DeepTutor: Lifelong Personalized Tutoring. https://deeptutor.info/.
Python ai-agentsai-tutorclawbotcli-tooldeepresearchinteractive-learninglearningmulti-agent-systemsrag
"use client";
import type { ReactNode } from "react";
import { ChevronRight } from "lucide-react";
/**
* One curve and one duration for every disclosure in the activity
* vocabulary, so a header and the fold hanging from it settle on the same
* beat. The curve is the fast-out/long-settle shape the sidebar's "More"
* disclosure uses — crisper than `ease-out` at this size.
*
* Reduced motion needs no opt-out here: `globals.css` shortens every
* transition to a near-imperceptible duration under that setting.
*/
export const FOLD_EASE = "duration-[220ms] ease-[cubic-bezier(0.32,0.72,0,1)]";
/**
* The caret that says a line is a disclosure.
*
* Kept at the same weight as the metadata beside it ("· 33s · 4 tool calls"):
* on a line that is mostly read and occasionally clicked, the caret is there
* to promise the content is recoverable, not to ask to be pressed. It lifts
* to full contrast with the rest of the line on hover.
*/
export function FoldCaret({ open }: { open: boolean }) {
return (
<ChevronRight
aria-hidden
className={`size-3 shrink-0 text-[var(--muted-foreground)]/45 transition-[transform,color] group-hover/act:text-current ${FOLD_EASE} ${
open ? "rotate-90" : ""
}`}
/>
);
}
/**
* Content that folds away without being unmounted.
*
* Animates on `grid-template-rows` between `0fr` and `1fr`, which is the one
* way to transition to a height the content decides — no measuring, no
* `max-height` guess that either clips a long trace or makes a short one
* crawl open.
*/
export function ActivityFold({
open,
children,
className = "",
}: {
open: boolean;
children: ReactNode;
className?: string;
}) {
return (
<div
className={`grid transition-[grid-template-rows,opacity] ${FOLD_EASE} ${
open ? "grid-rows-[1fr] opacity-100" : "grid-rows-[0fr] opacity-0"
} ${className}`}
>
<div className="overflow-hidden">{children}</div>
</div>
);
}