DeepTutor: Lifelong Personalized Tutoring. https://deeptutor.info/.
Python ai-agentsai-tutorclawbotcli-tooldeepresearchinteractive-learninglearningmulti-agent-systemsrag
import { render, waitFor } from "@testing-library/react";
import { describe, expect, it } from "vitest";
import { AskUserOptions } from "@/components/chat/home/AskUserOptions";
import { MasteryQuestionCard } from "@/components/chat/home/MasteryQuestionCard";
import MarkdownRenderer from "@/components/common/MarkdownRenderer";
import { initI18n } from "@/i18n/init";
import type { MasteryQuestion } from "@/lib/mastery-question";
initI18n("zh");
/**
* The cards a learner answers on are written by the same tutor that writes the
* prose above them, and on a maths or physics path that prose is LaTeX. Both
* cards printed it verbatim — a Bellman-equation stem reached the reader as
* `$G_t = R_{t+1} + \gamma G_{t+1}$`, and a bolded cue as `**听懂**`.
*/
function masteryQuestion(): MasteryQuestion {
return {
questionId: "q1",
prompt: "已知 $G_t = R_{t+1} + \\gamma G_{t+1}$,关于 $V^\\pi$ 的方程是?",
questionType: "multiple_choice",
objectiveName: "Bellman",
difficulty: "medium",
attempt: 1,
options: [
{ label: "A", body: "自举:$V^\\pi(s) = E_\\pi[R_{t+1}]$" },
{ label: "B", body: "蒙特卡洛采样" },
],
allowFreeText: false,
};
}
describe("learner-facing cards render the tutor's markup", () => {
it("renders KaTeX in a mastery question stem, its options and its verdict", async () => {
const { container } = render(
<MasteryQuestionCard
question={masteryQuestion()}
grade={{
questionId: "q1",
isCorrect: true,
learnerAnswer: "A",
correctLabel: "A",
correctBody: "",
explanation: "代入可得 $V^\\pi(s)$。",
}}
answered
submittedAnswer="A"
onSubmit={() => undefined}
/>,
);
await waitFor(() => {
expect(container.querySelectorAll(".katex").length).toBeGreaterThanOrEqual(
4,
);
});
expect(container).not.toHaveTextContent("$V^\\pi$");
});
it("renders math and emphasis on an ask_user card", async () => {
const { container } = render(
<AskUserOptions
data={{
resolved: false,
answers: [],
payload: {
intro: "先确认一件事:**听懂**了吗?",
questions: [
{
id: "goal",
prompt: "$e^{i\\pi} + 1 = 0$ 说明了什么?",
header: "Goal",
multi_select: false,
allow_free_text: false,
placeholder: null,
options: [
{ label: "$\\pi$ 与 $e$ 的关系", description: null },
{ label: "别的", description: null },
],
},
],
},
}}
onSubmit={() => undefined}
/>,
);
await waitFor(() => {
expect(container.querySelectorAll(".katex").length).toBeGreaterThanOrEqual(
2,
);
});
expect(container).not.toHaveTextContent("**听懂**");
});
it("renders math in a trace bubble", async () => {
const { container } = render(
<MarkdownRenderer
content="特征方程 $r^2 + 2r + 5 = 0$,判别式 $\\Delta = -16 < 0$。"
variant="trace"
/>,
);
await waitFor(() => {
expect(container.querySelectorAll(".katex")).toHaveLength(2);
});
expect(container).not.toHaveTextContent("$r^2 + 2r + 5 = 0$");
});
});