@kenz1117/dsh-ui-usage-billing 1.0.18 → 1.0.19

package/README.en.md

@@ -21,16 +21,6 @@
 
 ---
 
-> **Custom-price reliability + off-peak bands + Tencent TokenHub (v1.0.13, 2026-08-29)**: fixed the silent no-op of origin-bound (relay) custom prices — origins now compare after normalization (protocol optional, path stripped, case-insensitive), and a missing site grid falls back to the origin-bound price instead of keeping host cost; the model input gains a datalist of catalog keys and probed model ids. Custom prices accept optional off-peak bands (peak/off-peak blended 50/50). The balance table adds Tencent Cloud TokenHub token-plan quota (management API, TC3-signed; credential value `<SecretId>:<SecretKey>`).
-
-> **Compatibility floor raised to DSH `0.1.2-alpha.1` (v1.0.12, 2026-08-29)**: the host removed `dsh-client-runtime` and other client-half dependencies in that release; the plugin fully migrated (`dsh-client-store` / `dsh-api-session-controller` / the new LLM remote surface). On older hosts (`0.1.0-rc.8` ~ `0.1.1-rc.2`) stay on v1.0.11.
-
-> **Rate-table display update (2026-08-27)**: the models.dev supplement is **no longer rendered in full** — the rate table previously listed all ~5900 models from 158 gateway providers, drowning out the ones actually in use. The table now contains only the built-in catalog plus probed (configured and reachable) models. Billing is unaffected: models outside the catalog but priced on models.dev are still estimated at their official USD prices when hit; they just no longer appear in the table.
-
-> **Qwen family pricing update (2026-08-27)**: aligned with the latest Alibaba Cloud Model Studio list prices — **Qwen3.7-Max** corrected to official list price (input ¥12 / explicit-cache hit ¥1.2 / output ¥36, previously recorded at the discounted promo rate). The current official 50%-off promotion has **no announced end date**: the rate table shows the discounted price with a red promo badge (hover for details), and resumes list-price display once an end date is filled in after the announcement. The Qwen family also gains **supplementary pricing reference rows** (Batch File standing half-price tier, Batch Chat, explicit-cache create/hit — display-only, excluded from estimation); Qwen3-Coder Plus officially does not support batch inference, so only the explicit-cache rows are listed.
-
-> **Peak/off-peak pricing update (from 2026-08-23 (Sun) 00:00 Beijing)**: DeepSeek models follow the new official rule — **weekdays (Mon–Fri)** keep the original peak/off-peak split (peak 09:00–12:00 / 14:00–18:00, ×2); **weekends (Sat/Sun)** are no longer split and are billed at the **off-peak price** all day. The plugin's billing engine, rate table and per-turn peak/off-peak bands all reflect this.
-
 <div align="center">
   <img src="screenshots/demo.png" alt="dsh-ui-usage-billing — billing dashboard overview" width="80%">
 </div>
@@ -44,7 +34,7 @@
 - **Real usage, no fabricated samples** — the server aggregates from persisted session logs and estimates against live multi-provider official prices; it shows an empty snapshot until real data arrives.
 - **Everything on one screen** — a sidebar trigger card plus a full dashboard (Overview / Trends / Providers / Stats / Rates / Settings) across six tabs: month / today / projection / heatmap / trend.
 - **Subscriptions · balance · quota · reconcile** — plan quota, multi-provider balance, relay-station quota, declared endpoints and balance-delta reconciliation form a cross-verifiable billing loop.
-- **Peak/off-peak pricing + switch alerts** — weekday peak split and weekend all-day off-peak, with a popover / system notification before a tier switch, configurable lead time.
+- **Peak/off-peak pricing + switch alerts** — weekday peak split and weekend all-day off-peak, **priced per official change boundary** (base price before 08-17, weekend peak hours 08-17~08-23, weekend all-day off-peak from 08-23), with a popover / system notification before a tier switch, configurable lead time.
 - **Offline & self-contained** — no chart library, no external CDN, pure design tokens; lightweight and ready to use.
 - **Multi-language + dual currency** — Chinese / English, ¥/$ toggle that only affects this plugin.
 
@@ -80,7 +70,7 @@
 ## 📈 Usage visualizations
 
 - **Session detail + cost spikes + heatmap**: sessions sorted by cost (title / project / calls / cost / last active); per-turn cost bars (peak/off-peak background bands, >2× spike flagged with attribution); month / year calendar heatmap (5-color scale, hover detail; the year view is ~52 weeks, GitHub-style), with active-day and streak counts on top.
-- **Performance metrics**: per-model first-token latency (TTFT) mean / P50 / P90, generation speed (tokens/s), total-latency mean, plus per-hour TTFT and speed curves aggregating by Beijing hour; rendered in the Stats tab as a per-model performance table and per-hour TTFT/speed dual line charts.
+- **Performance metrics**: per-model first-token latency (TTFT) mean / P50 / P90, generation speed (tokens/s), total-latency mean; per-hour × per-model comparison curve — metric tabs (TTFT / tok/s), clickable model chips (top-5 by samples lit by default, one-click select-all), hover snapping to the nearest hour with a crosshair and per-model tooltip, broken lines for missing-sample hours (never fabricated); view preferences persist locally.
 - **Token insights**: a dedicated "Tokens" tab — the daily token chart switches between two views: "Structure" stacks "input (cache miss) / input (cache hit) / output" (including reasoning), while "By model" stacks each day by model (same per-model colors as the trends tab; the toggle hides itself when the snapshot carries no per-day-per-model detail); the hover tooltip shows the day's exact breakdown (per-bucket in structure view, per-model "hit / miss / output" in model view, thousand-separated); clicking a legend swatch or a model-table row focuses that model (other segments dim, y-axis unchanged, click again to release); per-model totals and share, structural KPIs (cache-hit rate / reasoning share / input-output ratio / peak day); per-day token CSV and JSON export (JSON includes the per-day-per-model detail).
 
   ![Trends: daily cost trend, per-turn costs and peak/off-peak share](screenshots/2.png)
@@ -154,7 +144,7 @@ cost (CNY) = (missInput × p_input + cacheHit × p_cacheHit + output × p_output
 
 | Provider  | Models                                                                                       |
 | -------- | ------------------------------------------------------------------------------------------- |
-| DeepSeek | V4 Flash, V4 Flash Vision (Exp), V4 Pro (peak/off-peak billing: weekdays peak 09:00-12:00 / 14:00-18:00 Beijing = 2× off-peak; weekends off-peak all day) |
+| DeepSeek | V4 Flash, V4 Flash Vision (Exp), V4 Pro (priced per official change boundary: base tier → peak/off-peak v1 → weekend off-peak) |
 | Zhipu AI  | GLM-5.3, GLM-5.2, GLM-5.1, GLM-5-Turbo, GLM-4.7, GLM-4.6, GLM-4.5-Air, GLM-5V-Turbo                                       |
 | Aliyun    | Qwen3.8 Max, Qwen3.7-Max, Qwen3.5-Plus, Qwen3.5-Flash                                        |
 | Doubao    | Seed-2.0 Pro, Seed-2.0 Mini, Seed-1.6                                                         |

package/README.md

@@ -21,16 +21,6 @@
 
 ---
 
-> **自定义单价可靠性 + 峰谷价 + 腾讯云 TokenHub（v1.0.13，2026-08-29）**：修复带「来源（中转站）」的自定义价静默失效——来源现按规范化 origin（补协议 / 去路径 / 忽略大小写）宽松比对，三维站点数据缺失时回落为该价重估；模型输入新增下拉候选（目录键 + 探活模型 id）。自定义价新增可选「低谷价」三栏（峰/谷 50% 比例混合估算）。余额表接入腾讯云 TokenHub Token Plan 余量（云 API 管控面 TC3 签名，凭据填 `<SecretId>:<SecretKey>`）。
-
-> **兼容底线提升至 DSH `0.1.2-alpha.1`（v1.0.12，2026-08-29）**：宿主在该版本移除了 `dsh-client-runtime` 等客户端半区依赖，插件已完整迁移（`dsh-client-store` / `dsh-api-session-controller` / 新 LLM Remote 面）；旧宿主（`0.1.0-rc.8` ~ `0.1.1-rc.2`）请停留在 v1.0.11。
-
-> **费率表展示规则更新（2026-08-27）**：models.dev 聚合的补充价目**不再整表渲染**——此前费率表会把 158 个网关厂商的全部约 5900 个模型一并铺出，淹没实际在用的条目；现在表格只包含内置目录与系统探活命中的模型。计费能力不受影响：目录外但 models.dev 有价的模型命中时仍按其官方 USD 价正常估算，只是不再出现在表格里。
-
-> **千问（Qwen）系列价格更新（2026-08-27）**：对齐阿里云百炼最新官方刊例——**Qwen3.7-Max** 更正为官方原价（输入 ¥12 / 显式缓存命中 ¥1.2 / 输出 ¥36，此前误录活动折后价），当前官方限时 5 折且**未公布截止时间**，费率表按折后价显示并挂红色促销徽章（悬停提示活动性质），公告截止后补填日期即自动恢复原价展示；Qwen 全系补充**附加计价参考行**（Batch File 长期半价档、Batch Chat、显式缓存创建/命中，纯展示不参与估算计费）；Qwen3-Coder Plus 官方明确不支持批量推理，仅列显式缓存两行。
-
-> **峰谷计费规则更新（自 2026-08-23（周日）00:00 起）**：DeepSeek 模型按官方新规计费——**工作日（周一至周五）** 继续执行原峰谷分段计费（高峰 09:00–12:00 / 14:00–18:00，×2）；**周末（周六、周日）** 全天不再区分峰谷时段，统一按**低谷价**计费。插件计费引擎、费率表与每轮费用峰谷分带均已同步生效。
-
 <div align="center">
   <img src="screenshots/demo.png" alt="dsh-ui-usage-billing — 计费仪表盘总览" width="80%">
 </div>
@@ -42,9 +32,9 @@
 ## ✨ 核心亮点
 
 - **真实用量，不伪造样本** — 服务端从持久化会话日志实时聚合，按实时多厂商官方价估算；数据到达前显示空快照，绝不展示假数据。
-- **一屏看懂一切** — 侧边栏触发卡 + 全屏仪表盘（概览 / 趋势 / 明细 / 统计 / 费率 / 设置）六区，本月/今日/预计/热力图/趋势全在。
+- **一屏看懂一切** — 侧边栏触发卡 + 全屏仪表盘（概览 / Token / 用量 / 趋势 / 费率 / 设置）六区，本月/今日/预计/热力图/趋势全在。
 - **订阅 · 余额 · 额度 · 对账** — 订阅套餐额度、多厂商余额、中转站额度、声明端点、余额差对账，形成可交叉验证的计费闭环。
-- **峰谷计价 + 切换提醒** — 工作日峰谷分时、周末全天低谷，切档前弹窗/系统通知，提前量可配。
+- **峰谷计价 + 切换提醒** — 工作日峰谷分时、周末全天低谷，**按官方变更节点分段计价**（8-17 前基础价、8-17~8-23 周末计峰、8-23 起周末全谷），切档前弹窗/系统通知，提前量可配。
 - **离线自包含** — 无图表库、无外部 CDN、纯设计令牌；依赖极轻，随装随用。
 - **多语种 + 双币种** — 中文/English、¥/≈$ 切换，只对本插件生效。
 
@@ -60,7 +50,7 @@
 
 ## 💰 计费引擎
 
-- **实时定价费率表**：models.dev 抓价 + 探活模型对标——系统实际配置模型全纳入；峰谷分时（工作日 9-12 / 14-18 高峰 ×2，周末全天低谷）+ 实时汇率（USD→CNY），每 6 小时刷新。
+- **实时定价费率表**：models.dev 抓价 + 探活模型对标——系统实际配置模型全纳入；峰谷分时（工作日 9-12 / 14-18 高峰 ×2，周末全天低谷；历史费用按官方变更节点分段计价，见下方「计费细节」）+ 实时汇率（USD→CNY），每 6 小时刷新。
 - **自定义单价**：设置面板为未收录或变价模型填入实付价（未命中 / 缓存命中 / 输出，可选 USD 与低谷价三栏），总览与日趋势按用户价重估显示；支持按中转站来源绑定同模型不同价（origin 规范化宽松匹配），目录外模型填价即生效。
 
   ![费率：模型单价表（峰谷分时与实时汇率）](screenshots/5.png)
@@ -81,7 +71,7 @@
 ## 📈 用量可视化
 
 - **会话明细 + 成本突增 + 热力图**：按会话费用倒序（标题 / 项目 / 调用 / 费用 / 最后活跃）；每轮费用柱状图（峰谷背景分带、超 2 倍红标归因）；月 / 年日历热力图（5 档色阶、悬停明细；年视图近 52 周、GitHub 风格），头部显示活跃天数 / 连续使用天数。
-- **性能指标**：每个模型首字延时（TTFT）均值 / P50 / P90、生成速度（tokens/s）、总延迟均值，另按北京时间小时聚合 TTFT 与速度曲线；统计 Tab 渲染为按模型性能表 + 按小时 TTFT/速度双折线。
+- **性能指标**：每个模型首字延时（TTFT）均值 / P50 / P90、生成速度（tokens/s）、总延迟均值；按小时×模型对比曲线——指标 tab 切换（首字延时 / 生成速度），模型 chip 点击开/关曲线（默认点亮样本数前 5，全选一键），悬停吸附最近小时显示十字线与逐模型数值，缺失样本小时断线不造假；视图偏好本地持久化。
 - **Token 统计洞察**：独立「Token」分区——每日 token 堆叠双视角切换：「按结构」按「输入（缓存未命中）/ 输入（缓存命中）/ 输出」三桶分色（含 reasoning 思考），「按模型」每天按模型堆叠（分色与趋势页一致，旧快照缺按日×模型明细时自动隐藏切换）；悬停柱状图显示当日精确明细（结构视角给三桶逐项，模型视角给逐模型「命中/未命中/输出」，千分位不缩写）；点击图例色块或模型 Token 表行可聚焦单个模型（其余段弱化、y 轴不变，再点解除），模型 token 总量与占比，结构 KPI（缓存命中率 / 思考占比 / 输入输出比 / 峰值日）；按日 token CSV 与 JSON 导出（JSON 含按日×模型明细）。
 
   ![趋势：每日费用趋势、每轮费用与峰谷时段占比](screenshots/2.png)
@@ -161,7 +151,7 @@ cost（CNY）= (missInput × p_input + cacheHit × p_cacheHit + output × p_outp
 
 | 厂商       | 代表模型                                        |
 | -------- | -------------------------------------------- |
-| DeepSeek | V4 Flash、V4 Pro（等 3 款；峰谷计费：工作日高峰 ×2、周末全天低谷） |
+| DeepSeek | V4 Flash、V4 Pro（等 3 款；按官方变更节点分段计价：基础价 → 峰谷 v1 → 周末全谷） |
 | 智谱 AI    | GLM-5.3、GLM-5.2（等 10 款）                     |
 | 阿里通义     | Qwen3.8 Max、Qwen3-Coder 480B（等 7 款）         |
 | 字节豆包     | Doubao Seed-2.1 Pro、Doubao-Seed-Evolving（等 8 款） |