admin/CC-Switch
1
0
Fork 0
mirror of https://github.com/farion1231/cc-switch.git synced 2026-07-28 08:44:41 +08:00
Code Issues Packages Projects Releases Wiki Activity

fix(usage): account for cache-write tokens across schema versions

Browse Source 
Parse cache_write_tokens from OpenAI usage details and preserve cache creation data across Chat, Responses, and Anthropic conversion paths.

Add explicit input-token semantics to request logs and rollups so legacy rows subtract cache reads only while new total-inclusive rows subtract both cache reads and writes. Migrate v12 databases, normalize rollups to fresh input, and cover historical backfill behavior with regression tests.
This commit is contained in:
Jason
2026-07-11 12:27:39 +08:00
parent 06039540ff
commit f991726ff0
13 changed files with 443 additions and 102 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src-tauri/src/database/dao/usage_rollup.rs
+42 -4
View File
@@ -4,6 +4,7 @@
use crate::database::{lock_conn, Database};
use crate::error::AppError;
use crate::services::sql_helpers::{fresh_input_sql, INPUT_TOKEN_SEMANTICS_FRESH};
use crate::services::usage_stats::effective_usage_log_filter;
use chrono::{Duration, Local, TimeZone};
@@ -115,6 +116,8 @@ impl Database {
fn do_rollup_and_prune(conn: &rusqlite::Connection, cutoff: i64) -> Result<u64, AppError> {
// Aggregate old logs, merging with any pre-existing rollup rows via LEFT JOIN.
let effective_filter = effective_usage_log_filter("l");
let fresh_detail_input = fresh_input_sql("l");
let fresh_old_input = fresh_input_sql("old");
// request_model 维度保留路由接管的「客户端别名 → 真实模型」映射，
// pricing_model 维度保留写入时的计价基准（request 计价模式下与 model 分叉）；
// 明细行的这两列可能为 NULL（历史/手工数据），归一为 ''。
@@ -124,15 +127,16 @@ impl Database {
request_count, success_count,
input_tokens, output_tokens,
cache_read_tokens, cache_creation_tokens,
total_cost_usd, avg_latency_ms)
input_token_semantics, total_cost_usd, avg_latency_ms)
SELECT
d, a, p, m, rm, pm,
COALESCE(old.request_count, 0) + new_req,
COALESCE(old.success_count, 0) + new_succ,
COALESCE(old.input_tokens, 0) + new_in,
COALESCE({fresh_old_input}, 0) + new_in,
COALESCE(old.output_tokens, 0) + new_out,
COALESCE(old.cache_read_tokens, 0) + new_cr,
COALESCE(old.cache_creation_tokens, 0) + new_cc,
{INPUT_TOKEN_SEMANTICS_FRESH},
CAST(COALESCE(CAST(old.total_cost_usd AS REAL), 0) + new_cost AS TEXT),
CASE WHEN COALESCE(old.request_count, 0) + new_req > 0
THEN (COALESCE(old.avg_latency_ms, 0) * COALESCE(old.request_count, 0)
@@ -147,7 +151,7 @@ impl Database {
COALESCE(l.pricing_model, '') as pm,
COUNT(*) as new_req,
SUM(CASE WHEN l.status_code >= 200 AND l.status_code < 300 THEN 1 ELSE 0 END) as new_succ,
COALESCE(SUM(l.input_tokens), 0) as new_in,
COALESCE(SUM({fresh_detail_input}), 0) as new_in,
COALESCE(SUM(l.output_tokens), 0) as new_out,
COALESCE(SUM(l.cache_read_tokens), 0) as new_cr,
COALESCE(SUM(l.cache_creation_tokens), 0) as new_cc,
@@ -327,7 +331,7 @@ mod tests {
let (provider_id, request_count, input_tokens, output_tokens, cache_read_tokens) = &rows[0];
assert_eq!(provider_id, "openai");
assert_eq!(*request_count, 1);
assert_eq!(*input_tokens, 100);
assert_eq!(*input_tokens, 90, "rollup stores normalized fresh input");
assert_eq!(*output_tokens, 20);
assert_eq!(*cache_read_tokens, 10);
@@ -340,6 +344,40 @@ mod tests {
Ok(())
}
#[test]
fn test_rollup_normalizes_total_cache_semantics_to_fresh() -> Result<(), AppError> {
let db = Database::memory()?;
let old_ts = chrono::Utc::now().timestamp() - 40 * 86400;
{
let conn = crate::database::lock_conn!(db.conn);
conn.execute(
"INSERT INTO proxy_request_logs (
request_id, provider_id, app_type, model,
input_tokens, output_tokens, cache_read_tokens, cache_creation_tokens,
input_token_semantics, total_cost_usd,
latency_ms, status_code, created_at
) VALUES ('total-semantics-rollup', 'p1', 'codex', 'gpt-5.5',
100, 5, 10, 20, 1, '0.10', 100, 200, ?1)",
[old_ts],
)?;
}
assert_eq!(db.rollup_and_prune(30)?, 1);
let conn = crate::database::lock_conn!(db.conn);
let row: (i64, i64, i64, i64) = conn.query_row(
"SELECT input_tokens, cache_read_tokens, cache_creation_tokens,
input_token_semantics
FROM usage_daily_rollups WHERE model = 'gpt-5.5'",
[],
|row| Ok((row.get(0)?, row.get(1)?, row.get(2)?, row.get(3)?)),
)?;
assert_eq!(row, (70, 10, 20, 2));
Ok(())
}
#[test]
fn test_rollup_preserves_request_model_dimension() -> Result<(), AppError> {
let db = Database::memory()?;
src-tauri/src/database/mod.rs
+1 -1
View File
@@ -50,7 +50,7 @@ use std::sync::Mutex;
/// 当前 Schema 版本号
/// 每次修改表结构时递增，并在 schema.rs 中添加相应的迁移逻辑
pub(crate) const SCHEMA_VERSION: i32 = 12;
pub(crate) const SCHEMA_VERSION: i32 = 13;
/// 安全地序列化 JSON，避免 unwrap panic
pub(crate) fn to_json_string<T: Serialize>(value: &T) -> Result<String, AppError> {
src-tauri/src/database/schema.rs
+74
View File
@@ -189,6 +189,7 @@ impl Database {
pricing_model TEXT,
input_tokens INTEGER NOT NULL DEFAULT 0, output_tokens INTEGER NOT NULL DEFAULT 0,
cache_read_tokens INTEGER NOT NULL DEFAULT 0, cache_creation_tokens INTEGER NOT NULL DEFAULT 0,
input_token_semantics INTEGER NOT NULL DEFAULT 0,
input_cost_usd TEXT NOT NULL DEFAULT '0', output_cost_usd TEXT NOT NULL DEFAULT '0',
cache_read_cost_usd TEXT NOT NULL DEFAULT '0', cache_creation_cost_usd TEXT NOT NULL DEFAULT '0',
total_cost_usd TEXT NOT NULL DEFAULT '0', latency_ms INTEGER NOT NULL, first_token_ms INTEGER,
@@ -275,6 +276,7 @@ impl Database {
output_tokens INTEGER NOT NULL DEFAULT 0,
cache_read_tokens INTEGER NOT NULL DEFAULT 0,
cache_creation_tokens INTEGER NOT NULL DEFAULT 0,
input_token_semantics INTEGER NOT NULL DEFAULT 0,
total_cost_usd TEXT NOT NULL DEFAULT '0',
avg_latency_ms INTEGER NOT NULL DEFAULT 0,
PRIMARY KEY (date, app_type, provider_id, model, request_model, pricing_model)
@@ -478,6 +480,11 @@ impl Database {
Self::migrate_v11_to_v12(conn)?;
Self::set_user_version(conn, 12)?;
}
12 => {
log::info!("迁移数据库从 v12 到 v13（记录输入 token 缓存语义）");
Self::migrate_v12_to_v13(conn)?;
Self::set_user_version(conn, 13)?;
}
_ => {
return Err(AppError::Database(format!(
"未知的数据库版本 {version}，无法迁移到 {SCHEMA_VERSION}"
@@ -650,6 +657,7 @@ impl Database {
request_model TEXT,
input_tokens INTEGER NOT NULL DEFAULT 0, output_tokens INTEGER NOT NULL DEFAULT 0,
cache_read_tokens INTEGER NOT NULL DEFAULT 0, cache_creation_tokens INTEGER NOT NULL DEFAULT 0,
input_token_semantics INTEGER NOT NULL DEFAULT 0,
input_cost_usd TEXT NOT NULL DEFAULT '0', output_cost_usd TEXT NOT NULL DEFAULT '0',
cache_read_cost_usd TEXT NOT NULL DEFAULT '0', cache_creation_cost_usd TEXT NOT NULL DEFAULT '0',
total_cost_usd TEXT NOT NULL DEFAULT '0', latency_ms INTEGER NOT NULL, first_token_ms INTEGER,
@@ -1322,6 +1330,30 @@ impl Database {
Ok(())
}
/// v12 -> v13：记录 input_tokens 是否包含缓存写入。
///
/// 默认 0 表示旧版/未知语义；旧 Codex 行只包含 cache read，不包含
/// cache creation。新代理行会显式写入 1(total-inclusive) 或 2(fresh)。
fn migrate_v12_to_v13(conn: &Connection) -> Result<(), AppError> {
if Self::table_exists(conn, "proxy_request_logs")? {
Self::add_column_if_missing(
conn,
"proxy_request_logs",
"input_token_semantics",
"INTEGER NOT NULL DEFAULT 0",
)?;
}
if Self::table_exists(conn, "usage_daily_rollups")? {
Self::add_column_if_missing(
conn,
"usage_daily_rollups",
"input_token_semantics",
"INTEGER NOT NULL DEFAULT 0",
)?;
}
Ok(())
}
/// 插入默认模型定价数据
/// 格式: (model_id, display_name, input, output, cache_read, cache_creation)
/// 注意: model_id 使用短横线格式（如 claude-haiku-4-5），与 API 返回的模型名称标准化后一致
@@ -2661,3 +2693,45 @@ impl Database {
Ok(true)
}
}
#[cfg(test)]
mod tests {
use super::*;
#[test]
fn migrate_v12_to_v13_adds_input_token_semantics_columns() -> Result<(), AppError> {
let conn = Connection::open_in_memory()?;
conn.execute(
"CREATE TABLE proxy_request_logs (request_id TEXT PRIMARY KEY)",
[],
)?;
conn.execute(
"CREATE TABLE usage_daily_rollups (date TEXT PRIMARY KEY)",
[],
)?;
Database::set_user_version(&conn, 12)?;
Database::apply_schema_migrations_on_conn(&conn)?;
assert_eq!(Database::get_user_version(&conn)?, 13);
assert!(Database::has_column(
&conn,
"proxy_request_logs",
"input_token_semantics"
)?);
assert!(Database::has_column(
&conn,
"usage_daily_rollups",
"input_token_semantics"
)?);
let log_default: i64 = conn.query_row(
"SELECT dflt_value = '0' FROM pragma_table_info('proxy_request_logs')
WHERE name = 'input_token_semantics'",
[],
|row| row.get(0),
)?;
assert_eq!(log_default, 1);
Ok(())
}
}