scenes/s09.js19 KB0262e7dcf8e2upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex/scenes/s10.js21 KB2b9555ed798fupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex/scenes/s11.js27 KB5a2cec6f0bf0upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex/scenes/s12.js26 KB9d86fa906088upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex/timing.js15 KBaed95b06b1acupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex/tm.js1.0 KBa0014191409fupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_lin/common.js40 KB90ff436aeaafupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_lin/presenter_meta.js6.2 KB4d4f5362a251upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_lin/scenes/s01.js15 KB5cbd1606961bupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_lin/scenes/s02.js6.5 KB44133d041cf6upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_lin/scenes/s03.js12 KB8e42c9374656upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_lin/scenes/s04.js8.1 KB51126a8e77ffupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_lin/scenes/s05.js12 KB281126f55b04upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_lin/scenes/s06.js6.8 KBe2d9f28d5444upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_lin/scenes/s07.js9.3 KBb7b529915539upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_lin/scenes/s08.js14 KBd01900b2181fupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_lin/scenes/s09.js9.2 KBc01b7bbd1875upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_lin/scenes/s10.js11 KB046741d1ad0fupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_lin/scenes/s11.js23 KB7ee58fed7c01upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_lin/scenes/s12.js11 KB6d5221da7d05upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_lin/tools/prep_presenter.py3.1 KB892ab71d1e1fupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_vox/common.js37 KB4b09a7be4882upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_vox/scenes/s01.js10 KBee62f128b13cupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_vox/scenes/s02.js5.9 KBf793f04212cbupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_vox/scenes/s03.js8.0 KB53260a25ffc9upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_vox/scenes/s04.js9.6 KB4de6ccd3a1dbupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_vox/scenes/s05.js11 KB2a21c48ac81aupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_vox/scenes/s06.js7.3 KB785b2fbc3587upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_vox/scenes/s07.js8.1 KBa770a78dc618upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_vox/scenes/s08.js9.8 KB2a5fff5b798eupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_vox/scenes/s09.js6.1 KBa9722d1b1f07upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_vox/scenes/s10.js12 KB6ac534e4bd80upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_vox/scenes/s11.js13 KBa3a8426965edupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_vox/scenes/s12.js10 KBf6210fbbf42aupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_wb/assets/ARPHICPL.TXT6.7 KBde7491ae3d4eupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_wb/assets/hand_1.png152 KBa8146eb91ba1upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_wb/assets/hand_2.png154 KB29e942086392upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_wb/assets/hand_3.png161 KB1a611d7f02d0upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_wb/assets/make_hand.py1.6 KB81d0b8cb2170upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_wb/assets/strokes.js55 KBa449d407a7c8upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_wb/common.js53 KB007a2636404bupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_wb/scenes/s01.js6.7 KB643bcf0b09dbupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_wb/scenes/s02.js4.7 KB795e20d7bf5fupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_wb/scenes/s03.js7.0 KB4a1b35e449e7upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_wb/scenes/s04.js4.0 KB69f23e823bf7upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_wb/scenes/s05.js6.4 KB649bfaf67892upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_wb/scenes/s06.js4.9 KBad08d87324d2upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_wb/scenes/s07.js5.9 KB5f4f13845c47upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_wb/scenes/s08.js8.6 KBb2a08827fa93upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_wb/scenes/s09.js4.4 KB1d3776fb8203upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_wb/scenes/s10.js11 KB0b05bb2b7e0aupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_wb/scenes/s11.js7.4 KB4996fc9e9dfdupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/spacex_wb/scenes/s12.js5.9 KB31616baba7e0upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/transitions/cardFlip.js2.3 KB93a219c60ef4upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/transitions/deskPullOut.js535 B3f852ea0fe10upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/transitions/kzBloom.js1.8 KB0bf6c0958485upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/transitions/lin_page.js5.9 KB82a9ed939196upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/transitions/paperSlide.js1.1 KBac12cade7585upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/transitions/vox_pull.js591 Bb7580da060f2upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/reference_films/spacex/transitions/vox_slap.js771 Bc4c128ab4a39upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/render.py10 KB1fb5dc1a7bfeupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/01_cave.js31 KBd9b9dae22060upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/02_egypt.js49 KBb562fdbdf867upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/03_greek.js40 KB731f61622246upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/04_roman.js46 KBa0c813416a55upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/05_gothic.js39 KB5df30d9435e7upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/06_renaissance.js31 KBd6df5f9eba63upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/08_impressionism.js20 KBfaaf347a12cdupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/09_postimp.js16 KB483ecbfc9166upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/10_nouveau.js30 KB0cff9a9cb978upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/11_cubism.js27 KBe26e76a50c7cupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/12_bauhaus.js17 KBfeade9f50aedupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/13_pop.js24 KB04c1fc844a9cupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/14_8bit.js17 KB8418bef5b6ffupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/15_raytrace.js25 KB11838972b0d5upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/16_2026.js49 KB2d344cc46db9upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/17_ink.js22 KB59cc47c3f641upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/18_klimt.js20 KB3ebe5dce8b79upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/19_munch.js13 KBc9d35197e3a2upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/20_dunhuang.js21 KB376bcc52c205upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/21_kusama.js12 KBa0925da636bfupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/22_constructivism.js16 KBca468b5cf06dupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/23_dali.js20 KB2b02d6e9eb44upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/24_hopper.js13 KB4f9eb1116d49upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/25_ghibli.js22 KBbc9255fce349upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/26_vaporwave.js17 KB02eb85d089ceupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/27_kirby.js21 KB16b54bc666f3upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/28_monet.js17 KB424e9a165c70upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/29_seurat.js11 KB3cb8617479c0upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/30_matisse.js13 KB2617fd7b55d3upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/31_haring.js10 KBaca68222ae44upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/32_rembrandt.js20 KB264178b24fc6upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/33_rubberhose.js18 KBb434105cc667upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/34_shadowpuppet.js18 KB02d9f91148feupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/35_shinkai.js22 KB8485c2925e4dupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/36_picasso_blue.js11 KB5db471d78b54upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/38_giotto.js14 KB6f0c5fd32181upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/39_piero.js5.9 KB9d9232b26193upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/40_raphael.js11 KB4887be8ca7a8upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/41_michelangelo.js13 KBb23f33f2d889upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/42_guardi.js8.3 KBd34ef1e0ea68upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/43_lorenzetti.js16 KB7e8ecdf35383upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/44_cortona.js11 KB96dff0b80553upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/45_renoir.js10 KB20dace328bc0upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/46_pissarro.js11 KB25a435cf0fcaupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/47_sisley.js11 KB8306dfe282c3upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/48_gauguin.js13 KB9e2e84c147c2upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/49_chagall.js8.8 KB8df4273d9ebdupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/50_ma_yuan.js13 KBf23f62f41a5cupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/51_persian_miniature.js17 KB1be77edb80cfupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/52_south_park.js10 KB367c8b2711a6upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/53_pixar.js9.7 KB98b765e2b779upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/54_ghibli_lab.js12 KBec61d94e7ea7upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/55_dunhuang.js20 KB29c688d80bdaupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/scenes/index.js2.4 KB498c2b583c73upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/spec_assets.py3.6 KB2936dd108e31upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/transitions.js68 KBb3b2520f0598upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/engine/时代作者规范.md4.6 KBadb4060e0518upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/film_gate.py43 KB45cf2a2cadf0upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/font_subset.py7.4 KB33614ba5b3e7upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/images.py15 KB839c057f7062upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/key_green.py1.9 KBe2271d008207upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/key_split.py1.6 KB357a4f25dc0fupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/koubo.py18 KBe6bbe77162b7upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/qa.py26 KB98758fc3c707upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/storyboard_lint.py32 KB07c8e2ede018upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/subzone_gate.py1.8 KB03add14ba713upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
scripts/voice_review.py1.9 KBc85d27c0de52upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/01-拆解.md4.6 KB0133872dddf7upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/02-机制.md5.0 KB316567657b10upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/03-一帧先行四条路线.md4.6 KBe459acd1fc76upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/04-纯代码绘制.md11 KB832d59eddb10upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/05-节奏与配乐.md1.5 KB27b893f68b79upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/06-口播驱动的艺术短片.md2.7 KBd575b989ab3cupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/07-正面经验.md21 KB8740be767957upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/08-风格作者规范.md5.6 KB88a1f31b770bupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/09-视频动画语法.md53 KB0ecd580c1667upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/10-角色.md3.7 KBf6bd1a94d8ceupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/11-长卷穿越片.md4.2 KBf1c22d0dab83upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/12-口播整片与经验回流.md2.2 KB0ad71ec640a5upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/13-口播与语音复刻.md587 B56fc1521502cupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/15-跨场景风格提炼.md4.9 KB0d5b0cb60beeupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/16-风格演示95分协议.md8.8 KB1dffc895f592upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/about-author.md9.7 KB171ab35fc0aeupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/capabilities.md7.2 KB35a8872afe41upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/compatibility.md2.9 KB9d0d9e3268b5upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/images.md5.7 KB78fbb82b06b6upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/voice-model-review.md2.0 KB01a3dc174cb7upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/动画语法/t1_3b1b.md14 KB9d65e7586641upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/动画语法/t2_keynote_ui.md14 KBd923fd02db4aupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/动画语法/t3_finance_chart.md13 KBe340c36a81afupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/动画语法/y1_kurzgesagt.md17 KBfacfc8228b10upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/动画语法/y2_vox.md16 KBffcdb0c8a704upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/动画语法/y3_whiteboard.md15 KB1ceef5f81762upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/动画语法/y4_storytime.md18 KB6e99470145c5upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/动画语法/y5_kinetic_type.md17 KB5583e154912cupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/动画语法/y6_presenter_explainer.md1.9 KB94a88fcbc122upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/口播API.md5.5 KBc3177bd8c791upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/色板.md3.3 KB78e7759a5c67upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/角色精细度.md5.2 KB8a7222d074c6upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/镜头表模板.md9.3 KBe0e565b817acupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/镜头语言.md10 KB8d6d7e7ccd5cupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/01_cave.md7.0 KB281849af2ec2upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/02_egypt.md8.7 KB0816d71077f5upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/03_greek.md8.7 KBa4b3e6c7fd1eupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/04_roman.md12 KB7131869ee44aupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/05_gothic.md4.1 KB0ed79e239d6eupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/06_renaissance.md4.0 KB8405bcd6bea3upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/08_impressionism.md4.1 KBa2593e510315upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/09_postimp.md1.5 KB50d7c8d8711fupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/10_nouveau.md5.0 KB71ac0d6a18dcupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/11_cubism.md5.3 KB40a647b8cc1aupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/12_bauhaus.md4.3 KBcbc229877659upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/13_pop.md4.2 KBa940e8512e82upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/14_8bit.md2.8 KB042c11185a9dupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/15_raytrace.md3.1 KB2a1638c282deupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/16_2026.md5.7 KBe092f569a849upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/17_ink.md4.3 KB6a8e4c418821upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/18_klimt.md2.8 KB29178ca219f6upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/19_munch.md2.8 KB09111285f852upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/20_dunhuang.md3.0 KBcd301f2fab6dupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/21_kusama.md2.2 KBb95a193aa1b6upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/22_constructivism.md3.9 KBea5e95da0a2dupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/23_dali.md3.4 KB4b4a54a86479upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/24_hopper.md2.8 KBb48b743d9a5dupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/25_ghibli.md6.6 KBb7af3b709818upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/26_vaporwave.md3.1 KB1087ae6cf953upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/27_kirby.md3.3 KBd213d6e2eadbupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/28_monet.md2.8 KBcef93afef282upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/29_seurat.md2.9 KB5b5cf9405b6fupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/30_matisse.md2.5 KB2f9ada95b11fupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/31_haring.md1.9 KB147e6969b8c1upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/32_rembrandt.md2.8 KB80793558e729upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/33_rubberhose.md2.4 KBa2e4c90725a2upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/34_shadowpuppet.md3.1 KB2fe632891b66upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/35_shinkai.md3.3 KB584b6c653e83upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/36_picasso_blue.md2.7 KBfa017a3cb04aupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/37_rick_morty.md4.5 KB6dccdb9dfbc7upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/38_giotto.md5.1 KB53a0568f2ec5upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/39_piero.md3.9 KB0c5f6291c195upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/40_raphael.md4.6 KB7a546c9f1d06upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/41_michelangelo.md4.8 KB8517df15c3d0upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/42_guardi.md4.3 KB015c11990f78upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/43_lorenzetti.md5.0 KB9541f2a8fb08upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/44_cortona.md4.8 KB1f3d66ffafceupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/45_renoir.md4.8 KB7f87cc6a77a2upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/46_pissarro.md4.5 KBc2fffa92ff45upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/47_sisley.md4.5 KB2c227901c701upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/48_gauguin.md4.3 KB781e1022b40fupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/49_chagall.md3.8 KB5a668e8f8917upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/50_ma_yuan.md5.0 KBd94b1343ee90upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/51_persian_miniature.md4.7 KBf70e8334aa10upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/52_south_park.md4.6 KB9e22a52e3025upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/53_pixar.md4.4 KBd5c33cee0588upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/55_dunhuang.md3.4 KBd25e3f27c1a4upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/_转场_迁移测试.md6.9 KB4f05627ad08fupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/_重做进度.md4.7 KB9ceff0130cd1upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/_音轨_艺术史速通.md12 KB070a0fe0f88bupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
references/风格配方/INDEX.md13 KB6d0b8906c24cupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
assets/showcase/clawd-clips.md1.7 KB056af58ced26upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
assets/showcase/clawd-dawn.gif3.3 MBc8d061705d1fupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
assets/showcase/clawd-trex.gif7.2 MBbfd1cae6340dupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
assets/showcase/clawd-vangogh.gif7.3 MB07e25083376cupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
assets/showcase/egypt.gif4.7 MBfac4f54e69bbupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
assets/showcase/hero.png1.6 MB1caac91f2de6upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
assets/showcase/mario-claude.gif2.0 MB4d949644aed9upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
assets/showcase/mario-clips.md1.1 KBb7f7df7fce8fupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
assets/showcase/mario-combo.gif3.0 MB331e39379e54upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
assets/showcase/mario-pipe.gif3.9 MB265214302343upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
assets/showcase/monet.gif5.1 MB9a70a6f0a2d8upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
assets/showcase/pixel.gif2.5 MB1e6568d98df0upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
assets/全风格总览.jpg1.5 MB9f834ab2eb4aupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
assets/动画语法/t1_3b1b_总览.jpg224 KB822f07d52d2aupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
assets/动画语法/t2_keynote_ui_总览.jpg198 KB2db960251fb0upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
assets/动画语法/t3_finance_chart_总览.jpg181 KB8ae77d02479dupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
assets/动画语法/y1_kurzgesagt_总览.jpg172 KBdf0f5dea278bupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
assets/动画语法/y2_vox_总览.jpg311 KB92f6230119afupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
assets/动画语法/y3_whiteboard_总览.jpg140 KB0e19c2cb4605upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
assets/动画语法/y4_storytime_总览.jpg149 KBb2d9dedbc867upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
assets/动画语法/y5_kinetic_type_总览.jpg132 KBb1a547c25039upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
assets/新风格v2_证据/38_giotto.jpg390 KBa2c4d4255509upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
assets/新风格v2_证据/40_raphael.jpg500 KB24b37de5ceddupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
assets/新风格v2_证据/41_michelangelo.jpg300 KB8f833f4a8f58upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
assets/新风格v2_证据/43_lorenzetti.jpg448 KB7ba37029c0a6upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
assets/新风格v2_证据/44_cortona.jpg316 KB2d54093e6a1bupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
assets/新风格v2_证据/46_pissarro.jpg490 KBb9a4ae6036caupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
assets/新风格v2_证据/47_sisley.jpg355 KB8d553e83b900upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
assets/新风格v2_证据/48_gauguin.jpg445 KBb0b4aefdc8d6upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
assets/新风格v2_证据/50_ma_yuan.jpg360 KB8ebe70225f35upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
assets/新风格v2_证据/51_persian_miniature.jpg382 KB477988a7c4e3upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
assets/新风格v2_证据/总览.jpg574 KB42845352791fupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
assets/角色/花叔形象_对比.jpg291 KB5460db2eb48eupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
assets/长卷穿越_总览.jpg472 KB26c83738d1ecupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
.github/workflows/tests.yml1.4 KBb1f4db143d63upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
.gitignore388 Bb8f8fa40a91dupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
CHANGELOG.md15 KB9403eedcdd56upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
CONTRIBUTING.md2.2 KB4cf71c00a98fupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
defaults/media.json916 Ba04ecc4dc9f0upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
LICENSE1.1 KBed67368a2a79upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
README.en.md19 KBcc7256b605dcupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
release-manifest.json59 KB33994a4b473eupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
schemas/media.schema.json4.8 KB7b88da369cb8upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
tests/test_capabilities.py12 KBff19beaf94e1upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
tests/test_clip_theme.py9.2 KB450389d12bbaupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
tests/test_clip_world.py4.1 KBe926d10264c3upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
tests/test_film_gate.py11 KBd25a80e076e5upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
tests/test_film_gate_blank.py5.2 KBc4a887b2022fupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
tests/test_film_gate_color.py3.0 KBf8cceeb2c2c7upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
tests/test_images.py8.2 KB172b6822372bupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
tests/test_local_server.py4.1 KB28e3591e9fdcupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
tests/test_release.py4.2 KB3e5d6513765dupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
tests/test_storyboard_lint.py18 KB211ae80fdd3fupstream ↗
Contents are not inlined in the catalog. Open the upstream file.
tests/test_voice_review.py1.4 KBd70ac9d8c0f4upstream ↗
Contents are not inlined in the catalog. Open the upstream file.
Plain-text copy, no login: /skill-bundles/alchaincyf-huashu-art-motion.md