xinyin025/new-api
1
0
Fork 0
mirror of https://github.com/linux-do/new-api.git synced 2025-12-25 08:05:56 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity

feat: 统一错误提示

Browse Source
This commit is contained in:
CaIon
2024-03-20 20:36:55 +08:00
parent eb6257a8d8
commit a232afe9fd
8 changed files with 35 additions and 25 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
relay/channel/claude/relay-claude.go
View File

@@ -313,7 +313,7 @@ func claudeHandler(requestMode int, c *gin.Context, resp *http.Response, promptT
}, nil
}
fullTextResponse := responseClaude2OpenAI(requestMode, &claudeResponse)
completionTokens, err := service.CountTokenText(claudeResponse.Completion, model, constant.ShouldCheckCompletionSensitive())
completionTokens, err, _ := service.CountTokenText(claudeResponse.Completion, model, constant.ShouldCheckCompletionSensitive())
if err != nil {
return service.OpenAIErrorWrapper(err, "count_token_text_failed", http.StatusInternalServerError), nil
}

2
relay/channel/gemini/relay-gemini.go
View File

@@ -257,7 +257,7 @@ func geminiChatHandler(c *gin.Context, resp *http.Response, promptTokens int, mo
}, nil
}
fullTextResponse := responseGeminiChat2OpenAI(&geminiResponse)
completionTokens, _ := service.CountTokenText(geminiResponse.GetResponseText(), model, constant.ShouldCheckCompletionSensitive())
completionTokens, _, _ := service.CountTokenText(geminiResponse.GetResponseText(), model, constant.ShouldCheckCompletionSensitive())
usage := dto.Usage{
PromptTokens: promptTokens,
CompletionTokens: completionTokens,

2
relay/channel/openai/relay-openai.go
View File

@@ -154,7 +154,7 @@ func OpenaiHandler(c *gin.Context, resp *http.Response, promptTokens int, model
completionTokens := 0
for _, choice := range textResponse.Choices {
stringContent := string(choice.Message.Content)
ctkm, _ := service.CountTokenText(stringContent, model, false)
ctkm, _, _ := service.CountTokenText(stringContent, model, false)
completionTokens += ctkm
if checkSensitive {
sensitive, words, stringContent := service.SensitiveWordReplace(stringContent, false)

2
relay/channel/palm/relay-palm.go
View File

@@ -157,7 +157,7 @@ func palmHandler(c *gin.Context, resp *http.Response, promptTokens int, model st
}, nil
}
fullTextResponse := responsePaLM2OpenAI(&palmResponse)
completionTokens, _ := service.CountTokenText(palmResponse.Candidates[0].Content, model, constant.ShouldCheckCompletionSensitive())
completionTokens, _, _ := service.CountTokenText(palmResponse.Candidates[0].Content, model, constant.ShouldCheckCompletionSensitive())
usage := dto.Usage{
PromptTokens: promptTokens,
CompletionTokens: completionTokens,

4
relay/relay-audio.go
View File

@@ -67,7 +67,7 @@ func AudioHelper(c *gin.Context, relayMode int) *dto.OpenAIErrorWithStatusCode {
promptTokens := 0
preConsumedTokens := common.PreConsumedQuota
if strings.HasPrefix(audioRequest.Model, "tts-1") {
promptTokens, err = service.CountAudioToken(audioRequest.Input, audioRequest.Model, constant.ShouldCheckPromptSensitive())
promptTokens, err, _ = service.CountAudioToken(audioRequest.Input, audioRequest.Model, constant.ShouldCheckPromptSensitive())
if err != nil {
return service.OpenAIErrorWrapper(err, "count_audio_token_failed", http.StatusInternalServerError)
}
@@ -173,7 +173,7 @@ func AudioHelper(c *gin.Context, relayMode int) *dto.OpenAIErrorWithStatusCode {
if strings.HasPrefix(audioRequest.Model, "tts-1") {
quota = promptTokens
} else {
quota, err = service.CountAudioToken(audioResponse.Text, audioRequest.Model, constant.ShouldCheckCompletionSensitive())
quota, err, _ = service.CountAudioToken(audioResponse.Text, audioRequest.Model, constant.ShouldCheckCompletionSensitive())
}
quota = int(float64(quota) * ratio)
if ratio != 0 && quota <= 0 {

18
relay/relay-text.go
View File

@@ -98,10 +98,13 @@ func TextHelper(c *gin.Context) *dto.OpenAIErrorWithStatusCode {
var ratio float64
var modelRatio float64
//err := service.SensitiveWordsCheck(textRequest)
promptTokens, err := getPromptTokens(textRequest, relayInfo)
promptTokens, err, sensitiveTrigger := getPromptTokens(textRequest, relayInfo)
// count messages token error 计算promptTokens错误
if err != nil {
if sensitiveTrigger {
return service.OpenAIErrorWrapper(err, "sensitive_words_detected", http.StatusBadRequest)
}
return service.OpenAIErrorWrapper(err, "count_token_messages_failed", http.StatusInternalServerError)
}
@@ -180,25 +183,26 @@ func TextHelper(c *gin.Context) *dto.OpenAIErrorWithStatusCode {
return nil
}
func getPromptTokens(textRequest *dto.GeneralOpenAIRequest, info *relaycommon.RelayInfo) (int, error) {
func getPromptTokens(textRequest *dto.GeneralOpenAIRequest, info *relaycommon.RelayInfo) (int, error, bool) {
var promptTokens int
var err error
var sensitiveTrigger bool
checkSensitive := constant.ShouldCheckPromptSensitive()
switch info.RelayMode {
case relayconstant.RelayModeChatCompletions:
promptTokens, err = service.CountTokenMessages(textRequest.Messages, textRequest.Model, checkSensitive)
promptTokens, err, sensitiveTrigger = service.CountTokenMessages(textRequest.Messages, textRequest.Model, checkSensitive)
case relayconstant.RelayModeCompletions:
promptTokens, err = service.CountTokenInput(textRequest.Prompt, textRequest.Model, checkSensitive)
promptTokens, err, sensitiveTrigger = service.CountTokenInput(textRequest.Prompt, textRequest.Model, checkSensitive)
case relayconstant.RelayModeModerations:
promptTokens, err = service.CountTokenInput(textRequest.Input, textRequest.Model, checkSensitive)
promptTokens, err, sensitiveTrigger = service.CountTokenInput(textRequest.Input, textRequest.Model, checkSensitive)
case relayconstant.RelayModeEmbeddings:
promptTokens, err = service.CountTokenInput(textRequest.Input, textRequest.Model, checkSensitive)
promptTokens, err, sensitiveTrigger = service.CountTokenInput(textRequest.Input, textRequest.Model, checkSensitive)
default:
err = errors.New("unknown relay mode")
promptTokens = 0
}
info.PromptTokens = promptTokens
return promptTokens, err
return promptTokens, err, sensitiveTrigger
}
// 预扣费并返回用户剩余配额