feat: 扩展 cache creation token 细分统计与 effective_input_tokens 计费逻辑

- 新增 cache_creation_ephemeral_5m/1h_input_tokens 字段，区分不同 TTL 的缓存写入 token
- 引入 effective_input_tokens（扣除 cache read 后的有效输入 token），暴露给 usage 接口
- billing 规则生成器支持 5m/1h ephemeral cache 独立定价与分级计费
- usage_mapper 增加 Claude/Anthropic 格式映射，修复 OpenAI responses 格式字段兼容性
- 迁移逻辑增强：支持 checksum 容错、applied/pending 数量日志、逐步执行信息输出
- executor 抽离 LocalExecutionRequestOutcome 类型，统一 sync/stream 路径返回语义
- provider-transport auth 层新增 complete passthrough headers 构建逻辑
- 前端 usage 类型全面补充 effective_input_tokens、cache_creation_tokens、total_input_context 字段

crates/aether-data/schema/README.md

@@ -0,0 +1,22 @@
+# Public Schema Snapshot
+
+This directory stores a searchable snapshot of the current local Postgres `public` schema.
+
+Purpose:
+- Make legacy Python-era columns discoverable without reverse-tracing every migration.
+- Keep `migrations/20260403000000_baseline.sql` as a no-op handoff point instead of stuffing the full legacy schema into a fake baseline.
+
+Files:
+- `current-public-tables.tsv`: `table_name`, `column_count`
+- `current-public-columns.tsv`: `table_name`, `ordinal_position`, `column_name`, `data_type`, `is_nullable`, `column_default`, `column_comment`
+
+Source:
+- Generated from the local `aether` Postgres database on `2026-04-10`.
+
+Refresh commands:
+```bash
+docker compose exec -T postgres psql -U postgres -d aether -At -F $'\t' -c "SELECT table_name, COUNT(*) FROM information_schema.columns WHERE table_schema = 'public' GROUP BY table_name ORDER BY table_name;"
+docker compose exec -T postgres psql -U postgres -d aether -At -F $'\t' -c "SELECT c.table_name, c.ordinal_position, c.column_name, c.data_type, c.is_nullable, COALESCE(c.column_default, ''), COALESCE(pg_catalog.col_description((quote_ident(c.table_schema) || '.' || quote_ident(c.table_name))::regclass::oid, c.ordinal_position), '') FROM information_schema.columns c WHERE c.table_schema = 'public' ORDER BY c.table_name, c.ordinal_position;"
+```
+
+This snapshot is documentation only. Runtime code must still treat the actual database as source of truth.