@mars-sea/dsh-commandcode-provider 0.14.1 → 0.14.3

package/CHANGELOG.md

@@ -4,7 +4,29 @@ All notable changes to this project will be documented in this file.
 
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/), and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
-## [Unreleased]
+## [0.14.3] - 2026-10-19
+
+### Changed
+
+- **发布包不再携带 sourcemap，体积减少约 16%（tarball 2.9 MB → 2.5 MB，解包 5.1 MB → 3.5 MB；文件数 18 → 16）。** `tsdown.config.ts` 的 `sourcemap: true` 会产出 `lib/index.js.map` 与 `lib/client.js.map`，而 `files` 直接收整个 `lib`，两份 map（1.63 MiB）因此进了 tarball。它们没有任何读取方：`lib/client.js` 是 WebView 直接执行的 CJS 包，`//# sourceMappingURL` 只是尾部注释；`scripts/verify-engine-load.mjs` 的审计用正则匹配 `import`／`require` **语句**，不匹配注释。本地 `npm run build` 照旧产出 map，调试能力不受影响。**用户侧无需做任何事**，运行时行为与公开导出面零变化（`lib/index.d.ts` 的 `export { ... }` 一行未改）。
+
+- **长会话的历史分析报告更快（约 −28%～−33%）。** `buildAnalytics` 原先有两处对「事实 × 模型」的重复扫描：每模型小计对整表过滤 N 次，场景矩阵对每个（事实 × 目录模型）都查一次候选价格行（96 个模型共 480,000 次查找，查的是同一批行）。现改为先按模型分桶一次，并把定价算式收进 `cost-facts.requestCost()`、让场景循环把查价提到内层循环之外；`requestRates` 增加可选的高峰时段 memo（由调用方持有，不会活得比它的价格表更久）。两套独立测量给出同一量级：PR 作者脚本 5,000 条事实 × 96 个目录模型 42.17 ms → 30.36 ms，维护者本机用另一套合成数据复测 14.87 ms → 9.90 ms。**输出逐位不变**：分桶顺序与原 `Set` 顺序一致，排序本来就是按唯一 id 的全序（维护者另用 3,000 条确定性事实 × 96 个模型对比两版 JSON 输出，147,944 字节完全一致）。
+
+- **图片请求里同一张图只编码一次 base64。** `readImageRequest()` 对每个（附件，编码目标）返回同一个缓存对象，但一条请求里同一张图会出现在多条历史消息中；三个 wire 编解码器原先各自按「出现次数」执行 `Buffer.from(version.data).toString('base64')`（1 MiB 载荷 0.738 ms／次）。现改为以该缓存对象为键的 `WeakMap`，保存原始 base64 与 data URL，副本生命周期跟随宿主自己的版本缓存。命中约 21 ns。
+
+- **「查看模型用量、成本与性能分析」入口移到「账户用量」分组标题行**，紧邻「刷新」。它原先裸放在页面说明段与第一个分组之间，既无外边距也不属于任何分组，视觉上像是游离在版式之外；移到「账户用量」标题行后与服务着同一批用量／成本数据的「刷新」并列，也复用了该行既有的 `cc-spacer` 布局，不新增样式类。分析入口仍只在 Host 暴露面板接缝时出现，行为不变（切到面板的「模型分析」页签并打开面板）。
+
+- **多账户轮换说明精简 2 处冗余用词**（中文 47 → 43 字，英文同步）。`账户操作立即生效。当前账户达到用量限额或密钥失效时，请求会按列表顺序自动切换到下一个可用账户。` → 去掉「用量」与「自动」后为：`账户操作立即生效。当前账户达到限额或密钥失效时，请求会按列表顺序切换到下一个可用账户。` 「按列表顺序切换」本身已表达自动性，`达到限额` 在账户上下文里仍是额度含意；英文侧 `usage limit` → `limit`，`switch` 原本就没有 `automatically`。**已知取舍**：同页单账户文案 `accountsHint` 仍用「用量限额」，两条说明的术语因此不完全一致。
+
+### Added
+
+- **新增 `.gitattributes`：全部文本文件统一按 LF 检出与提交。** 仓库此前没有该文件，Windows 贡献者按 CRLF 检出后，sourcemap 的 `sourcesContent` 会内嵌带 CRLF 的源码，重建出的 `.map` 与提交里的不再逐字节相同，CI 的 `git diff --exit-code -- lib/` 守卫因此误报「产物过期」（2026-10-17 那次 CI 失败的根因）。实测对照：无该文件时 `core.autocrlf=true` 检出后 `src/adapter.ts` 有 3,482 行 CRLF、`lib/index.js` 有 11,817 行；加上后两者均为 0。图片与压缩包按 `binary` 排除在行尾转换之外。**本次加入时已跟踪文件 100% 是 LF，`git add --renormalize .` 未改写任何文件。**
+
+### Removed
+
+- 两份 sourcemap（`lib/index.js.map`、`lib/client.js.map`）不再纳入版本控制。它们是 tsdown 产物里唯一在 CRLF 工作树上无法逐字节重建的文件，跟踪它们只会让「产物过期」守卫依赖工作树的行尾；移出 tarball 与版本控制后，该守卫在任意平台都只反映源码与产物的真实一致性。
+
+## [0.14.2] - 2026-10-18
 
 ### Changed
 
@@ -21,6 +43,9 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
 - 复核**无变化**的项：峰谷计价仍是 5 个 DeepSeek 模型、窗口 `01–04 & 06–10 UTC, Mon–Fri` 与本地常量逐值一致；订阅条目与每月额度零变化；`reflection/beam:free` 支持图像输入这一点未被目录宣告，图像能力表保持 70 条；effort 词汇表仍是 `off｜low｜medium｜high｜xhigh｜max`。
 
 - **证据边界**：本轮 `scripts/` 巡检工具的状态为 `incomplete`、`readyToApply=false`、`protocol=not-verified`；唯一一条 error 是 `reflection/beam:free: absent from CLI registry`——该模型只出现在两个目录视图里，`1.84.0` 产物注册表完全不认它，因此 `KNOWN_EFFORTS` 与 `KNOWN_NON_ZDR_MODELS` 里这一行属**目录侧单源证据**（与 `1.82.4` 时 `stealth/cat-jam:free` 的处境相同）。16 个数据源全部返回 HTTP 200。**未做**：真实账号登录、付费生成、浏览器渲染验证，也没有在真实账号上验证价格与订阅额度的页面读数；`?view=client` 与定价页均为公开数据采集。
+- **设置页「模型与上下文」分组改为可折叠，默认收起（issue #94）。**该分组是页面上最长的一组（套餐过滤开关 → 可见模型白名单 → 自定义上下文长度 → 按模型覆盖行），折起后「隐私与安全」与「集成与显示」两个分组不必翻过白名单即可到达。折叠头与「高级设置」共用同一实现（`DisclosureHead`／`CollapsibleGroup`），三条行为两边逐字一致：整行可点、箭头在行尾、状态不跨刷新记忆（组件内 `useState`，既不进配置文档也不进 localStorage——折叠是呈现偏好，不该由设置文档承载）。**「高级设置」呈现与交互零变化**，只是换成共用实现；`cc-groupHead`／`cc-disclosure`／`cc-chevron`／`cc-badgeWarn` 四个既有类原样复用，样式表零改动。**边界：纯呈现层开关**——`filterModelsByPlan`／`visibleModels`／`contextWindow`／`maxOutputTokens`／`contextWindowOverrides` 五个字段的暂存、校验、整批提交与写入所有权一律不变（沿用[《决策记录》第六章第1、2条](docs/决策记录.md#六设置写入所有权)），收起只决定组内 DOM 是否渲染，已暂存的编辑不会丢。**连带处理**：默认收起会把组内的无效数字一并藏起，而保存按钮会因此禁用，故折叠时在标题行显示警告徽标（新 locale 键 `modelsInvalid`，与「高级设置」的既有徽标同形）；不新增「已选 N 个模型」式摘要徽标——诉求只是可收起，「非默认」口径需要另行定义。**已知取舍**：折叠头的标题仍是 `span` 而非 `h3`（该分组在标题导航里不再是一个 heading），沿用「高级设置」的既有写法（`button` 内不能放 heading），以换取与它逐字一致的 DOM 与零 CSS 布局风险。新增 `tests/section.test.ts` 12 项 React-free 用例固定两个分组「组内是否有未填好的数字」的判定（含按模型覆盖行分支）；检查集与保存守卫走的**同一份 staged plan** 对齐——连 UI 上填不错的 `filterModelsByPlan` 开关也纳入，免得呈现判定比保存条件窄。两组变异对照（删 rows 分支、删 `filterModelsByPlan`）各自精准命中对应用例、恢复后 12／12 全通过，证明用例有判别力。
+
+- **设置页「自定义上下文长度」与「单轮输出上限」的界面说明精简约一半。**两段 hint 只留「填什么 → 会怎样」：中文 124 → 61 字、240 → 98 字，英文 325 → 228、686 → 223 字符；删掉的是机制叙述（为何影响压缩与选择器标注、`0` 的动机、二进制换算示例、宿主限定预算的辅助调用），这些在 README 的对应段落里原样保留（**用户选定：README 保持不动**）。**两个信息点不可删**——「影响上下文计量与自动压缩」与「`0` 只在 CLI 与 OpenAI 传输生效，Claude 的 Messages 端点必填该字段」；后者由 `tests/settings.test.ts` 的既有文案护栏守着：首轮精简把它删掉后护栏立即失败（`zh hint names the route that requires the field`），遂以精简措辞加回，同组英文断言随之从 `/sends no such field/` 改为 `/omits the field/`（断言语义不变：文案必须说明 `0` 不发送该字段）。顺带去掉 hint 里的反引号——它经 `.cc-rowDesc` 纯文本渲染，反引号会在界面上原样显示。
 
 ## [0.14.1] - 2026-10-18
 

package/README.md

@@ -210,7 +210,7 @@ llm-commandcode:
 
 ## Configure
 
-**Settings → Command Code** is organized as **Accounts** (keys, sign-in, live quota, pinning, dedicated models), **Models & context** (hide out-of-plan models, visible models, then the **Custom context length** field and the **Per-model overrides** rows right after the allowlist — global and per model, see below), **Privacy & security** (zero data retention, off by default, see below), **Integrations & display** (serve web search with Command Code, show the quota card in the sidebar, off by default), and a collapsed **Advanced** section (API base URL, request/stream timeouts, transport retries, cache-aware image offload). The working directory is no longer on the page; `workingDir` in config still works.
+**Settings → Command Code** is organized as **Accounts** (keys, sign-in, live quota, pinning, dedicated models), a collapsed **Models & context** group (click its heading to unfold: hide out-of-plan models, visible models, then the **Custom context length** field and the **Per-model overrides** rows right after the allowlist — global and per model, see below), **Privacy & security** (zero data retention, off by default, see below), **Integrations & display** (serve web search with Command Code, show the quota card in the sidebar, off by default), and a collapsed **Advanced** section (API base URL, request/stream timeouts, transport retries, cache-aware image offload). Both foldable groups toggle from their whole heading row, and neither remembers its expanded state across visits. The working directory is no longer on the page; `workingDir` in config still works.
 
 The same options live in `$DSH_HOME/settings.yaml` (changes apply immediately, no restart):
 

package/README.zh-CN.md

@@ -191,7 +191,7 @@ llm-commandcode:
 
 ## 配置
 
-**设置 → Command Code** 分为：**账户**（密钥、网页登录、实时额度、固定账户、专用模型）、**模型与上下文**（隐藏套餐外模型、可见模型，以及紧随白名单之后的**自定义上下文长度**字段与**按模型覆盖**列表，全局与按模型，见下）、**隐私与安全**（零数据保留 ZDR，默认关闭，见下）、**集成与显示**（用 Command Code 承载联网搜索、在侧边栏显示额度卡片，默认关闭），以及默认折叠的**高级设置**（API 地址、请求/流超时、传输重试次数、历史图片缓存优化）。工作目录已不在页面上显示，配置中的 `workingDir` 仍然有效。
+**设置 → Command Code** 分为：**账户**（密钥、网页登录、实时额度、固定账户、专用模型）、**模型与上下文**（默认折叠，点标题展开：隐藏套餐外模型、可见模型，以及紧随白名单之后的**自定义上下文长度**字段与**按模型覆盖**列表，全局与按模型，见下）、**隐私与安全**（零数据保留 ZDR，默认关闭，见下）、**集成与显示**（用 Command Code 承载联网搜索、在侧边栏显示额度卡片，默认关闭），以及默认折叠的**高级设置**（API 地址、请求/流超时、传输重试次数、历史图片缓存优化）。两个可折叠分组都点标题整行展开/收起，展开状态不跨刷新记忆。工作目录已不在页面上显示，配置中的 `workingDir` 仍然有效。
 
 同一组选项也位于 `$DSH_HOME/settings.yaml`（修改即刻生效，无需重启）：
 

package/lib/client.js

@@ -3129,7 +3129,7 @@ window.__ModuleLoader__.load({
 			visibleModelsStaleHint: "有 {count} 个已选模型在目录中找不到了（可能已下架），不影响其他模型；可清理或保留。",
 			visibleModelsCleanStale: "清理失效（{count}）",
 			contextWindow: "自定义上下文长度",
-			contextWindowHint: "覆盖模型目录公布的上下文窗口（token）：上下文计量、自动压缩时机与模型选择器的上下文标注都按此值计算，对未单独设置的模型生效。留空则跟随目录；支持 32K、1M 这样的简写（二进制换算：32K = 32 768，1M = 1 048 576）。",
+			contextWindowHint: "覆盖模型目录公布的上下文窗口（对未单独覆盖的模型生效），影响上下文计量与自动压缩；留空跟随目录。支持 32K、1M 简写。",
 			contextOverridesTitle: "按模型覆盖",
 			contextOverridesHint: "为单个模型单独设置上下文长度，优先于上面的全局值；模型下拉只列出模型白名单中的可见模型（显示全部时即整个目录）。",
 			contextOverrideAdd: "添加模型",
@@ -3141,7 +3141,7 @@ window.__ModuleLoader__.load({
 			contextWindowTooLarge: "上下文长度不能大于 10,000,000",
 			contextWindowInvalid: "请输入有效的整数 token 数",
 			maxOutputTokens: "单轮输出上限",
-			maxOutputTokensHint: "每次请求发送的输出预算（token）。留空＝按模型自身输出上限推导（默认，行为不变）；填正整数＝与推导值取较小者，永远不会超过模型上限或网关上限 131072。填 `0`＝请求中不发送该字段，把单轮输出规模交回平台自行限制——这是「长回答被硬切」的解法，但仅 CLI 与 OpenAI 两条传输支持；Claude 走的 Messages 端点必填该字段，因此那里仍按推导值发送，填 0 对它无效。会话标题这类由宿主明确限定预算的辅助调用也不受此设置影响。支持 64K 这样的简写。",
+			maxOutputTokensHint: "留空＝按模型推导（默认）；正整数＝不超过模型上限与网关上限 131072；0＝不发送该字段（仅 CLI 与 OpenAI 有效，Claude 的 Messages 端点必填）。支持 64K 简写。",
 			maxOutputTokensTooSmall: "输出上限不能小于 0（0 表示不发送该字段）",
 			maxOutputTokensTooLarge: "输出上限不能大于 131072（网关上限）",
 			maxOutputTokensInvalid: "请输入 0 或有效的整数 token 数",
@@ -3237,7 +3237,7 @@ window.__ModuleLoader__.load({
 			cardRouteActive: "已启用",
 			cardLoadingHint: "正在读取 Command Code 配置…",
 			cardRegistrationHint: "此卡片由 Command Code 插件提供，正在等待设置页装配，稍后即可显示完整内容。",
-			accountsRotationHint: "账户操作立即生效。当前账户达到用量限额或密钥失效时，请求会按列表顺序自动切换到下一个可用账户。",
+			accountsRotationHint: "账户操作立即生效。当前账户达到限额或密钥失效时，请求会按列表顺序切换到下一个可用账户。",
 			accountModeAuto: "使用方式：自动切换（按列表顺序）",
 			accountModePinned: "使用方式：固定使用「{name}」，不可用时仍会自动切换",
 			accountStatusPinned: "已固定",
@@ -3279,6 +3279,7 @@ window.__ModuleLoader__.load({
 			modelCount: "已选 {count} 个模型",
 			modelCatalogFailed: "模型目录获取失败，暂时无法选择模型；已保存的设置仍会生效。",
 			modelsTitle: "模型与上下文",
+			modelsInvalid: "模型与上下文中有未填好的数字，请展开修正后再保存。",
 			privacyTitle: "隐私与安全",
 			integrationsTitle: "集成与显示",
 			enrollmentCreating: "正在创建账号",
@@ -3354,7 +3355,7 @@ window.__ModuleLoader__.load({
 			visibleModelsStaleHint: "{count} selected model(s) are no longer in the catalog (possibly retired); other models are unaffected. Clean them up or keep them.",
 			visibleModelsCleanStale: "Clean stale ({count})",
 			contextWindow: "Custom context length",
-			contextWindowHint: "Overrides the context window published by the model catalog (in tokens): the context meter, auto-compaction timing and the model-picker annotation all follow it, for every model without a per-model override below. Empty follows the catalog; shorthand like 32K or 1M is accepted (binary units: 32K is 32,768, 1M is 1,048,576).",
+			contextWindowHint: "Overrides the context window published by the model catalog, for every model without an override of its own below; the context meter and auto-compaction follow it. Empty follows the catalog. Shorthand like 32K or 1M is accepted.",
 			contextOverridesTitle: "Per-model overrides",
 			contextOverridesHint: "Set a context length for one model; it wins over the global value above. The picker lists only the models kept by the visible-models allowlist (the whole catalog while showing all).",
 			contextOverrideAdd: "Add model",
@@ -3366,7 +3367,7 @@ window.__ModuleLoader__.load({
 			contextWindowTooLarge: "Context length must not exceed 10,000,000",
 			contextWindowInvalid: "Enter a valid whole-token count",
 			maxOutputTokens: "Output limit per turn",
-			maxOutputTokensHint: "The output budget sent with each request, in tokens. Empty derives it from the model’s own ceiling (the default; nothing changes). A positive number is taken as a cap alongside that derived value, so it can never exceed the model ceiling or the gateway’s 131,072. `0` sends no such field at all and leaves the per-turn output size to the platform — the fix for answers being cut short — but only the CLI and OpenAI transports support it: the Messages endpoint Claude uses requires the field, so Claude keeps its derived budget and `0` has no effect there. Auxiliary calls whose budget the host fixes itself, such as the session title, are unaffected too. Shorthand like 64K is accepted.",
+			maxOutputTokensHint: "Empty derives it from the model (default). A positive number never exceeds the model ceiling or the gateway’s 131,072. 0 omits the field (CLI and OpenAI only; Claude’s Messages endpoint requires it). 64K shorthand accepted.",
 			maxOutputTokensTooSmall: "Output limit must be at least 0 (0 omits the field)",
 			maxOutputTokensTooLarge: "Output limit must not exceed 131,072 (the gateway cap)",
 			maxOutputTokensInvalid: "Enter 0 or a valid whole-token count",
@@ -3462,7 +3463,7 @@ window.__ModuleLoader__.load({
 			cardRouteActive: "Active",
 			cardLoadingHint: "Loading the Command Code configuration…",
 			cardRegistrationHint: "This card is contributed by the Command Code plugin and is waiting for the settings page to finish assembling.",
-			accountsRotationHint: "Account actions apply immediately. When the current account hits its usage limit or its key stops working, requests switch to the next usable account in list order.",
+			accountsRotationHint: "Account actions apply immediately. When the current account hits its limit or its key stops working, requests switch to the next usable account in list order.",
 			accountModeAuto: "Mode: switch automatically (list order)",
 			accountModePinned: "Mode: pinned to \"{name}\", still switching when it is unavailable",
 			accountStatusPinned: "Pinned",
@@ -3504,6 +3505,7 @@ window.__ModuleLoader__.load({
 			modelCount: "{count} models selected",
 			modelCatalogFailed: "The model catalog could not be loaded; saved settings still apply.",
 			modelsTitle: "Models & context",
+			modelsInvalid: "A number in Models & context is not ready to save; expand to fix it.",
 			privacyTitle: "Privacy & security",
 			integrationsTitle: "Integrations & display",
 			enrollmentCreating: "Creating the account",
@@ -3702,7 +3704,7 @@ window.__ModuleLoader__.load({
 		}
 		//#endregion
 		//#region package.json
-		var version = "0.14.1";
+		var version = "0.14.3";
 		var repository = {
 			"type": "git",
 			"url": "git+https://github.com/Mars-Sea/dsh-commandcode-provider.git"
@@ -3892,6 +3894,13 @@ window.__ModuleLoader__.load({
 		* below Accounts is a flat group of rows (title/description left, control
 		* right) inside the staged form the floating save bar writes.
 		*
+		* Two of those groups fold from their own heading — Models & context (issue
+		* #94) and Advanced — and both start collapsed on every visit, with no
+		* remembered state. While a group is collapsed its rows are not on screen, so
+		* anything inside it that blocks the save is reported on the heading itself
+		* (see `CollapsibleGroup`); the save bar alone would leave the disabled Save
+		* button unexplained.
+		*
 		* Styles are injected once by the client entry (see src/client/index.ts) and
 		* class-prefixed `cc-` to stay local.
 		*/
@@ -3902,6 +3911,28 @@ window.__ModuleLoader__.load({
 			"transportMaxRetries",
 			"offloadSeenImagesForCache"
 		];
+		const MODELS_FIELDS = [
+			"filterModelsByPlan",
+			"contextWindow",
+			"maxOutputTokens"
+		];
+		/**
+		* Whether anything the collapsed Advanced card owns is blocking the save. The
+		* card is folded on arrival, so this drives the warning badge on its heading.
+		*/
+		function advancedGroupInvalid(state) {
+			return ADVANCED_FIELDS.some((field) => state[field].invalid);
+		}
+		/**
+		* Whether anything the collapsed Models & context group owns is blocking the
+		* save (issue #94). Besides its staged drafts — the plan switch and the two
+		* numbers — that group owns the per-model context rows, whose drafts are
+		* validated one by one: a bad row must surface on the heading just like a bad
+		* number, or Save would sit disabled with nothing on screen explaining why.
+		*/
+		function modelsGroupInvalid(state) {
+			return MODELS_FIELDS.some((field) => state[field].invalid) || state.contextWindowOverrides.some((row) => row.invalid);
+		}
 		/** The "customized" tag and reset link a staged field shows once it differs from the default. */
 		function FieldOverride({ label, state, disabled, t, onReset }) {
 			if (!state.overridden && !state.clear) return null;
@@ -4453,17 +4484,86 @@ window.__ModuleLoader__.load({
 			});
 		}
 		/**
+		* The heading of a foldable group: the whole row is the toggle and the chevron
+		* sits at its trailing edge. Shared by Models & context (issue #94: the
+		* reporter asked for it to fold "like Advanced below") and Advanced itself, so
+		* the two cannot drift apart in either markup or behaviour.
+		*
+		* The title is a `span`, not the `h3` a static group uses, because a heading
+		* element cannot sit inside a button — the same trade-off Advanced has always
+		* made on its own.
+		*/
+		function DisclosureHead({ title, bodyId, expanded, onToggle, badges }) {
+			return /* @__PURE__ */ (0, react_jsx_runtime.jsxs)("button", {
+				type: "button",
+				className: "cc-groupHead cc-disclosure",
+				"aria-expanded": expanded,
+				"aria-controls": bodyId,
+				onClick: onToggle,
+				children: [
+					/* @__PURE__ */ (0, react_jsx_runtime.jsx)("span", {
+						className: "cc-groupTitle",
+						children: title
+					}),
+					badges?.(expanded),
+					/* @__PURE__ */ (0, react_jsx_runtime.jsx)("span", { className: "cc-spacer" }),
+					/* @__PURE__ */ (0, react_jsx_runtime.jsx)("span", {
+						className: expanded ? "cc-chevron cc-chevronUp" : "cc-chevron",
+						"aria-hidden": "true"
+					})
+				]
+			});
+		}
+		/**
+		* A titled group that folds from its own heading and starts collapsed. The
+		* state is deliberately component-local — not persisted, not lifted into the
+		* controller — so every visit opens at the same, predictable page length, and
+		* folding can never turn into a stored preference the settings document would
+		* have to carry.
+		*/
+		function CollapsibleGroup({ title, bodyId, divided = false, badges, footer, children }) {
+			const [expanded, setExpanded] = (0, react.useState)(false);
+			return /* @__PURE__ */ (0, react_jsx_runtime.jsxs)("section", {
+				className: divided ? "cc-group cc-groupDivided" : "cc-group",
+				"aria-label": title,
+				children: [
+					/* @__PURE__ */ (0, react_jsx_runtime.jsx)(DisclosureHead, {
+						title,
+						bodyId,
+						expanded,
+						onToggle: () => setExpanded((value) => !value),
+						badges
+					}),
+					expanded ? /* @__PURE__ */ (0, react_jsx_runtime.jsx)("div", {
+						id: bodyId,
+						className: "cc-rows",
+						children
+					}) : null,
+					expanded ? footer : null
+				]
+			});
+		}
+		/**
 		* Models & context: the plan filter and the allowlist sit together because both
 		* decide what the model picker lists (the allowlist applies after the plan
 		* filter), and the context-length fields follow them because they are the same
 		* subject seen from the other side — the window each listed model advertises.
 		* Those fields carry no heading or divider of their own: they are rows of this
-		* group, not a section inside it.
+		* group, not a section inside it. The group folds from its heading and starts
+		* collapsed (issue #94, "same as Advanced below"): it is the longest group on
+		* the page and its rows are set-and-forget, so folding keeps Accounts and the
+		* two switches below it reachable without scrolling past the allowlist.
 		*/
 		function ModelsCard({ state, disabled, t, onEdit, onReset, onSelect, onClear, onOverrides }) {
-			return /* @__PURE__ */ (0, react_jsx_runtime.jsxs)(SettingsGroup, {
+			const invalid = modelsGroupInvalid(state);
+			return /* @__PURE__ */ (0, react_jsx_runtime.jsxs)(CollapsibleGroup, {
 				title: t("modelsTitle"),
+				bodyId: "cc-models-body",
 				divided: true,
+				badges: (expanded) => !expanded && invalid ? /* @__PURE__ */ (0, react_jsx_runtime.jsx)("span", {
+					className: "cc-badge cc-badgeWarn",
+					children: t("modelsInvalid")
+				}) : null,
 				children: [
 					/* @__PURE__ */ (0, react_jsx_runtime.jsx)(ToggleField, {
 						id: "cc-filter-models-by-plan",
@@ -4699,116 +4799,92 @@ window.__ModuleLoader__.load({
 			});
 		}
 		/**
-		* The collapsed "Advanced" card: the API base and the network limits. Starts
-		* collapsed on every visit; while collapsed, a badge names the customized
+		* The collapsed "Advanced" card: the API base and the network limits. Folds
+		* through the same `CollapsibleGroup` the Models & context group uses, so the
+		* two disclosures stay identical; while collapsed, a badge names the customized
 		* count and an invalid number (which blocks save) is surfaced on the header.
 		*/
 		function AdvancedSection({ state, disabled, t, onEdit, onReset }) {
-			const [expanded, setExpanded] = (0, react.useState)(false);
 			const overridden = ADVANCED_FIELDS.filter((field) => state[field].overridden).length;
-			const invalid = ADVANCED_FIELDS.some((field) => state[field].invalid);
-			return /* @__PURE__ */ (0, react_jsx_runtime.jsxs)("section", {
-				className: "cc-group cc-groupDivided",
-				"aria-label": t("advancedSettings"),
+			const invalid = advancedGroupInvalid(state);
+			return /* @__PURE__ */ (0, react_jsx_runtime.jsxs)(CollapsibleGroup, {
+				title: t("advancedSettings"),
+				bodyId: "cc-advanced-body",
+				divided: true,
+				badges: (expanded) => /* @__PURE__ */ (0, react_jsx_runtime.jsxs)(react_jsx_runtime.Fragment, { children: [overridden > 0 ? /* @__PURE__ */ (0, react_jsx_runtime.jsx)("span", {
+					className: "cc-badge",
+					children: overridden === 1 ? t("advancedOverriddenOne") : t("advancedOverriddenMany", { count: overridden })
+				}) : null, !expanded && invalid ? /* @__PURE__ */ (0, react_jsx_runtime.jsx)("span", {
+					className: "cc-badge cc-badgeWarn",
+					children: t("advancedInvalid")
+				}) : null] }),
+				footer: invalid ? /* @__PURE__ */ (0, react_jsx_runtime.jsx)("p", {
+					className: "cc-rowError",
+					role: "status",
+					children: t("advancedInvalid")
+				}) : null,
 				children: [
-					/* @__PURE__ */ (0, react_jsx_runtime.jsxs)("button", {
-						type: "button",
-						className: "cc-groupHead cc-disclosure",
-						"aria-expanded": expanded,
-						"aria-controls": "cc-advanced-body",
-						onClick: () => setExpanded((value) => !value),
-						children: [
-							/* @__PURE__ */ (0, react_jsx_runtime.jsx)("span", {
-								className: "cc-groupTitle",
-								children: t("advancedSettings")
-							}),
-							overridden > 0 ? /* @__PURE__ */ (0, react_jsx_runtime.jsx)("span", {
-								className: "cc-badge",
-								children: overridden === 1 ? t("advancedOverriddenOne") : t("advancedOverriddenMany", { count: overridden })
-							}) : null,
-							!expanded && invalid ? /* @__PURE__ */ (0, react_jsx_runtime.jsx)("span", {
-								className: "cc-badge cc-badgeWarn",
-								children: t("advancedInvalid")
-							}) : null,
-							/* @__PURE__ */ (0, react_jsx_runtime.jsx)("span", { className: "cc-spacer" }),
-							/* @__PURE__ */ (0, react_jsx_runtime.jsx)("span", {
-								className: expanded ? "cc-chevron cc-chevronUp" : "cc-chevron",
-								"aria-hidden": "true"
-							})
-						]
+					/* @__PURE__ */ (0, react_jsx_runtime.jsx)("p", {
+						className: "cc-groupDesc",
+						children: t("advancedSettingsHint")
 					}),
-					expanded ? /* @__PURE__ */ (0, react_jsx_runtime.jsxs)("div", {
-						id: "cc-advanced-body",
-						className: "cc-rows",
-						children: [
-							/* @__PURE__ */ (0, react_jsx_runtime.jsx)("p", {
-								className: "cc-groupDesc",
-								children: t("advancedSettingsHint")
-							}),
-							/* @__PURE__ */ (0, react_jsx_runtime.jsx)(Field, {
-								id: "cc-api-base",
-								label: t("apiBase"),
-								hint: t("apiBaseHint"),
-								state: state.apiBase,
-								disabled,
-								wide: true,
-								onEdit: (text) => onEdit("apiBase", text),
-								onReset: () => onReset("apiBase"),
-								t
-							}),
-							/* @__PURE__ */ (0, react_jsx_runtime.jsx)(Field, {
-								id: "cc-request-timeout",
-								label: t("requestTimeoutMs"),
-								hint: t("requestTimeoutMsHint"),
-								state: state.requestTimeoutMs,
-								disabled,
-								numeric: true,
-								decimal: true,
-								onEdit: (text) => onEdit("requestTimeoutMs", text),
-								onReset: () => onReset("requestTimeoutMs"),
-								t
-							}),
-							/* @__PURE__ */ (0, react_jsx_runtime.jsx)(Field, {
-								id: "cc-stream-idle-timeout",
-								label: t("streamIdleTimeoutMs"),
-								hint: t("streamIdleTimeoutMsHint"),
-								state: state.streamIdleTimeoutMs,
-								disabled,
-								numeric: true,
-								decimal: true,
-								onEdit: (text) => onEdit("streamIdleTimeoutMs", text),
-								onReset: () => onReset("streamIdleTimeoutMs"),
-								t
-							}),
-							/* @__PURE__ */ (0, react_jsx_runtime.jsx)(Field, {
-								id: "cc-transport-max-retries",
-								label: t("transportMaxRetries"),
-								hint: t("transportMaxRetriesHint"),
-								state: state.transportMaxRetries,
-								disabled,
-								numeric: true,
-								onEdit: (text) => onEdit("transportMaxRetries", text),
-								onReset: () => onReset("transportMaxRetries"),
-								t
-							}),
-							/* @__PURE__ */ (0, react_jsx_runtime.jsx)(ToggleField, {
-								id: "cc-offload-seen-images-for-cache",
-								label: t("offloadSeenImagesForCache"),
-								hint: t("offloadSeenImagesForCacheHint"),
-								state: state.offloadSeenImagesForCache,
-								disabled,
-								defaultChecked: false,
-								onEdit: (text) => onEdit("offloadSeenImagesForCache", text),
-								onReset: () => onReset("offloadSeenImagesForCache"),
-								t
-							})
-						]
-					}) : null,
-					expanded && invalid ? /* @__PURE__ */ (0, react_jsx_runtime.jsx)("p", {
-						className: "cc-rowError",
-						role: "status",
-						children: t("advancedInvalid")
-					}) : null
+					/* @__PURE__ */ (0, react_jsx_runtime.jsx)(Field, {
+						id: "cc-api-base",
+						label: t("apiBase"),
+						hint: t("apiBaseHint"),
+						state: state.apiBase,
+						disabled,
+						wide: true,
+						onEdit: (text) => onEdit("apiBase", text),
+						onReset: () => onReset("apiBase"),
+						t
+					}),
+					/* @__PURE__ */ (0, react_jsx_runtime.jsx)(Field, {
+						id: "cc-request-timeout",
+						label: t("requestTimeoutMs"),
+						hint: t("requestTimeoutMsHint"),
+						state: state.requestTimeoutMs,
+						disabled,
+						numeric: true,
+						decimal: true,
+						onEdit: (text) => onEdit("requestTimeoutMs", text),
+						onReset: () => onReset("requestTimeoutMs"),
+						t
+					}),
+					/* @__PURE__ */ (0, react_jsx_runtime.jsx)(Field, {
+						id: "cc-stream-idle-timeout",
+						label: t("streamIdleTimeoutMs"),
+						hint: t("streamIdleTimeoutMsHint"),
+						state: state.streamIdleTimeoutMs,
+						disabled,
+						numeric: true,
+						decimal: true,
+						onEdit: (text) => onEdit("streamIdleTimeoutMs", text),
+						onReset: () => onReset("streamIdleTimeoutMs"),
+						t
+					}),
+					/* @__PURE__ */ (0, react_jsx_runtime.jsx)(Field, {
+						id: "cc-transport-max-retries",
+						label: t("transportMaxRetries"),
+						hint: t("transportMaxRetriesHint"),
+						state: state.transportMaxRetries,
+						disabled,
+						numeric: true,
+						onEdit: (text) => onEdit("transportMaxRetries", text),
+						onReset: () => onReset("transportMaxRetries"),
+						t
+					}),
+					/* @__PURE__ */ (0, react_jsx_runtime.jsx)(ToggleField, {
+						id: "cc-offload-seen-images-for-cache",
+						label: t("offloadSeenImagesForCache"),
+						hint: t("offloadSeenImagesForCacheHint"),
+						state: state.offloadSeenImagesForCache,
+						disabled,
+						defaultChecked: false,
+						onEdit: (text) => onEdit("offloadSeenImagesForCache", text),
+						onReset: () => onReset("offloadSeenImagesForCache"),
+						t
+					})
 				]
 			});
 		}
@@ -5478,7 +5554,7 @@ window.__ModuleLoader__.load({
 		* each, with its status and quota inline — so one account's facts no longer
 		* have to be cross-read across separate surfaces.
 		*/
-		function AccountsCard({ t, state, usage, login, disabled, actions, onRefresh, createEnrollment }) {
+		function AccountsCard({ t, state, usage, login, disabled, actions, onRefresh, openAnalytics, createEnrollment }) {
 			const [adding, setAdding] = (0, react.useState)(false);
 			const [enrollment] = (0, react.useState)(createEnrollment);
 			const snapshot = (0, react.useSyncExternalStore)(enrollment.store.subscribe, enrollment.store.getSnapshot);
@@ -5506,6 +5582,12 @@ window.__ModuleLoader__.load({
 								children: t("accountsTitle")
 							}),
 							/* @__PURE__ */ (0, react_jsx_runtime.jsx)("span", { className: "cc-spacer" }),
+							openAnalytics ? /* @__PURE__ */ (0, react_jsx_runtime.jsx)(_deepseek_ai_dsh_client_ui_primitives.Button, {
+								variant: "ghost",
+								size: "sm",
+								onClick: openAnalytics,
+								children: t("openAnalytics")
+							}) : null,
 							/* @__PURE__ */ (0, react_jsx_runtime.jsx)("button", {
 								type: "button",
 								className: "cc-linkButton",
@@ -5853,12 +5935,6 @@ window.__ModuleLoader__.load({
 						className: "cc-intro",
 						children: t("intro")
 					}),
-					props.openAnalytics ? /* @__PURE__ */ (0, react_jsx_runtime.jsx)(_deepseek_ai_dsh_client_ui_primitives.Button, {
-						variant: "ghost",
-						size: "sm",
-						onClick: props.openAnalytics,
-						children: t("openAnalytics")
-					}) : null,
 					!state.writable ? /* @__PURE__ */ (0, react_jsx_runtime.jsx)("p", {
 						className: "cc-readOnly",
 						role: "status",
@@ -5872,6 +5948,7 @@ window.__ModuleLoader__.load({
 						disabled,
 						actions,
 						onRefresh: props.refreshUsage,
+						openAnalytics: props.openAnalytics,
 						createEnrollment: props.createEnrollment
 					}),
 					/* @__PURE__ */ (0, react_jsx_runtime.jsx)(ModelsCard, {
@@ -8300,10 +8377,16 @@ window.__ModuleLoader__.load({
 		* today, so the combination becoming real is visible before it reaches a user.
 		* This ordering is a latent choice, not a verified upstream rule.
 		*/
-		function requestRates(price, at, contextTokens, table) {
+		function requestRates(price, at, contextTokens, table, peakHourCache) {
 			const tier = price.contextTiers?.find((tier) => tier.maxContext === void 0 || contextTokens <= tier.maxContext);
 			if (tier !== void 0) return tier;
-			return price.peak !== void 0 && at !== null && peakHour(at, table.peakHours) ? price.peak : price;
+			if (price.peak === void 0 || at === null) return price;
+			let peak = peakHourCache?.get(at);
+			if (peak === void 0) {
+				peak = peakHour(at, table.peakHours);
+				peakHourCache?.set(at, peak);
+			}
+			return peak ? price.peak : price;
 		}
 		/** Cache version and wire guard: changing bands/rates must refold historical groups. */
 		function pricingKey(table) {

package/lib/index.d.ts

@@ -207,7 +207,13 @@ interface CommandCodeConnectionOptions {
  * text-only request never depends on the attachment seam.
  */
 type ResolveAttachments = () => AttachmentStore | undefined;
-/** What the rotation hook knows about the request it is rotating within. */
+/**
+ * What the rotation hook knows about the request it is rotating within.
+ *
+ * Not exported: the only consumer is the `rotateApiKey` shape below, so the
+ * name never reaches the published `.d.ts` and the package's `exports` map
+ * leaves no deep-import path to it.
+ */
 interface AccountRotationContext {
   /** Every API key this request has already used, just-rejected key included. */
   tried: readonly string[];
@@ -265,13 +271,13 @@ interface CommandCodeAdapterDeps<C extends CommandCodeConnectionOptions = Comman
    */
   imageOffload?: SurfaceImagePolicy;
 }
-/** Account identity from `/alpha/whoami`. */
+/** Account identity from `/alpha/whoami`; reached only through `CommandCodeUsageReport`. */
 interface CommandCodeAccount {
   id: string;
   name: string;
   userName: string;
 }
-/** Usage summary from `/alpha/usage/summary`. */
+/** Usage summary from `/alpha/usage/summary`; reached only through `CommandCodeUsageReport`. */
 interface CommandCodeUsage {
   totalCount: number;
   totalCost: number;
@@ -326,7 +332,7 @@ interface CommandCodeCredits {
     resetAt: number;
   };
 }
-/** Subscription plan state from `/alpha/billing/subscriptions`. */
+/** Subscription plan state from `/alpha/billing/subscriptions`; reached only through `CommandCodeUsageReport`. */
 interface CommandCodePlan {
   /** Raw subscription plan id (e.g. `individual-pro`); empty when unreported. */
   planId: string;