opencode-go-usage-tui 1.3.0 → 1.3.2

package/README.md

@@ -12,10 +12,12 @@
 
 ## 功能
 
-- **双独立面板**：`Go 用量`（额度百分比 + 进度条）+ `模型价格`（请求限额表），可独立折叠
-- **模型请求限额**：从 opencode.ai/docs/go 实时抓取各模型 5小时/每周/每月 请求数上限（无需登录）
+- **双独立面板**：`Go 用量`（额度百分比 + 进度条）+ `模型限额`，可独立折叠
+- **模型限额一览**：折叠行显示各模型用量上限（$），点击行内展开查看价格明细与完整请求数限额（每5小时/每周/每月）
+- **多档价格分档展示**：支持按上下文分档计价的模型（如 GPT 5.6 Luna ≤272K / >272K、DeepSeek V4 Off-Peak / Peak），逐档列出输入/输出/缓存价
+- **变化高亮与记录**：新增模型=绿色、价格变化=蓝色、限额变化=橙色；展开可见同档位涨跌箭头；面板底部显示最近一次变化的明细清单
 - **关注模型**：`/go-settings` 设置关注的白名单，只显示你关心的模型
-- **刷新间隔可调**：用量刷新和价格刷新间隔均可通过 `/go-settings` 设置（秒）
+- **刷新间隔可调**：用量刷新和限额刷新间隔均可通过 `/go-settings` 设置（秒）
 - **安全加密**：auth cookie 用 AES-256-GCM 加密存储，旧明文自动迁移
 - **自动提醒**：价格/限额变化时弹 toast 通知，变化历史保留最近 10 次
 
@@ -100,10 +102,10 @@ npm install -g opencode-go-usage-tui
 | 字段 | 说明 |
 | ---- | ---- |
 | `ui.usage_panel` | 显示/隐藏用量面板（默认 true） |
-| `ui.price_panel` | 显示/隐藏价格面板（默认 true） |
+| `ui.price_panel` | 显示/隐藏限额面板（默认 true） |
 | `focus_models` | 关注模型白名单（空数组显示全部） |
 | `refresh_interval_sec` | 用量刷新间隔秒数（默认 60） |
-| `pricing_interval_sec` | 价格刷新间隔秒数（默认 1800） |
+| `pricing_interval_sec` | 限额刷新间隔秒数（默认 1800） |
 
 ## 获取 auth cookie
 
@@ -116,18 +118,25 @@ Cookie 有效期约 1 年，过期后重新获取。
 ## 使用
 
 - **`/go-config`**：设置 workspace_id 和 auth cookie
-- **`/go-settings`**：设置关注模型、用量刷新间隔、价格刷新间隔
+- **`/go-settings`**：设置关注模型、用量刷新间隔、限额刷新间隔
 - **`/go-lang`**：切换 中文 / English
 - 用量超过阈值（默认 80%，可用 `OPENCODE_GO_WARN_THRESHOLD` 调整）显示为红色
 - 点击面板标题可折叠/展开
-- 面板标题显示当前插件版本号（如 `v1.3.0`）
+- 面板标题显示当前插件版本号（如 `v1.3.2`）
 
 ## 模型请求限额面板
 
-- 独立面板显示各模型的 5小时/每周/每月 请求数上限，表格风格两端对齐
-- 从 opencode.ai/docs/go 实时抓取公开数据（无需登录）
+- 折叠行：`模型名 + 用量上限($)`，两端对齐、永不折行；点击任意行展开详情
+- 展开详情（逐行清单）：
+  - 价格明细——多档模型按档位逐块列出输入/输出/缓存读/缓存写（$/MTok），价格变化时带 ↑/↓ 涨跌箭头
+  - 请求限额——每5小时 / 每周 / 每月的完整精确请求数（千分位显示，如 `2,050`）
+  - 元数据——SDK、数据保留期、训练用途
+- 数据从 opencode.ai/docs/go 实时抓取公开页面（无需登录）
 - 默认每 30 分钟自动刷新（可用 `/go-settings` 或 `pricing_interval_sec` 调整）
-- 每次刷新自动对比上次数据，变化时 toast 提醒，历史保留最近 10 次到 `go-pricing-data.json`
+- 每次刷新自动对比上次数据：
+  - 变化行着色：新增=绿、价格变化=蓝、限额变化=橙，点击展开查看细节
+  - toast 提醒 + 面板底部"限额变化记录"清单，历史保留最近 10 次到 `go-pricing-data.json`
+- 额度达到 100%（`rate-limited` 状态）时仍正常显示用量与重置时间，不会出现"查询失败"
 
 ## 语言
 

package/README_EN.md

@@ -12,10 +12,12 @@ Bilingual UI (Chinese / English) — switch with `/go-lang`, or pick a language
 
 ## Features
 
-- **Two independent panels**: `Go Usage` (quota percentage + progress bar) + `Model Pricing` (request limits table), each collapsible independently
-- **Model request limits**: scrapes real-time 5h / weekly / monthly request caps per model from opencode.ai/docs/go (no login required)
+- **Two independent panels**: `Go Usage` (quota percentage + progress bar) + `Model Limits`, each collapsible independently
+- **Model limits at a glance**: collapsed rows show each model's usage cap ($); click a row to expand pricing details and full request caps (5h / weekly / monthly)
+- **Tiered pricing display**: models priced by context tiers (e.g. GPT 5.6 Luna ≤272K / >272K, DeepSeek V4 Off-Peak / Peak) list input/output/cache prices per tier
+- **Change highlights & changelog**: added=green, price change=blue, limit change=orange; expand to see ↑/↓ arrows per tier; recent change details listed at panel bottom
 - **Focused models**: use `/go-settings` to whitelist only the models you care about
-- **Adjustable refresh intervals**: set usage and pricing refresh intervals via `/go-settings` (in seconds)
+- **Adjustable refresh intervals**: set usage and limit refresh intervals via `/go-settings` (in seconds)
 - **Secure storage**: auth cookie encrypted with AES-256-GCM, legacy plaintext auto-migrated
 - **Change alerts**: toast notification when prices/limits change; last 10 changes kept
 
@@ -100,10 +102,10 @@ Create `~/.config/opencode/go-usage-config.json` manually:
 | Field | Description |
 | ----- | ----------- |
 | `ui.usage_panel` | Show/hide usage panel (default true) |
-| `ui.price_panel` | Show/hide pricing panel (default true) |
+| `ui.price_panel` | Show/hide limits panel (default true) |
 | `focus_models` | Focused model whitelist (empty array shows all) |
 | `refresh_interval_sec` | Usage refresh interval in seconds (default 60) |
-| `pricing_interval_sec` | Pricing refresh interval in seconds (default 1800) |
+| `pricing_interval_sec` | Limit refresh interval in seconds (default 1800) |
 
 ## Getting the auth cookie
 
@@ -116,18 +118,25 @@ The cookie is valid for about 1 year; get a new one after it expires.
 ## Usage
 
 - **`/go-config`**: set workspace_id and auth cookie
-- **`/go-settings`**: set focused models, usage refresh interval, pricing refresh interval
+- **`/go-settings`**: set focused models, usage refresh interval, limit refresh interval
 - **`/go-lang`**: switch between 中文 / English
 - Usage above the threshold (default 80%, adjustable via `OPENCODE_GO_WARN_THRESHOLD`) is shown in red
 - Click the panel title to collapse/expand
-- The panel title shows the current plugin version (e.g. `v1.3.0`)
+- The panel title shows the current plugin version (e.g. `v1.3.2`)
 
 ## Model request limits panel
 
-- Independent panel showing each model's 5h / weekly / monthly request caps, table style with two-end alignment
+- Collapsed rows: `model name + usage cap ($)`, justified and never wrapped; click any row to expand details
+- Expanded details (line-by-line list):
+  - Pricing — tiered models list input/output/cache-read/cache-write ($/MTok) per tier, with ↑/↓ arrows on price changes
+  - Request caps — full exact request counts for per-5h / weekly / monthly (thousands separators, e.g. `2,050`)
+  - Metadata — SDK, data retention, training usage
 - Scraped live from opencode.ai/docs/go (public data, no login)
 - Auto-refreshes every 30 minutes by default (adjustable via `/go-settings` or `pricing_interval_sec`)
-- Each refresh diffs against the last snapshot; toast on change; last 10 changes kept in `go-pricing-data.json`
+- Each refresh diffs against the last snapshot:
+  - Changed rows are color-coded: added=green, price change=blue, limit change=orange; click to expand details
+  - Toast alert + "limit change log" at the panel bottom; last 10 changes kept in `go-pricing-data.json`
+- When quota hits 100% (`rate-limited` status), usage and reset time still display normally - no "query failed" error
 
 ## Language