@kenz1117/dsh-ui-usage-billing 1.4.33 → 1.4.34

package/README.en.md

@@ -251,7 +251,7 @@ None. This plugin is a pure UI surface: it registers no tools, injects no system
 ## ⚠️ Known Limitations and Deferred Work
 
 - **Balance queries cover DeepSeek / Moonshot (Kimi) / StepFun / SiliconFlow / xAI / Zhipu GLM (Z.ai CN region) / Tencent Cloud TokenHub (token-plan quota)**: the first six use a standard Bearer API key; Tencent Cloud uses the cloud-API key pair (`<SecretId>:<SecretKey>`, TC3-signed management API). Other providers expose no public balance API or need non-Bearer auth (Xiaomi MiMo via console Cookie, SenseTime via AccessKey signing, MiniMax/Doubao via quota or AK/SK), so they currently show "not configured"; the extension point is `src/balance.ts` (add a querier per provider balance API).
-- **Relay quota depends on upstream private schemas**: New API / Sub2API interface fields are not public, so an unreadable station is labeled "no quota" rather than fabricating an amount; if a station's response fields differ, extend the parsers in `src/relay.ts`. An "unknown route" means that route no longer exists in the current provider config (renamed / deleted); historical call data is not lost — re-adding the same-named route restores attribution automatically. Cost in the unknown-route bucket is always 0 (issue #82): the spending party cannot be verified — same posture as uncatalogued models and subscription channels — while tokens are still counted, and the panel shows "—" for the group cost with an explanatory note.
+- **Relay quota depends on upstream private schemas**: New API / Sub2API interface fields are not public, so an unreadable station is labeled "no quota" rather than fabricating an amount; if a station's response fields differ, extend the parsers in `src/relay.ts`. An "unknown route" means that route exists neither in the current provider config nor in the host llm registry (deleted config, logs older than the install, …); historical call data is not lost — re-adding the same-named route restores attribution automatically. Cost in the unknown-route bucket is always 0 (issue #82): the spending party cannot be verified — same posture as uncatalogued models and subscription channels — while tokens are still counted, and the panel shows "—" for the group cost with an explanatory note. Channels registered by another plugin through `ctx.llm.registerAdapter` (e.g. CommandCode) are **not** unknown routes: the plugin discovers them from the host llm registry and bills them in the `direct:<provider>` bucket at the model's official catalog price (issue #91); the registry is read through `ctx.get('llm')` (no `inject` requirement), and an unreadable or empty registry warns once (panel numbers and the unknown policy are unchanged) so this class of failure is never silent again.
 - **Overspend notifications rely on the browser Notification API**: when permission is denied or the platform lacks support, only the in-UI red-pulse fallback remains — no host-level notification channel; notifications are capped at once per day.
 - **Session rows are not navigable**: clicking a session row does not open that session (cross-plugin navigation needs a host session-selection channel); sessions are capped at 100 rows and the panel shows the top 20.
 - **Cost is a catalog estimate**: models without published per-token pricing (iFlytek, SenseTime, Xiaomi) use estimates (feature-list footnote ¹); official billing is authoritative.

package/README.md

@@ -264,7 +264,7 @@ npm publish --access public
 ## ⚠️ Known Limitations and Deferred Work
 
 - **余额查询已接入 DeepSeek / 月之暗面（Kimi）/ 阶跃星辰（StepFun）/ 硅基流动 / xAI / 智谱 GLM（Z.ai 国内域）/ 腾讯云 TokenHub（Token Plan 余量）**：前六家用标准 Bearer API key，腾讯云用云 API 密钥对（TC3 签名）。其余厂商因无公开余额接口或需非 Bearer 鉴权（小米 MiMo 走控制台 Cookie、商汤走 AccessKey 签名、MiniMax/字节豆包走额度制或 AK/SK），暂显示「未配置」；扩展点在 `src/balance.ts`（按厂商余额 API 增加查询器）。
-- **中转站额度依赖上游私有 schema**：New API / Sub2API 的接口字段未公开，读不出时标「未读出额度」而非臆造金额；若某中转站响应字段不同，需按 `src/relay.ts` 的解析器扩展。未知路由表示该路由在当前 provider 配置与宿主 llm 注册表里都不存在（配置已删、日志比安装更早等），历史调用数据未丢，重新配置同名路由即可自动归位。未知路由桶的费用一律记 0（issue #82）：支出主体无法核实，与未收录模型、订阅通道同姿态——token 照常统计，面板组头费用显示「—」并附说明。由独立插件经 `ctx.llm.registerAdapter` 注册的通道（如 CommandCode）不属于未知路由：插件从宿主 llm 注册表发现它们，按 `direct:<provider>` 桶以模型官方目录价计费（issue #91）。
+- **中转站额度依赖上游私有 schema**：New API / Sub2API 的接口字段未公开，读不出时标「未读出额度」而非臆造金额；若某中转站响应字段不同，需按 `src/relay.ts` 的解析器扩展。未知路由表示该路由在当前 provider 配置与宿主 llm 注册表里都不存在（配置已删、日志比安装更早等），历史调用数据未丢，重新配置同名路由即可自动归位。未知路由桶的费用一律记 0（issue #82）：支出主体无法核实，与未收录模型、订阅通道同姿态——token 照常统计，面板组头费用显示「—」并附说明。由独立插件经 `ctx.llm.registerAdapter` 注册的通道（如 CommandCode）不属于未知路由：插件从宿主 llm 注册表发现它们，按 `direct:<provider>` 桶以模型官方目录价计费（issue #91）；注册表读取走 `ctx.get('llm')`（免 `inject` 声明），读不到/读到空集合时告警一次（面板数字不变，unknown 口径不变），避免这类失效无声无息。
 - **超支通知依赖浏览器 Notification**：权限被拒绝或平台不支持时只有界面红色脉冲兜底，没有宿主级通知通道；通知上限为每天一次。
 - **会话明细不可跳转**：点击会话行不会打开对应会话（跨插件导航需要宿主会话选择通道）；会话数封顶 100 行、面板只显示前 20 行。
 - **费用为目录价估算**：讯飞 / 商汤 / 小米等未公布按量单价的模型使用估算价（特性表脚注 ¹），正式定价以厂商账单为准。

package/lib/index.js

@@ -7063,6 +7063,53 @@ function readPiAiProviderRoutes(settings) {
 	}
 }
 /**
+* 读取宿主 llm 注册表里「已注册 adapter」的 provider id（issue #91）。
+*
+* 必须经 `ctx.get('llm')`，**不能**用属性访问 `ctx.llm`：cordis 的服务代理要求
+* 服务名出现在插件 `inject` 里，否则抛 `cannot get property "llm" without inject`
+* ——1.4.31 把这个异常吞掉后退化成空集合，结果是「修复发版了但完全不生效」且无任何
+* 可观察信号（issue #91 follow-up）。本函数把「读不到」与「读到了空集合」都显式
+* 回报给调用方，让上层能告警而不是静默降级。
+* @param get - 宿主上下文的 `get(name)` 访问器（cordis 的免 inject 读取口）。
+* @returns `ids` = 已注册 provider id 集合（读不到时为空集）；
+*  `reason` = 读取失败或结果不可用的原因；成功时缺省。
+*/
+function readRegisteredProviderIds(get) {
+	if (typeof get !== "function") return {
+		ids: /* @__PURE__ */ new Set(),
+		reason: "the host context exposes no get() accessor"
+	};
+	try {
+		const llm = get("llm");
+		if (llm === void 0) return {
+			ids: /* @__PURE__ */ new Set(),
+			reason: "the host llm service is not registered"
+		};
+		if (typeof llm.listProviders !== "function") return {
+			ids: /* @__PURE__ */ new Set(),
+			reason: "the host llm service exposes no listProviders()"
+		};
+		const providers = llm.listProviders();
+		if (providers.length === 0) return {
+			ids: /* @__PURE__ */ new Set(),
+			reason: "the host llm registry reports no registered providers"
+		};
+		const ids = /* @__PURE__ */ new Set();
+		for (const provider of providers) if (typeof provider?.id === "string" && provider.id !== "") ids.add(provider.id);
+		if (ids.size === 0) return {
+			ids,
+			reason: "the host llm registry returned no usable provider ids"
+		};
+		return { ids };
+	} catch (error) {
+		const detail = error instanceof Error ? error.message : String(error);
+		return {
+			ids: /* @__PURE__ */ new Set(),
+			reason: `the host llm registry could not be read (${detail})`
+		};
+	}
+}
+/**
 * 构造「cwd → 工作区标题」解析器（host 的 `workspaceRegistry` 为可选依赖）。
 * 匹配与 TokenLedger 同口径：会话 cwd 等于某工作区 path、或位于其子目录时，用
 * 工作区标题命名该项目（子目录的会话也计入）；否则返回 undefined（回退到目录名）。
@@ -7249,17 +7296,16 @@ function apply(ctx, config = {}) {
 	applyUserModelAliases(config.modelKeyAliases);
 	let llmRegistryWarned = false;
 	const registeredProvidersOf = () => {
-		try {
-			const providers = (ctx.get?.("llm"))?.listProviders() ?? [];
+		const { ids, reason } = readRegisteredProviderIds(ctx.get?.bind(ctx));
+		if (reason === void 0) {
 			llmRegistryWarned = false;
-			return new Set(providers.map((provider) => provider.id));
-		} catch (error) {
-			if (!llmRegistryWarned) {
-				llmRegistryWarned = true;
-				console.warn("[usage-billing] llm registry unavailable; plugin-registered adapter channels stay unknown (issue #91):", error);
-			}
-			return /* @__PURE__ */ new Set();
+			return ids;
+		}
+		if (!llmRegistryWarned) {
+			llmRegistryWarned = true;
+			console.warn(`[usage-billing] ${reason}; plugin-registered adapter channels (e.g. commandcode) stay in the unknown bucket and are billed at 0 (issue #91). Channel grouping falls back to llm-pi-ai routes only.`);
 		}
+		return ids;
 	};
 	const aggregator = createUsageAggregator(adaptSessionPersistence(ctx.sessionPersistence), {
 		...config.subscriptionProviders === void 0 ? {} : { subscriptionProviders: config.subscriptionProviders },
@@ -7875,4 +7921,4 @@ function apply(ctx, config = {}) {
 	}), "usage-billing: usage-stats route");
 }
 //#endregion
-export { Config, adaptSessionPersistence, apply, createFileUsageLedgerStore, guardLoopback, inject, normalizeTrustedHosts, readPiAiProviderRoutes, resolveSubscriptionKeys };
+export { Config, adaptSessionPersistence, apply, createFileUsageLedgerStore, guardLoopback, inject, normalizeTrustedHosts, readPiAiProviderRoutes, readRegisteredProviderIds, resolveSubscriptionKeys };

package/lib/types/index.d.ts

@@ -154,6 +154,22 @@ export interface PiAiProviderRoute {
 export declare function readPiAiProviderRoutes(settings: SettingsReader): Readonly<Record<string, {
     baseURL?: string;
 }>>;
+/**
+ * 读取宿主 llm 注册表里「已注册 adapter」的 provider id（issue #91）。
+ *
+ * 必须经 `ctx.get('llm')`，**不能**用属性访问 `ctx.llm`：cordis 的服务代理要求
+ * 服务名出现在插件 `inject` 里，否则抛 `cannot get property "llm" without inject`
+ * ——1.4.31 把这个异常吞掉后退化成空集合，结果是「修复发版了但完全不生效」且无任何
+ * 可观察信号（issue #91 follow-up）。本函数把「读不到」与「读到了空集合」都显式
+ * 回报给调用方，让上层能告警而不是静默降级。
+ * @param get - 宿主上下文的 `get(name)` 访问器（cordis 的免 inject 读取口）。
+ * @returns `ids` = 已注册 provider id 集合（读不到时为空集）；
+ *  `reason` = 读取失败或结果不可用的原因；成功时缺省。
+ */
+export declare function readRegisteredProviderIds(get: ((name: string) => unknown) | undefined): {
+    ids: ReadonlySet<string>;
+    reason?: string;
+};
 /**
   * 解析订阅适配器需要的 API Key：从 llm-pi-ai 设置的 `providers.<id>.apiKeyEnv`
  * 读引用（如 kimi-coding → KIMI_CODING_API_KEY），再经凭据 seam 解析成实际值。

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@kenz1117/dsh-ui-usage-billing",
   "description": "Usage billing dashboard for DeepSeek Harness: sidebar cost metrics plus a full dashboard modal, priced from a current multi-provider catalog with real usage aggregated from session logs.",
-  "version": "1.4.33",
+  "version": "1.4.34",
   "keywords": [
     "deepseek-harness",
     "deepseek",