feat(usage): filter-driven Hero with cache-normalized totals

- Normalize OpenAI/Gemini input_tokens semantics in SQL via the new
  fresh_input_sql helper (cache_read subtracted at query time, no data
  migration). Recovers correct cache hit rates for Codex/Gemini.
- Add get_usage_summary_by_app endpoint for per-app split (single
  UNION ALL + GROUP BY, avoids N+1).
- Replace UsageSummaryCards + AppBreakdownRail with a single
  filter-driven UsageHero card; clicking a filter button now truly
  changes the displayed numbers and the title accent color.
- Tighten KNOWN_APP_TYPES to the 3 app_types whose token data is
  reliably collected (claude/codex/gemini); hide claude-desktop,
  hermes, opencode, openclaw filter buttons and i18n keys.
- Flag cache_creation as N/A for OpenAI-style protocols (Codex,
  Gemini); show a "partial" tooltip when the All view mixes both
  protocol families.

src/i18n/locales/en.json

@@ -1135,8 +1135,13 @@
     "rangeToday": "Last 24 hours (hourly)",
     "rangeLast7Days": "Last 7 days",
     "rangeLast30Days": "Last 30 days",
-    "totalTokens": "Total Tokens",
+    "totalTokens": "New Tokens (Input+Output)",
     "cacheTokens": "Cache Tokens",
+    "realTotal": "Tokens Processed",
+    "realTotalHint": "Input + Output + Cache Write + Cache Read",
+    "tokensSuffix": "tokens",
+    "freshInput": "Fresh Input",
+    "cacheHitRate": "Cache Hit Rate",
     "requestLogs": "Request Logs",
     "providerStats": "Provider Stats",
     "modelStats": "Model Stats",
@@ -1165,6 +1170,7 @@
       "codex": "Codex",
       "gemini": "Gemini"
     },
+    "rawInputLabel": "Raw",
     "dataSources": "Data Sources",
     "dataSource": {
       "proxy": "Routing",
@@ -1225,6 +1231,8 @@
     "input": "Input",
     "output": "Output",
     "cacheWrite": "Creation",
+    "cacheWriteNotReported": "OpenAI protocol doesn't distinguish cache creation; only cache hits are reported",
+    "cacheWritePartial": "Some protocols (e.g. OpenAI) don't report cache creation; the value may be incomplete",
     "cacheRead": "Hit",
     "baseCost": "Base",
     "costMultiplier": "Cost Multiplier",

src/i18n/locales/ja.json

@@ -1135,8 +1135,13 @@
     "rangeToday": "直近24時間 (時間別)",
     "rangeLast7Days": "過去7日間",
     "rangeLast30Days": "過去30日間",
-    "totalTokens": "総トークン数",
+    "totalTokens": "新規トークン (Input+Output)",
     "cacheTokens": "キャッシュトークン",
+    "realTotal": "実消費トークン",
+    "realTotalHint": "Input + Output + キャッシュ書込 + キャッシュ命中",
+    "tokensSuffix": "tokens",
+    "freshInput": "新規入力",
+    "cacheHitRate": "キャッシュヒット率",
     "requestLogs": "リクエストログ",
     "providerStats": "プロバイダー統計",
     "modelStats": "モデル統計",
@@ -1165,6 +1170,7 @@
       "codex": "Codex",
       "gemini": "Gemini"
     },
+    "rawInputLabel": "原始",
     "dataSources": "データソース",
     "dataSource": {
       "proxy": "ルーティング",
@@ -1225,6 +1231,8 @@
     "input": "Input",
     "output": "Output",
     "cacheWrite": "作成",
+    "cacheWriteNotReported": "OpenAI プロトコルはキャッシュ作成を区別せず、キャッシュヒットのみ報告します",
+    "cacheWritePartial": "一部のプロトコル（OpenAIなど）はキャッシュ作成を報告しないため、値は不完全な可能性があります",
     "cacheRead": "ヒット",
     "baseCost": "基本",
     "costMultiplier": "コスト倍率",

src/i18n/locales/zh.json

@@ -1135,8 +1135,13 @@
     "rangeToday": "过去 24 小时 (按小时)",
     "rangeLast7Days": "过去 7 天",
     "rangeLast30Days": "过去 30 天",
-    "totalTokens": "总 Token 数",
+    "totalTokens": "新生成 Token (Input+Output)",
     "cacheTokens": "缓存 Token",
+    "realTotal": "真实消耗 Tokens",
+    "realTotalHint": "Input + Output + 缓存写入 + 缓存命中",
+    "tokensSuffix": "tokens",
+    "freshInput": "新增输入",
+    "cacheHitRate": "缓存命中率",
     "requestLogs": "请求日志",
     "providerStats": "Provider 统计",
     "modelStats": "模型统计",
@@ -1165,6 +1170,7 @@
       "codex": "Codex",
       "gemini": "Gemini"
     },
+    "rawInputLabel": "原始",
     "dataSources": "数据来源",
     "dataSource": {
       "proxy": "路由",
@@ -1225,6 +1231,8 @@
     "input": "Input",
     "output": "Output",
     "cacheWrite": "创建",
+    "cacheWriteNotReported": "OpenAI 协议不区分缓存写入，仅上报缓存命中",
+    "cacheWritePartial": "部分协议（如 OpenAI）不上报缓存写入，数值可能偏低",
     "cacheRead": "命中",
     "baseCost": "基础",
     "costMultiplier": "成本倍率",