DeepTutor: Lifelong Personalized Tutoring. https://deeptutor.info/.
Python ai-agentsai-tutorclawbotcli-tooldeepresearchinteractive-learninglearningmulti-agent-systemsrag
"use client";
import { useEffect, useState } from "react";
import { Check } from "lucide-react";
import { useTranslation } from "react-i18next";
import {
listDiscussionModes,
type DiscussionMode,
} from "@/lib/partner-groups-api";
/**
* How the group talks.
*
* The backend's own label/description are written for the model, so the copy
* here is the product's: what actually happens, plus how long it takes —
* a group turn costs one model call per member, and the modes differ by a
* factor of two in both calls and waiting. Hiding that behind a bare name
* would let someone pick "debate" without knowing they doubled the wait.
*
* Unknown modes fall back to the server's own strings, so a mode added later
* still renders before this table knows about it.
*/
function useModeCopy() {
const { t } = useTranslation();
return (
mode: DiscussionMode,
): { label: string; hint: string; cost: string } => {
switch (mode.name) {
case "panel_parallel":
return {
label: t("Parallel answers"),
hint: t("Everyone answers the same question at once, independently."),
cost: t("Fastest"),
};
case "sequential":
return {
label: t("One after another"),
hint: t(
"Each member sees what the others already said and builds on it.",
),
cost: t("Slower · they wait for each other"),
};
case "debate":
return {
label: t("Cross-examination"),
hint: t(
"First everyone states a position, then each responds to the disagreements.",
),
cost: t("Two rounds · about twice the wait"),
};
default:
return { label: mode.label, hint: mode.description, cost: "" };
}
};
}
export default function DiscussionModePicker({
value,
onChange,
title,
note,
disabled = false,
}: {
value: string;
onChange: (name: string) => void;
/** Rendered with the options, so a single-mode install shows neither. */
title: string;
note?: string;
disabled?: boolean;
}) {
const [modes, setModes] = useState<DiscussionMode[]>([]);
const copyFor = useModeCopy();
useEffect(() => {
void listDiscussionModes()
.then(setModes)
.catch(() => setModes([]));
}, []);
// One mode is not a choice. The heading goes with it — a lone title over an
// empty area is worse than not raising the subject at all.
if (modes.length < 2) return null;
return (
<section>
<h2 className="mb-2 text-[12px] font-medium text-[var(--foreground)]">
{title}
</h2>
<div className="space-y-2">
{modes.map((mode) => {
const active = mode.name === value;
const copy = copyFor(mode);
return (
<button
key={mode.name}
type="button"
disabled={disabled}
onClick={() => onChange(mode.name)}
className={`flex w-full items-start gap-2.5 rounded-xl border p-3 text-left transition-colors disabled:opacity-50 ${
active
? "border-[var(--primary)] bg-[var(--primary)]/[0.05]"
: "border-[var(--border)] hover:border-[var(--ring)]"
}`}
>
<span
className={`mt-0.5 flex h-4 w-4 shrink-0 items-center justify-center rounded-full border ${
active
? "border-[var(--primary)] bg-[var(--primary)] text-[var(--primary-foreground)]"
: "border-[var(--border)]"
}`}
>
{active ? <Check size={10} /> : null}
</span>
<span className="min-w-0 flex-1">
<span className="flex items-baseline gap-2">
<span className="text-[12.5px] font-medium text-[var(--foreground)]">
{copy.label}
</span>
{copy.cost ? (
<span className="text-[10px] text-[var(--muted-foreground)]">
{copy.cost}
</span>
) : null}
</span>
<span className="mt-0.5 block text-[11px] leading-relaxed text-[var(--muted-foreground)]">
{copy.hint}
</span>
</span>
</button>
);
})}
</div>
{note ? (
<p className="mt-2 text-[11px] leading-relaxed text-[var(--muted-foreground)]">
{note}
</p>
) : null}
</section>
);
}
/** The short name for a mode, for headers and list rows. */
export function useDiscussionModeLabel() {
const { t } = useTranslation();
return (name: string): string => {
switch (name) {
case "panel_parallel":
return t("Parallel answers");
case "sequential":
return t("One after another");
case "debate":
return t("Cross-examination");
default:
return name;
}
};
}