feat: add user info to topup link

* 修正百度模型名称

* 更新百度模型名称，并保留旧版兼容以及修正单价

* chore: add more model and adjust order

---------

Co-authored-by: JustSong <songquanpeng@foxmail.com>

README.md

@@ -109,6 +109,7 @@ _✨ 通过标准的 OpenAI API 格式访问所有的大模型，开箱即用
     + 微信公众号授权（需要额外部署 [WeChat Server](https://github.com/songquanpeng/wechat-server)）。
 23. 支持主题切换，设置环境变量 `THEME` 即可，默认为 `default`，欢迎 PR 更多主题，具体参考[此处](./web/README.md)。
 24. 配合 [Message Pusher](https://github.com/songquanpeng/message-pusher) 可将报警信息推送到多种 App 上。
+25. 支持**扩展**，详情请参考此处 [API 文档](./docs/API.md)。
 
 ## 部署
 ### 基于 Docker 进行部署

common/model-ratio.go

@@ -72,14 +72,22 @@ var ModelRatio = map[string]float64{
 	"claude-3-sonnet-20240229": 3.0 / 1000 * USD,
 	"claude-3-opus-20240229":   15.0 / 1000 * USD,
 	// https://cloud.baidu.com/doc/WENXINWORKSHOP/s/hlrk4akp7
-	"ERNIE-Bot":       0.8572,     // ￥0.012 / 1k tokens
-	"ERNIE-Bot-turbo": 0.5715,     // ￥0.008 / 1k tokens
-	"ERNIE-Bot-4":     0.12 * RMB, // ￥0.12 / 1k tokens
-	"ERNIE-Bot-8K":    0.024 * RMB,
-	"Embedding-V1":    0.1429, // ￥0.002 / 1k tokens
-	"bge-large-zh":    0.002 * RMB,
-	"bge-large-en":    0.002 * RMB,
-	"bge-large-8k":    0.002 * RMB,
+	"ERNIE-4.0-8K":       0.120 * RMB,
+	"ERNIE-Bot-8K-0922":  0.024 * RMB,
+	"ERNIE-3.5-8K":       0.012 * RMB,
+	"ERNIE-Lite-8K-0922": 0.008 * RMB,
+	"ERNIE-Speed-8K":     0.004 * RMB,
+	"ERNIE-3.5-4K-0205":  0.012 * RMB,
+	"ERNIE-3.5-8K-0205":  0.024 * RMB,
+	"ERNIE-3.5-8K-1222":  0.012 * RMB,
+	"ERNIE-Lite-8K":      0.003 * RMB,
+	"ERNIE-Speed-128K":   0.004 * RMB,
+	"ERNIE-Tiny-8K":      0.001 * RMB,
+	"BLOOMZ-7B":          0.004 * RMB,
+	"Embedding-V1":       0.002 * RMB,
+	"bge-large-zh":       0.002 * RMB,
+	"bge-large-en":       0.002 * RMB,
+	"tao-8k":             0.002 * RMB,
 	// https://ai.google.dev/pricing
 	"PaLM-2":                    1,
 	"gemini-pro":                1, // $0.00025 / 1k characters -> $0.001 / 1k tokens
@@ -91,6 +99,7 @@ var ModelRatio = map[string]float64{
 	"glm-4":                     0.1 * RMB,
 	"glm-4v":                    0.1 * RMB,
 	"glm-3-turbo":               0.005 * RMB,
+	"embedding-2":               0.0005 * RMB,
 	"chatglm_turbo":             0.3572, // ￥0.005 / 1k tokens
 	"chatglm_pro":               0.7143, // ￥0.01 / 1k tokens
 	"chatglm_std":               0.3572, // ￥0.005 / 1k tokens

controller/model.go

@@ -4,12 +4,14 @@ import (
 	"fmt"
 	"github.com/gin-gonic/gin"
 	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/model"
 	"github.com/songquanpeng/one-api/relay/channel/openai"
 	"github.com/songquanpeng/one-api/relay/constant"
 	"github.com/songquanpeng/one-api/relay/helper"
 	relaymodel "github.com/songquanpeng/one-api/relay/model"
 	"github.com/songquanpeng/one-api/relay/util"
 	"net/http"
+	"strings"
 )
 
 // https://platform.openai.com/docs/api-reference/models/list
@@ -120,9 +122,41 @@ func DashboardListModels(c *gin.Context) {
 }
 
 func ListModels(c *gin.Context) {
+	ctx := c.Request.Context()
+	var availableModels []string
+	if c.GetString("available_models") != "" {
+		availableModels = strings.Split(c.GetString("available_models"), ",")
+	} else {
+		userId := c.GetInt("id")
+		userGroup, _ := model.CacheGetUserGroup(userId)
+		availableModels, _ = model.CacheGetGroupModels(ctx, userGroup)
+	}
+	modelSet := make(map[string]bool)
+	for _, availableModel := range availableModels {
+		modelSet[availableModel] = true
+	}
+	availableOpenAIModels := make([]OpenAIModels, 0)
+	for _, model := range openAIModels {
+		if _, ok := modelSet[model.Id]; ok {
+			modelSet[model.Id] = false
+			availableOpenAIModels = append(availableOpenAIModels, model)
+		}
+	}
+	for modelName, ok := range modelSet {
+		if ok {
+			availableOpenAIModels = append(availableOpenAIModels, OpenAIModels{
+				Id:      modelName,
+				Object:  "model",
+				Created: 1626777600,
+				OwnedBy: "custom",
+				Root:    modelName,
+				Parent:  nil,
+			})
+		}
+	}
 	c.JSON(200, gin.H{
 		"object": "list",
-		"data":   openAIModels,
+		"data":   availableOpenAIModels,
 	})
 }
 
@@ -142,3 +176,30 @@ func RetrieveModel(c *gin.Context) {
 		})
 	}
 }
+
+func GetUserAvailableModels(c *gin.Context) {
+	ctx := c.Request.Context()
+	id := c.GetInt("id")
+	userGroup, err := model.CacheGetUserGroup(id)
+	if err != nil {
+		c.JSON(http.StatusOK, gin.H{
+			"success": false,
+			"message": err.Error(),
+		})
+		return
+	}
+	models, err := model.CacheGetGroupModels(ctx, userGroup)
+	if err != nil {
+		c.JSON(http.StatusOK, gin.H{
+			"success": false,
+			"message": err.Error(),
+		})
+		return
+	}
+	c.JSON(http.StatusOK, gin.H{
+		"success": true,
+		"message": "",
+		"data":    models,
+	})
+	return
+}

controller/token.go

@@ -130,6 +130,7 @@ func AddToken(c *gin.Context) {
 		ExpiredTime:    token.ExpiredTime,
 		RemainQuota:    token.RemainQuota,
 		UnlimitedQuota: token.UnlimitedQuota,
+		Models:         token.Models,
 	}
 	err = cleanToken.Insert()
 	if err != nil {
@@ -216,6 +217,7 @@ func UpdateToken(c *gin.Context) {
 		cleanToken.ExpiredTime = token.ExpiredTime
 		cleanToken.RemainQuota = token.RemainQuota
 		cleanToken.UnlimitedQuota = token.UnlimitedQuota
+		cleanToken.Models = token.Models
 	}
 	err = cleanToken.Update()
 	if err != nil {

controller/user.go

@@ -180,27 +180,27 @@ func Register(c *gin.Context) {
 }
 
 func GetAllUsers(c *gin.Context) {
-    p, _ := strconv.Atoi(c.Query("p"))
-    if p < 0 {
-        p = 0
-    }
-    
-    order := c.DefaultQuery("order", "")
-    users, err := model.GetAllUsers(p*config.ItemsPerPage, config.ItemsPerPage, order)
-	
-    if err != nil {
-        c.JSON(http.StatusOK, gin.H{
-            "success": false,
-            "message": err.Error(),
-        })
-        return
-    }
-    
-    c.JSON(http.StatusOK, gin.H{
-        "success": true,
-        "message": "",
-        "data":    users,
-    })
+	p, _ := strconv.Atoi(c.Query("p"))
+	if p < 0 {
+		p = 0
+	}
+
+	order := c.DefaultQuery("order", "")
+	users, err := model.GetAllUsers(p*config.ItemsPerPage, config.ItemsPerPage, order)
+
+	if err != nil {
+		c.JSON(http.StatusOK, gin.H{
+			"success": false,
+			"message": err.Error(),
+		})
+		return
+	}
+
+	c.JSON(http.StatusOK, gin.H{
+		"success": true,
+		"message": "",
+		"data":    users,
+	})
 }
 
 func SearchUsers(c *gin.Context) {
@@ -770,3 +770,38 @@ func TopUp(c *gin.Context) {
 	})
 	return
 }
+
+type adminTopUpRequest struct {
+	UserId int    `json:"user_id"`
+	Quota  int    `json:"quota"`
+	Remark string `json:"remark"`
+}
+
+func AdminTopUp(c *gin.Context) {
+	req := adminTopUpRequest{}
+	err := c.ShouldBindJSON(&req)
+	if err != nil {
+		c.JSON(http.StatusOK, gin.H{
+			"success": false,
+			"message": err.Error(),
+		})
+		return
+	}
+	err = model.IncreaseUserQuota(req.UserId, int64(req.Quota))
+	if err != nil {
+		c.JSON(http.StatusOK, gin.H{
+			"success": false,
+			"message": err.Error(),
+		})
+		return
+	}
+	if req.Remark == "" {
+		req.Remark = fmt.Sprintf("通过 API 充值 %s", common.LogQuota(int64(req.Quota)))
+	}
+	model.RecordTopupLog(req.UserId, req.Remark, req.Quota)
+	c.JSON(http.StatusOK, gin.H{
+		"success": true,
+		"message": "",
+	})
+	return
+}

docs/API.md

@@ -0,0 +1,44 @@
+# 使用 API 操控 & 扩展 One API
+> 欢迎提交 PR 在此放上你的拓展项目。
+
+例如，虽然 One API 本身没有直接支持支付，但是你可以通过系统扩展的 API 来实现支付功能。
+
+又或者你想自定义渠道管理策略，也可以通过 API 来实现渠道的禁用与启用。
+
+## 鉴权
+One API 支持两种鉴权方式：Cookie 和 Token，对于 Token，参照下图获取：
+
+![image](https://github.com/songquanpeng/songquanpeng.github.io/assets/39998050/c15281a7-83ed-47cb-a1f6-913cb6bf4a7c)
+
+之后，将 Token 作为请求头的 Authorization 字段的值即可，例如下面使用 Token 调用测试渠道的 API：
+![image](https://github.com/songquanpeng/songquanpeng.github.io/assets/39998050/1273b7ae-cb60-4c0d-93a6-b1cbc039c4f8)
+
+## 请求格式与响应格式
+One API 使用 JSON 格式进行请求和响应。
+
+对于响应体，一般格式如下：
+```json
+{
+  "message": "请求信息",
+  "success": true,
+  "data": {}
+}
+```
+
+## API 列表
+> 当前 API 列表不全，请自行通过浏览器抓取前端请求
+
+如果现有的 API 没有办法满足你的需求，欢迎提交 issue 讨论。
+
+### 获取当前登录用户信息
+**GET** `/api/user/self`
+
+### 为给定用户充值额度
+**POST** `/api/topup`
+```json
+{
+  "user_id": 1,
+  "quota": 100000,
+  "remark": "充值 100000 额度"
+}
+```

middleware/auth.go

@@ -1,6 +1,7 @@
 package middleware
 
 import (
+	"fmt"
 	"github.com/gin-contrib/sessions"
 	"github.com/gin-gonic/gin"
 	"github.com/songquanpeng/one-api/common"
@@ -107,6 +108,19 @@ func TokenAuth() func(c *gin.Context) {
 			abortWithMessage(c, http.StatusForbidden, "用户已被封禁")
 			return
 		}
+		requestModel, err := getRequestModel(c)
+		if err != nil {
+			abortWithMessage(c, http.StatusBadRequest, err.Error())
+			return
+		}
+		c.Set("request_model", requestModel)
+		if token.Models != nil && *token.Models != "" {
+			c.Set("available_models", *token.Models)
+			if requestModel != "" && !isModelInList(requestModel, *token.Models) {
+				abortWithMessage(c, http.StatusForbidden, fmt.Sprintf("该令牌无权使用模型：%s", requestModel))
+				return
+			}
+		}
 		c.Set("id", token.UserId)
 		c.Set("token_id", token.Id)
 		c.Set("token_name", token.Name)

middleware/distributor.go

@@ -2,14 +2,12 @@ package middleware
 
 import (
 	"fmt"
+	"github.com/gin-gonic/gin"
 	"github.com/songquanpeng/one-api/common"
 	"github.com/songquanpeng/one-api/common/logger"
 	"github.com/songquanpeng/one-api/model"
 	"net/http"
 	"strconv"
-	"strings"
-
-	"github.com/gin-gonic/gin"
 )
 
 type ModelRequest struct {
@@ -40,37 +38,11 @@ func Distribute() func(c *gin.Context) {
 				return
 			}
 		} else {
-			// Select a channel for the user
-			var modelRequest ModelRequest
-			err := common.UnmarshalBodyReusable(c, &modelRequest)
+			requestModel := c.GetString("request_model")
+			var err error
+			channel, err = model.CacheGetRandomSatisfiedChannel(userGroup, requestModel, false)
 			if err != nil {
-				abortWithMessage(c, http.StatusBadRequest, "无效的请求")
-				return
-			}
-			if strings.HasPrefix(c.Request.URL.Path, "/v1/moderations") {
-				if modelRequest.Model == "" {
-					modelRequest.Model = "text-moderation-stable"
-				}
-			}
-			if strings.HasSuffix(c.Request.URL.Path, "embeddings") {
-				if modelRequest.Model == "" {
-					modelRequest.Model = c.Param("model")
-				}
-			}
-			if strings.HasPrefix(c.Request.URL.Path, "/v1/images/generations") {
-				if modelRequest.Model == "" {
-					modelRequest.Model = "dall-e-2"
-				}
-			}
-			if strings.HasPrefix(c.Request.URL.Path, "/v1/audio/transcriptions") || strings.HasPrefix(c.Request.URL.Path, "/v1/audio/translations") {
-				if modelRequest.Model == "" {
-					modelRequest.Model = "whisper-1"
-				}
-			}
-			requestModel = modelRequest.Model
-			channel, err = model.CacheGetRandomSatisfiedChannel(userGroup, modelRequest.Model, false)
-			if err != nil {
-				message := fmt.Sprintf("当前分组 %s 下对于模型 %s 无可用渠道", userGroup, modelRequest.Model)
+				message := fmt.Sprintf("当前分组 %s 下对于模型 %s 无可用渠道", userGroup, requestModel)
 				if channel != nil {
 					logger.SysError(fmt.Sprintf("渠道不存在：%d", channel.Id))
 					message = "数据库一致性已被破坏，请联系管理员"

middleware/utils.go

@@ -1,9 +1,12 @@
 package middleware
 
 import (
+	"fmt"
 	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common"
 	"github.com/songquanpeng/one-api/common/helper"
 	"github.com/songquanpeng/one-api/common/logger"
+	"strings"
 )
 
 func abortWithMessage(c *gin.Context, statusCode int, message string) {
@@ -16,3 +19,42 @@ func abortWithMessage(c *gin.Context, statusCode int, message string) {
 	c.Abort()
 	logger.Error(c.Request.Context(), message)
 }
+
+func getRequestModel(c *gin.Context) (string, error) {
+	var modelRequest ModelRequest
+	err := common.UnmarshalBodyReusable(c, &modelRequest)
+	if err != nil {
+		return "", fmt.Errorf("common.UnmarshalBodyReusable failed: %w", err)
+	}
+	if strings.HasPrefix(c.Request.URL.Path, "/v1/moderations") {
+		if modelRequest.Model == "" {
+			modelRequest.Model = "text-moderation-stable"
+		}
+	}
+	if strings.HasSuffix(c.Request.URL.Path, "embeddings") {
+		if modelRequest.Model == "" {
+			modelRequest.Model = c.Param("model")
+		}
+	}
+	if strings.HasPrefix(c.Request.URL.Path, "/v1/images/generations") {
+		if modelRequest.Model == "" {
+			modelRequest.Model = "dall-e-2"
+		}
+	}
+	if strings.HasPrefix(c.Request.URL.Path, "/v1/audio/transcriptions") || strings.HasPrefix(c.Request.URL.Path, "/v1/audio/translations") {
+		if modelRequest.Model == "" {
+			modelRequest.Model = "whisper-1"
+		}
+	}
+	return modelRequest.Model, nil
+}
+
+func isModelInList(modelName string, models string) bool {
+	modelList := strings.Split(models, ",")
+	for _, model := range modelList {
+		if modelName == model {
+			return true
+		}
+	}
+	return false
+}

model/ability.go

@@ -1,8 +1,10 @@
 package model
 
 import (
+	"context"
 	"github.com/songquanpeng/one-api/common"
 	"gorm.io/gorm"
+	"sort"
 	"strings"
 )
 
@@ -88,3 +90,19 @@ func (channel *Channel) UpdateAbilities() error {
 func UpdateAbilityStatus(channelId int, status bool) error {
 	return DB.Model(&Ability{}).Where("channel_id = ?", channelId).Select("enabled").Update("enabled", status).Error
 }
+
+func GetGroupModels(ctx context.Context, group string) ([]string, error) {
+	groupCol := "`group`"
+	trueVal := "1"
+	if common.UsingPostgreSQL {
+		groupCol = `"group"`
+		trueVal = "true"
+	}
+	var models []string
+	err := DB.Model(&Ability{}).Distinct("model").Where(groupCol+" = ? and enabled = "+trueVal, group).Pluck("model", &models).Error
+	if err != nil {
+		return nil, err
+	}
+	sort.Strings(models)
+	return models, err
+}

model/cache.go

@@ -21,6 +21,7 @@ var (
 	UserId2GroupCacheSeconds  = config.SyncFrequency
 	UserId2QuotaCacheSeconds  = config.SyncFrequency
 	UserId2StatusCacheSeconds = config.SyncFrequency
+	GroupModelsCacheSeconds   = config.SyncFrequency
 )
 
 func CacheGetTokenByKey(key string) (*Token, error) {
@@ -146,6 +147,25 @@ func CacheIsUserEnabled(userId int) (bool, error) {
 	return userEnabled, err
 }
 
+func CacheGetGroupModels(ctx context.Context, group string) ([]string, error) {
+	if !common.RedisEnabled {
+		return GetGroupModels(ctx, group)
+	}
+	modelsStr, err := common.RedisGet(fmt.Sprintf("group_models:%s", group))
+	if err == nil {
+		return strings.Split(modelsStr, ","), nil
+	}
+	models, err := GetGroupModels(ctx, group)
+	if err != nil {
+		return nil, err
+	}
+	err = common.RedisSet(fmt.Sprintf("group_models:%s", group), strings.Join(models, ","), time.Duration(GroupModelsCacheSeconds)*time.Second)
+	if err != nil {
+		logger.SysError("Redis set group models error: " + err.Error())
+	}
+	return models, nil
+}
+
 var group2model2channels map[string]map[string][]*Channel
 var channelSyncLock sync.RWMutex
 

model/log.go

@@ -51,6 +51,21 @@ func RecordLog(userId int, logType int, content string) {
 	}
 }
 
+func RecordTopupLog(userId int, content string, quota int) {
+	log := &Log{
+		UserId:    userId,
+		Username:  GetUsernameById(userId),
+		CreatedAt: helper.GetTimestamp(),
+		Type:      LogTypeTopup,
+		Content:   content,
+		Quota:     quota,
+	}
+	err := LOG_DB.Create(log).Error
+	if err != nil {
+		logger.SysError("failed to record log: " + err.Error())
+	}
+}
+
 func RecordConsumeLog(ctx context.Context, userId int, channelId int, promptTokens int, completionTokens int, modelName string, tokenName string, quota int64, content string) {
 	logger.Info(ctx, fmt.Sprintf("record consume log: userId=%d, channelId=%d, promptTokens=%d, completionTokens=%d, modelName=%s, tokenName=%s, quota=%d, content=%s", userId, channelId, promptTokens, completionTokens, modelName, tokenName, quota, content))
 	if !config.LogConsumeEnabled {

model/token.go

@@ -12,24 +12,25 @@ import (
 )
 
 type Token struct {
-	Id             int    `json:"id"`
-	UserId         int    `json:"user_id"`
-	Key            string `json:"key" gorm:"type:char(48);uniqueIndex"`
-	Status         int    `json:"status" gorm:"default:1"`
-	Name           string `json:"name" gorm:"index" `
-	CreatedTime    int64  `json:"created_time" gorm:"bigint"`
-	AccessedTime   int64  `json:"accessed_time" gorm:"bigint"`
-	ExpiredTime    int64  `json:"expired_time" gorm:"bigint;default:-1"` // -1 means never expired
-	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
+	Id             int     `json:"id"`
+	UserId         int     `json:"user_id"`
+	Key            string  `json:"key" gorm:"type:char(48);uniqueIndex"`
+	Status         int     `json:"status" gorm:"default:1"`
+	Name           string  `json:"name" gorm:"index" `
+	CreatedTime    int64   `json:"created_time" gorm:"bigint"`
+	AccessedTime   int64   `json:"accessed_time" gorm:"bigint"`
+	ExpiredTime    int64   `json:"expired_time" gorm:"bigint;default:-1"` // -1 means never expired
+	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:''"`
 }
 
 func GetAllUserTokens(userId int, startIdx int, num int, order string) ([]*Token, error) {
 	var tokens []*Token
 	var err error
 	query := DB.Where("user_id = ?", userId)
-	
+
 	switch order {
 	case "remain_quota":
 		query = query.Order("unlimited_quota desc, remain_quota desc")
@@ -38,7 +39,7 @@ func GetAllUserTokens(userId int, startIdx int, num int, order string) ([]*Token
 	default:
 		query = query.Order("id desc")
 	}
-	
+
 	err = query.Limit(num).Offset(startIdx).Find(&tokens).Error
 	return tokens, err
 }
@@ -121,7 +122,7 @@ func (token *Token) Insert() error {
 // Update Make sure your token's fields is completed, because this will update non-zero values
 func (token *Token) Update() error {
 	var err error
-	err = DB.Model(token).Select("name", "status", "expired_time", "remain_quota", "unlimited_quota").Updates(token).Error
+	err = DB.Model(token).Select("name", "status", "expired_time", "remain_quota", "unlimited_quota", "models").Updates(token).Error
 	return err
 }
 

relay/channel/ali/main.go

@@ -50,8 +50,8 @@ func ConvertRequest(request model.GeneralOpenAIRequest) *ChatRequest {
 			TopP:              request.TopP,
 			TopK:              request.TopK,
 			ResultFormat:      "message",
+			Tools:             request.Tools,
 		},
-		Tools: request.Tools,
 	}
 }
 

relay/channel/ali/model.go

@@ -16,21 +16,21 @@ type Input struct {
 }
 
 type Parameters struct {
-	TopP              float64 `json:"top_p,omitempty"`
-	TopK              int     `json:"top_k,omitempty"`
-	Seed              uint64  `json:"seed,omitempty"`
-	EnableSearch      bool    `json:"enable_search,omitempty"`
-	IncrementalOutput bool    `json:"incremental_output,omitempty"`
-	MaxTokens         int     `json:"max_tokens,omitempty"`
-	Temperature       float64 `json:"temperature,omitempty"`
-	ResultFormat      string  `json:"result_format,omitempty"`
+	TopP              float64      `json:"top_p,omitempty"`
+	TopK              int          `json:"top_k,omitempty"`
+	Seed              uint64       `json:"seed,omitempty"`
+	EnableSearch      bool         `json:"enable_search,omitempty"`
+	IncrementalOutput bool         `json:"incremental_output,omitempty"`
+	MaxTokens         int          `json:"max_tokens,omitempty"`
+	Temperature       float64      `json:"temperature,omitempty"`
+	ResultFormat      string       `json:"result_format,omitempty"`
+	Tools             []model.Tool `json:"tools,omitempty"`
 }
 
 type ChatRequest struct {
-	Model      string       `json:"model"`
-	Input      Input        `json:"input"`
-	Parameters Parameters   `json:"parameters,omitempty"`
-	Tools      []model.Tool `json:"tools,omitempty"`
+	Model      string     `json:"model"`
+	Input      Input      `json:"input"`
+	Parameters Parameters `json:"parameters,omitempty"`
 }
 
 type EmbeddingRequest struct {

relay/channel/baidu/adaptor.go

@@ -38,16 +38,26 @@ func (a *Adaptor) GetRequestURL(meta *util.RelayMeta) (string, error) {
 		suffix += "completions_pro"
 	case "ERNIE-Bot-4":
 		suffix += "completions_pro"
-	case "ERNIE-3.5-8K":
-		suffix += "completions"
-	case "ERNIE-Bot-8K":
-		suffix += "ernie_bot_8k"
 	case "ERNIE-Bot":
 		suffix += "completions"
-	case "ERNIE-Speed":
-		suffix += "ernie_speed"
 	case "ERNIE-Bot-turbo":
 		suffix += "eb-instant"
+	case "ERNIE-Speed":
+		suffix += "ernie_speed"
+	case "ERNIE-Bot-8K":
+		suffix += "ernie_bot_8k"
+	case "ERNIE-4.0-8K":
+		suffix += "completions_pro"
+	case "ERNIE-3.5-8K":
+		suffix += "completions"
+	case "ERNIE-Speed-8K":
+		suffix += "ernie_speed"
+	case "ERNIE-Speed-128K":
+		suffix += "ernie-speed-128k"
+	case "ERNIE-Lite-8K":
+		suffix += "ernie-lite-8k"
+	case "ERNIE-Tiny-8K":
+		suffix += "ernie-tiny-8k"
 	case "BLOOMZ-7B":
 		suffix += "bloomz_7b1"
 	case "Embedding-V1":
@@ -59,7 +69,7 @@ func (a *Adaptor) GetRequestURL(meta *util.RelayMeta) (string, error) {
 	case "tao-8k":
 		suffix += "tao_8k"
 	default:
-		suffix += meta.ActualModelName
+		suffix += strings.ToLower(meta.ActualModelName)
 	}
 	fullRequestURL := fmt.Sprintf("%s/rpc/2.0/ai_custom/v1/wenxinworkshop/%s", meta.BaseURL, suffix)
 	var accessToken string

relay/channel/baidu/constants.go

@@ -1,11 +1,18 @@
 package baidu
 
 var ModelList = []string{
-	"ERNIE-Bot-4",
-	"ERNIE-Bot-8K",
-	"ERNIE-Bot",
-	"ERNIE-Speed",
-	"ERNIE-Bot-turbo",
+	"ERNIE-4.0-8K",
+	"ERNIE-Bot-8K-0922",
+	"ERNIE-3.5-8K",
+	"ERNIE-Lite-8K-0922",
+	"ERNIE-Speed-8K",
+	"ERNIE-3.5-4K-0205",
+	"ERNIE-3.5-8K-0205",
+	"ERNIE-3.5-8K-1222",
+	"ERNIE-Lite-8K",
+	"ERNIE-Speed-128K",
+	"ERNIE-Tiny-8K",
+	"BLOOMZ-7B",
 	"Embedding-V1",
 	"bge-large-zh",
 	"bge-large-en",

relay/channel/xunfei/model.go

@@ -28,7 +28,7 @@ type ChatRequest struct {
 		} `json:"message"`
 		Functions struct {
 			Text []model.Function `json:"text,omitempty"`
-		} `json:"functions"`
+		} `json:"functions,omitempty"`
 	} `json:"payload"`
 }
 

relay/channel/zhipu/adaptor.go

@@ -6,6 +6,7 @@ import (
 	"github.com/gin-gonic/gin"
 	"github.com/songquanpeng/one-api/relay/channel"
 	"github.com/songquanpeng/one-api/relay/channel/openai"
+	"github.com/songquanpeng/one-api/relay/constant"
 	"github.com/songquanpeng/one-api/relay/model"
 	"github.com/songquanpeng/one-api/relay/util"
 	"io"
@@ -35,6 +36,9 @@ func (a *Adaptor) GetRequestURL(meta *util.RelayMeta) (string, error) {
 	if a.APIVersion == "v4" {
 		return fmt.Sprintf("%s/api/paas/v4/chat/completions", meta.BaseURL), nil
 	}
+	if meta.Mode == constant.RelayModeEmbeddings {
+		return fmt.Sprintf("%s/api/paas/v4/embeddings", meta.BaseURL), nil
+	}
 	method := "invoke"
 	if meta.IsStream {
 		method = "sse-invoke"
@@ -53,18 +57,24 @@ func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.G
 	if request == nil {
 		return nil, errors.New("request is nil")
 	}
-	// TopP (0.0, 1.0)
-	request.TopP = math.Min(0.99, request.TopP)
-	request.TopP = math.Max(0.01, request.TopP)
+	switch relayMode {
+	case constant.RelayModeEmbeddings:
+		baiduEmbeddingRequest := ConvertEmbeddingRequest(*request)
+		return baiduEmbeddingRequest, nil
+	default:
+		// TopP (0.0, 1.0)
+		request.TopP = math.Min(0.99, request.TopP)
+		request.TopP = math.Max(0.01, request.TopP)
 
-	// Temperature (0.0, 1.0)
-	request.Temperature = math.Min(0.99, request.Temperature)
-	request.Temperature = math.Max(0.01, request.Temperature)
-	a.SetVersionByModeName(request.Model)
-	if a.APIVersion == "v4" {
-		return request, nil
+		// Temperature (0.0, 1.0)
+		request.Temperature = math.Min(0.99, request.Temperature)
+		request.Temperature = math.Max(0.01, request.Temperature)
+		a.SetVersionByModeName(request.Model)
+		if a.APIVersion == "v4" {
+			return request, nil
+		}
+		return ConvertRequest(*request), nil
 	}
-	return ConvertRequest(*request), nil
 }
 
 func (a *Adaptor) DoRequest(c *gin.Context, meta *util.RelayMeta, requestBody io.Reader) (*http.Response, error) {
@@ -84,14 +94,26 @@ func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *util.Rel
 	if a.APIVersion == "v4" {
 		return a.DoResponseV4(c, resp, meta)
 	}
+
 	if meta.IsStream {
 		err, usage = StreamHandler(c, resp)
 	} else {
-		err, usage = Handler(c, resp)
+		if meta.Mode == constant.RelayModeEmbeddings {
+			err, usage = EmbeddingsHandler(c, resp)
+		} else {
+			err, usage = Handler(c, resp)
+		}
 	}
 	return
 }
 
+func ConvertEmbeddingRequest(request model.GeneralOpenAIRequest) *EmbeddingRequest {
+	return &EmbeddingRequest{
+		Model: "embedding-2",
+		Input: request.Input.(string),
+	}
+}
+
 func (a *Adaptor) GetModelList() []string {
 	return ModelList
 }

relay/channel/zhipu/constants.go

@@ -2,5 +2,5 @@ package zhipu
 
 var ModelList = []string{
 	"chatglm_turbo", "chatglm_pro", "chatglm_std", "chatglm_lite",
-	"glm-4", "glm-4v", "glm-3-turbo",
+	"glm-4", "glm-4v", "glm-3-turbo", "embedding-2",
 }

relay/channel/zhipu/main.go

@@ -254,3 +254,50 @@ func Handler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, *
 	_, err = c.Writer.Write(jsonResponse)
 	return nil, &fullTextResponse.Usage
 }
+
+func EmbeddingsHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, *model.Usage) {
+	var zhipuResponse EmbeddingRespone
+	responseBody, err := io.ReadAll(resp.Body)
+	if err != nil {
+		return openai.ErrorWrapper(err, "read_response_body_failed", http.StatusInternalServerError), nil
+	}
+	err = resp.Body.Close()
+	if err != nil {
+		return openai.ErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), nil
+	}
+	err = json.Unmarshal(responseBody, &zhipuResponse)
+	if err != nil {
+		return openai.ErrorWrapper(err, "unmarshal_response_body_failed", http.StatusInternalServerError), nil
+	}
+	fullTextResponse := embeddingResponseZhipu2OpenAI(&zhipuResponse)
+	jsonResponse, err := json.Marshal(fullTextResponse)
+	if err != nil {
+		return openai.ErrorWrapper(err, "marshal_response_body_failed", http.StatusInternalServerError), nil
+	}
+	c.Writer.Header().Set("Content-Type", "application/json")
+	c.Writer.WriteHeader(resp.StatusCode)
+	_, err = c.Writer.Write(jsonResponse)
+	return nil, &fullTextResponse.Usage
+}
+
+func embeddingResponseZhipu2OpenAI(response *EmbeddingRespone) *openai.EmbeddingResponse {
+	openAIEmbeddingResponse := openai.EmbeddingResponse{
+		Object: "list",
+		Data:   make([]openai.EmbeddingResponseItem, 0, len(response.Embeddings)),
+		Model:  response.Model,
+		Usage: model.Usage{
+			PromptTokens:     response.PromptTokens,
+			CompletionTokens: response.CompletionTokens,
+			TotalTokens:      response.Usage.TotalTokens,
+		},
+	}
+
+	for _, item := range response.Embeddings {
+		openAIEmbeddingResponse.Data = append(openAIEmbeddingResponse.Data, openai.EmbeddingResponseItem{
+			Object:    `embedding`,
+			Index:     item.Index,
+			Embedding: item.Embedding,
+		})
+	}
+	return &openAIEmbeddingResponse
+}

relay/channel/zhipu/model.go

@@ -44,3 +44,21 @@ type tokenData struct {
 	Token      string
 	ExpiryTime time.Time
 }
+
+type EmbeddingRequest struct {
+	Model string `json:"model"`
+	Input string `json:"input"`
+}
+
+type EmbeddingRespone struct {
+	Model       string          `json:"model"`
+	Object      string          `json:"object"`
+	Embeddings  []EmbeddingData `json:"data"`
+	model.Usage `json:"usage"`
+}
+
+type EmbeddingData struct {
+	Index     int       `json:"index"`
+	Object    string    `json:"object"`
+	Embedding []float64 `json:"embedding"`
+}

relay/util/common.go

@@ -46,6 +46,15 @@ func ShouldDisableChannel(err *relaymodel.Error, statusCode int) bool {
 	} else if strings.HasPrefix(err.Message, "This organization has been disabled.") {
 		return true
 	}
+	if strings.Contains(err.Message, "quota") {
+		return true
+	}
+	if strings.Contains(err.Message, "credit") {
+		return true
+	}
+	if strings.Contains(err.Message, "balance") {
+		return true
+	}
 	return false
 }
 

router/api-router.go

@@ -26,6 +26,7 @@ func SetApiRouter(router *gin.Engine) {
 		apiRouter.GET("/oauth/wechat", middleware.CriticalRateLimit(), controller.WeChatAuth)
 		apiRouter.GET("/oauth/wechat/bind", middleware.CriticalRateLimit(), middleware.UserAuth(), controller.WeChatBind)
 		apiRouter.GET("/oauth/email/bind", middleware.CriticalRateLimit(), middleware.UserAuth(), controller.EmailBind)
+		apiRouter.POST("/topup", middleware.AdminAuth(), controller.AdminTopUp)
 
 		userRoute := apiRouter.Group("/user")
 		{
@@ -43,6 +44,7 @@ func SetApiRouter(router *gin.Engine) {
 				selfRoute.GET("/token", controller.GenerateAccessToken)
 				selfRoute.GET("/aff", controller.GetAffCode)
 				selfRoute.POST("/topup", controller.TopUp)
+				selfRoute.GET("/available_models", controller.GetUserAvailableModels)
 			}
 
 			adminRoute := userRoute.Group("/")

web/default/src/pages/Token/EditToken.js

@@ -1,19 +1,21 @@
 import React, { useEffect, useState } from 'react';
 import { Button, Form, Header, Message, Segment } from 'semantic-ui-react';
-import { useParams, useNavigate } from 'react-router-dom';
-import { API, showError, showSuccess, timestamp2string } from '../../helpers';
-import { renderQuota, renderQuotaWithPrompt } from '../../helpers/render';
+import { useNavigate, useParams } from 'react-router-dom';
+import { API, copy, showError, showSuccess, timestamp2string } from '../../helpers';
+import { renderQuotaWithPrompt } from '../../helpers/render';
 
 const EditToken = () => {
   const params = useParams();
   const tokenId = params.id;
   const isEdit = tokenId !== undefined;
   const [loading, setLoading] = useState(isEdit);
+  const [modelOptions, setModelOptions] = useState([]);
   const originInputs = {
     name: '',
     remain_quota: isEdit ? 0 : 500000,
     expired_time: -1,
-    unlimited_quota: false
+    unlimited_quota: false,
+    models: []
   };
   const [inputs, setInputs] = useState(originInputs);
   const { name, remain_quota, expired_time, unlimited_quota } = inputs;
@@ -22,8 +24,8 @@ const EditToken = () => {
     setInputs((inputs) => ({ ...inputs, [name]: value }));
   };
   const handleCancel = () => {
-    navigate("/token");
-  }
+    navigate('/token');
+  };
   const setExpiredTime = (month, day, hour, minute) => {
     let now = new Date();
     let timestamp = now.getTime() / 1000;
@@ -50,6 +52,11 @@ const EditToken = () => {
       if (data.expired_time !== -1) {
         data.expired_time = timestamp2string(data.expired_time);
       }
+      if (data.models === '') {
+        data.models = [];
+      } else {
+        data.models = data.models.split(',');
+      }
       setInputs(data);
     } else {
       showError(message);
@@ -60,8 +67,26 @@ const EditToken = () => {
     if (isEdit) {
       loadToken().then();
     }
+    loadAvailableModels().then();
   }, []);
 
+  const loadAvailableModels = async () => {
+    let res = await API.get(`/api/user/available_models`);
+    const { success, message, data } = res.data;
+    if (success) {
+      let options = data.map((model) => {
+        return {
+          key: model,
+          text: model,
+          value: model
+        };
+      });
+      setModelOptions(options);
+    } else {
+      showError(message);
+    }
+  };
+
   const submit = async () => {
     if (!isEdit && inputs.name === '') return;
     let localInputs = inputs;
@@ -74,6 +99,7 @@ const EditToken = () => {
       }
       localInputs.expired_time = Math.ceil(time / 1000);
     }
+    localInputs.models = localInputs.models.join(',');
     let res;
     if (isEdit) {
       res = await API.put(`/api/token/`, { ...localInputs, id: parseInt(tokenId) });
@@ -109,6 +135,24 @@ const EditToken = () => {
               required={!isEdit}
             />
           </Form.Field>
+          <Form.Field>
+            <Form.Dropdown
+              label='模型范围'
+              placeholder={'请选择允许使用的模型，留空则不进行限制'}
+              name='models'
+              fluid
+              multiple
+              search
+              onLabelClick={(e, { value }) => {
+                copy(value).then();
+              }}
+              selection
+              onChange={handleInputChange}
+              value={inputs.models}
+              autoComplete='new-password'
+              options={modelOptions}
+            />
+          </Form.Field>
           <Form.Field>
             <Form.Input
               label='过期时间'

web/default/src/pages/TopUp/index.js

@@ -8,6 +8,7 @@ const TopUp = () => {
   const [topUpLink, setTopUpLink] = useState('');
   const [userQuota, setUserQuota] = useState(0);
   const [isSubmitting, setIsSubmitting] = useState(false);
+  const [user, setUser] = useState({});
 
   const topUp = async () => {
     if (redemptionCode === '') {
@@ -41,7 +42,14 @@ const TopUp = () => {
       showError('超级管理员未设置充值链接！');
       return;
     }
-    window.open(topUpLink, '_blank');
+    let url = new URL(topUpLink);
+    let username = user.username;
+    let user_id = user.id;
+    // add  username and user_id to the topup link
+    url.searchParams.append('username', username);
+    url.searchParams.append('user_id', user_id);
+    url.searchParams.append('transaction_id', crypto.randomUUID());
+    window.open(url.toString(), '_blank');
   };
 
   const getUserQuota = async ()=>{
@@ -49,6 +57,7 @@ const TopUp = () => {
     const {success, message, data} = res.data;
     if (success) {
       setUserQuota(data.quota);
+      setUser(data);
     } else {
       showError(message);
     }
@@ -80,7 +89,7 @@ const TopUp = () => {
               }}
             />
             <Button color='green' onClick={openTopUpLink}>
-              获取兑换码
+              充值
             </Button>
             <Button color='yellow' onClick={topUp} disabled={isSubmitting}>
                 {isSubmitting ? '兑换中...' : '兑换'}