Codebase intelligence for AI and humans: code health scores, auto-generated docs, git analytics, dead code detection, and architectural decisions via MCP.
Python aicode-complexitycode-healthcode-intelligencecode-qualitydead-codedeveloper-toolsdocumentationgit-analyticsmcpopen-sourcepythonrefactoringstatic-analysistechnical-debt
"use client";
import { useTheme } from "next-themes";
import { ThinkingOrb, type OrbSize, type OrbState } from "thinking-orbs";
import { cn } from "../lib/cn";
export type { OrbSize, OrbState };
/**
* One orb animation per kind of work, so the same motion always means the
* same thing across the product. Pick by what the wait is doing:
*
* - `thinking` breathing - model reasoning before any output
* - `searching` searching - search / index lookups
* - `tracing` connecting - graph, call-path and architecture layout
* - `working` working - generic tool or job work (inline default)
* - `writing` composing - generating text or pages (region default)
* - `analysing` solving - health, risk and other analysis passes
* - `workspace` weaving - cross-repo / workspace work
* - `page` shaping - a whole page or panel assembling
*/
export const ORB_STATE = {
thinking: "breathing",
searching: "searching",
tracing: "connecting",
working: "working",
writing: "composing",
analysing: "solving",
workspace: "weaving",
page: "shaping",
} as const satisfies Record<string, OrbState>;
export type OrbIntent = keyof typeof ORB_STATE;
export interface OrbLoaderProps {
/**
* The animation, ideally from `ORB_STATE` (e.g. `ORB_STATE.tracing`).
* Defaults to `composing` at 64px and `working` at 20/32px.
*/
state?: OrbState;
/**
* 64 for a region or canvas, 20 for inline next to text. 32 is an
* interpolated in-between for tight panels. Default 64.
*/
size?: OrbSize;
/** Accessible name, and the visible caption at 64px. Default "Loading". */
label?: string;
/** Show `label` under the orb. Defaults to true at 64px only. */
showLabel?: boolean;
/**
* Centre the orb in all the space its parent gives it. Off by default;
* the loader never invents a height of its own.
*/
fill?: boolean;
className?: string;
}
/**
* The loader for waits with no known shape (canvases, graphs, streaming).
* Known shapes use `PageSkeleton` / `SkeletonRegion`; buttons use `Spinner`.
*
* A monochrome dotted orb on a 2D canvas: no network, no WASM, pauses
* offscreen, and shows a still frame under reduced motion. The canvas paints
* client-side only, at a fixed size, so SSR reserves the box without a shift.
* The theme comes from the app's resolved theme rather than the library's
* own detection, which falls back to the OS setting when no ancestor carries
* a `dark`/`light` class (the VS Code webview's light theme).
*/
export function OrbLoader({
state,
size = 64,
label = "Loading",
showLabel,
fill = false,
className,
}: OrbLoaderProps) {
const { resolvedTheme } = useTheme();
const theme = resolvedTheme === "dark" || resolvedTheme === "light" ? resolvedTheme : "auto";
const region = size === 64;
const caption = showLabel ?? region;
const orb = (
<ThinkingOrb
state={state ?? (region ? ORB_STATE.writing : ORB_STATE.working)}
size={size}
theme={theme}
aria-hidden="true"
className="shrink-0"
/>
);
if (!region) {
return (
<span
role="status"
aria-label={label}
data-orb-loader=""
className={cn(
"inline-flex items-center gap-2 text-xs text-[var(--color-text-tertiary)]",
fill && "h-full w-full justify-center",
className,
)}
>
{orb}
{caption && <span>{label}</span>}
</span>
);
}
return (
<div
role="status"
aria-label={label}
data-orb-loader=""
className={cn(
"flex flex-col items-center justify-center gap-3",
fill && "h-full min-h-0 w-full flex-1",
className,
)}
>
{orb}
{caption && <span className="text-xs text-[var(--color-text-tertiary)]">{label}</span>}
</div>
);
}