@dataimago/ui 0.1.0-alpha.4 → 0.1.0-alpha.43

package/README.md

@@ -2,27 +2,49 @@
 
 React components for the dataimago ethical AI design system.
 
-> **Registry:** published to **public npm** (`registry.npmjs.org`) under the `@dataimago` scope alongside `@dataimago/tokens` and `@dataimago/css`. No scoped `.npmrc` or `GITHUB_PACKAGES_TOKEN` is required — a plain `pnpm install @dataimago/ui` works.
+> **Registry:** published to **public npm** (`registry.npmjs.org`) under the `@dataimago` scope alongside the rest of the design-system packages. No scoped `.npmrc` or `GITHUB_PACKAGES_TOKEN` is required — a plain `pnpm install @dataimago/ui` works.
 
 ## Peer dependencies
 
 - `react` ≥ 18
 - `react-dom` ≥ 18
+- `@dataimago/brand` ≥ 0.1.0-alpha.0
 - `@dataimago/tokens` ≥ 0.1.0-alpha.0
 - `@dataimago/css` ≥ 0.1.0-alpha.0 (visual styling consumes the semantic CSS variables emitted by this package)
 
-## Alpha scope
+## Alpha scope and exports
 
-Just one smoke component for now:
+Current public exports:
+
+- `<Button>` — smoke-level button using semantic CSS variables.
+- `<LogoMonogram>`, `<LogoMonogramBg>`, `<LogoWordmark>` — legacy outlined SVG brand marks.
+- `<Logo>` — manifest-driven mark renderer backed by `@dataimago/brand`.
+- `<BrandAnimation>` — decorative brand animation with reduced-motion poster fallback.
+- `<NavbarLogo>` — animated navbar mark with reduced-motion poster fallback.
+- `<CaptionedVideo>` — a video with the video-foundry's caption treatment drawn over it (frosted panel, word tone, MathML), from a release's `captions.<lang>.json`; the WebVTT file stays as the hidden native track and the fallback text. Fullscreen takes the whole player so the captions come along. `<CaptionLayer>` draws one moment of it (pure presentation), and the caption functions (`pageAt`, `wordTone`, `captionLook`, …) match the foundry's burned-in treatment (the foundry imports them once this ships; #407). Needs `@dataimago/css` loaded for the palette, `--font-family-sans` and the self-hosted Noto Sans Math. `poster` is required and nothing autoplays. Under `prefers-reduced-motion: reduce` the player plays `reducedSrc` if given; otherwise it says the video is animated throughout and offers `storyboard` (a link) first, and plays the full cut only if the reader chooses it.
 
 ```tsx
-import { Button } from '@dataimago/ui';
+<CaptionedVideo
+  src="/research/liwld/video/liwld-draft-16x9.mp4"
+  poster="/research/liwld/video/liwld-draft-16x9.poster.png"
+  captions="/research/liwld/video/liwld-draft-16x9.captions.en.json"
+  fallbackCaptions="/research/liwld/video/liwld-draft-16x9.captions.en.vtt"
+  chapters="/research/liwld/video/liwld-draft-16x9.chapters.en.vtt"
+  title="Growth without linkage"
+/>
+```
+
+```tsx
+import { BrandAnimation, Button, Logo, NavbarLogo } from '@dataimago/ui';
 
 <Button variant="primary">Generate</Button>
 <Button variant="secondary" size="sm">Cancel</Button>
+<Logo name="dataimago_dash_ai" variant="PATH" title="dataimago-ai" />
+<NavbarLogo name="dataimago-ai" />
+<BrandAnimation name="dataimago-ai" />
 ```
 
-Additional components get harvested from consumer prototypes (`dataimago-ai`, `dissertation.ai`, etc.) via the prototype-in-consumer workflow.
+Additional components get harvested from consumer prototypes (`dataimago-ai`, `dissertation-ai`, etc.) via the prototype-in-consumer workflow. Broad new primitives are deferred until the substrate is stable and L1/L2 usage identifies reusable candidates.
 
 ## Building
 
@@ -46,3 +68,7 @@ use semantic CSS variables rather than app-local Tailwind classes. When the dev
 server is running, the Storybook MCP addon is available at
 `http://localhost:6006/mcp` for AI-assisted component discovery and story
 testing.
+
+Accessibility checks are progressively hardened: current stories surface a11y
+findings for review, and the gate should move to blocking only after the existing
+story set is clean.