feat: initial support of Dall-E (#148, #266)

* feat: initial support of Dall-E

* fix: fix N not timed

---------

Co-authored-by: JustSong <songquanpeng@foxmail.com>
Co-authored-by: JustSong <39998050+songquanpeng@users.noreply.github.com>

README.md

@@ -51,6 +51,8 @@ _✨ All in one 的 OpenAI 接口，整合各种 API 访问方式，开箱即用
   <a href="https://iamazing.cn/page/reward">赞赏支持</a>
 </p>
 
+> **Note**：本项目为开源项目，使用者必须在遵循 OpenAI 的[使用条款](https://openai.com/policies/terms-of-use)以及**法律法规**的情况下使用，不得用于非法用途。
+
 > **Note**：使用 Docker 拉取的最新镜像可能是 `alpha` 版本，如果追求稳定性请手动指定版本。
 
 > **Warning**：从 `v0.3` 版本升级到 `v0.4` 版本需要手动迁移数据库，请手动执行[数据库迁移脚本](./bin/migration_v0.3-v0.4.sql)。
@@ -64,6 +66,7 @@ _✨ All in one 的 OpenAI 接口，整合各种 API 访问方式，开箱即用
    + [x] [API2D](https://api2d.com/r/197971)
    + [x] [OhMyGPT](https://aigptx.top?aff=uFpUl2Kf)
    + [x] [AI Proxy](https://aiproxy.io/?i=OneAPI) （邀请码：`OneAPI`）
+   + [x] [CloseAI](https://console.closeai-asia.com/r/2412)
    + [x] 自定义渠道：例如各种未收录的第三方代理服务
 2. 支持通过**负载均衡**的方式访问多个渠道。
 3. 支持 **stream 模式**，可以通过流式传输实现打字机效果。
@@ -78,21 +81,26 @@ _✨ All in one 的 OpenAI 接口，整合各种 API 访问方式，开箱即用
 12. 支持以美元为单位显示额度。
 13. 支持发布公告，设置充值链接，设置新用户初始额度。
 14. 支持模型映射，重定向用户的请求模型。
-15. 支持丰富的**自定义**设置，
+15. 支持绘图接口。
+16. 支持丰富的**自定义**设置，
     1. 支持自定义系统名称，logo 以及页脚。
     2. 支持自定义首页和关于页面，可以选择使用 HTML & Markdown 代码进行自定义，或者使用一个单独的网页通过 iframe 嵌入。
-16. 支持通过系统访问令牌访问管理 API。
-17. 支持 Cloudflare Turnstile 用户校验。
-18. 支持用户管理，支持**多种用户登录注册方式**：
+17. 支持通过系统访问令牌访问管理 API。
+18. 支持 Cloudflare Turnstile 用户校验。
+19. 支持用户管理，支持**多种用户登录注册方式**：
     + 邮箱登录注册以及通过邮箱进行密码重置。
     + [GitHub 开放授权](https://github.com/settings/applications/new)。
     + 微信公众号授权（需要额外部署 [WeChat Server](https://github.com/songquanpeng/wechat-server)）。
-19. 未来其他大模型开放 API 后，将第一时间支持，并将其封装成同样的 API 访问方式。
+20. 未来其他大模型开放 API 后，将第一时间支持，并将其封装成同样的 API 访问方式。
 
 ## 部署
 ### 基于 Docker 进行部署
 部署命令：`docker run --name one-api -d --restart always -p 3000:3000 -e TZ=Asia/Shanghai -v /home/ubuntu/data/one-api:/data justsong/one-api`
 
+如果上面的镜像无法拉取，可以尝试使用 GitHub 的 Docker 镜像，将上面的 `justsong/one-api` 替换为 `ghcr.io/songquanpeng/one-api` 即可。
+
+如果你的并发量较大，推荐设置 `SQL_DSN`，详见下面[环境变量](#环境变量)一节。
+
 更新命令：`docker run --rm -v /var/run/docker.sock:/var/run/docker.sock containrrr/watchtower -cR`
 
 `-p 3000:3000` 中的第一个 `3000` 是宿主机的端口，可以根据需要进行修改。
@@ -158,8 +166,8 @@ sudo service nginx restart
 ### 多机部署
 1. 所有服务器 `SESSION_SECRET` 设置一样的值。
 2. 必须设置 `SQL_DSN`，使用 MySQL 数据库而非 SQLite，所有服务器连接同一个数据库。
-3. 所有从服务器必须设置 `NODE_TYPE` 为 `slave`。
-4. 设置 `SYNC_FREQUENCY` 后服务器将定期从数据库同步配置。
+3. 所有从服务器必须设置 `NODE_TYPE` 为 `slave`，不设置则默认为主服务器。
+4. 设置 `SYNC_FREQUENCY` 后服务器将定期从数据库同步配置，在使用远程数据库的情况下，推荐设置该项并启用 Redis，无论主从。
 5. 从服务器可以选择设置 `FRONTEND_BASE_URL`，以重定向页面请求到主服务器。
 6. 从服务器上**分别**装好 Redis，设置好 `REDIS_CONN_STRING`，这样可以做到在缓存未过期的情况下数据库零访问，可以减少延迟。
 7. 如果主服务器访问数据库延迟也比较高，则也需要启用 Redis，并设置 `SYNC_FREQUENCY`，以定期从数据库同步配置。
@@ -182,7 +190,7 @@ sudo service nginx restart
 docker run --name chat-next-web -d -p 3001:3000 yidadaa/chatgpt-next-web
 ```
 
-注意修改端口号和 `BASE_URL`。
+注意修改端口号，之后在页面上设置接口地址（例如：https://openai.justsong.cn/ ）和 API Key 即可。
 
 #### ChatGPT Web
 项目主页：https://github.com/Chanzhaoyu/chatgpt-web
@@ -231,6 +239,8 @@ docker run --name chatgpt-web -d -p 3002:3002 -e OPENAI_API_BASE_URL=https://ope
 
 等到系统启动后，使用 `root` 用户登录系统并做进一步的配置。
 
+**Note**：如果你不知道某个配置项的含义，可以临时删掉值以看到进一步的提示文字。
+
 ## 使用方法
 在`渠道`页面中添加你的 API Key，之后在`令牌`页面中新增访问令牌。
 
@@ -261,7 +271,10 @@ graph LR
    + 例子：`SESSION_SECRET=random_string`
 3. `SQL_DSN`：设置之后将使用指定数据库而非 SQLite，请使用 MySQL 8.0 版本。
    + 例子：`SQL_DSN=root:123456@tcp(localhost:3306)/oneapi`
-4. `FRONTEND_BASE_URL`：设置之后将使用指定的前端地址，而非后端地址，仅限从服务器设置。
+   + 注意需要提前建立数据库 `oneapi`，无需手动建表，程序将自动建表。
+   + 如果使用本地数据库：部署命令可添加 `--network="host"` 以使得容器内的程序可以访问到宿主机上的 MySQL。
+   + 如果使用云数据库：如果云服务器需要验证身份，需要在连接参数中添加 `?tls=skip-verify`。
+4. `FRONTEND_BASE_URL`：设置之后将重定向页面请求到指定的地址，仅限从服务器设置。
    + 例子：`FRONTEND_BASE_URL=https://openai.justsong.cn`
 5. `SYNC_FREQUENCY`：设置之后将定期与数据库同步配置，单位为秒，未设置则不进行同步。
    + 例子：`SYNC_FREQUENCY=60`
@@ -315,7 +328,6 @@ https://openai.justsong.cn
 [FastGPT](https://github.com/c121914yu/FastGPT): 三分钟搭建 AI 知识库
 
 ## 注意
-本项目为开源项目，请在遵循 OpenAI 的[使用条款](https://openai.com/policies/terms-of-use)以及**法律法规**的情况下使用，不得用于非法用途。
 
 本项目使用 MIT 协议进行开源，**在此基础上**，必须在页面底部保留署名以及指向本项目的链接。如果不想保留署名，必须首先获得授权。
 

bin/time_test.sh

@@ -1,13 +1,15 @@
 #!/bin/bash
 
-if [ $# -ne 3 ]; then
-  echo "Usage: time_test.sh <domain> <key> <count>"
+if [ $# -lt 3 ]; then
+  echo "Usage: time_test.sh <domain> <key> <count> [<model>]"
   exit 1
 fi
 
 domain=$1
 key=$2
 count=$3
+model=${4:-"gpt-3.5-turbo"} # 设置默认模型为 gpt-3.5-turbo
+
 total_time=0
 times=()
 
@@ -16,7 +18,7 @@ for ((i=1; i<=count; i++)); do
            https://"$domain"/v1/chat/completions \
            -H "Content-Type: application/json" \
            -H "Authorization: Bearer $key" \
-           -d '{"messages": [{"content": "echo hi", "role": "user"}], "model": "gpt-3.5-turbo", "stream": false, "max_tokens": 1}')
+           -d '{"messages": [{"content": "echo hi", "role": "user"}], "model": "'"$model"'", "stream": false, "max_tokens": 1}')
   http_code=$(echo "$result" | awk '{print $1}')
   time=$(echo "$result" | awk '{print $2}')
   echo "HTTP status code: $http_code, Time taken: $time"

common/constants.go

@@ -67,6 +67,7 @@ var ChannelDisableThreshold = 5.0
 var AutomaticDisableChannelEnabled = false
 var QuotaRemindThreshold = 1000
 var PreConsumedQuota = 500
+var ApproximateTokenEnabled = false
 
 var RootUserEmail = ""
 
@@ -152,18 +153,18 @@ const (
 )
 
 var ChannelBaseURLs = []string{
-	"",                             // 0
-	"https://api.openai.com",       // 1
-	"https://oa.api2d.net",         // 2
-	"",                             // 3
-	"https://api.openai-proxy.org", // 4
-	"https://api.openai-sb.com",    // 5
-	"https://api.openaimax.com",    // 6
-	"https://api.ohmygpt.com",      // 7
-	"",                             // 8
-	"https://api.caipacity.com",    // 9
-	"https://api.aiproxy.io",       // 10
-	"",                             // 11
-	"https://api.api2gpt.com",      // 12
-	"https://api.aigc2d.com",       // 13
+	"",                              // 0
+	"https://api.openai.com",        // 1
+	"https://oa.api2d.net",          // 2
+	"",                              // 3
+	"https://api.closeai-proxy.xyz", // 4
+	"https://api.openai-sb.com",     // 5
+	"https://api.openaimax.com",     // 6
+	"https://api.ohmygpt.com",       // 7
+	"",                              // 8
+	"https://api.caipacity.com",     // 9
+	"https://api.aiproxy.io",        // 10
+	"",                              // 11
+	"https://api.api2gpt.com",       // 12
+	"https://api.aigc2d.com",        // 13
 }

common/model-ratio.go

@@ -35,6 +35,7 @@ var ModelRatio = map[string]float64{
 	"text-search-ada-doc-001": 10,
 	"text-moderation-stable":  0.1,
 	"text-moderation-latest":  0.1,
+	"dall-e":                  8,
 }
 
 func ModelRatio2JSONString() string {

controller/channel-billing.go

@@ -32,6 +32,13 @@ type OpenAIUsageDailyCost struct {
 	}
 }
 
+type OpenAICreditGrants struct {
+	Object         string  `json:"object"`
+	TotalGranted   float64 `json:"total_granted"`
+	TotalUsed      float64 `json:"total_used"`
+	TotalAvailable float64 `json:"total_available"`
+}
+
 type OpenAIUsageResponse struct {
 	Object string `json:"object"`
 	//DailyCosts []OpenAIUsageDailyCost `json:"daily_costs"`
@@ -100,6 +107,22 @@ func GetResponseBody(method, url string, channel *model.Channel, headers http.He
 	return body, nil
 }
 
+func updateChannelCloseAIBalance(channel *model.Channel) (float64, error) {
+	url := fmt.Sprintf("%s/dashboard/billing/credit_grants", channel.BaseURL)
+	body, err := GetResponseBody("GET", url, channel, GetAuthHeader(channel.Key))
+
+	if err != nil {
+		return 0, err
+	}
+	response := OpenAICreditGrants{}
+	err = json.Unmarshal(body, &response)
+	if err != nil {
+		return 0, err
+	}
+	channel.UpdateBalance(response.TotalAvailable)
+	return response.TotalAvailable, nil
+}
+
 func updateChannelOpenAISBBalance(channel *model.Channel) (float64, error) {
 	url := fmt.Sprintf("https://api.openai-sb.com/sb-api/user/status?api_key=%s", channel.Key)
 	body, err := GetResponseBody("GET", url, channel, GetAuthHeader(channel.Key))
@@ -175,6 +198,9 @@ func updateChannelAIGC2DBalance(channel *model.Channel) (float64, error) {
 
 func updateChannelBalance(channel *model.Channel) (float64, error) {
 	baseURL := common.ChannelBaseURLs[channel.Type]
+	if channel.BaseURL == "" {
+		channel.BaseURL = baseURL
+	}
 	switch channel.Type {
 	case common.ChannelTypeOpenAI:
 		if channel.BaseURL != "" {
@@ -184,6 +210,8 @@ func updateChannelBalance(channel *model.Channel) (float64, error) {
 		return 0, errors.New("尚未实现")
 	case common.ChannelTypeCustom:
 		baseURL = channel.BaseURL
+	case common.ChannelTypeCloseAI:
+		return updateChannelCloseAIBalance(channel)
 	case common.ChannelTypeOpenAISB:
 		return updateChannelOpenAISBBalance(channel)
 	case common.ChannelTypeAIProxy:

controller/model.go

@@ -2,6 +2,7 @@ package controller
 
 import (
 	"fmt"
+
 	"github.com/gin-gonic/gin"
 )
 
@@ -53,6 +54,15 @@ func init() {
 	})
 	// https://platform.openai.com/docs/models/model-endpoint-compatibility
 	openAIModels = []OpenAIModels{
+		{
+			Id:         "dall-e",
+			Object:     "model",
+			Created:    1677649963,
+			OwnedBy:    "openai",
+			Permission: permission,
+			Root:       "dall-e",
+			Parent:     nil,
+		},
 		{
 			Id:         "gpt-3.5-turbo",
 			Object:     "model",

controller/relay-image.go

@@ -1,34 +1,181 @@
 package controller
 
 import (
-	"github.com/gin-gonic/gin"
+	"bytes"
+	"encoding/json"
+	"errors"
+	"fmt"
 	"io"
 	"net/http"
+	"one-api/common"
+	"one-api/model"
+
+	"github.com/gin-gonic/gin"
 )
 
 func relayImageHelper(c *gin.Context, relayMode int) *OpenAIErrorWithStatusCode {
-	// TODO: this part is not finished
-	req, err := http.NewRequest(c.Request.Method, c.Request.RequestURI, c.Request.Body)
+	imageModel := "dall-e"
+
+	tokenId := c.GetInt("token_id")
+	channelType := c.GetInt("channel")
+	userId := c.GetInt("id")
+	consumeQuota := c.GetBool("consume_quota")
+	group := c.GetString("group")
+
+	var imageRequest ImageRequest
+	if consumeQuota {
+		err := common.UnmarshalBodyReusable(c, &imageRequest)
+		if err != nil {
+			return errorWrapper(err, "bind_request_body_failed", http.StatusBadRequest)
+		}
+	}
+
+	// Prompt validation
+	if imageRequest.Prompt == "" {
+		return errorWrapper(errors.New("prompt is required"), "required_field_missing", http.StatusBadRequest)
+	}
+
+	// Not "256x256", "512x512", or "1024x1024"
+	if imageRequest.Size != "" && imageRequest.Size != "256x256" && imageRequest.Size != "512x512" && imageRequest.Size != "1024x1024" {
+		return errorWrapper(errors.New("size must be one of 256x256, 512x512, or 1024x1024"), "invalid_field_value", http.StatusBadRequest)
+	}
+
+	// N should between 1 and 10
+	if imageRequest.N != 0 && (imageRequest.N < 1 || imageRequest.N > 10) {
+		return errorWrapper(errors.New("n must be between 1 and 10"), "invalid_field_value", http.StatusBadRequest)
+	}
+
+	// map model name
+	modelMapping := c.GetString("model_mapping")
+	isModelMapped := false
+	if modelMapping != "" {
+		modelMap := make(map[string]string)
+		err := json.Unmarshal([]byte(modelMapping), &modelMap)
+		if err != nil {
+			return errorWrapper(err, "unmarshal_model_mapping_failed", http.StatusInternalServerError)
+		}
+		if modelMap[imageModel] != "" {
+			imageModel = modelMap[imageModel]
+			isModelMapped = true
+		}
+	}
+
+	baseURL := common.ChannelBaseURLs[channelType]
+	requestURL := c.Request.URL.String()
+
+	if c.GetString("base_url") != "" {
+		baseURL = c.GetString("base_url")
+	}
+
+	fullRequestURL := fmt.Sprintf("%s%s", baseURL, requestURL)
+
+	var requestBody io.Reader
+	if isModelMapped {
+		jsonStr, err := json.Marshal(imageRequest)
+		if err != nil {
+			return errorWrapper(err, "marshal_text_request_failed", http.StatusInternalServerError)
+		}
+		requestBody = bytes.NewBuffer(jsonStr)
+	} else {
+		requestBody = c.Request.Body
+	}
+
+	modelRatio := common.GetModelRatio(imageModel)
+	groupRatio := common.GetGroupRatio(group)
+	ratio := modelRatio * groupRatio
+	userQuota, err := model.CacheGetUserQuota(userId)
+
+	sizeRatio := 1.0
+	// Size
+	if imageRequest.Size == "256x256" {
+		sizeRatio = 1
+	} else if imageRequest.Size == "512x512" {
+		sizeRatio = 1.125
+	} else if imageRequest.Size == "1024x1024" {
+		sizeRatio = 1.25
+	}
+	quota := int(ratio*sizeRatio*1000) * imageRequest.N
+
+	if consumeQuota && userQuota-quota < 0 {
+		return errorWrapper(err, "insufficient_user_quota", http.StatusForbidden)
+	}
+
+	req, err := http.NewRequest(c.Request.Method, fullRequestURL, requestBody)
+	if err != nil {
+		return errorWrapper(err, "new_request_failed", http.StatusInternalServerError)
+	}
+	req.Header.Set("Authorization", c.Request.Header.Get("Authorization"))
+
+	req.Header.Set("Content-Type", c.Request.Header.Get("Content-Type"))
+	req.Header.Set("Accept", c.Request.Header.Get("Accept"))
+
 	client := &http.Client{}
 	resp, err := client.Do(req)
 	if err != nil {
-		return errorWrapper(err, "do_request_failed", http.StatusOK)
+		return errorWrapper(err, "do_request_failed", http.StatusInternalServerError)
 	}
+
 	err = req.Body.Close()
 	if err != nil {
-		return errorWrapper(err, "close_request_body_failed", http.StatusOK)
+		return errorWrapper(err, "close_request_body_failed", http.StatusInternalServerError)
 	}
+	err = c.Request.Body.Close()
+	if err != nil {
+		return errorWrapper(err, "close_request_body_failed", http.StatusInternalServerError)
+	}
+	var textResponse ImageResponse
+
+	defer func() {
+		if consumeQuota {
+			err := model.PostConsumeTokenQuota(tokenId, quota)
+			if err != nil {
+				common.SysError("error consuming token remain quota: " + err.Error())
+			}
+			err = model.CacheUpdateUserQuota(userId)
+			if err != nil {
+				common.SysError("error update user quota cache: " + err.Error())
+			}
+			if quota != 0 {
+				tokenName := c.GetString("token_name")
+				logContent := fmt.Sprintf("模型倍率 %.2f，分组倍率 %.2f", modelRatio, groupRatio)
+				model.RecordConsumeLog(userId, 0, 0, imageModel, tokenName, quota, logContent)
+				model.UpdateUserUsedQuotaAndRequestCount(userId, quota)
+				channelId := c.GetInt("channel_id")
+				model.UpdateChannelUsedQuota(channelId, quota)
+			}
+		}
+	}()
+
+	if consumeQuota {
+		responseBody, err := io.ReadAll(resp.Body)
+
+		if err != nil {
+			return errorWrapper(err, "read_response_body_failed", http.StatusInternalServerError)
+		}
+		err = resp.Body.Close()
+		if err != nil {
+			return errorWrapper(err, "close_response_body_failed", http.StatusInternalServerError)
+		}
+		err = json.Unmarshal(responseBody, &textResponse)
+		if err != nil {
+			return errorWrapper(err, "unmarshal_response_body_failed", http.StatusInternalServerError)
+		}
+
+		resp.Body = io.NopCloser(bytes.NewBuffer(responseBody))
+	}
+
 	for k, v := range resp.Header {
 		c.Writer.Header().Set(k, v[0])
 	}
 	c.Writer.WriteHeader(resp.StatusCode)
+
 	_, err = io.Copy(c.Writer, resp.Body)
 	if err != nil {
-		return errorWrapper(err, "copy_response_body_failed", http.StatusOK)
+		return errorWrapper(err, "copy_response_body_failed", http.StatusInternalServerError)
 	}
 	err = resp.Body.Close()
 	if err != nil {
-		return errorWrapper(err, "close_response_body_failed", http.StatusOK)
+		return errorWrapper(err, "close_response_body_failed", http.StatusInternalServerError)
 	}
 	return nil
 }

controller/relay-text.go

@@ -6,12 +6,13 @@ import (
 	"encoding/json"
 	"errors"
 	"fmt"
-	"github.com/gin-gonic/gin"
 	"io"
 	"net/http"
 	"one-api/common"
 	"one-api/model"
 	"strings"
+
+	"github.com/gin-gonic/gin"
 )
 
 func relayTextHelper(c *gin.Context, relayMode int) *OpenAIErrorWithStatusCode {
@@ -30,6 +31,9 @@ func relayTextHelper(c *gin.Context, relayMode int) *OpenAIErrorWithStatusCode {
 	if relayMode == RelayModeModerations && textRequest.Model == "" {
 		textRequest.Model = "text-moderation-latest"
 	}
+	if relayMode == RelayModeEmbeddings && textRequest.Model == "" {
+		textRequest.Model = c.Param("model")
+	}
 	// request validation
 	if textRequest.Model == "" {
 		return errorWrapper(errors.New("model is required"), "required_field_missing", http.StatusBadRequest)
@@ -151,7 +155,7 @@ func relayTextHelper(c *gin.Context, relayMode int) *OpenAIErrorWithStatusCode {
 	}
 	req.Header.Set("Content-Type", c.Request.Header.Get("Content-Type"))
 	req.Header.Set("Accept", c.Request.Header.Get("Accept"))
-	req.Header.Set("Connection", c.Request.Header.Get("Connection"))
+	//req.Header.Set("Connection", c.Request.Header.Get("Connection"))
 	client := &http.Client{}
 	resp, err := client.Do(req)
 	if err != nil {

controller/relay-utils.go

@@ -4,7 +4,6 @@ import (
 	"fmt"
 	"github.com/pkoukk/tiktoken-go"
 	"one-api/common"
-	"strings"
 )
 
 var tokenEncoderMap = map[string]*tiktoken.Tiktoken{}
@@ -25,6 +24,13 @@ func getTokenEncoder(model string) *tiktoken.Tiktoken {
 	return tokenEncoder
 }
 
+func getTokenNum(tokenEncoder *tiktoken.Tiktoken, text string) int {
+	if common.ApproximateTokenEnabled {
+		return int(float64(len(text)) * 0.38)
+	}
+	return len(tokenEncoder.Encode(text, nil, nil))
+}
+
 func countTokenMessages(messages []Message, model string) int {
 	tokenEncoder := getTokenEncoder(model)
 	// Reference:
@@ -34,12 +40,9 @@ func countTokenMessages(messages []Message, model string) int {
 	// Every message follows <|start|>{role/name}\n{content}<|end|>\n
 	var tokensPerMessage int
 	var tokensPerName int
-	if strings.HasPrefix(model, "gpt-3.5") {
+	if model == "gpt-3.5-turbo-0301" {
 		tokensPerMessage = 4
 		tokensPerName = -1 // If there's a name, the role is omitted
-	} else if strings.HasPrefix(model, "gpt-4") {
-		tokensPerMessage = 3
-		tokensPerName = 1
 	} else {
 		tokensPerMessage = 3
 		tokensPerName = 1
@@ -47,11 +50,11 @@ func countTokenMessages(messages []Message, model string) int {
 	tokenNum := 0
 	for _, message := range messages {
 		tokenNum += tokensPerMessage
-		tokenNum += len(tokenEncoder.Encode(message.Content, nil, nil))
-		tokenNum += len(tokenEncoder.Encode(message.Role, nil, nil))
+		tokenNum += getTokenNum(tokenEncoder, message.Content)
+		tokenNum += getTokenNum(tokenEncoder, message.Role)
 		if message.Name != nil {
 			tokenNum += tokensPerName
-			tokenNum += len(tokenEncoder.Encode(*message.Name, nil, nil))
+			tokenNum += getTokenNum(tokenEncoder, *message.Name)
 		}
 	}
 	tokenNum += 3 // Every reply is primed with <|start|>assistant<|message|>
@@ -74,8 +77,7 @@ func countTokenInput(input any, model string) int {
 
 func countTokenText(text string, model string) int {
 	tokenEncoder := getTokenEncoder(model)
-	token := tokenEncoder.Encode(text, nil, nil)
-	return len(token)
+	return getTokenNum(tokenEncoder, text)
 }
 
 func errorWrapper(err error, code string, statusCode int) *OpenAIErrorWithStatusCode {

controller/relay.go

@@ -2,10 +2,11 @@ package controller
 
 import (
 	"fmt"
-	"github.com/gin-gonic/gin"
 	"net/http"
 	"one-api/common"
 	"strings"
+
+	"github.com/gin-gonic/gin"
 )
 
 type Message struct {
@@ -37,6 +38,7 @@ type GeneralOpenAIRequest struct {
 	N           int       `json:"n,omitempty"`
 	Input       any       `json:"input,omitempty"`
 	Instruction string    `json:"instruction,omitempty"`
+	Size        string    `json:"size,omitempty"`
 }
 
 type ChatRequest struct {
@@ -53,6 +55,12 @@ type TextRequest struct {
 	//Stream   bool      `json:"stream"`
 }
 
+type ImageRequest struct {
+	Prompt string `json:"prompt"`
+	N      int    `json:"n"`
+	Size   string `json:"size"`
+}
+
 type Usage struct {
 	PromptTokens     int `json:"prompt_tokens"`
 	CompletionTokens int `json:"completion_tokens"`
@@ -76,6 +84,13 @@ type TextResponse struct {
 	Error OpenAIError `json:"error"`
 }
 
+type ImageResponse struct {
+	Created int `json:"created"`
+	Data    []struct {
+		Url string `json:"url"`
+	}
+}
+
 type ChatCompletionsStreamResponse struct {
 	Choices []struct {
 		Delta struct {
@@ -100,6 +115,8 @@ func Relay(c *gin.Context) {
 		relayMode = RelayModeCompletions
 	} else if strings.HasPrefix(c.Request.URL.Path, "/v1/embeddings") {
 		relayMode = RelayModeEmbeddings
+	} else if strings.HasSuffix(c.Request.URL.Path, "embeddings") {
+		relayMode = RelayModeEmbeddings
 	} else if strings.HasPrefix(c.Request.URL.Path, "/v1/moderations") {
 		relayMode = RelayModeModerations
 	} else if strings.HasPrefix(c.Request.URL.Path, "/v1/images/generations") {
@@ -124,7 +141,7 @@ func Relay(c *gin.Context) {
 		channelId := c.GetInt("channel_id")
 		common.SysError(fmt.Sprintf("relay error (channel #%d): %s", channelId, err.Message))
 		// https://platform.openai.com/docs/guides/error-codes/api-errors
-		if common.AutomaticDisableChannelEnabled && (err.Type == "insufficient_quota" || err.Code == "invalid_api_key") {
+		if common.AutomaticDisableChannelEnabled && (err.Type == "insufficient_quota" || err.Code == "invalid_api_key" || err.Code == "account_deactivated") {
 			channelId := c.GetInt("channel_id")
 			channelName := c.GetString("channel_name")
 			disableChannel(channelId, channelName, err.Message)

i18n/en.json

@@ -107,6 +107,11 @@
   "已禁用": "Disabled",
   "未知状态": "Unknown status",
   " 秒": "s",
+  " 分钟 ": " m ",
+  " 小时 ": " h ",
+  " 天 ": " d ",
+  " 个月 ": " M ",
+  " 年 ": " y ",
   "未测试": "Not tested",
   "通道 ${name} 测试成功，耗时 ${time.toFixed(2)} 秒。": "Channel ${name} test succeeded, time consumed ${time.toFixed(2)} s.",
   "已成功开始测试所有已启用通道，请刷新页面查看结果。": "All enabled channels have been successfully tested, please refresh the page to view the results.",
@@ -458,5 +463,45 @@
   "消耗额度": "Used Quota",
   "可选值": "Optional Values",
   "渠道不存在：%d": "Channel does not exist: %d",
-  "数据库一致性已被破坏，请联系管理员": "Database consistency has been broken, please contact the administrator"
+  "数据库一致性已被破坏，请联系管理员": "Database consistency has been broken, please contact the administrator",
+  "使用近似的方式估算 token 数以减少计算量": "Estimate the number of tokens in an approximate way to reduce computational load",
+  "请填写ChannelName和ChannelKey！": "Please fill in the ChannelName and ChannelKey!",
+  "请至少选择一个Model！": "Please select at least one Model!",
+  "加载首页内容失败": "Failed to load the homepage content",
+  "加载关于内容失败": "Failed to load the About content",
+  "兑换码更新成功！": "Redemption code updated successfully!",
+  "兑换码创建成功！": "Redemption code created successfully!",
+  "用户账户创建成功！": "User account created successfully!",
+  "生成数量": "Generate quantity",
+  "请输入生成数量": "Please enter the quantity to generate",
+  "创建新用户账户": "Create new user account",
+  "渠道更新成功！": "Channel updated successfully!",
+  "渠道创建成功！": "Channel created successfully!",
+  "请选择分组": "Please select a group",
+  "更新兑换码信息": "Update redemption code information",
+  "创建新的兑换码": "Create a new redemption code",
+  "请在系统设置页面编辑分组倍率以添加新的分组：": "Please edit the group ratio in the system settings page to add a new group:",
+  "未找到所请求的页面": "The requested page was not found",
+  "过期时间格式错误！": "Expiration time format error!",
+  "请输入过期时间，格式为 yyyy-MM-dd HH:mm:ss，-1 表示无限制": "Please enter the expiration time, the format is yyyy-MM-dd HH:mm:ss, -1 means no limit",
+  "此项可选，为一个 JSON 文本，键为用户请求的模型名称，值为要替换的模型名称，例如：": "This is optional, it's a JSON text, the key is the model name requested by the user, and the value is the model name to be replaced, for example:",
+  "此项可选，输入镜像站地址，格式为：": "This is optional, enter the mirror site address, the format is:",
+  "模型映射": "Model mapping",
+  "请输入默认 API 版本，例如：2023-03-15-preview，该配置可以被实际的请求查询参数所覆盖": "Please enter the default API version, for example: 2023-03-15-preview, this configuration can be overridden by the actual request query parameters",
+  "默认": "Default",
+  "图片演示": "Image demo",
+  "参数替换为你的部署名称（模型名称中的点会被剔除）": "Replace the parameter with your deployment name (dots in the model name will be removed)",
+  "模型映射必须是合法的 JSON 格式！": "Model mapping must be in valid JSON format!",
+  "取消无限额度": "Cancel unlimited quota",
+  "请输入新的剩余额度": "Please enter the new remaining quota",
+  "请输入单个兑换码中包含的额度": "Please enter the quota included in a single redemption code",
+  "请输入用户名": "Please enter username",
+  "请输入显示名称": "Please enter display name",
+  "请输入密码": "Please enter password",
+  "模型部署名称必须和模型名称保持一致": "The model deployment name must be consistent with the model name",
+  "，因为 One API 会把请求体中的 model": ", because One API will take the model in the request body",
+  "请输入 AZURE_OPENAI_ENDPOINT": "Please enter AZURE_OPENAI_ENDPOINT",
+  "请输入自定义渠道的 Base URL": "Please enter the Base URL of the custom channel",
+  "Homepage URL 填": "Fill in the Homepage URL",
+  "Authorization callback URL 填": "Fill in the Authorization callback URL"
 }

middleware/distributor.go

@@ -2,12 +2,13 @@ package middleware
 
 import (
 	"fmt"
-	"github.com/gin-gonic/gin"
 	"net/http"
 	"one-api/common"
 	"one-api/model"
 	"strconv"
 	"strings"
+
+	"github.com/gin-gonic/gin"
 )
 
 type ModelRequest struct {
@@ -73,6 +74,16 @@ func Distribute() func(c *gin.Context) {
 					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"
+				}
+			}
 			channel, err = model.CacheGetRandomSatisfiedChannel(userGroup, modelRequest.Model)
 			if err != nil {
 				message := "无可用渠道"

model/option.go

@@ -34,6 +34,7 @@ func InitOptionMap() {
 	common.OptionMap["TurnstileCheckEnabled"] = strconv.FormatBool(common.TurnstileCheckEnabled)
 	common.OptionMap["RegisterEnabled"] = strconv.FormatBool(common.RegisterEnabled)
 	common.OptionMap["AutomaticDisableChannelEnabled"] = strconv.FormatBool(common.AutomaticDisableChannelEnabled)
+	common.OptionMap["ApproximateTokenEnabled"] = strconv.FormatBool(common.ApproximateTokenEnabled)
 	common.OptionMap["LogConsumeEnabled"] = strconv.FormatBool(common.LogConsumeEnabled)
 	common.OptionMap["DisplayInCurrencyEnabled"] = strconv.FormatBool(common.DisplayInCurrencyEnabled)
 	common.OptionMap["DisplayTokenStatEnabled"] = strconv.FormatBool(common.DisplayTokenStatEnabled)
@@ -141,6 +142,8 @@ func updateOptionMap(key string, value string) (err error) {
 			common.RegisterEnabled = boolValue
 		case "AutomaticDisableChannelEnabled":
 			common.AutomaticDisableChannelEnabled = boolValue
+		case "ApproximateTokenEnabled":
+			common.ApproximateTokenEnabled = boolValue
 		case "LogConsumeEnabled":
 			common.LogConsumeEnabled = boolValue
 		case "DisplayInCurrencyEnabled":

model/redemption.go

@@ -3,6 +3,7 @@ package model
 import (
 	"errors"
 	"fmt"
+	"gorm.io/gorm"
 	"one-api/common"
 )
 
@@ -48,26 +49,27 @@ func Redeem(key string, userId int) (quota int, err error) {
 		return 0, errors.New("无效的 user id")
 	}
 	redemption := &Redemption{}
-	err = DB.Where("`key` = ?", key).First(redemption).Error
-	if err != nil {
-		return 0, errors.New("无效的兑换码")
-	}
-	if redemption.Status != common.RedemptionCodeStatusEnabled {
-		return 0, errors.New("该兑换码已被使用")
-	}
-	err = IncreaseUserQuota(userId, redemption.Quota)
-	if err != nil {
-		return 0, err
-	}
-	go func() {
+
+	err = DB.Transaction(func(tx *gorm.DB) error {
+		err := DB.Where("`key` = ?", key).First(redemption).Error
+		if err != nil {
+			return errors.New("无效的兑换码")
+		}
+		if redemption.Status != common.RedemptionCodeStatusEnabled {
+			return errors.New("该兑换码已被使用")
+		}
+		err = DB.Model(&User{}).Where("id = ?", userId).Update("quota", gorm.Expr("quota + ?", redemption.Quota)).Error
+		if err != nil {
+			return err
+		}
 		redemption.RedeemedTime = common.GetTimestamp()
 		redemption.Status = common.RedemptionCodeStatusUsed
-		err := redemption.SelectUpdate()
-		if err != nil {
-			common.SysError("failed to update redemption status: " + err.Error())
-		}
-		RecordLog(userId, LogTypeTopup, fmt.Sprintf("通过兑换码充值 %s", common.LogQuota(redemption.Quota)))
-	}()
+		return redemption.SelectUpdate()
+	})
+	if err != nil {
+		return 0, errors.New("兑换失败，" + err.Error())
+	}
+	RecordLog(userId, LogTypeTopup, fmt.Sprintf("通过兑换码充值 %s", common.LogQuota(redemption.Quota)))
 	return redemption.Quota, nil
 }
 

router/relay-router.go

@@ -1,9 +1,10 @@
 package router
 
 import (
-	"github.com/gin-gonic/gin"
 	"one-api/controller"
 	"one-api/middleware"
+
+	"github.com/gin-gonic/gin"
 )
 
 func SetRelayRouter(router *gin.Engine) {
@@ -20,10 +21,11 @@ func SetRelayRouter(router *gin.Engine) {
 		relayV1Router.POST("/completions", controller.Relay)
 		relayV1Router.POST("/chat/completions", controller.Relay)
 		relayV1Router.POST("/edits", controller.Relay)
-		relayV1Router.POST("/images/generations", controller.RelayNotImplemented)
+		relayV1Router.POST("/images/generations", controller.Relay)
 		relayV1Router.POST("/images/edits", controller.RelayNotImplemented)
 		relayV1Router.POST("/images/variations", controller.RelayNotImplemented)
 		relayV1Router.POST("/embeddings", controller.Relay)
+		relayV1Router.POST("/engines/:model/embeddings", controller.Relay)
 		relayV1Router.POST("/audio/transcriptions", controller.RelayNotImplemented)
 		relayV1Router.POST("/audio/translations", controller.RelayNotImplemented)
 		relayV1Router.GET("/files", controller.RelayNotImplemented)

web/src/components/ChannelsTable.js

@@ -30,6 +30,9 @@ function renderType(type) {
 function renderBalance(type, balance) {
   switch (type) {
     case 1: // OpenAI
+      return <span>${balance.toFixed(2)}</span>;
+    case 4: // CloseAI
+      return <span>¥{balance.toFixed(2)}</span>;
     case 8: // 自定义
       return <span>${balance.toFixed(2)}</span>;
     case 5: // OpenAI-SB

web/src/components/OperationSetting.js

@@ -18,7 +18,8 @@ const OperationSetting = () => {
     ChannelDisableThreshold: 0,
     LogConsumeEnabled: '',
     DisplayInCurrencyEnabled: '',
-    DisplayTokenStatEnabled: ''
+    DisplayTokenStatEnabled: '',
+    ApproximateTokenEnabled: '',
   });
   const [originInputs, setOriginInputs] = useState({});
   let [loading, setLoading] = useState(false);
@@ -181,6 +182,12 @@ const OperationSetting = () => {
               name='DisplayTokenStatEnabled'
               onChange={handleInputChange}
             />
+            <Form.Checkbox
+              checked={inputs.ApproximateTokenEnabled === 'true'}
+              label='使用近似的方式估算 token 数以减少计算量'
+              name='ApproximateTokenEnabled'
+              onChange={handleInputChange}
+            />
           </Form.Group>
           <Form.Button onClick={() => {
             submitConfig('general').then();

web/src/components/UsersTable.js

@@ -183,14 +183,6 @@ const UsersTable = () => {
             >
               分组
             </Table.HeaderCell>
-            <Table.HeaderCell
-              style={{ cursor: 'pointer' }}
-              onClick={() => {
-                sortUser('email');
-              }}
-            >
-              邮箱地址
-            </Table.HeaderCell>
             <Table.HeaderCell
               style={{ cursor: 'pointer' }}
               onClick={() => {
@@ -233,20 +225,20 @@ const UsersTable = () => {
                   <Table.Cell>
                     <Popup
                       content={user.email ? user.email : '未绑定邮箱地址'}
-                      key={user.display_name}
+                      key={user.username}
                       header={user.display_name ? user.display_name : user.username}
                       trigger={<span>{renderText(user.username, 10)}</span>}
                       hoverable
                     />
                   </Table.Cell>
                   <Table.Cell>{renderGroup(user.group)}</Table.Cell>
+                  {/*<Table.Cell>*/}
+                  {/*  {user.email ? <Popup hoverable content={user.email} trigger={<span>{renderText(user.email, 24)}</span>} /> : '无'}*/}
+                  {/*</Table.Cell>*/}
                   <Table.Cell>
-                    {user.email ? <Popup hoverable content={user.email} trigger={<span>{renderText(user.email, 24)}</span>} /> : '无'}
-                  </Table.Cell>
-                  <Table.Cell>
-                    <Popup content='剩余额度' trigger={<Label>{renderQuota(user.quota)}</Label>} />
-                    <Popup content='已用额度' trigger={<Label>{renderQuota(user.used_quota)}</Label>} />
-                    <Popup content='请求次数' trigger={<Label>{renderNumber(user.request_count)}</Label>} />
+                    <Popup content='剩余额度' trigger={<Label basic>{renderQuota(user.quota)}</Label>} />
+                    <Popup content='已用额度' trigger={<Label basic>{renderQuota(user.used_quota)}</Label>} />
+                    <Popup content='请求次数' trigger={<Label basic>{renderNumber(user.request_count)}</Label>} />
                   </Table.Cell>
                   <Table.Cell>{renderRole(user.role)}</Table.Cell>
                   <Table.Cell>{renderStatus(user.status)}</Table.Cell>
@@ -320,7 +312,7 @@ const UsersTable = () => {
 
         <Table.Footer>
           <Table.Row>
-            <Table.HeaderCell colSpan='8'>
+            <Table.HeaderCell colSpan='7'>
               <Button size='small' as={Link} to='/user/add' loading={loading}>
                 添加新的用户
               </Button>

web/src/pages/Channel/EditChannel.js

@@ -145,6 +145,7 @@ const EditChannel = () => {
             <Form.Select
               label='类型'
               name='type'
+              required
               options={CHANNEL_OPTIONS}
               value={inputs.type}
               onChange={handleInputChange}
@@ -201,7 +202,7 @@ const EditChannel = () => {
                 <Form.Input
                   label='镜像'
                   name='base_url'
-                  placeholder={'请输入镜像站地址，格式为：https://domain.com，可不填，不填则使用渠道默认值'}
+                  placeholder={'此项可选，输入镜像站地址，格式为：https://domain.com'}
                   onChange={handleInputChange}
                   value={inputs.base_url}
                   autoComplete='new-password'
@@ -212,6 +213,7 @@ const EditChannel = () => {
           <Form.Field>
             <Form.Input
               label='名称'
+              required
               name='name'
               placeholder={'请输入名称'}
               onChange={handleInputChange}
@@ -224,6 +226,7 @@ const EditChannel = () => {
               label='分组'
               placeholder={'请选择分组'}
               name='groups'
+              required
               fluid
               multiple
               selection
@@ -240,6 +243,7 @@ const EditChannel = () => {
               label='模型'
               placeholder={'请选择该通道所支持的模型'}
               name='models'
+              required
               fluid
               multiple
               selection
@@ -263,7 +267,7 @@ const EditChannel = () => {
           <Form.Field>
             <Form.TextArea
               label='模型映射'
-              placeholder={`为一个 JSON 文本，键为用户请求的模型名称，值为要替换的模型名称，例如：\n${JSON.stringify(MODEL_MAPPING_EXAMPLE, null, 2)}`}
+              placeholder={`此项可选，为一个 JSON 文本，键为用户请求的模型名称，值为要替换的模型名称，例如：\n${JSON.stringify(MODEL_MAPPING_EXAMPLE, null, 2)}`}
               name='model_mapping'
               onChange={handleInputChange}
               value={inputs.model_mapping}
@@ -276,6 +280,7 @@ const EditChannel = () => {
               <Form.TextArea
                 label='密钥'
                 name='key'
+                required
                 placeholder={'请输入密钥，一行一个'}
                 onChange={handleInputChange}
                 value={inputs.key}
@@ -286,6 +291,7 @@ const EditChannel = () => {
               <Form.Input
                 label='密钥'
                 name='key'
+                required
                 placeholder={'请输入密钥'}
                 onChange={handleInputChange}
                 value={inputs.key}

web/src/pages/Token/EditToken.js

@@ -11,7 +11,7 @@ const EditToken = () => {
   const [loading, setLoading] = useState(isEdit);
   const originInputs = {
     name: '',
-    remain_quota: 0,
+    remain_quota: isEdit ? 0 : 500000,
     expired_time: -1,
     unlimited_quota: false
   };