DefaultScheduler

DefaultScheduler 是非同步工廠，回傳一個可直接使用的排程器實例。它把模型（預設 FSRS-7）與 dateChrono 以及標準排程策略組合在一起，因此你只需設定它的選項。內部它與自訂 defineScheduler 使用同一套「定義與實例」機制——預設只是把各部分裝配起來的便利方式。

選項

從 FSRS-3~6 升級至 FSRS-7 時，省略 weights 以使用 FSRS-7 預設權重，或提供為 FSRS-7 訓練的權重。重播歷史用於重建卡片狀態，不會轉換舊版權重。

FSRS-7 卡片除了 stability 與 difficulty，還包含 stabilityFast。舊版 FSRS-3~6 的記憶狀態需透過 Reschedule 重播複習歷史來重建；參數遷移不會遷移卡片狀態，不應以 0.8 * stability 推測演化後的快速軌跡。預設為 FSRS-7 啟用小數天數；enableShortTerm: false 只停用明確的學習步驟，不會停用模型的快速軌跡。

需要成本自適應保留率時，自行組合定義，以 schedulerCostAdrMiddleware 取代 schedulerDesiredRetentionMiddleware。Cost ADR 不是 DefaultScheduler 選項。

scheduler.nextInterval(memoryState, desiredRetention, { card, grade, elapsedDays }) 查詢已計算的評分後狀態所對應的間隔，只執行 handlers.nextInterval，不推進卡片或 chrono；共用間隔策略應放在這個掛鉤中。

選項	預設	用途
`version`	`'FSRS-7'`	選擇 FSRS 模型：`'FSRS-3'`、`'FSRS-4'`、`'FSRS-4.5'`、`'FSRS-5'`, `'FSRS-6'` 或 `'FSRS-7'`。FSRS-7 無法遷移舊版 FSRS-3~6 的權重。
`weights`	預設權重	選用的訓練權重；省略時使用所選版本的預設值。FSRS-7 需要恰好 34 個 FSRS-7 權重；傳入舊版權重會在建立排程器時拋出例外。
`desiredRetention`	`0.9`	在計算間隔前套用的目標保留率。
`enableShortTerm`	`true`	啟用學習與再學習步驟。
`learningSteps`	`['1m', '10m']`	`enableShortTerm` 開啟時套用的學習步驟。
`relearningSteps`	`['10m']`	`enableShortTerm` 開啟時套用的再學習步驟。
`graduatingInterval`	`1`	卡片處於 `Review` 時的最短間隔（天）。最終間隔更短時，卡片維持 `Learning`/`Relearning`。設定了步驟時，走完步驟以及 `Review` 卡片按 Hard、Good、Easy 都至少使用這個間隔。
`allowModelOverride`	FSRS-7: `true`; 其他版本: `false`	模型間隔比目前步驟長時，改用模型間隔。Again 不超過 `graduatingInterval`；Hard、Good 的間隔達到 `graduatingInterval` 時提前畢業。
`enableFuzz`	`false`	開啟時分散間隔以增加變化。
`maximumInterval`	`36500`	間隔的天數上限。
`clearStatsOnForget`	`true`	`forget` 是否清空 `reps` 與 `lapses`。

DefaultScheduler 始終使用 dateChrono，並依中介軟體頁面所述的順序組合標準策略：目標保留率、模糊間隔、統計、scheduled days、學習步驟、最大間隔、單調間隔。選擇 version 只改變模型及其參數，不改變這套策略。

import { function DefaultScheduler<Version extends DefaultSchedulerVersion = "FSRS-7">(options?: DefaultSchedulerOptions<Version>): Promise<DefaultScheduler<Version>>
type DefaultScheduler<Version extends DefaultSchedulerVersion = "FSRS-7"> = (Version extends "FSRS-7" ? Omit<SchedulerCore<SchedulerCoreEnv<SchedulerEnvFor<FSRS7DefaultSchedulerModel, Chrono<{
    readonly time: SRSSchema<{
        input: Date;
        output: Date;
    }>;
    readonly config: SRSSchema<{
        input: {
            readonly fractionalDays?: boolean | undefined;
        };
        output: {
            readonly fractionalDays: boolean;
        };
    }>;
    readonly fields: {
        readonly card: SRSSchema<{
            input: {
                dueAt: Date;
                lastReviewAt?: (Date | null) | undefined;
            };
            output: {
                dueAt: Date;
                lastReviewAt: Date | null;
            };
        }>;
        readonly revlog: SRSSchema<{
            input: {
                dueAt: Date;
                lastReviewAt?: (Date | null) | undefined;
                reviewTime: Date;
            };
            output: {
                dueAt: Date;
                lastReviewAt: Date | null;
                reviewTime: Date;
            };
        }>;
    };
}>, readonly [...]>>, ModelCore<...>, ChronoCore<...>>, "definition"> : Omit<...> & {
    ...;
}) & {
    ...;
}
DefaultScheduler, type Rating = 0 | 1 | 2 | 3 | 4
const Rating: Readonly<{
    readonly Manual: 0;
    readonly Again: 1;
    readonly Hard: 2;
    readonly Good: 3;
    readonly Easy: 4;
}>
Rating } from 'ts-fsrs'

const now = new Date('2026-01-01T00:00:00.000Z')
const const scheduler: DefaultScheduler<"FSRS-7">scheduler = await DefaultScheduler<"FSRS-7">(options?: DefaultSchedulerOptions<"FSRS-7"> | undefined): Promise<DefaultScheduler<"FSRS-7">>DefaultScheduler({
  DefaultSchedulerOptions<"FSRS-7">.version?: "FSRS-7" | undefinedFSRS model and parameter migration version; defaults to FSRS-7.version: 'FSRS-7',
  DefaultSchedulerOptions<Version extends DefaultSchedulerVersion = DefaultSchedulerVersion>.desiredRetention?: number | undefineddesiredRetention: 0.9,
  DefaultSchedulerOptions<Version extends DefaultSchedulerVersion = DefaultSchedulerVersion>.enableShortTerm?: boolean | undefinedEnable explicit learning steps. FSRS-7 always retains its intrinsic fast trace.enableShortTerm: true,
  DefaultSchedulerOptions<"FSRS-7">.learningSteps?: readonly (`${number}d` | `${number}m` | `${number}h`)[] | undefinedDefaults to ['1m', '10m'].learningSteps: ['1m', '10m'],
  DefaultSchedulerOptions<"FSRS-7">.relearningSteps?: readonly (`${number}d` | `${number}m` | `${number}h`)[] | undefinedDefaults to ['10m'].relearningSteps: ['10m'],
  DefaultSchedulerOptions<Version extends DefaultSchedulerVersion = DefaultSchedulerVersion>.enableFuzz?: boolean | undefinedenableFuzz: false,
  DefaultSchedulerOptions<Version extends DefaultSchedulerVersion = DefaultSchedulerVersion>.maximumInterval?: number | undefinedmaximumInterval: 36_500,
})
const const card: {
    cardId: string | number;
    stabilityFast: number;
    stability: number;
    difficulty: number;
    dueAt: Date;
    lastReviewAt: Date | null;
    reps: number;
    lapses: number;
    scheduledDays: number;
    learningStep: number;
    state: State;
    scheduleStatus: "review" | "new" | "learning";
}
card = const scheduler: DefaultScheduler<"FSRS-7">scheduler.newCard: (options?: ({
    readonly cardId?: string | number | undefined;
} & {
    readonly now?: Date | undefined;
}) | undefined) => {
    cardId: string | number;
    stabilityFast: number;
    stability: number;
    difficulty: number;
    dueAt: Date;
    lastReviewAt: Date | null;
    reps: number;
    lapses: number;
    scheduledDays: number;
    learningStep: number;
    state: State;
    scheduleStatus: "review" | "new" | "learning";
}
newCard({ cardId: 'quick-start', now })
const const result: {
    card: {
        cardId: string | number;
        stabilityFast: number;
        stability: number;
        difficulty: number;
        dueAt: Date;
        lastReviewAt: Date | null;
        reps: number;
        lapses: number;
        scheduledDays: number;
        learningStep: number;
        state: State;
        scheduleStatus: "review" | "new" | "learning";
    };
    revlog: {
        cardId: string | number;
        stabilityFast: number;
        stability: number;
        difficulty: number;
        dueAt: Date;
        lastReviewAt: Date | null;
        scheduledDays: number;
        learningStep: number;
        reviewTime: Date;
        state: State;
        scheduleStatus: "review" | "new" | "learning";
        rating: Grade;
    };
}
result = const scheduler: DefaultScheduler<"FSRS-7">scheduler.review: (input: {
    readonly card: Readonly<Omit<Readonly<{
        stabilityFast: number;
        stability: number;
        difficulty: number;
    } & {
        dueAt: Date;
        lastReviewAt: Date | null;
    } & {
        readonly cardId: string | number;
        readonly reps: number;
        readonly lapses: number;
        readonly scheduledDays: number;
        readonly learningStep?: number | undefined;
    } & SchedulerCoreFields$1<"review" | "new" | "learning">>, keyof SchedulerCoreFields$1<"review" | "new" | "learning">> & SchedulerCoreFields$1<"review" | "new" | "learning">>;
    readonly grade: Grade;
    readonly now?: Date | undefined;
}) => {
    ...;
}
review({ card, grade: const Rating: Readonly<{
    readonly Manual: 0;
    readonly Again: 1;
    readonly Hard: 2;
    readonly Good: 3;
    readonly Easy: 4;
}>
Rating.Good, now })

console.log(
  JSON.stringify(
    {
      state: const result: {
    card: {
        cardId: string | number;
        stabilityFast: number;
        stability: number;
        difficulty: number;
        dueAt: Date;
        lastReviewAt: Date | null;
        reps: number;
        lapses: number;
        scheduledDays: number;
        learningStep: number;
        state: State;
        scheduleStatus: "review" | "new" | "learning";
    };
    revlog: {
        cardId: string | number;
        stabilityFast: number;
        stability: number;
        difficulty: number;
        dueAt: Date;
        lastReviewAt: Date | null;
        scheduledDays: number;
        learningStep: number;
        reviewTime: Date;
        state: State;
        scheduleStatus: "review" | "new" | "learning";
        rating: Grade;
    };
}
result.card: {
    cardId: string | number;
    stabilityFast: number;
    stability: number;
    difficulty: number;
    dueAt: Date;
    lastReviewAt: Date | null;
    reps: number;
    lapses: number;
    scheduledDays: number;
    learningStep: number;
    state: State;
    scheduleStatus: "review" | "new" | "learning";
}
card.state,
      dueAt: const result: {
    card: {
        cardId: string | number;
        stabilityFast: number;
        stability: number;
        difficulty: number;
        dueAt: Date;
        lastReviewAt: Date | null;
        reps: number;
        lapses: number;
        scheduledDays: number;
        learningStep: number;
        state: State;
        scheduleStatus: "review" | "new" | "learning";
    };
    revlog: {
        cardId: string | number;
        stabilityFast: number;
        stability: number;
        difficulty: number;
        dueAt: Date;
        lastReviewAt: Date | null;
        scheduledDays: number;
        learningStep: number;
        reviewTime: Date;
        state: State;
        scheduleStatus: "review" | "new" | "learning";
        rating: Grade;
    };
}
result.card: {
    cardId: string | number;
    stabilityFast: number;
    stability: number;
    difficulty: number;
    dueAt: Date;
    lastReviewAt: Date | null;
    reps: number;
    lapses: number;
    scheduledDays: number;
    learningStep: number;
    state: State;
    scheduleStatus: "review" | "new" | "learning";
}
card.dueAt,
      stability: const result: {
    card: {
        cardId: string | number;
        stabilityFast: number;
        stability: number;
        difficulty: number;
        dueAt: Date;
        lastReviewAt: Date | null;
        reps: number;
        lapses: number;
        scheduledDays: number;
        learningStep: number;
        state: State;
        scheduleStatus: "review" | "new" | "learning";
    };
    revlog: {
        cardId: string | number;
        stabilityFast: number;
        stability: number;
        difficulty: number;
        dueAt: Date;
        lastReviewAt: Date | null;
        scheduledDays: number;
        learningStep: number;
        reviewTime: Date;
        state: State;
        scheduleStatus: "review" | "new" | "learning";
        rating: Grade;
    };
}
result.card: {
    cardId: string | number;
    stabilityFast: number;
    stability: number;
    difficulty: number;
    dueAt: Date;
    lastReviewAt: Date | null;
    reps: number;
    lapses: number;
    scheduledDays: number;
    learningStep: number;
    state: State;
    scheduleStatus: "review" | "new" | "learning";
}
card.stability,
      stabilityFast: const result: {
    card: {
        cardId: string | number;
        stabilityFast: number;
        stability: number;
        difficulty: number;
        dueAt: Date;
        lastReviewAt: Date | null;
        reps: number;
        lapses: number;
        scheduledDays: number;
        learningStep: number;
        state: State;
        scheduleStatus: "review" | "new" | "learning";
    };
    revlog: {
        cardId: string | number;
        stabilityFast: number;
        stability: number;
        difficulty: number;
        dueAt: Date;
        lastReviewAt: Date | null;
        scheduledDays: number;
        learningStep: number;
        reviewTime: Date;
        state: State;
        scheduleStatus: "review" | "new" | "learning";
        rating: Grade;
    };
}
result.card: {
    cardId: string | number;
    stabilityFast: number;
    stability: number;
    difficulty: number;
    dueAt: Date;
    lastReviewAt: Date | null;
    reps: number;
    lapses: number;
    scheduledDays: number;
    learningStep: number;
    state: State;
    scheduleStatus: "review" | "new" | "learning";
}
card.stabilityFast,
      difficulty: const result: {
    card: {
        cardId: string | number;
        stabilityFast: number;
        stability: number;
        difficulty: number;
        dueAt: Date;
        lastReviewAt: Date | null;
        reps: number;
        lapses: number;
        scheduledDays: number;
        learningStep: number;
        state: State;
        scheduleStatus: "review" | "new" | "learning";
    };
    revlog: {
        cardId: string | number;
        stabilityFast: number;
        stability: number;
        difficulty: number;
        dueAt: Date;
        lastReviewAt: Date | null;
        scheduledDays: number;
        learningStep: number;
        reviewTime: Date;
        state: State;
        scheduleStatus: "review" | "new" | "learning";
        rating: Grade;
    };
}
result.card: {
    cardId: string | number;
    stabilityFast: number;
    stability: number;
    difficulty: number;
    dueAt: Date;
    lastReviewAt: Date | null;
    reps: number;
    lapses: number;
    scheduledDays: number;
    learningStep: number;
    state: State;
    scheduleStatus: "review" | "new" | "learning";
}
card.difficulty,
      rating: const result: {
    card: {
        cardId: string | number;
        stabilityFast: number;
        stability: number;
        difficulty: number;
        dueAt: Date;
        lastReviewAt: Date | null;
        reps: number;
        lapses: number;
        scheduledDays: number;
        learningStep: number;
        state: State;
        scheduleStatus: "review" | "new" | "learning";
    };
    revlog: {
        cardId: string | number;
        stabilityFast: number;
        stability: number;
        difficulty: number;
        dueAt: Date;
        lastReviewAt: Date | null;
        scheduledDays: number;
        learningStep: number;
        reviewTime: Date;
        state: State;
        scheduleStatus: "review" | "new" | "learning";
        rating: Grade;
    };
}
result.revlog: {
    cardId: string | number;
    stabilityFast: number;
    stability: number;
    difficulty: number;
    dueAt: Date;
    lastReviewAt: Date | null;
    scheduledDays: number;
    learningStep: number;
    reviewTime: Date;
    state: State;
    scheduleStatus: "review" | "new" | "learning";
    rating: Grade;
}
revlog.rating,
    },
    null,
    2
  )
)

該工廠是非同步的，使用回傳的實例前要先 await DefaultScheduler(...)。範例把常用預設值寫得很明確：version、desiredRetention、學習與再學習步驟、模糊間隔和 maximumInterval。省略 weights 即用所選 FSRS 版本的預設。

預設還是自訂定義

當預設的 FSRS 模型、dateChrono 與標準策略集契合你的產品時，用 DefaultScheduler——這是最快且正確的起點。當需要不同的模型或 chrono、增減策略，或需要為每個使用者驗證設定時，用 defineScheduler。二者最終都落到同一套排程器實例 API（newCard、review、rollback、preview 與 forget）。定義與實例模型見排程器概覽。

#DefaultScheduler

#選項

#預設還是自訂定義

DefaultScheduler

選項

預設還是自訂定義