oadmin/new-api
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat: Implement cache token ratio for more precise token pricing

Browse Source
This commit is contained in:
1808837298@qq.com
2025-03-08 01:30:50 +08:00
parent 81137e0533
commit 4f194f4e6a
18 changed files with 258 additions and 71 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
relay/websocket.go
View File

@@ -11,6 +11,7 @@ import (
relaycommon "one-api/relay/common"
"one-api/service"
"one-api/setting"
"one-api/setting/operation_setting"
)
func WssHelper(c *gin.Context, ws *websocket.Conn) (openaiErr *dto.OpenAIErrorWithStatusCode) {
@@ -39,7 +40,7 @@ func WssHelper(c *gin.Context, ws *websocket.Conn) (openaiErr *dto.OpenAIErrorWi
}
}
//relayInfo.UpstreamModelName = textRequest.Model
modelPrice, getModelPriceSuccess := setting.GetModelPrice(relayInfo.UpstreamModelName, false)
modelPrice, getModelPriceSuccess := operation_setting.GetModelPrice(relayInfo.UpstreamModelName, false)
groupRatio := setting.GetGroupRatio(relayInfo.Group)
var preConsumedQuota int
@@ -65,7 +66,7 @@ func WssHelper(c *gin.Context, ws *websocket.Conn) (openaiErr *dto.OpenAIErrorWi
//if realtimeEvent.Session.MaxResponseOutputTokens != 0 {
// preConsumedTokens = promptTokens + int(realtimeEvent.Session.MaxResponseOutputTokens)
//}
modelRatio, _ = setting.GetModelRatio(relayInfo.UpstreamModelName)
modelRatio, _ = operation_setting.GetModelRatio(relayInfo.UpstreamModelName)
ratio = modelRatio * groupRatio
preConsumedQuota = int(float64(preConsumedTokens) * ratio)
} else {