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

12
service/log_info_generate.go
View File

@@ -1,16 +1,20 @@
package service
import (
"github.com/gin-gonic/gin"
"one-api/dto"
relaycommon "one-api/relay/common"
"github.com/gin-gonic/gin"
)
func GenerateTextOtherInfo(ctx *gin.Context, relayInfo *relaycommon.RelayInfo, modelRatio, groupRatio, completionRatio, modelPrice float64) map[string]interface{} {
func GenerateTextOtherInfo(ctx *gin.Context, relayInfo *relaycommon.RelayInfo, modelRatio, groupRatio, completionRatio float64,
cacheTokens int, cacheRatio float64, modelPrice float64) map[string]interface{} {
other := make(map[string]interface{})
other["model_ratio"] = modelRatio
other["group_ratio"] = groupRatio
other["completion_ratio"] = completionRatio
other["cache_tokens"] = cacheTokens
other["cache_ratio"] = cacheRatio
other["model_price"] = modelPrice
other["frt"] = float64(relayInfo.FirstResponseTime.UnixMilli() - relayInfo.StartTime.UnixMilli())
if relayInfo.ReasoningEffort != "" {
@@ -27,7 +31,7 @@ func GenerateTextOtherInfo(ctx *gin.Context, relayInfo *relaycommon.RelayInfo, m
}
func GenerateWssOtherInfo(ctx *gin.Context, relayInfo *relaycommon.RelayInfo, usage *dto.RealtimeUsage, modelRatio, groupRatio, completionRatio, audioRatio, audioCompletionRatio, modelPrice float64) map[string]interface{} {
info := GenerateTextOtherInfo(ctx, relayInfo, modelRatio, groupRatio, completionRatio, modelPrice)
info := GenerateTextOtherInfo(ctx, relayInfo, modelRatio, groupRatio, completionRatio, 0, 0.0, modelPrice)
info["ws"] = true
info["audio_input"] = usage.InputTokenDetails.AudioTokens
info["audio_output"] = usage.OutputTokenDetails.AudioTokens
@@ -39,7 +43,7 @@ func GenerateWssOtherInfo(ctx *gin.Context, relayInfo *relaycommon.RelayInfo, us
}
func GenerateAudioOtherInfo(ctx *gin.Context, relayInfo *relaycommon.RelayInfo, usage *dto.Usage, modelRatio, groupRatio, completionRatio, audioRatio, audioCompletionRatio, modelPrice float64) map[string]interface{} {
info := GenerateTextOtherInfo(ctx, relayInfo, modelRatio, groupRatio, completionRatio, modelPrice)
info := GenerateTextOtherInfo(ctx, relayInfo, modelRatio, groupRatio, completionRatio, 0, 0.0, modelPrice)
info["audio"] = true
info["audio_input"] = usage.PromptTokensDetails.AudioTokens
info["audio_output"] = usage.CompletionTokenDetails.AudioTokens

21
service/quota.go
View File

@@ -12,6 +12,7 @@ import (
relaycommon "one-api/relay/common"
"one-api/relay/helper"
"one-api/setting"
"one-api/setting/operation_setting"
"strings"
"time"
@@ -38,9 +39,9 @@ func calculateAudioQuota(info QuotaInfo) int {
return int(info.ModelPrice * common.QuotaPerUnit * info.GroupRatio)
}
completionRatio := setting.GetCompletionRatio(info.ModelName)
audioRatio := setting.GetAudioRatio(info.ModelName)
audioCompletionRatio := setting.GetAudioCompletionRatio(info.ModelName)
completionRatio := operation_setting.GetCompletionRatio(info.ModelName)
audioRatio := operation_setting.GetAudioRatio(info.ModelName)
audioCompletionRatio := operation_setting.GetAudioCompletionRatio(info.ModelName)
ratio := info.GroupRatio * info.ModelRatio
quota := info.InputDetails.TextTokens + int(math.Round(float64(info.OutputDetails.TextTokens)*completionRatio))
@@ -75,7 +76,7 @@ func PreWssConsumeQuota(ctx *gin.Context, relayInfo *relaycommon.RelayInfo, usag
audioInputTokens := usage.InputTokenDetails.AudioTokens
audioOutTokens := usage.OutputTokenDetails.AudioTokens
groupRatio := setting.GetGroupRatio(relayInfo.Group)
modelRatio, _ := setting.GetModelRatio(modelName)
modelRatio, _ := operation_setting.GetModelRatio(modelName)
quotaInfo := QuotaInfo{
InputDetails: TokenDetails{
@@ -122,9 +123,9 @@ func PostWssConsumeQuota(ctx *gin.Context, relayInfo *relaycommon.RelayInfo, mod
audioOutTokens := usage.OutputTokenDetails.AudioTokens
tokenName := ctx.GetString("token_name")
completionRatio := setting.GetCompletionRatio(modelName)
audioRatio := setting.GetAudioRatio(relayInfo.OriginModelName)
audioCompletionRatio := setting.GetAudioCompletionRatio(modelName)
completionRatio := operation_setting.GetCompletionRatio(modelName)
audioRatio := operation_setting.GetAudioRatio(relayInfo.OriginModelName)
audioCompletionRatio := operation_setting.GetAudioCompletionRatio(modelName)
quotaInfo := QuotaInfo{
InputDetails: TokenDetails{
@@ -184,9 +185,9 @@ func PostAudioConsumeQuota(ctx *gin.Context, relayInfo *relaycommon.RelayInfo,
audioOutTokens := usage.CompletionTokenDetails.AudioTokens
tokenName := ctx.GetString("token_name")
completionRatio := setting.GetCompletionRatio(relayInfo.OriginModelName)
audioRatio := setting.GetAudioRatio(relayInfo.OriginModelName)
audioCompletionRatio := setting.GetAudioCompletionRatio(relayInfo.OriginModelName)
completionRatio := operation_setting.GetCompletionRatio(relayInfo.OriginModelName)
audioRatio := operation_setting.GetAudioRatio(relayInfo.OriginModelName)
audioCompletionRatio := operation_setting.GetAudioCompletionRatio(relayInfo.OriginModelName)
modelRatio := priceData.ModelRatio
groupRatio := priceData.GroupRatio

4
service/token_counter.go
View File

@@ -10,7 +10,7 @@ import (
"one-api/constant"
"one-api/dto"
relaycommon "one-api/relay/common"
"one-api/setting"
"one-api/setting/operation_setting"
"strings"
"unicode/utf8"
@@ -33,7 +33,7 @@ func InitTokenEncoders() {
if err != nil {
common.FatalLog(fmt.Sprintf("failed to get gpt-4o token encoder: %s", err.Error()))
}
for model, _ := range setting.GetDefaultModelRatioMap() {
for model, _ := range operation_setting.GetDefaultModelRatioMap() {
if strings.HasPrefix(model, "gpt-3.5") {
tokenEncoderMap[model] = cl100TokenEncoder
} else if strings.HasPrefix(model, "gpt-4") {