Merge 8726729ade into 629378691b

Co-authored-by: oliang <oliang@tencent.com>

Dockerfile

@@ -5,15 +5,15 @@ COPY ./VERSION .
 COPY ./web .
 
 WORKDIR /web/default
-RUN npm install
+RUN npm config set registry https://mirrors.huaweicloud.com/repository/npm/ && npm install
 RUN DISABLE_ESLINT_PLUGIN='true' REACT_APP_VERSION=$(cat VERSION) npm run build
 
 WORKDIR /web/berry
-RUN npm install
+RUN npm config set registry https://mirrors.huaweicloud.com/repository/npm/ && npm install
 RUN DISABLE_ESLINT_PLUGIN='true' REACT_APP_VERSION=$(cat VERSION) npm run build
 
 WORKDIR /web/air
-RUN npm install
+RUN npm config set registry https://mirrors.huaweicloud.com/repository/npm/ && npm install
 RUN DISABLE_ESLINT_PLUGIN='true' REACT_APP_VERSION=$(cat VERSION) npm run build
 
 FROM golang:alpine AS builder2

common/gin.go

@@ -31,15 +31,15 @@ func UnmarshalBodyReusable(c *gin.Context, v any) error {
 	contentType := c.Request.Header.Get("Content-Type")
 	if strings.HasPrefix(contentType, "application/json") {
 		err = json.Unmarshal(requestBody, &v)
+		c.Request.Body = io.NopCloser(bytes.NewBuffer(requestBody))
 	} else {
-		// skip for now
-		// TODO: someday non json request have variant model, we will need to implementation this
+		c.Request.Body = io.NopCloser(bytes.NewBuffer(requestBody))
+		err = c.ShouldBind(&v)
 	}
 	if err != nil {
 		return err
 	}
 	// Reset request body
-	c.Request.Body = io.NopCloser(bytes.NewBuffer(requestBody))
 	return nil
 }
 

controller/channel-billing.go

@@ -81,6 +81,26 @@ type APGC2DGPTUsageResponse struct {
 	TotalUsed      float64 `json:"total_used"`
 }
 
+type SiliconFlowUsageResponse struct {
+	Code    int    `json:"code"`
+	Message string `json:"message"`
+	Status  bool   `json:"status"`
+	Data    struct {
+		ID            string `json:"id"`
+		Name          string `json:"name"`
+		Image         string `json:"image"`
+		Email         string `json:"email"`
+		IsAdmin       bool   `json:"isAdmin"`
+		Balance       string `json:"balance"`
+		Status        string `json:"status"`
+		Introduction  string `json:"introduction"`
+		Role          string `json:"role"`
+		ChargeBalance string `json:"chargeBalance"`
+		TotalBalance  string `json:"totalBalance"`
+		Category      string `json:"category"`
+	} `json:"data"`
+}
+
 // GetAuthHeader get auth header
 func GetAuthHeader(token string) http.Header {
 	h := http.Header{}
@@ -203,6 +223,28 @@ func updateChannelAIGC2DBalance(channel *model.Channel) (float64, error) {
 	return response.TotalAvailable, nil
 }
 
+func updateChannelSiliconFlowBalance(channel *model.Channel) (float64, error) {
+	url := "https://api.siliconflow.cn/v1/user/info"
+	body, err := GetResponseBody("GET", url, channel, GetAuthHeader(channel.Key))
+	if err != nil {
+		return 0, err
+	}
+	response := SiliconFlowUsageResponse{}
+	err = json.Unmarshal(body, &response)
+	if err != nil {
+		return 0, err
+	}
+	if response.Code != 20000 {
+		return 0, fmt.Errorf("code: %d, message: %s", response.Code, response.Message)
+	}
+	balance, err := strconv.ParseFloat(response.Data.Balance, 64)
+	if err != nil {
+		return 0, err
+	}
+	channel.UpdateBalance(balance)
+	return balance, nil
+}
+
 func updateChannelBalance(channel *model.Channel) (float64, error) {
 	baseURL := channeltype.ChannelBaseURLs[channel.Type]
 	if channel.GetBaseURL() == "" {
@@ -227,6 +269,8 @@ func updateChannelBalance(channel *model.Channel) (float64, error) {
 		return updateChannelAPI2GPTBalance(channel)
 	case channeltype.AIGC2D:
 		return updateChannelAIGC2DBalance(channel)
+	case channeltype.SiliconFlow:
+		return updateChannelSiliconFlowBalance(channel)
 	default:
 		return 0, errors.New("尚未实现")
 	}

controller/channel-test.go

@@ -76,9 +76,9 @@ func testChannel(channel *model.Channel, request *relaymodel.GeneralOpenAIReques
 		if len(modelNames) > 0 {
 			modelName = modelNames[0]
 		}
-		if modelMap != nil && modelMap[modelName] != "" {
-			modelName = modelMap[modelName]
-		}
+	}
+	if modelMap != nil && modelMap[modelName] != "" {
+		modelName = modelMap[modelName]
 	}
 	meta.OriginModelName, meta.ActualModelName = request.Model, modelName
 	request.Model = modelName

middleware/distributor.go

@@ -12,7 +12,7 @@ import (
 )
 
 type ModelRequest struct {
-	Model string `json:"model"`
+	Model string `json:"model" form:"model"`
 }
 
 func Distribute() func(c *gin.Context) {

model/token.go

@@ -30,7 +30,7 @@ type Token struct {
 	RemainQuota    int64   `json:"remain_quota" gorm:"bigint;default:0"`
 	UnlimitedQuota bool    `json:"unlimited_quota" gorm:"default:false"`
 	UsedQuota      int64   `json:"used_quota" gorm:"bigint;default:0"` // used quota
-	Models         *string `json:"models" gorm:"default:''"`           // allowed models
+	Models         *string `json:"models" gorm:"type:text"`            // allowed models
 	Subnet         *string `json:"subnet" gorm:"default:''"`           // allowed subnet
 }
 
@@ -121,30 +121,40 @@ func GetTokenById(id int) (*Token, error) {
 	return &token, err
 }
 
-func (token *Token) Insert() error {
+func (t *Token) Insert() error {
 	var err error
-	err = DB.Create(token).Error
+	err = DB.Create(t).Error
 	return err
 }
 
 // Update Make sure your token's fields is completed, because this will update non-zero values
-func (token *Token) Update() error {
+func (t *Token) Update() error {
 	var err error
-	err = DB.Model(token).Select("name", "status", "expired_time", "remain_quota", "unlimited_quota", "models", "subnet").Updates(token).Error
+	err = DB.Model(t).Select("name", "status", "expired_time", "remain_quota", "unlimited_quota", "models", "subnet").Updates(t).Error
 	return err
 }
 
-func (token *Token) SelectUpdate() error {
+func (t *Token) SelectUpdate() error {
 	// This can update zero values
-	return DB.Model(token).Select("accessed_time", "status").Updates(token).Error
+	return DB.Model(t).Select("accessed_time", "status").Updates(t).Error
 }
 
-func (token *Token) Delete() error {
+func (t *Token) Delete() error {
 	var err error
-	err = DB.Delete(token).Error
+	err = DB.Delete(t).Error
 	return err
 }
 
+func (t *Token) GetModels() string {
+	if t == nil {
+		return ""
+	}
+	if t.Models == nil {
+		return ""
+	}
+	return *t.Models
+}
+
 func DeleteTokenById(id int, userId int) (err error) {
 	// Why we need userId here? In case user want to delete other's token.
 	if id == 0 || userId == 0 {

relay/adaptor/ali/main.go

@@ -149,7 +149,24 @@ func responseAli2OpenAI(response *ChatResponse) *openai.TextResponse {
 	return &fullTextResponse
 }
 
-func streamResponseAli2OpenAI(aliResponse *ChatResponse) *openai.ChatCompletionsStreamResponse {
+func streamResponseAli2OpenAI(aliResponse *ChatResponse) interface{} {
+	if aliResponse.Code != "" {
+		var choice openai.ChatCompletionsStreamResponseChoice
+		choice.Index = 0
+		choice.Delta = model.Message{
+			Role:    "assistant",
+			Content: "",
+		}
+		response := openai.ChatCompletionsErrorStreamResponse{
+			Id:        aliResponse.RequestId,
+			Object:    "chat.completion.chunk",
+			Created:   helper.GetTimestamp(),
+			Model:     "qwen",
+			ErrorCode: aliResponse.Code,
+			Choices:   []openai.ChatCompletionsStreamResponseChoice{choice},
+		}
+		return &response
+	}
 	if len(aliResponse.Output.Choices) == 0 {
 		return nil
 	}
@@ -201,6 +218,19 @@ func StreamHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusC
 			logger.SysError("error unmarshalling stream response: " + err.Error())
 			continue
 		}
+
+		// Check for known error codes and handle accordingly
+		if aliResponse.Code != "" {
+			response := streamResponseAli2OpenAI(&aliResponse)
+
+			err = render.ObjectData(c, response)
+			if err != nil {
+				logger.SysError(err.Error())
+			}
+			render.Done(c)
+			return nil, nil
+		}
+
 		if aliResponse.Usage.OutputTokens != 0 {
 			usage.PromptTokens = aliResponse.Usage.InputTokens
 			usage.CompletionTokens = aliResponse.Usage.OutputTokens
@@ -245,6 +275,8 @@ func Handler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, *
 	if err != nil {
 		return openai.ErrorWrapper(err, "unmarshal_response_body_failed", http.StatusInternalServerError), nil
 	}
+
+	// Check for known error codes and handle accordingly
 	if aliResponse.Code != "" {
 		return &model.ErrorWithStatusCode{
 			Error: model.Error{
@@ -256,6 +288,7 @@ func Handler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, *
 			StatusCode: resp.StatusCode,
 		}, nil
 	}
+
 	fullTextResponse := responseAli2OpenAI(&aliResponse)
 	fullTextResponse.Model = "qwen"
 	jsonResponse, err := json.Marshal(fullTextResponse)

relay/adaptor/groq/constants.go

@@ -4,14 +4,21 @@ package groq
 
 var ModelList = []string{
 	"gemma-7b-it",
-	"mixtral-8x7b-32768",
-	"llama3-8b-8192",
-	"llama3-70b-8192",
 	"gemma2-9b-it",
-	"llama-3.1-405b-reasoning",
 	"llama-3.1-70b-versatile",
 	"llama-3.1-8b-instant",
+	"llama-3.2-11b-text-preview",
+	"llama-3.2-11b-vision-preview",
+	"llama-3.2-1b-preview",
+	"llama-3.2-3b-preview",
+	"llama-3.2-90b-text-preview",
+	"llama-guard-3-8b",
+	"llama3-70b-8192",
+	"llama3-8b-8192",
 	"llama3-groq-70b-8192-tool-use-preview",
 	"llama3-groq-8b-8192-tool-use-preview",
+	"llava-v1.5-7b-4096-preview",
+	"mixtral-8x7b-32768",
+	"distil-whisper-large-v3-en",
 	"whisper-large-v3",
 }

relay/adaptor/openai/main.go

@@ -55,8 +55,8 @@ func StreamHandler(c *gin.Context, resp *http.Response, relayMode int) (*model.E
 				render.StringData(c, data) // if error happened, pass the data to client
 				continue                   // just ignore the error
 			}
-			if len(streamResponse.Choices) == 0 {
-				// but for empty choice, we should not pass it to client, this is for azure
+			if len(streamResponse.Choices) == 0 && streamResponse.Usage == nil {
+				// but for empty choice and no usage, we should not pass it to client, this is for azure
 				continue // just ignore empty choice
 			}
 			render.StringData(c, data)

relay/adaptor/openai/model.go

@@ -97,6 +97,16 @@ type TextResponse struct {
 	model.Usage `json:"usage"`
 }
 
+type ErrorTextResponse struct {
+	Id          string               `json:"id"`
+	Model       string               `json:"model,omitempty"`
+	Object      string               `json:"object"`
+	ErrorCode   string               `json:"error_code"`
+	Created     int64                `json:"created"`
+	Choices     []TextResponseChoice `json:"choices"`
+	model.Usage `json:"usage"`
+}
+
 type EmbeddingResponseItem struct {
 	Object    string    `json:"object"`
 	Index     int       `json:"index"`
@@ -137,6 +147,16 @@ type ChatCompletionsStreamResponse struct {
 	Usage   *model.Usage                          `json:"usage,omitempty"`
 }
 
+type ChatCompletionsErrorStreamResponse struct {
+	Id        string                                `json:"id"`
+	Object    string                                `json:"object"`
+	Created   int64                                 `json:"created"`
+	ErrorCode string                                `json:"error_code"`
+	Model     string                                `json:"model"`
+	Choices   []ChatCompletionsStreamResponseChoice `json:"choices"`
+	Usage     *model.Usage                          `json:"usage,omitempty"`
+}
+
 type CompletionsStreamResponse struct {
 	Choices []struct {
 		Text         string `json:"text"`

relay/adaptor/tencent/constants.go

@@ -5,4 +5,5 @@ var ModelList = []string{
 	"hunyuan-standard",
 	"hunyuan-standard-256K",
 	"hunyuan-pro",
+	"hunyuan-vision",
 }

relay/adaptor/vertexai/gemini/adapter.go

@@ -15,7 +15,7 @@ import (
 )
 
 var ModelList = []string{
-	"gemini-1.5-pro-001", "gemini-1.5-flash-001", "gemini-pro", "gemini-pro-vision",
+	"gemini-1.5-pro-001", "gemini-1.5-flash-001", "gemini-pro", "gemini-pro-vision", "gemini-1.5-pro-002", "gemini-1.5-flash-002", 
 }
 
 type Adaptor struct {

relay/adaptor/xunfei/constants.go

@@ -5,6 +5,8 @@ var ModelList = []string{
 	"SparkDesk-v1.1",
 	"SparkDesk-v2.1",
 	"SparkDesk-v3.1",
+	"SparkDesk-v3.1-128K",
 	"SparkDesk-v3.5",
+	"SparkDesk-v3.5-32K",
 	"SparkDesk-v4.0",
 }

relay/adaptor/xunfei/main.go

@@ -272,9 +272,9 @@ func xunfeiMakeRequest(textRequest model.GeneralOpenAIRequest, domain, authUrl,
 }
 
 func parseAPIVersionByModelName(modelName string) string {
-	parts := strings.Split(modelName, "-")
-	if len(parts) == 2 {
-		return parts[1]
+	index := strings.IndexAny(modelName, "-")
+	if index != -1 {
+		return modelName[index+1:]
 	}
 	return ""
 }
@@ -288,8 +288,12 @@ func apiVersion2domain(apiVersion string) string {
 		return "generalv2"
 	case "v3.1":
 		return "generalv3"
+	case "v3.1-128K":
+		return "pro-128k"
 	case "v3.5":
 		return "generalv3.5"
+	case "v3.5-32K":
+		return "max-32k"
 	case "v4.0":
 		return "4.0Ultra"
 	}
@@ -297,7 +301,17 @@ func apiVersion2domain(apiVersion string) string {
 }
 
 func getXunfeiAuthUrl(apiVersion string, apiKey string, apiSecret string) (string, string) {
+	var authUrl string
 	domain := apiVersion2domain(apiVersion)
-	authUrl := buildXunfeiAuthUrl(fmt.Sprintf("wss://spark-api.xf-yun.com/%s/chat", apiVersion), apiKey, apiSecret)
+	switch apiVersion {
+	case "v3.1-128K":
+		authUrl = buildXunfeiAuthUrl(fmt.Sprintf("wss://spark-api.xf-yun.com/chat/pro-128k"), apiKey, apiSecret)
+		break
+	case "v3.5-32K":
+		authUrl = buildXunfeiAuthUrl(fmt.Sprintf("wss://spark-api.xf-yun.com/chat/max-32k"), apiKey, apiSecret)
+		break
+	default:
+		authUrl = buildXunfeiAuthUrl(fmt.Sprintf("wss://spark-api.xf-yun.com/%s/chat", apiVersion), apiKey, apiSecret)
+	}
 	return domain, authUrl
 }

relay/billing/ratio/model.go

@@ -128,7 +128,9 @@ var ModelRatio = map[string]float64{
 	"SparkDesk-v1.1":            1.2858, // ￥0.018 / 1k tokens
 	"SparkDesk-v2.1":            1.2858, // ￥0.018 / 1k tokens
 	"SparkDesk-v3.1":            1.2858, // ￥0.018 / 1k tokens
+	"SparkDesk-v3.1-128K":       1.2858, // ￥0.018 / 1k tokens
 	"SparkDesk-v3.5":            1.2858, // ￥0.018 / 1k tokens
+	"SparkDesk-v3.5-32K":        1.2858, // ￥0.018 / 1k tokens
 	"SparkDesk-v4.0":            1.2858, // ￥0.018 / 1k tokens
 	"360GPT_S2_V9":              0.8572, // ¥0.012 / 1k tokens
 	"embedding-bert-512-v1":     0.0715, // ¥0.001 / 1k tokens
@@ -160,15 +162,21 @@ var ModelRatio = map[string]float64{
 	"mistral-embed":         0.1 / 1000 * USD,
 	// https://wow.groq.com/#:~:text=inquiries%C2%A0here.-,Model,-Current%20Speed
 	"gemma-7b-it":                           0.07 / 1000000 * USD,
-	"mixtral-8x7b-32768":                    0.24 / 1000000 * USD,
-	"llama3-8b-8192":                        0.05 / 1000000 * USD,
-	"llama3-70b-8192":                       0.59 / 1000000 * USD,
 	"gemma2-9b-it":                          0.20 / 1000000 * USD,
-	"llama-3.1-405b-reasoning":              0.89 / 1000000 * USD,
 	"llama-3.1-70b-versatile":               0.59 / 1000000 * USD,
 	"llama-3.1-8b-instant":                  0.05 / 1000000 * USD,
+	"llama-3.2-11b-text-preview":            0.05 / 1000000 * USD,
+	"llama-3.2-11b-vision-preview":          0.05 / 1000000 * USD,
+	"llama-3.2-1b-preview":                  0.05 / 1000000 * USD,
+	"llama-3.2-3b-preview":                  0.05 / 1000000 * USD,
+	"llama-3.2-90b-text-preview":            0.59 / 1000000 * USD,
+	"llama-guard-3-8b":                      0.05 / 1000000 * USD,
+	"llama3-70b-8192":                       0.59 / 1000000 * USD,
+	"llama3-8b-8192":                        0.05 / 1000000 * USD,
 	"llama3-groq-70b-8192-tool-use-preview": 0.89 / 1000000 * USD,
 	"llama3-groq-8b-8192-tool-use-preview":  0.19 / 1000000 * USD,
+	"mixtral-8x7b-32768":                    0.24 / 1000000 * USD,
+
 	// https://platform.lingyiwanwu.com/docs#-计费单元
 	"yi-34b-chat-0205": 2.5 / 1000 * RMB,
 	"yi-34b-chat-200k": 12.0 / 1000 * RMB,

web/air/src/components/TokensTable.js

@@ -395,7 +395,7 @@ const TokensTable = () => {
         url = mjLink + `/#/?settings={"key":"sk-${key}","url":"${serverAddress}"}`;
         break;
       case 'lobechat':
-        url = chatLink + `/?settings={"keyVaults":{"openai":{"apiKey":"sk-${key}","baseURL":"${serverAddress}"/v1"}}}`;
+        url = chatLink + `/?settings={"keyVaults":{"openai":{"apiKey":"sk-${key}","baseURL":"${serverAddress}/v1"}}}`;
         break;
       default:
         if (!chatLink) {

web/air/src/pages/Channel/EditChannel.js

@@ -78,7 +78,7 @@ const EditChannel = (props) => {
                     localModels = ['chatglm_pro', 'chatglm_std', 'chatglm_lite'];
                     break;
                 case 18:
-                    localModels = ['SparkDesk', 'SparkDesk-v1.1', 'SparkDesk-v2.1', 'SparkDesk-v3.1', 'SparkDesk-v3.5', 'SparkDesk-v4.0'];
+                    localModels = ['SparkDesk', 'SparkDesk-v1.1', 'SparkDesk-v2.1', 'SparkDesk-v3.1', 'SparkDesk-v3.1-128K', 'SparkDesk-v3.5', 'SparkDesk-v3.5-32K', 'SparkDesk-v4.0'];
                     break;
                 case 19:
                     localModels = ['360GPT_S2_V9', 'embedding-bert-512-v1', 'embedding_s1_v1', 'semantic_similarity_s1_v1'];

web/berry/src/views/Channel/component/TableRow.js

@@ -268,6 +268,8 @@ function renderBalance(type, balance) {
       return <span>¥{balance.toFixed(2)}</span>;
     case 13: // AIGC2D
       return <span>{renderNumber(balance)}</span>;
+    case 44: // SiliconFlow
+      return <span>¥{balance.toFixed(2)}</span>;
     default:
       return <span>不支持</span>;
   }

web/berry/src/views/Channel/type/Config.js

@@ -91,7 +91,7 @@ const typeConfig = {
       other: '版本号'
     },
     input: {
-      models: ['SparkDesk', 'SparkDesk-v1.1', 'SparkDesk-v2.1', 'SparkDesk-v3.1', 'SparkDesk-v3.5', 'SparkDesk-v4.0']
+      models: ['SparkDesk', 'SparkDesk-v1.1', 'SparkDesk-v2.1', 'SparkDesk-v3.1', 'SparkDesk-v3.1-128K', 'SparkDesk-v3.5', 'SparkDesk-v3.5-32K', 'SparkDesk-v4.0']
     },
     prompt: {
       key: '按照如下格式输入：APPID|APISecret|APIKey',

web/berry/src/views/Token/component/TableRow.js

@@ -33,7 +33,7 @@ const COPY_OPTIONS = [
   },
   { key: 'ama', text: 'BotGem', url: 'ama://set-api-key?server={serverAddress}&key=sk-{key}', encode: true },
   { key: 'opencat', text: 'OpenCat', url: 'opencat://team/join?domain={serverAddress}&token=sk-{key}', encode: true },
-  { key: 'lobechat', text: 'LobeChat', url: 'https://lobehub.com/?settings={"keyVaults":{"openai":{"apiKey":"user-key","baseURL":"https://your-proxy.com/v1"}}}', encode: true }
+  { key: 'lobechat', text: 'LobeChat', url: 'https://lobehub.com/?settings={"keyVaults":{"openai":{"apiKey":"sk-{key}","baseURL":"{serverAddress}"}}}', encode: true }
 ];
 
 function replacePlaceholders(text, key, serverAddress) {

web/default/src/components/ChannelsTable.js

@@ -52,6 +52,8 @@ function renderBalance(type, balance) {
       return <span>¥{balance.toFixed(2)}</span>;
     case 13: // AIGC2D
       return <span>{renderNumber(balance)}</span>;
+    case 44: // SiliconFlow
+      return <span>¥{balance.toFixed(2)}</span>;
     default:
       return <span>不支持</span>;
   }

web/default/src/components/TokensTable.js

@@ -117,7 +117,7 @@ const TokensTable = () => {
         url = nextUrl;
         break;
       case 'lobechat':
-        url = nextLink + `/?settings={"keyVaults":{"openai":{"apiKey":"sk-${key}","baseURL":"${serverAddress}"/v1"}}}`;
+        url = nextLink + `/?settings={"keyVaults":{"openai":{"apiKey":"sk-${key}","baseURL":"${serverAddress}/v1"}}}`;
         break;
       default:
         url = `sk-${key}`;
@@ -160,7 +160,7 @@ const TokensTable = () => {
         break;
 
       case 'lobechat':
-        url = chatLink + `/?settings={"keyVaults":{"openai":{"apiKey":"sk-${key}","baseURL":"${serverAddress}"/v1"}}}`;
+        url = chatLink + `/?settings={"keyVaults":{"openai":{"apiKey":"sk-${key}","baseURL":"${serverAddress}/v1"}}}`;
         break;
 
       default: