revert: feat: fast build linux/arm64 frontend (#1645)

This reverts commit 1c44d7e1cd.

.github/workflows/ci.yml

@@ -36,21 +36,9 @@ jobs:
       # in the next step as well as the next job.
       - name: Test
         run: go test -cover -coverprofile=coverage.txt ./...
-
+      - uses: codecov/codecov-action@v4
-      - name: Archive code coverage results
-        uses: actions/upload-artifact@v4
         with:
-          name: code-coverage
+          token: ${{ secrets.CODECOV_TOKEN }}
-          path: coverage.txt # Make sure to use the same file name you chose for the "-coverprofile" in the "Test" step
-
-  code_coverage:
-    name: "Code coverage report"
-    runs-on: ubuntu-latest
-    needs: unit_tests # Depends on the artifact uploaded by the "unit_tests" job
-    steps:
-    - uses: codecov/codecov-action@v4
-      with:
-        token: ${{ secrets.CODECOV_TOKEN }}
 
   commit_lint:
     runs-on: ubuntu-latest

Dockerfile

@@ -16,7 +16,9 @@ WORKDIR /web/air
 RUN npm install
 RUN DISABLE_ESLINT_PLUGIN='true' REACT_APP_VERSION=$(cat VERSION) npm run build
 
-FROM golang AS builder2
+FROM golang:alpine AS builder2
+
+RUN apk add --no-cache g++
 
 ENV GO111MODULE=on \
     CGO_ENABLED=1 \
@@ -27,7 +29,7 @@ ADD go.mod go.sum ./
 RUN go mod download
 COPY . .
 COPY --from=builder /web/build ./web/build
-RUN go build -ldflags "-s -w -X 'github.com/songquanpeng/one-api/common.Version=$(cat VERSION)' -extldflags '-static'" -o one-api
+RUN go build -trimpath -ldflags "-s -w -X 'github.com/songquanpeng/one-api/common.Version=$(cat VERSION)' -extldflags '-static'" -o one-api
 
 FROM alpine
 

README.en.md

@@ -101,7 +101,7 @@ Nginx reference configuration:
 ```
 server{
    server_name openai.justsong.cn;  # Modify your domain name accordingly
-   
+
    location / {
           client_max_body_size  64m;
           proxy_http_version 1.1;
@@ -132,12 +132,12 @@ The initial account username is `root` and password is `123456`.
 1. Download the executable file from [GitHub Releases](https://github.com/songquanpeng/one-api/releases/latest) or compile from source:
    ```shell
    git clone https://github.com/songquanpeng/one-api.git
-   
+
    # Build the frontend
    cd one-api/web/default
    npm install
    npm run build
-   
+
    # Build the backend
    cd ../..
    go mod download
@@ -245,16 +245,41 @@ If the channel ID is not provided, load balancing will be used to distribute the
     + Example: `LOG_SQL_DSN=root:123456@tcp(localhost:3306)/oneapi-logs`
 5. `FRONTEND_BASE_URL`: When set, the specified frontend address will be used instead of the backend address.
     + Example: `FRONTEND_BASE_URL=https://openai.justsong.cn`
-6. `SYNC_FREQUENCY`: When set, the system will periodically sync configurations from the database, with the unit in seconds. If not set, no sync will happen.
+6. 'MEMORY_CACHE_ENABLED': Enabling memory caching can cause a certain delay in updating user quotas, with optional values of 'true' and 'false'. If not set, it defaults to 'false'.
+7. `SYNC_FREQUENCY`: When set, the system will periodically sync configurations from the database, with the unit in seconds. If not set, no sync will happen.
     + Example: `SYNC_FREQUENCY=60`
-7. `NODE_TYPE`: When set, specifies the node type. Valid values are `master` and `slave`. If not set, it defaults to `master`.
+8. `NODE_TYPE`: When set, specifies the node type. Valid values are `master` and `slave`. If not set, it defaults to `master`.
     + Example: `NODE_TYPE=slave`
-8. `CHANNEL_UPDATE_FREQUENCY`: When set, it periodically updates the channel balances, with the unit in minutes. If not set, no update will happen.
+9. `CHANNEL_UPDATE_FREQUENCY`: When set, it periodically updates the channel balances, with the unit in minutes. If not set, no update will happen.
     + Example: `CHANNEL_UPDATE_FREQUENCY=1440`
-9. `CHANNEL_TEST_FREQUENCY`: When set, it periodically tests the channels, with the unit in minutes. If not set, no test will happen.
+10. `CHANNEL_TEST_FREQUENCY`: When set, it periodically tests the channels, with the unit in minutes. If not set, no test will happen.
     + Example: `CHANNEL_TEST_FREQUENCY=1440`
-10. `POLLING_INTERVAL`: The time interval (in seconds) between requests when updating channel balances and testing channel availability. Default is no interval.
+11. `POLLING_INTERVAL`: The time interval (in seconds) between requests when updating channel balances and testing channel availability. Default is no interval.
     + Example: `POLLING_INTERVAL=5`
+12. `BATCH_UPDATE_ENABLED`: Enabling batch database update aggregation can cause a certain delay in updating user quotas. The optional values are 'true' and 'false', but if not set, it defaults to 'false'.
+    +Example: ` BATCH_UPDATE_ENABLED=true`
+    +If you encounter an issue with too many database connections, you can try enabling this option.
+13. `BATCH_UPDATE_INTERVAL=5`: The time interval for batch updating aggregates, measured in seconds, defaults to '5'.
+    +Example: ` BATCH_UPDATE_INTERVAL=5`
+14. Request frequency limit:
+    + `GLOBAL_API_RATE_LIMIT`: Global API rate limit (excluding relay requests), the maximum number of requests within three minutes per IP, default to 180.
+    + `GLOBAL_WEL_RATE_LIMIT`: Global web speed limit, the maximum number of requests within three minutes per IP, default to 60.
+15. Encoder cache settings:
+    +`TIKTOKEN_CACHE_DIR`: By default, when the program starts, it will download the encoding of some common word elements online, such as' gpt-3.5 turbo '. In some unstable network environments or offline situations, it may cause startup problems. This directory can be configured to cache data and can be migrated to an offline environment.
+    +`DATA_GYM_CACHE_DIR`: Currently, this configuration has the same function as' TIKTOKEN-CACHE-DIR ', but its priority is not as high as it.
+16. `RELAY_TIMEOUT`: Relay timeout setting, measured in seconds, with no default timeout time set.
+17. `RELAY_PROXY`: After setting up, use this proxy to request APIs.
+18. `USER_CONTENT_REQUEST_TIMEOUT`: The timeout period for users to upload and download content, measured in seconds.
+19. `USER_CONTENT_REQUEST_PROXY`: After setting up, use this agent to request content uploaded by users, such as images.
+20. `SQLITE_BUSY_TIMEOUT`: SQLite lock wait timeout setting, measured in milliseconds, default to '3000'.
+21. `GEMINI_SAFETY_SETTING`: Gemini's security settings are set to 'BLOCK-NONE' by default.
+22. `GEMINI_VERSION`: The Gemini version used by the One API, which defaults to 'v1'.
+23. `THE`: The system's theme setting, default to 'default', specific optional values refer to [here] (./web/README. md).
+24. `ENABLE_METRIC`: Whether to disable channels based on request success rate, default not enabled, optional values are 'true' and 'false'.
+25. `METRIC_QUEUE_SIZE`: Request success rate statistics queue size, default to '10'.
+26. `METRIC_SUCCESS_RATE_THRESHOLD`: Request success rate threshold, default to '0.8'.
+27. `INITIAL_ROOT_TOKEN`: If this value is set, a root user token with the value of the environment variable will be automatically created when the system starts for the first time.
+28. `INITIAL_ROOT_ACCESS_TOKEN`: If this value is set, a system management token will be automatically created for the root user with a value of the environment variable when the system starts for the first time.
 
 ### Command Line Parameters
 1. `--port <port_number>`: Specifies the port number on which the server listens. Defaults to `3000`.
@@ -287,7 +312,9 @@ If the channel ID is not provided, load balancing will be used to distribute the
     + Double-check that your interface address and API Key are correct.
 
 ## Related Projects
-[FastGPT](https://github.com/labring/FastGPT): Knowledge question answering system based on the LLM
+* [FastGPT](https://github.com/labring/FastGPT): Knowledge question answering system based on the LLM
+* [VChart](https://github.com/VisActor/VChart):  More than just a cross-platform charting library, but also an expressive data storyteller.
+* [VMind](https://github.com/VisActor/VMind):  Not just automatic, but also fantastic. Open-source solution for intelligent visualization.
 
 ## Note
 This project is an open-source project. Please use it in compliance with OpenAI's [Terms of Use](https://openai.com/policies/terms-of-use) and **applicable laws and regulations**. It must not be used for illegal purposes.

README.md

@@ -53,7 +53,7 @@ _✨ 通过标准的 OpenAI API 格式访问所有的大模型，开箱即用
 
 > [!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。
-> 
+>
 > 根据[《生成式人工智能服务管理暂行办法》](http://www.cac.gov.cn/2023-07/13/c_1690898327029107.htm)的要求，请勿对中国地区公众提供一切未经备案的生成式人工智能服务。
 
 > [!WARNING]
@@ -88,6 +88,7 @@ _✨ 通过标准的 OpenAI API 格式访问所有的大模型，开箱即用
    + [x] [Cloudflare Workers AI](https://developers.cloudflare.com/workers-ai/)
    + [x] [DeepL](https://www.deepl.com/)
    + [x] [together.ai](https://www.together.ai/)
+   + [x] [novita.ai](https://www.novita.ai/)
 2. 支持配置镜像以及众多[第三方代理服务](https://iamazing.cn/page/openai-api-third-party-services)。
 3. 支持通过**负载均衡**的方式访问多个渠道。
 4. 支持 **stream 模式**，可以通过流式传输实现打字机效果。
@@ -144,7 +145,7 @@ Nginx 的参考配置：
 ```
 server{
    server_name openai.justsong.cn;  # 请根据实际情况修改你的域名
-   
+
    location / {
           client_max_body_size  64m;
           proxy_http_version 1.1;
@@ -189,12 +190,12 @@ docker-compose ps
 1. 从 [GitHub Releases](https://github.com/songquanpeng/one-api/releases/latest) 下载可执行文件或者从源码编译：
    ```shell
    git clone https://github.com/songquanpeng/one-api.git
-   
+
    # 构建前端
    cd one-api/web/default
    npm install
    npm run build
-   
+
    # 构建后端
    cd ../..
    go mod download
@@ -321,7 +322,7 @@ Render 可以直接部署 docker 镜像，不需要 fork 仓库：https://dashbo
 例如对于 OpenAI 的官方库：
 ```bash
 OPENAI_API_KEY="sk-xxxxxx"
-OPENAI_API_BASE="https://<HOST>:<PORT>/v1" 
+OPENAI_API_BASE="https://<HOST>:<PORT>/v1"
 ```
 
 ```mermaid
@@ -369,33 +370,34 @@ graph LR
    + 例子：`NODE_TYPE=slave`
 9. `CHANNEL_UPDATE_FREQUENCY`：设置之后将定期更新渠道余额，单位为分钟，未设置则不进行更新。
    + 例子：`CHANNEL_UPDATE_FREQUENCY=1440`
-10. `CHANNEL_TEST_FREQUENCY`：设置之后将定期检查渠道，单位为分钟，未设置则不进行检查。
+10. `CHANNEL_TEST_FREQUENCY`：设置之后将定期检查渠道，单位为分钟，未设置则不进行检查。 
-11. 例子：`CHANNEL_TEST_FREQUENCY=1440`
+   +例子：`CHANNEL_TEST_FREQUENCY=1440`
-12. `POLLING_INTERVAL`：批量更新渠道余额以及测试可用性时的请求间隔，单位为秒，默认无间隔。
+11. `POLLING_INTERVAL`：批量更新渠道余额以及测试可用性时的请求间隔，单位为秒，默认无间隔。
     + 例子：`POLLING_INTERVAL=5`
-13. `BATCH_UPDATE_ENABLED`：启用数据库批量更新聚合，会导致用户额度的更新存在一定的延迟可选值为 `true` 和 `false`，未设置则默认为 `false`。
+12. `BATCH_UPDATE_ENABLED`：启用数据库批量更新聚合，会导致用户额度的更新存在一定的延迟可选值为 `true` 和 `false`，未设置则默认为 `false`。
     + 例子：`BATCH_UPDATE_ENABLED=true`
     + 如果你遇到了数据库连接数过多的问题，可以尝试启用该选项。
-14. `BATCH_UPDATE_INTERVAL=5`：批量更新聚合的时间间隔，单位为秒，默认为 `5`。
+13. `BATCH_UPDATE_INTERVAL=5`：批量更新聚合的时间间隔，单位为秒，默认为 `5`。
     + 例子：`BATCH_UPDATE_INTERVAL=5`
-15. 请求频率限制：
+14. 请求频率限制：
     + `GLOBAL_API_RATE_LIMIT`：全局 API 速率限制（除中继请求外），单 ip 三分钟内的最大请求数，默认为 `180`。
     + `GLOBAL_WEB_RATE_LIMIT`：全局 Web 速率限制，单 ip 三分钟内的最大请求数，默认为 `60`。
-16. 编码器缓存设置：
+15. 编码器缓存设置：
     + `TIKTOKEN_CACHE_DIR`：默认程序启动时会联网下载一些通用的词元的编码，如：`gpt-3.5-turbo`，在一些网络环境不稳定，或者离线情况，可能会导致启动有问题，可以配置此目录缓存数据，可迁移到离线环境。
     + `DATA_GYM_CACHE_DIR`：目前该配置作用与 `TIKTOKEN_CACHE_DIR` 一致，但是优先级没有它高。
-17. `RELAY_TIMEOUT`：中继超时设置，单位为秒，默认不设置超时时间。
+16. `RELAY_TIMEOUT`：中继超时设置，单位为秒，默认不设置超时时间。
-18. `RELAY_PROXY`：设置后使用该代理来请求 API。
+17. `RELAY_PROXY`：设置后使用该代理来请求 API。
-19. `USER_CONTENT_REQUEST_TIMEOUT`：用户上传内容下载超时时间，单位为秒。
+18. `USER_CONTENT_REQUEST_TIMEOUT`：用户上传内容下载超时时间，单位为秒。
-20. `USER_CONTENT_REQUEST_PROXY`：设置后使用该代理来请求用户上传的内容，例如图片。
+19. `USER_CONTENT_REQUEST_PROXY`：设置后使用该代理来请求用户上传的内容，例如图片。
-21. `SQLITE_BUSY_TIMEOUT`：SQLite 锁等待超时设置，单位为毫秒，默认 `3000`。
+20. `SQLITE_BUSY_TIMEOUT`：SQLite 锁等待超时设置，单位为毫秒，默认 `3000`。
-22. `GEMINI_SAFETY_SETTING`：Gemini 的安全设置，默认 `BLOCK_NONE`。
+21. `GEMINI_SAFETY_SETTING`：Gemini 的安全设置，默认 `BLOCK_NONE`。
-23. `GEMINI_VERSION`：One API 所使用的 Gemini 版本，默认为 `v1`。
+22. `GEMINI_VERSION`：One API 所使用的 Gemini 版本，默认为 `v1`。
-24. `THEME`：系统的主题设置，默认为 `default`，具体可选值参考[此处](./web/README.md)。
+23. `THEME`：系统的主题设置，默认为 `default`，具体可选值参考[此处](./web/README.md)。
-25. `ENABLE_METRIC`：是否根据请求成功率禁用渠道，默认不开启，可选值为 `true` 和 `false`。
+24. `ENABLE_METRIC`：是否根据请求成功率禁用渠道，默认不开启，可选值为 `true` 和 `false`。
-26. `METRIC_QUEUE_SIZE`：请求成功率统计队列大小，默认为 `10`。
+25. `METRIC_QUEUE_SIZE`：请求成功率统计队列大小，默认为 `10`。
-27. `METRIC_SUCCESS_RATE_THRESHOLD`：请求成功率阈值，默认为 `0.8`。
+26. `METRIC_SUCCESS_RATE_THRESHOLD`：请求成功率阈值，默认为 `0.8`。
-28. `INITIAL_ROOT_TOKEN`：如果设置了该值，则在系统首次启动时会自动创建一个值为该环境变量值的 root 用户令牌。
+27. `INITIAL_ROOT_TOKEN`：如果设置了该值，则在系统首次启动时会自动创建一个值为该环境变量值的 root 用户令牌。
+28. `INITIAL_ROOT_ACCESS_TOKEN`：如果设置了该值，则在系统首次启动时会自动创建一个值为该环境变量的 root 用户创建系统管理令牌。
 
 ### 命令行参数
 1. `--port <port_number>`: 指定服务器监听的端口号，默认为 `3000`。
@@ -448,6 +450,8 @@ https://openai.justsong.cn
 ## 相关项目
 * [FastGPT](https://github.com/labring/FastGPT): 基于 LLM 大语言模型的知识库问答系统
 * [ChatGPT Next Web](https://github.com/Yidadaa/ChatGPT-Next-Web):  一键拥有你自己的跨平台 ChatGPT 应用
+* [VChart](https://github.com/VisActor/VChart):  不只是开箱即用的多端图表库，更是生动灵活的数据故事讲述者。
+* [VMind](https://github.com/VisActor/VMind):  不仅自动，还很智能。开源智能可视化解决方案。
 
 ## 注意
 

common/config/config.go

@@ -143,8 +143,12 @@ var MetricFailChanSize = env.Int("METRIC_FAIL_CHAN_SIZE", 128)
 
 var InitialRootToken = os.Getenv("INITIAL_ROOT_TOKEN")
 
+var InitialRootAccessToken = os.Getenv("INITIAL_ROOT_ACCESS_TOKEN")
+
 var GeminiVersion = env.String("GEMINI_VERSION", "v1")
 
+var OnlyOneLogFile = env.Bool("ONLY_ONE_LOG_FILE", false)
+
 var RelayProxy = env.String("RELAY_PROXY", "")
 var UserContentRequestProxy = env.String("USER_CONTENT_REQUEST_PROXY", "")
 var UserContentRequestTimeout = env.Int("USER_CONTENT_REQUEST_TIMEOUT", 30)

common/logger/logger.go

@@ -27,7 +27,12 @@ var setupLogOnce sync.Once
 func SetupLogger() {
 	setupLogOnce.Do(func() {
 		if LogDir != "" {
-			logPath := filepath.Join(LogDir, fmt.Sprintf("oneapi-%s.log", time.Now().Format("20060102")))
+			var logPath string
+			if config.OnlyOneLogFile {
+				logPath = filepath.Join(LogDir, "oneapi.log")
+			} else {
+				logPath = filepath.Join(LogDir, fmt.Sprintf("oneapi-%s.log", time.Now().Format("20060102")))
+			}
 			fd, err := os.OpenFile(logPath, os.O_APPEND|os.O_CREATE|os.O_WRONLY, 0644)
 			if err != nil {
 				log.Fatal("failed to open log file")

common/message/email.go

@@ -6,11 +6,16 @@ import (
 	"encoding/base64"
 	"fmt"
 	"github.com/songquanpeng/one-api/common/config"
+	"net"
 	"net/smtp"
 	"strings"
 	"time"
 )
 
+func shouldAuth() bool {
+	return config.SMTPAccount != "" || config.SMTPToken != ""
+}
+
 func SendEmail(subject string, receiver string, content string) error {
 	if receiver == "" {
 		return fmt.Errorf("receiver is empty")
@@ -41,16 +46,24 @@ func SendEmail(subject string, receiver string, content string) error {
 		"Date: %s\r\n"+
 		"Content-Type: text/html; charset=UTF-8\r\n\r\n%s\r\n",
 		receiver, config.SystemName, config.SMTPFrom, encodedSubject, messageId, time.Now().Format(time.RFC1123Z), content))
+
 	auth := smtp.PlainAuth("", config.SMTPAccount, config.SMTPToken, config.SMTPServer)
 	addr := fmt.Sprintf("%s:%d", config.SMTPServer, config.SMTPPort)
 	to := strings.Split(receiver, ";")
 
-	if config.SMTPPort == 465 {
+	if config.SMTPPort == 465 || !shouldAuth() {
-		tlsConfig := &tls.Config{
+		// need advanced client
-			InsecureSkipVerify: true,
+		var conn net.Conn
-			ServerName:         config.SMTPServer,
+		var err error
+		if config.SMTPPort == 465 {
+			tlsConfig := &tls.Config{
+				InsecureSkipVerify: true,
+				ServerName:         config.SMTPServer,
+			}
+			conn, err = tls.Dial("tcp", fmt.Sprintf("%s:%d", config.SMTPServer, config.SMTPPort), tlsConfig)
+		} else {
+			conn, err = net.Dial("tcp", fmt.Sprintf("%s:%d", config.SMTPServer, config.SMTPPort))
 		}
-		conn, err := tls.Dial("tcp", fmt.Sprintf("%s:%d", config.SMTPServer, config.SMTPPort), tlsConfig)
 		if err != nil {
 			return err
 		}
@@ -59,8 +72,10 @@ func SendEmail(subject string, receiver string, content string) error {
 			return err
 		}
 		defer client.Close()
-		if err = client.Auth(auth); err != nil {
+		if shouldAuth() {
-			return err
+			if err = client.Auth(auth); err != nil {
+				return err
+			}
 		}
 		if err = client.Mail(config.SMTPFrom); err != nil {
 			return err

controller/channel-test.go

@@ -14,6 +14,7 @@ import (
 	"sync"
 	"time"
 
+	"github.com/gin-gonic/gin"
 	"github.com/songquanpeng/one-api/common/config"
 	"github.com/songquanpeng/one-api/common/ctxkey"
 	"github.com/songquanpeng/one-api/common/logger"
@@ -27,15 +28,15 @@ import (
 	"github.com/songquanpeng/one-api/relay/meta"
 	relaymodel "github.com/songquanpeng/one-api/relay/model"
 	"github.com/songquanpeng/one-api/relay/relaymode"
-
-	"github.com/gin-gonic/gin"
 )
 
-func buildTestRequest() *relaymodel.GeneralOpenAIRequest {
+func buildTestRequest(model string) *relaymodel.GeneralOpenAIRequest {
+	if model == "" {
+		model = "gpt-3.5-turbo"
+	}
 	testRequest := &relaymodel.GeneralOpenAIRequest{
 		MaxTokens: 2,
-		Stream:    false,
+		Model:     model,
-		Model:     "gpt-3.5-turbo",
 	}
 	testMessage := relaymodel.Message{
 		Role:    "user",
@@ -45,7 +46,7 @@ func buildTestRequest() *relaymodel.GeneralOpenAIRequest {
 	return testRequest
 }
 
-func testChannel(channel *model.Channel) (err error, openaiErr *relaymodel.Error) {
+func testChannel(channel *model.Channel, request *relaymodel.GeneralOpenAIRequest) (err error, openaiErr *relaymodel.Error) {
 	w := httptest.NewRecorder()
 	c, _ := gin.CreateTestContext(w)
 	c.Request = &http.Request{
@@ -68,12 +69,8 @@ func testChannel(channel *model.Channel) (err error, openaiErr *relaymodel.Error
 		return fmt.Errorf("invalid api type: %d, adaptor is nil", apiType), nil
 	}
 	adaptor.Init(meta)
-	var modelName string
+	modelName := request.Model
-	modelList := adaptor.GetModelList()
 	modelMap := channel.GetModelMapping()
-	if len(modelList) != 0 {
-		modelName = modelList[0]
-	}
 	if modelName == "" || !strings.Contains(channel.Models, modelName) {
 		modelNames := strings.Split(channel.Models, ",")
 		if len(modelNames) > 0 {
@@ -83,9 +80,8 @@ func testChannel(channel *model.Channel) (err error, openaiErr *relaymodel.Error
 			modelName = modelMap[modelName]
 		}
 	}
-	request := buildTestRequest()
+	meta.OriginModelName, meta.ActualModelName = request.Model, modelName
 	request.Model = modelName
-	meta.OriginModelName, meta.ActualModelName = modelName, modelName
 	convertedRequest, err := adaptor.ConvertRequest(c, relaymode.ChatCompletions, request)
 	if err != nil {
 		return err, nil
@@ -139,10 +135,15 @@ func TestChannel(c *gin.Context) {
 		})
 		return
 	}
+	model := c.Query("model")
+	testRequest := buildTestRequest(model)
 	tik := time.Now()
-	err, _ = testChannel(channel)
+	err, _ = testChannel(channel, testRequest)
 	tok := time.Now()
 	milliseconds := tok.Sub(tik).Milliseconds()
+	if err != nil {
+		milliseconds = 0
+	}
 	go channel.UpdateResponseTime(milliseconds)
 	consumedTime := float64(milliseconds) / 1000.0
 	if err != nil {
@@ -150,6 +151,7 @@ func TestChannel(c *gin.Context) {
 			"success": false,
 			"message": err.Error(),
 			"time":    consumedTime,
+			"model":   model,
 		})
 		return
 	}
@@ -157,6 +159,7 @@ func TestChannel(c *gin.Context) {
 		"success": true,
 		"message": "",
 		"time":    consumedTime,
+		"model":   model,
 	})
 	return
 }
@@ -187,11 +190,12 @@ func testChannels(notify bool, scope string) error {
 		for _, channel := range channels {
 			isChannelEnabled := channel.Status == model.ChannelStatusEnabled
 			tik := time.Now()
-			err, openaiErr := testChannel(channel)
+			testRequest := buildTestRequest("")
+			err, openaiErr := testChannel(channel, testRequest)
 			tok := time.Now()
 			milliseconds := tok.Sub(tik).Milliseconds()
 			if isChannelEnabled && milliseconds > disableThreshold {
-				err = errors.New(fmt.Sprintf("响应时间 %.2fs 超过阈值 %.2fs", float64(milliseconds)/1000.0, float64(disableThreshold)/1000.0))
+				err = fmt.Errorf("响应时间 %.2fs 超过阈值 %.2fs", float64(milliseconds)/1000.0, float64(disableThreshold)/1000.0)
 				if config.AutomaticDisableChannelEnabled {
 					monitor.DisableChannel(channel.Id, channel.Name, err.Error())
 				} else {

go.mod

@@ -4,6 +4,7 @@ module github.com/songquanpeng/one-api
 go 1.20
 
 require (
+	cloud.google.com/go/iam v1.1.10
 	github.com/aws/aws-sdk-go-v2 v1.27.0
 	github.com/aws/aws-sdk-go-v2/credentials v1.17.15
 	github.com/aws/aws-sdk-go-v2/service/bedrockruntime v1.8.3
@@ -19,12 +20,14 @@ require (
 	github.com/gorilla/websocket v1.5.1
 	github.com/jinzhu/copier v0.4.0
 	github.com/joho/godotenv v1.5.1
+	github.com/patrickmn/go-cache v2.1.0+incompatible
 	github.com/pkg/errors v0.9.1
 	github.com/pkoukk/tiktoken-go v0.1.7
 	github.com/smartystreets/goconvey v1.8.1
 	github.com/stretchr/testify v1.9.0
-	golang.org/x/crypto v0.23.0
+	golang.org/x/crypto v0.24.0
 	golang.org/x/image v0.18.0
+	google.golang.org/api v0.187.0
 	gorm.io/driver/mysql v1.5.6
 	gorm.io/driver/postgres v1.5.7
 	gorm.io/driver/sqlite v1.5.5
@@ -32,6 +35,9 @@ require (
 )
 
 require (
+	cloud.google.com/go/auth v0.6.1 // indirect
+	cloud.google.com/go/auth/oauth2adapt v0.2.2 // indirect
+	cloud.google.com/go/compute/metadata v0.3.0 // indirect
 	filippo.io/edwards25519 v1.1.0 // indirect
 	github.com/aws/aws-sdk-go-v2/aws/protocol/eventstream v1.6.2 // indirect
 	github.com/aws/aws-sdk-go-v2/internal/configsources v1.3.7 // indirect
@@ -45,13 +51,21 @@ require (
 	github.com/davecgh/go-spew v1.1.1 // indirect
 	github.com/dgryski/go-rendezvous v0.0.0-20200823014737-9f7001d12a5f // indirect
 	github.com/dlclark/regexp2 v1.11.0 // indirect
+	github.com/felixge/httpsnoop v1.0.4 // indirect
 	github.com/fsnotify/fsnotify v1.7.0 // indirect
 	github.com/gabriel-vasile/mimetype v1.4.3 // indirect
 	github.com/gin-contrib/sse v0.1.0 // indirect
+	github.com/go-logr/logr v1.4.1 // indirect
+	github.com/go-logr/stdr v1.2.2 // indirect
 	github.com/go-playground/locales v0.14.1 // indirect
 	github.com/go-playground/universal-translator v0.18.1 // indirect
 	github.com/go-sql-driver/mysql v1.8.1 // indirect
 	github.com/goccy/go-json v0.10.3 // indirect
+	github.com/golang/groupcache v0.0.0-20210331224755-41bb18bfe9da // indirect
+	github.com/golang/protobuf v1.5.4 // indirect
+	github.com/google/s2a-go v0.1.7 // indirect
+	github.com/googleapis/enterprise-certificate-proxy v0.3.2 // indirect
+	github.com/googleapis/gax-go/v2 v2.12.5 // indirect
 	github.com/gopherjs/gopherjs v1.17.2 // indirect
 	github.com/gorilla/context v1.1.2 // indirect
 	github.com/gorilla/securecookie v1.1.2 // indirect
@@ -68,7 +82,7 @@ require (
 	github.com/kr/text v0.2.0 // indirect
 	github.com/leodido/go-urn v1.4.0 // indirect
 	github.com/mattn/go-isatty v0.0.20 // indirect
-	github.com/mattn/go-sqlite3 v2.0.3+incompatible // indirect
+	github.com/mattn/go-sqlite3 v1.14.22 // indirect
 	github.com/modern-go/concurrent v0.0.0-20180306012644-bacd9c7ef1dd // indirect
 	github.com/modern-go/reflect2 v1.0.2 // indirect
 	github.com/pelletier/go-toml/v2 v2.2.2 // indirect
@@ -76,11 +90,22 @@ require (
 	github.com/smarty/assertions v1.15.0 // indirect
 	github.com/twitchyliquid64/golang-asm v0.15.1 // indirect
 	github.com/ugorji/go/codec v1.2.12 // indirect
+	go.opencensus.io v0.24.0 // indirect
+	go.opentelemetry.io/contrib/instrumentation/google.golang.org/grpc/otelgrpc v0.49.0 // indirect
+	go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp v0.49.0 // indirect
+	go.opentelemetry.io/otel v1.24.0 // indirect
+	go.opentelemetry.io/otel/metric v1.24.0 // indirect
+	go.opentelemetry.io/otel/trace v1.24.0 // indirect
 	golang.org/x/arch v0.8.0 // indirect
-	golang.org/x/net v0.25.0 // indirect
+	golang.org/x/net v0.26.0 // indirect
+	golang.org/x/oauth2 v0.21.0 // indirect
 	golang.org/x/sync v0.7.0 // indirect
-	golang.org/x/sys v0.20.0 // indirect
+	golang.org/x/sys v0.21.0 // indirect
 	golang.org/x/text v0.16.0 // indirect
-	google.golang.org/protobuf v1.34.1 // indirect
+	golang.org/x/time v0.5.0 // indirect
+	google.golang.org/genproto/googleapis/api v0.0.0-20240617180043-68d350f18fd4 // indirect
+	google.golang.org/genproto/googleapis/rpc v0.0.0-20240624140628-dc46fd24d27d // indirect
+	google.golang.org/grpc v1.64.1 // indirect
+	google.golang.org/protobuf v1.34.2 // indirect
 	gopkg.in/yaml.v3 v3.0.1 // indirect
 )

go.sum

@@ -1,5 +1,15 @@
+cloud.google.com/go v0.26.0/go.mod h1:aQUYkXzVsufM+DwF1aE+0xfcU+56JwCaLick0ClmMTw=
+cloud.google.com/go/auth v0.6.1 h1:T0Zw1XM5c1GlpN2HYr2s+m3vr1p2wy+8VN+Z1FKxW38=
+cloud.google.com/go/auth v0.6.1/go.mod h1:eFHG7zDzbXHKmjJddFG/rBlcGp6t25SwRUiEQSlO4x4=
+cloud.google.com/go/auth/oauth2adapt v0.2.2 h1:+TTV8aXpjeChS9M+aTtN/TjdQnzJvmzKFt//oWu7HX4=
+cloud.google.com/go/auth/oauth2adapt v0.2.2/go.mod h1:wcYjgpZI9+Yu7LyYBg4pqSiaRkfEK3GQcpb7C/uyF1Q=
+cloud.google.com/go/compute/metadata v0.3.0 h1:Tz+eQXMEqDIKRsmY3cHTL6FVaynIjX2QxYC4trgAKZc=
+cloud.google.com/go/compute/metadata v0.3.0/go.mod h1:zFmK7XCadkQkj6TtorcaGlCW1hT1fIilQDwofLpJ20k=
+cloud.google.com/go/iam v1.1.10 h1:ZSAr64oEhQSClwBL670MsJAW5/RLiC6kfw3Bqmd5ZDI=
+cloud.google.com/go/iam v1.1.10/go.mod h1:iEgMq62sg8zx446GCaijmA2Miwg5o3UbO+nI47WHJps=
 filippo.io/edwards25519 v1.1.0 h1:FNf4tywRC1HmFuKW5xopWpigGjJKiJSV0Cqo0cJWDaA=
 filippo.io/edwards25519 v1.1.0/go.mod h1:BxyFTGdWcka3PhytdK4V28tE5sGfRvvvRV7EaN4VDT4=
+github.com/BurntSushi/toml v0.3.1/go.mod h1:xHWCNGjB5oqiDr8zfno3MHue2Ht5sIBksp03qcyfWMU=
 github.com/aws/aws-sdk-go-v2 v1.27.0 h1:7bZWKoXhzI+mMR/HjdMx8ZCC5+6fY0lS5tr0bbgiLlo=
 github.com/aws/aws-sdk-go-v2 v1.27.0/go.mod h1:ffIFB97e2yNsv4aTSGkqtHnppsIJzw7G7BReUZ3jCXM=
 github.com/aws/aws-sdk-go-v2/aws/protocol/eventstream v1.6.2 h1:x6xsQXGSmW6frevwDA+vi/wqhp1ct18mVXYN08/93to=
@@ -18,12 +28,15 @@ github.com/bytedance/sonic v1.11.6 h1:oUp34TzMlL+OY1OUWxHqsdkgC/Zfc85zGqw9siXjrc
 github.com/bytedance/sonic v1.11.6/go.mod h1:LysEHSvpvDySVdC2f87zGWf6CIKJcAvqab1ZaiQtds4=
 github.com/bytedance/sonic/loader v0.1.1 h1:c+e5Pt1k/cy5wMveRDyk2X4B9hF4g7an8N3zCYjJFNM=
 github.com/bytedance/sonic/loader v0.1.1/go.mod h1:ncP89zfokxS5LZrJxl5z0UJcsk4M4yY2JpfqGeCtNLU=
+github.com/census-instrumentation/opencensus-proto v0.2.1/go.mod h1:f6KPmirojxKA12rnyqOA5BBL4O983OfeGPqjHWSTneU=
 github.com/cespare/xxhash/v2 v2.3.0 h1:UL815xU9SqsFlibzuggzjXhog7bL6oX9BbNZnL2UFvs=
 github.com/cespare/xxhash/v2 v2.3.0/go.mod h1:VGX0DQ3Q6kWi7AoAeZDth3/j3BFtOZR5XLFGgcrjCOs=
+github.com/client9/misspell v0.3.4/go.mod h1:qj6jICC3Q7zFZvVWo7KLAzC3yx5G7kyvSDkc90ppPyw=
 github.com/cloudwego/base64x v0.1.4 h1:jwCgWpFanWmN8xoIUHa2rtzmkd5J2plF/dnLS6Xd/0Y=
 github.com/cloudwego/base64x v0.1.4/go.mod h1:0zlkT4Wn5C6NdauXdJRhSKRlJvmclQ1hhJgA0rcu/8w=
 github.com/cloudwego/iasm v0.2.0 h1:1KNIy1I1H9hNNFEEH3DVnI4UujN+1zjpuk6gwHLTssg=
 github.com/cloudwego/iasm v0.2.0/go.mod h1:8rXZaNYT2n95jn+zTI1sDr+IgcD2GVs0nlbbQPiEFhY=
+github.com/cncf/udpa/go v0.0.0-20191209042840-269d4d468f6f/go.mod h1:M8M6+tZqaGXZJjfX53e64911xZQV5JYwmTeXPW+k8Sc=
 github.com/creack/pty v1.1.9/go.mod h1:oKZEueFk5CKHvIhNR5MUki03XCEU+Q6VDXinZuGJ33E=
 github.com/davecgh/go-spew v1.1.0/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
 github.com/davecgh/go-spew v1.1.1 h1:vj9j/u1bqnvCEfJOwUhtlOARqs3+rkHYY13jYWTU97c=
@@ -32,6 +45,12 @@ github.com/dgryski/go-rendezvous v0.0.0-20200823014737-9f7001d12a5f h1:lO4WD4F/r
 github.com/dgryski/go-rendezvous v0.0.0-20200823014737-9f7001d12a5f/go.mod h1:cuUVRXasLTGF7a8hSLbxyZXjz+1KgoB3wDUb6vlszIc=
 github.com/dlclark/regexp2 v1.11.0 h1:G/nrcoOa7ZXlpoa/91N3X7mM3r8eIlMBBJZvsz/mxKI=
 github.com/dlclark/regexp2 v1.11.0/go.mod h1:DHkYz0B9wPfa6wondMfaivmHpzrQ3v9q8cnmRbL6yW8=
+github.com/envoyproxy/go-control-plane v0.9.0/go.mod h1:YTl/9mNaCwkRvm6d1a2C3ymFceY/DCBVvsKhRF0iEA4=
+github.com/envoyproxy/go-control-plane v0.9.1-0.20191026205805-5f8ba28d4473/go.mod h1:YTl/9mNaCwkRvm6d1a2C3ymFceY/DCBVvsKhRF0iEA4=
+github.com/envoyproxy/go-control-plane v0.9.4/go.mod h1:6rpuAdCZL397s3pYoYcLgu1mIlRU8Am5FuJP05cCM98=
+github.com/envoyproxy/protoc-gen-validate v0.1.0/go.mod h1:iSmxcyjqTsJpI2R4NaDN7+kN2VEUnK/pcBlmesArF7c=
+github.com/felixge/httpsnoop v1.0.4 h1:NFTV2Zj1bL4mc9sqWACXbQFVBBg2W3GPvqp8/ESS2Wg=
+github.com/felixge/httpsnoop v1.0.4/go.mod h1:m8KPJKqk1gH5J9DgRY2ASl2lWCfGKXixSwevea8zH2U=
 github.com/fsnotify/fsnotify v1.7.0 h1:8JEhPFa5W2WU7YfeZzPNqzMP6Lwt7L2715Ggo0nosvA=
 github.com/fsnotify/fsnotify v1.7.0/go.mod h1:40Bi/Hjc2AVfZrqy+aj+yEI+/bRxZnMJyTJwOpGvigM=
 github.com/gabriel-vasile/mimetype v1.4.3 h1:in2uUcidCuFcDKtdcBxlR0rJ1+fsokWf+uqxgUFjbI0=
@@ -48,6 +67,11 @@ github.com/gin-contrib/static v1.1.2 h1:c3kT4bFkUJn2aoRU3s6XnMjJT8J6nNWJkR0Nglqm
 github.com/gin-contrib/static v1.1.2/go.mod h1:Fw90ozjHCmZBWbgrsqrDvO28YbhKEKzKp8GixhR4yLw=
 github.com/gin-gonic/gin v1.10.0 h1:nTuyha1TYqgedzytsKYqna+DfLos46nTv2ygFy86HFU=
 github.com/gin-gonic/gin v1.10.0/go.mod h1:4PMNQiOhvDRa013RKVbsiNwoyezlm2rm0uX/T7kzp5Y=
+github.com/go-logr/logr v1.2.2/go.mod h1:jdQByPbusPIv2/zmleS9BjJVeZ6kBagPoEUsqbVz/1A=
+github.com/go-logr/logr v1.4.1 h1:pKouT5E8xu9zeFC39JXRDukb6JFQPXM5p5I91188VAQ=
+github.com/go-logr/logr v1.4.1/go.mod h1:9T104GzyrTigFIr8wt5mBrctHMim0Nb2HLGrmQ40KvY=
+github.com/go-logr/stdr v1.2.2 h1:hSWxHoqTgW2S2qGc0LTAI563KZ5YKYRhT3MFKZMbjag=
+github.com/go-logr/stdr v1.2.2/go.mod h1:mMo/vtBO5dYbehREoey6XUKy/eSumjCCveDpRre4VKE=
 github.com/go-playground/assert/v2 v2.2.0 h1:JvknZsQTYeFEAhQwI4qEt9cyV5ONwRHC+lYKSsYSR8s=
 github.com/go-playground/locales v0.14.1 h1:EWaQ/wswjilfKLTECiXz7Rh+3BjFhfDFKv/oXslEjJA=
 github.com/go-playground/locales v0.14.1/go.mod h1:hxrqLVvrK65+Rwrd5Fc6F2O76J/NuW9t0sjnWqG1slY=
@@ -64,11 +88,40 @@ github.com/goccy/go-json v0.10.3 h1:KZ5WoDbxAIgm2HNbYckL0se1fHD6rz5j4ywS6ebzDqA=
 github.com/goccy/go-json v0.10.3/go.mod h1:oq7eo15ShAhp70Anwd5lgX2pLfOS3QCiwU/PULtXL6M=
 github.com/golang-jwt/jwt v3.2.2+incompatible h1:IfV12K8xAKAnZqdXVzCZ+TOjboZ2keLg81eXfW3O+oY=
 github.com/golang-jwt/jwt v3.2.2+incompatible/go.mod h1:8pz2t5EyA70fFQQSrl6XZXzqecmYZeUEB8OUGHkxJ+I=
-github.com/google/go-cmp v0.5.6 h1:BKbKCqvP6I+rmFHt06ZmyQtvB8xAkWdhFyr0ZUNZcxQ=
+github.com/golang/glog v0.0.0-20160126235308-23def4e6c14b/go.mod h1:SBH7ygxi8pfUlaOkMMuAQtPIUF8ecWP5IEl/CR7VP2Q=
+github.com/golang/groupcache v0.0.0-20200121045136-8c9f03a8e57e/go.mod h1:cIg4eruTrX1D+g88fzRXU5OdNfaM+9IcxsU14FzY7Hc=
+github.com/golang/groupcache v0.0.0-20210331224755-41bb18bfe9da h1:oI5xCqsCo564l8iNU+DwB5epxmsaqB+rhGL0m5jtYqE=
+github.com/golang/groupcache v0.0.0-20210331224755-41bb18bfe9da/go.mod h1:cIg4eruTrX1D+g88fzRXU5OdNfaM+9IcxsU14FzY7Hc=
+github.com/golang/mock v1.1.1/go.mod h1:oTYuIxOrZwtPieC+H1uAHpcLFnEyAGVDL/k47Jfbm0A=
+github.com/golang/protobuf v1.2.0/go.mod h1:6lQm79b+lXiMfvg/cZm0SGofjICqVBUtrP5yJMmIC1U=
+github.com/golang/protobuf v1.3.2/go.mod h1:6lQm79b+lXiMfvg/cZm0SGofjICqVBUtrP5yJMmIC1U=
+github.com/golang/protobuf v1.4.0-rc.1/go.mod h1:ceaxUfeHdC40wWswd/P6IGgMaK3YpKi5j83Wpe3EHw8=
+github.com/golang/protobuf v1.4.0-rc.1.0.20200221234624-67d41d38c208/go.mod h1:xKAWHe0F5eneWXFV3EuXVDTCmh+JuBKY0li0aMyXATA=
+github.com/golang/protobuf v1.4.0-rc.2/go.mod h1:LlEzMj4AhA7rCAGe4KMBDvJI+AwstrUpVNzEA03Pprs=
+github.com/golang/protobuf v1.4.0-rc.4.0.20200313231945-b860323f09d0/go.mod h1:WU3c8KckQ9AFe+yFwt9sWVRKCVIyN9cPHBJSNnbL67w=
+github.com/golang/protobuf v1.4.0/go.mod h1:jodUvKwWbYaEsadDk5Fwe5c77LiNKVO9IDvqG2KuDX0=
+github.com/golang/protobuf v1.4.1/go.mod h1:U8fpvMrcmy5pZrNK1lt4xCsGvpyWQ/VVv6QDs8UjoX8=
+github.com/golang/protobuf v1.4.3/go.mod h1:oDoupMAO8OvCJWAcko0GGGIgR6R6ocIYbsSw735rRwI=
+github.com/golang/protobuf v1.5.4 h1:i7eJL8qZTpSEXOPTxNKhASYpMn+8e5Q6AdndVa1dWek=
+github.com/golang/protobuf v1.5.4/go.mod h1:lnTiLA8Wa4RWRcIUkrtSVa5nRhsEGBg48fD6rSs7xps=
+github.com/google/go-cmp v0.2.0/go.mod h1:oXzfMopK8JAjlY9xF4vHSVASa0yLyX7SntLO5aqRK0M=
+github.com/google/go-cmp v0.3.0/go.mod h1:8QqcDgzrUqlUb/G2PQTWiueGozuR1884gddMywk6iLU=
+github.com/google/go-cmp v0.3.1/go.mod h1:8QqcDgzrUqlUb/G2PQTWiueGozuR1884gddMywk6iLU=
+github.com/google/go-cmp v0.4.0/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
+github.com/google/go-cmp v0.5.0/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
+github.com/google/go-cmp v0.5.3/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
+github.com/google/go-cmp v0.6.0 h1:ofyhxvXcZhMsU5ulbFiLKl/XBFqE1GSq7atu8tAmTRI=
 github.com/google/gofuzz v1.0.0/go.mod h1:dBl0BpW6vV/+mYPU4Po3pmUjxk6FQPldtuIdl/M65Eg=
 github.com/google/gofuzz v1.2.0 h1:xRy4A+RhZaiKjJ1bPfwQ8sedCA+YS2YcCHW6ec7JMi0=
+github.com/google/s2a-go v0.1.7 h1:60BLSyTrOV4/haCDW4zb1guZItoSq8foHCXrAnjBo/o=
+github.com/google/s2a-go v0.1.7/go.mod h1:50CgR4k1jNlWBu4UfS4AcfhVe1r6pdZPygJ3R8F0Qdw=
+github.com/google/uuid v1.1.2/go.mod h1:TIyPZe4MgqvfeYDBFedMoGGpEw/LqOeaOT+nhxU+yHo=
 github.com/google/uuid v1.6.0 h1:NIvaJDMOsjHA8n1jAhLSgzrAzy1Hgr+hNrb57e+94F0=
 github.com/google/uuid v1.6.0/go.mod h1:TIyPZe4MgqvfeYDBFedMoGGpEw/LqOeaOT+nhxU+yHo=
+github.com/googleapis/enterprise-certificate-proxy v0.3.2 h1:Vie5ybvEvT75RniqhfFxPRy3Bf7vr3h0cechB90XaQs=
+github.com/googleapis/enterprise-certificate-proxy v0.3.2/go.mod h1:VLSiSSBs/ksPL8kq3OBOQ6WRI2QnaFynd1DCjZ62+V0=
+github.com/googleapis/gax-go/v2 v2.12.5 h1:8gw9KZK8TiVKB6q3zHY3SBzLnrGp6HQjyfYBYGmXdxA=
+github.com/googleapis/gax-go/v2 v2.12.5/go.mod h1:BUDKcWo+RaKq5SC9vVYL0wLADa3VcfswbOMMRmB9H3E=
 github.com/gopherjs/gopherjs v1.17.2 h1:fQnZVsXk8uxXIStYb0N4bGk7jeyTalG/wsZjQ25dO0g=
 github.com/gopherjs/gopherjs v1.17.2/go.mod h1:pRRIvn/QzFLrKfvEz3qUuEhtE/zLCWfreZ6J5gM2i+k=
 github.com/gorilla/context v1.1.2 h1:WRkNAv2uoa03QNIc1A6u4O7DAGMUVoopZhkiXWA2V1o=
@@ -110,8 +163,8 @@ github.com/leodido/go-urn v1.4.0 h1:WT9HwE9SGECu3lg4d/dIA+jxlljEa1/ffXKmRjqdmIQ=
 github.com/leodido/go-urn v1.4.0/go.mod h1:bvxc+MVxLKB4z00jd1z+Dvzr47oO32F/QSNjSBOlFxI=
 github.com/mattn/go-isatty v0.0.20 h1:xfD0iDuEKnDkl03q4limB+vH+GxLEtL/jb4xVJSWWEY=
 github.com/mattn/go-isatty v0.0.20/go.mod h1:W+V8PltTTMOvKvAeJH7IuucS94S2C6jfK/D7dTCTo3Y=
-github.com/mattn/go-sqlite3 v2.0.3+incompatible h1:gXHsfypPkaMZrKbD5209QV9jbUTJKjyR5WD3HYQSd+U=
+github.com/mattn/go-sqlite3 v1.14.22 h1:2gZY6PC6kBnID23Tichd1K+Z0oS6nE/XwU+Vz/5o4kU=
-github.com/mattn/go-sqlite3 v2.0.3+incompatible/go.mod h1:FPy6KqzDD04eiIsT53CuJW3U88zkxoIYsOqkbpncsNc=
+github.com/mattn/go-sqlite3 v1.14.22/go.mod h1:Uh1q+B4BYcTPb+yiD3kU8Ct7aC0hY9fxUwlHK0RXw+Y=
 github.com/modern-go/concurrent v0.0.0-20180228061459-e0a39a4cb421/go.mod h1:6dJC0mAP4ikYIbvyc7fijjWJddQyLn8Ig3JB5CqoB9Q=
 github.com/modern-go/concurrent v0.0.0-20180306012644-bacd9c7ef1dd h1:TRLaZ9cD/w8PVh93nsPXa1VrQ6jlwL5oN8l14QlcNfg=
 github.com/modern-go/concurrent v0.0.0-20180306012644-bacd9c7ef1dd/go.mod h1:6dJC0mAP4ikYIbvyc7fijjWJddQyLn8Ig3JB5CqoB9Q=
@@ -120,6 +173,8 @@ github.com/modern-go/reflect2 v1.0.2/go.mod h1:yWuevngMOJpCy52FWWMvUC8ws7m/LJsjY
 github.com/nxadm/tail v1.4.8 h1:nPr65rt6Y5JFSKQO7qToXr7pePgD6Gwiw05lkbyAQTE=
 github.com/onsi/ginkgo v1.16.5 h1:8xi0RTUf59SOSfEtZMvwTvXYMzG4gV23XVHOZiXNtnE=
 github.com/onsi/gomega v1.18.1 h1:M1GfJqGRrBrrGGsbxzV5dqM2U2ApXefZCQpkukxYRLE=
+github.com/patrickmn/go-cache v2.1.0+incompatible h1:HRMgzkcYKYpi3C8ajMPV8OFXaaRUnok+kx1WdO15EQc=
+github.com/patrickmn/go-cache v2.1.0+incompatible/go.mod h1:3Qf8kWWT7OJRJbdiICTKqZju1ZixQ/KpMGzzAfe6+WQ=
 github.com/pelletier/go-toml/v2 v2.2.2 h1:aYUidT7k73Pcl9nb2gScu7NSrKCSHIDE89b3+6Wq+LM=
 github.com/pelletier/go-toml/v2 v2.2.2/go.mod h1:1t835xjRzz80PqgE6HHgN2JOsmgYu/h4qDAS4n929Rs=
 github.com/pkg/errors v0.9.1 h1:FEBLx1zS214owpjy7qsBeixbURkuhQAwrK5UwLGTwt4=
@@ -128,6 +183,7 @@ github.com/pkoukk/tiktoken-go v0.1.7 h1:qOBHXX4PHtvIvmOtyg1EeKlwFRiMKAcoMp4Q+bLQ
 github.com/pkoukk/tiktoken-go v0.1.7/go.mod h1:9NiV+i9mJKGj1rYOT+njbv+ZwA/zJxYdewGl6qVatpg=
 github.com/pmezard/go-difflib v1.0.0 h1:4DBwDE0NGyQoBHbLQYPwSUPoCMWR5BEzIk/f1lZbAQM=
 github.com/pmezard/go-difflib v1.0.0/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4=
+github.com/prometheus/client_model v0.0.0-20190812154241-14fe0d1b01d4/go.mod h1:xMI15A0UPsDsEKsMN9yxemIoYk6Tm2C1GtYGdfGttqA=
 github.com/rogpeppe/go-internal v1.8.0 h1:FCbCCtXNOY3UtUuHUYaghJg4y7Fd14rXifAYUAtL9R8=
 github.com/smarty/assertions v1.15.0 h1:cR//PqUBUiQRakZWqBiFFQ9wb8emQGDb0HeGdqGByCY=
 github.com/smarty/assertions v1.15.0/go.mod h1:yABtdzeQs6l1brC900WlRNwj6ZR55d7B+E8C6HtKdec=
@@ -149,26 +205,96 @@ github.com/twitchyliquid64/golang-asm v0.15.1 h1:SU5vSMR7hnwNxj24w34ZyCi/FmDZTkS
 github.com/twitchyliquid64/golang-asm v0.15.1/go.mod h1:a1lVb/DtPvCB8fslRZhAngC2+aY1QWCk3Cedj/Gdt08=
 github.com/ugorji/go/codec v1.2.12 h1:9LC83zGrHhuUA9l16C9AHXAqEV/2wBQ4nkvumAE65EE=
 github.com/ugorji/go/codec v1.2.12/go.mod h1:UNopzCgEMSXjBc6AOMqYvWC1ktqTAfzJZUZgYf6w6lg=
+go.opencensus.io v0.24.0 h1:y73uSU6J157QMP2kn2r30vwW1A2W2WFwSCGnAVxeaD0=
+go.opencensus.io v0.24.0/go.mod h1:vNK8G9p7aAivkbmorf4v+7Hgx+Zs0yY+0fOtgBfjQKo=
+go.opentelemetry.io/contrib/instrumentation/google.golang.org/grpc/otelgrpc v0.49.0 h1:4Pp6oUg3+e/6M4C0A/3kJ2VYa++dsWVTtGgLVj5xtHg=
+go.opentelemetry.io/contrib/instrumentation/google.golang.org/grpc/otelgrpc v0.49.0/go.mod h1:Mjt1i1INqiaoZOMGR1RIUJN+i3ChKoFRqzrRQhlkbs0=
+go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp v0.49.0 h1:jq9TW8u3so/bN+JPT166wjOI6/vQPF6Xe7nMNIltagk=
+go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp v0.49.0/go.mod h1:p8pYQP+m5XfbZm9fxtSKAbM6oIllS7s2AfxrChvc7iw=
+go.opentelemetry.io/otel v1.24.0 h1:0LAOdjNmQeSTzGBzduGe/rU4tZhMwL5rWgtp9Ku5Jfo=
+go.opentelemetry.io/otel v1.24.0/go.mod h1:W7b9Ozg4nkF5tWI5zsXkaKKDjdVjpD4oAt9Qi/MArHo=
+go.opentelemetry.io/otel/metric v1.24.0 h1:6EhoGWWK28x1fbpA4tYTOWBkPefTDQnb8WSGXlc88kI=
+go.opentelemetry.io/otel/metric v1.24.0/go.mod h1:VYhLe1rFfxuTXLgj4CBiyz+9WYBA8pNGJgDcSFRKBco=
+go.opentelemetry.io/otel/trace v1.24.0 h1:CsKnnL4dUAr/0llH9FKuc698G04IrpWV0MQA/Y1YELI=
+go.opentelemetry.io/otel/trace v1.24.0/go.mod h1:HPc3Xr/cOApsBI154IU0OI0HJexz+aw5uPdbs3UCjNU=
 golang.org/x/arch v0.0.0-20210923205945-b76863e36670/go.mod h1:5om86z9Hs0C8fWVUuoMHwpExlXzs5Tkyp9hOrfG7pp8=
 golang.org/x/arch v0.8.0 h1:3wRIsP3pM4yUptoR96otTUOXI367OS0+c9eeRi9doIc=
 golang.org/x/arch v0.8.0/go.mod h1:FEVrYAQjsQXMVJ1nsMoVVXPZg6p2JE2mx8psSWTDQys=
-golang.org/x/crypto v0.23.0 h1:dIJU/v2J8Mdglj/8rJ6UUOM3Zc9zLZxVZwwxMooUSAI=
+golang.org/x/crypto v0.0.0-20190308221718-c2843e01d9a2/go.mod h1:djNgcEr1/C05ACkg1iLfiJU5Ep61QUkGW8qpdssI0+w=
-golang.org/x/crypto v0.23.0/go.mod h1:CKFgDieR+mRhux2Lsu27y0fO304Db0wZe70UKqHu0v8=
+golang.org/x/crypto v0.0.0-20200622213623-75b288015ac9/go.mod h1:LzIPMQfyMNhhGPhUkYOs5KpL4U8rLKemX1yGLhDgUto=
+golang.org/x/crypto v0.24.0 h1:mnl8DM0o513X8fdIkmyFE/5hTYxbwYOjDS/+rK6qpRI=
+golang.org/x/crypto v0.24.0/go.mod h1:Z1PMYSOR5nyMcyAVAIQSKCDwalqy85Aqn1x3Ws4L5DM=
+golang.org/x/exp v0.0.0-20190121172915-509febef88a4/go.mod h1:CJ0aWSM057203Lf6IL+f9T1iT9GByDxfZKAQTCR3kQA=
 golang.org/x/image v0.18.0 h1:jGzIakQa/ZXI1I0Fxvaa9W7yP25TqT6cHIHn+6CqvSQ=
 golang.org/x/image v0.18.0/go.mod h1:4yyo5vMFQjVjUcVk4jEQcU9MGy/rulF5WvUILseCM2E=
-golang.org/x/net v0.25.0 h1:d/OCCoBEUq33pjydKrGQhw7IlUPI2Oylr+8qLx49kac=
+golang.org/x/lint v0.0.0-20181026193005-c67002cb31c3/go.mod h1:UVdnD1Gm6xHRNCYTkRU2/jEulfH38KcIWyp/GAMgvoE=
-golang.org/x/net v0.25.0/go.mod h1:JkAGAh7GEvH74S6FOH42FLoXpXbE/aqXSrIQjXgsiwM=
+golang.org/x/lint v0.0.0-20190227174305-5b3e6a55c961/go.mod h1:wehouNa3lNwaWXcvxsM5YxQ5yQlVC4a0KAMCusXpPoU=
+golang.org/x/lint v0.0.0-20190313153728-d0100b6bd8b3/go.mod h1:6SW0HCj/g11FgYtHlgUYUwCkIfeOF89ocIRzGO/8vkc=
+golang.org/x/net v0.0.0-20180724234803-3673e40ba225/go.mod h1:mL1N/T3taQHkDXs73rZJwtUhF3w3ftmwwsq0BUmARs4=
+golang.org/x/net v0.0.0-20180826012351-8a410e7b638d/go.mod h1:mL1N/T3taQHkDXs73rZJwtUhF3w3ftmwwsq0BUmARs4=
+golang.org/x/net v0.0.0-20190213061140-3a22650c66bd/go.mod h1:mL1N/T3taQHkDXs73rZJwtUhF3w3ftmwwsq0BUmARs4=
+golang.org/x/net v0.0.0-20190311183353-d8887717615a/go.mod h1:t9HGtf8HONx5eT2rtn7q6eTqICYqUVnKs3thJo3Qplg=
+golang.org/x/net v0.0.0-20190404232315-eb5bcb51f2a3/go.mod h1:t9HGtf8HONx5eT2rtn7q6eTqICYqUVnKs3thJo3Qplg=
+golang.org/x/net v0.0.0-20201110031124-69a78807bb2b/go.mod h1:sp8m0HH+o8qH0wwXwYZr8TS3Oi6o0r6Gce1SSxlDquU=
+golang.org/x/net v0.26.0 h1:soB7SVo0PWrY4vPW/+ay0jKDNScG2X9wFeYlXIvJsOQ=
+golang.org/x/net v0.26.0/go.mod h1:5YKkiSynbBIh3p6iOc/vibscux0x38BZDkn8sCUPxHE=
+golang.org/x/oauth2 v0.0.0-20180821212333-d2e6202438be/go.mod h1:N/0e6XlmueqKjAGxoOufVs8QHGRruUQn6yWY3a++T0U=
+golang.org/x/oauth2 v0.21.0 h1:tsimM75w1tF/uws5rbeHzIWxEqElMehnc+iW793zsZs=
+golang.org/x/oauth2 v0.21.0/go.mod h1:XYTD2NtWslqkgxebSiOHnXEap4TF09sJSc7H1sXbhtI=
+golang.org/x/sync v0.0.0-20180314180146-1d60e4601c6f/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
+golang.org/x/sync v0.0.0-20181108010431-42b317875d0f/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
+golang.org/x/sync v0.0.0-20190423024810-112230192c58/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
 golang.org/x/sync v0.7.0 h1:YsImfSBoP9QPYL0xyKJPq0gcaJdG3rInoqxTWbfQu9M=
 golang.org/x/sync v0.7.0/go.mod h1:Czt+wKu1gCyEFDUtn0jG5QVvpJ6rzVqr5aXyt9drQfk=
+golang.org/x/sys v0.0.0-20180830151530-49385e6e1522/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
+golang.org/x/sys v0.0.0-20190215142949-d0b11bdaac8a/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
+golang.org/x/sys v0.0.0-20190412213103-97732733099d/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200930185726-fdedc70b468f/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
 golang.org/x/sys v0.5.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
 golang.org/x/sys v0.6.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
-golang.org/x/sys v0.20.0 h1:Od9JTbYCk261bKm4M/mw7AklTlFYIa0bIp9BgSm1S8Y=
+golang.org/x/sys v0.21.0 h1:rF+pYz3DAGSQAxAu1CbC7catZg4ebC4UIeIhKxBZvws=
-golang.org/x/sys v0.20.0/go.mod h1:/VUhepiaJMQUp4+oa/7Zr1D23ma6VTLIYjOOTFZPUcA=
+golang.org/x/sys v0.21.0/go.mod h1:/VUhepiaJMQUp4+oa/7Zr1D23ma6VTLIYjOOTFZPUcA=
+golang.org/x/text v0.3.0/go.mod h1:NqM8EUOU14njkJ3fqMW+pc6Ldnwhi/IjpwHt7yyuwOQ=
+golang.org/x/text v0.3.3/go.mod h1:5Zoc/QRtKVWzQhOtBMvqHzDpF6irO9z98xDceosuGiQ=
 golang.org/x/text v0.16.0 h1:a94ExnEXNtEwYLGJSIUxnWoxoRz/ZcCsV63ROupILh4=
 golang.org/x/text v0.16.0/go.mod h1:GhwF1Be+LQoKShO3cGOHzqOgRrGaYc9AvblQOmPVHnI=
-golang.org/x/xerrors v0.0.0-20200804184101-5ec99f83aff1 h1:go1bK/D/BFZV2I8cIQd1NKEZ+0owSTG1fDTci4IqFcE=
+golang.org/x/time v0.5.0 h1:o7cqy6amK/52YcAKIPlM3a+Fpj35zvRj2TP+e1xFSfk=
-google.golang.org/protobuf v1.34.1 h1:9ddQBjfCyZPOHPUiPxpYESBLc+T8P3E+Vo4IbKZgFWg=
+golang.org/x/time v0.5.0/go.mod h1:3BpzKBy/shNhVucY/MWOyx10tF3SFh9QdLuxbVysPQM=
-google.golang.org/protobuf v1.34.1/go.mod h1:c6P6GXX6sHbq/GpV6MGZEdwhWPcYBgnhAHhKbcUYpos=
+golang.org/x/tools v0.0.0-20180917221912-90fa682c2a6e/go.mod h1:n7NCudcB/nEzxVGmLbDWY5pfWTLqBcC2KZ6jyYvM4mQ=
+golang.org/x/tools v0.0.0-20190114222345-bf090417da8b/go.mod h1:n7NCudcB/nEzxVGmLbDWY5pfWTLqBcC2KZ6jyYvM4mQ=
+golang.org/x/tools v0.0.0-20190226205152-f727befe758c/go.mod h1:9Yl7xja0Znq3iFh3HoIrodX9oNMXvdceNzlUR8zjMvY=
+golang.org/x/tools v0.0.0-20190311212946-11955173bddd/go.mod h1:LCzVGOaR6xXOjkQ3onu1FJEFr0SW1gC7cKk1uF8kGRs=
+golang.org/x/tools v0.0.0-20190524140312-2c0ae7006135/go.mod h1:RgjU9mgBXZiqYHBnxXauZ1Gv1EHHAz9KjViQ78xBX0Q=
+golang.org/x/xerrors v0.0.0-20191204190536-9bdfabe68543/go.mod h1:I/5z698sn9Ka8TeJc9MKroUUfqBBauWjQqLJ2OPfmY0=
+google.golang.org/api v0.187.0 h1:Mxs7VATVC2v7CY+7Xwm4ndkX71hpElcvx0D1Ji/p1eo=
+google.golang.org/api v0.187.0/go.mod h1:KIHlTc4x7N7gKKuVsdmfBXN13yEEWXWFURWY6SBp2gk=
+google.golang.org/appengine v1.1.0/go.mod h1:EbEs0AVv82hx2wNQdGPgUI5lhzA/G0D9YwlJXL52JkM=
+google.golang.org/appengine v1.4.0/go.mod h1:xpcJRLb0r/rnEns0DIKYYv+WjYCduHsrkT7/EB5XEv4=
+google.golang.org/genproto v0.0.0-20180817151627-c66870c02cf8/go.mod h1:JiN7NxoALGmiZfu7CAH4rXhgtRTLTxftemlI0sWmxmc=
+google.golang.org/genproto v0.0.0-20190819201941-24fa4b261c55/go.mod h1:DMBHOl98Agz4BDEuKkezgsaosCRResVns1a3J2ZsMNc=
+google.golang.org/genproto v0.0.0-20200526211855-cb27e3aa2013/go.mod h1:NbSheEEYHJ7i3ixzK3sjbqSGDJWnxyFXZblF3eUsNvo=
+google.golang.org/genproto/googleapis/api v0.0.0-20240617180043-68d350f18fd4 h1:MuYw1wJzT+ZkybKfaOXKp5hJiZDn2iHaXRw0mRYdHSc=
+google.golang.org/genproto/googleapis/api v0.0.0-20240617180043-68d350f18fd4/go.mod h1:px9SlOOZBg1wM1zdnr8jEL4CNGUBZ+ZKYtNPApNQc4c=
+google.golang.org/genproto/googleapis/rpc v0.0.0-20240624140628-dc46fd24d27d h1:k3zyW3BYYR30e8v3x0bTDdE9vpYFjZHK+HcyqkrppWk=
+google.golang.org/genproto/googleapis/rpc v0.0.0-20240624140628-dc46fd24d27d/go.mod h1:Ue6ibwXGpU+dqIcODieyLOcgj7z8+IcskoNIgZxtrFY=
+google.golang.org/grpc v1.19.0/go.mod h1:mqu4LbDTu4XGKhr4mRzUsmM4RtVoemTSY81AxZiDr8c=
+google.golang.org/grpc v1.23.0/go.mod h1:Y5yQAOtifL1yxbo5wqy6BxZv8vAUGQwXBOALyacEbxg=
+google.golang.org/grpc v1.25.1/go.mod h1:c3i+UQWmh7LiEpx4sFZnkU36qjEYZ0imhYfXVyQciAY=
+google.golang.org/grpc v1.27.0/go.mod h1:qbnxyOmOxrQa7FizSgH+ReBfzJrCY1pSN7KXBS8abTk=
+google.golang.org/grpc v1.33.2/go.mod h1:JMHMWHQWaTccqQQlmk3MJZS+GWXOdAesneDmEnv2fbc=
+google.golang.org/grpc v1.64.1 h1:LKtvyfbX3UGVPFcGqJ9ItpVWW6oN/2XqTxfAnwRRXiA=
+google.golang.org/grpc v1.64.1/go.mod h1:hiQF4LFZelK2WKaP6W0L92zGHtiQdZxk8CrSdvyjeP0=
+google.golang.org/protobuf v0.0.0-20200109180630-ec00e32a8dfd/go.mod h1:DFci5gLYBciE7Vtevhsrf46CRTquxDuWsQurQQe4oz8=
+google.golang.org/protobuf v0.0.0-20200221191635-4d8936d0db64/go.mod h1:kwYJMbMJ01Woi6D6+Kah6886xMZcty6N08ah7+eCXa0=
+google.golang.org/protobuf v0.0.0-20200228230310-ab0ca4ff8a60/go.mod h1:cfTl7dwQJ+fmap5saPgwCLgHXTUD7jkjRqWcaiX5VyM=
+google.golang.org/protobuf v1.20.1-0.20200309200217-e05f789c0967/go.mod h1:A+miEFZTKqfCUM6K7xSMQL9OKL/b6hQv+e19PK+JZNE=
+google.golang.org/protobuf v1.21.0/go.mod h1:47Nbq4nVaFHyn7ilMalzfO3qCViNmqZ2kzikPIcrTAo=
+google.golang.org/protobuf v1.22.0/go.mod h1:EGpADcykh3NcUnDUJcl1+ZksZNG86OlYog2l/sGQquU=
+google.golang.org/protobuf v1.23.0/go.mod h1:EGpADcykh3NcUnDUJcl1+ZksZNG86OlYog2l/sGQquU=
+google.golang.org/protobuf v1.23.1-0.20200526195155-81db48ad09cc/go.mod h1:EGpADcykh3NcUnDUJcl1+ZksZNG86OlYog2l/sGQquU=
+google.golang.org/protobuf v1.25.0/go.mod h1:9JNX74DMeImyA3h4bdi1ymwjUzf21/xIlbajtzgsN7c=
+google.golang.org/protobuf v1.34.2 h1:6xV6lTsCfpGD21XK49h7MhtcApnLqkfYgPcdHftf6hg=
+google.golang.org/protobuf v1.34.2/go.mod h1:qYOHts0dSfpeUzUFpOMr/WGzszTmLH+DiWniOlNbLDw=
 gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
 gopkg.in/check.v1 v1.0.0-20201130134442-10cb98267c6c h1:Hei/4ADfdWqJk1ZMxUNpqntNwaWcugrBjAiHlqqRiVk=
 gopkg.in/tomb.v1 v1.0.0-20141024135613-dd632973f1e7 h1:uRGJdciOHaEIrze2W8Q3AKkepLTh2hOroT7a+7czfdQ=
@@ -185,5 +311,7 @@ gorm.io/driver/sqlite v1.5.5/go.mod h1:6NgQ7sQWAIFsPrJJl1lSNSu2TABh0ZZ/zm5fosATa
 gorm.io/gorm v1.25.7/go.mod h1:hbnx/Oo0ChWMn1BIhpy1oYozzpM15i4YPuHDmfYtwg8=
 gorm.io/gorm v1.25.10 h1:dQpO+33KalOA+aFYGlK+EfxcI5MbO7EP2yYygwh9h+s=
 gorm.io/gorm v1.25.10/go.mod h1:hbnx/Oo0ChWMn1BIhpy1oYozzpM15i4YPuHDmfYtwg8=
+honnef.co/go/tools v0.0.0-20190102054323-c2f93a96b099/go.mod h1:rf3lG4BRIbNafJWhAfAdb/ePZxsR/4RtNHQocxwk9r4=
+honnef.co/go/tools v0.0.0-20190523083050-ea95bdfd59fc/go.mod h1:rf3lG4BRIbNafJWhAfAdb/ePZxsR/4RtNHQocxwk9r4=
 nullprogram.com/x/optparse v1.0.0/go.mod h1:KdyPE+Igbe0jQUrVfMqDMeJQIJZEuyV7pjYmp6pbG50=
 rsc.io/pdf v0.1.1/go.mod h1:n8OzWcQ6Sp37PL01nO98y4iUCRdTGarVfzxY20ICaU4=

main.go

@@ -27,27 +27,19 @@ func main() {
 	common.Init()
 	logger.SetupLogger()
 	logger.SysLogf("One API %s started", common.Version)
-	if os.Getenv("GIN_MODE") != "debug" {
+
+	if os.Getenv("GIN_MODE") != gin.DebugMode {
 		gin.SetMode(gin.ReleaseMode)
 	}
 	if config.DebugEnabled {
 		logger.SysLog("running in debug mode")
 	}
-	var err error
+
 	// Initialize SQL Database
-	model.DB, err = model.InitDB("SQL_DSN")
+	model.InitDB()
-	if err != nil {
+	model.InitLogDB()
-		logger.FatalLog("failed to initialize database: " + err.Error())
+
-	}
+	var err error
-	if os.Getenv("LOG_SQL_DSN") != "" {
-		logger.SysLog("using secondary database for table logs")
-		model.LOG_DB, err = model.InitDB("LOG_SQL_DSN")
-		if err != nil {
-			logger.FatalLog("failed to initialize secondary database: " + err.Error())
-		}
-	} else {
-		model.LOG_DB = model.DB
-	}
 	err = model.CreateRootAccountIfNeed()
 	if err != nil {
 		logger.FatalLog("database init error: " + err.Error())

middleware/rate-limit.go

@@ -3,11 +3,12 @@ package middleware
 import (
 	"context"
 	"fmt"
+	"net/http"
+	"time"
+
 	"github.com/gin-gonic/gin"
 	"github.com/songquanpeng/one-api/common"
 	"github.com/songquanpeng/one-api/common/config"
-	"net/http"
-	"time"
 )
 
 var timeFormat = "2006-01-02T15:04:05.000Z"
@@ -70,6 +71,11 @@ func memoryRateLimiter(c *gin.Context, maxRequestNum int, duration int64, mark s
 }
 
 func rateLimitFactory(maxRequestNum int, duration int64, mark string) func(c *gin.Context) {
+	if maxRequestNum == 0 {
+		return func(c *gin.Context) {
+			c.Next()
+		}
+	}
 	if common.RedisEnabled {
 		return func(c *gin.Context) {
 			redisRateLimiter(c, maxRequestNum, duration, mark)

model/channel.go

@@ -3,6 +3,7 @@ package model
 import (
 	"encoding/json"
 	"fmt"
+
 	"github.com/songquanpeng/one-api/common/config"
 	"github.com/songquanpeng/one-api/common/helper"
 	"github.com/songquanpeng/one-api/common/logger"
@@ -39,13 +40,15 @@ type Channel struct {
 }
 
 type ChannelConfig struct {
-	Region     string `json:"region,omitempty"`
+	Region            string `json:"region,omitempty"`
-	SK         string `json:"sk,omitempty"`
+	SK                string `json:"sk,omitempty"`
-	AK         string `json:"ak,omitempty"`
+	AK                string `json:"ak,omitempty"`
-	UserID     string `json:"user_id,omitempty"`
+	UserID            string `json:"user_id,omitempty"`
-	APIVersion string `json:"api_version,omitempty"`
+	APIVersion        string `json:"api_version,omitempty"`
-	LibraryID  string `json:"library_id,omitempty"`
+	LibraryID         string `json:"library_id,omitempty"`
-	Plugin     string `json:"plugin,omitempty"`
+	Plugin            string `json:"plugin,omitempty"`
+	VertexAIProjectID string `json:"vertex_ai_project_id,omitempty"`
+	VertexAIADC       string `json:"vertex_ai_adc,omitempty"`
 }
 
 func GetAllChannels(startIdx int, num int, scope string) ([]*Channel, error) {

model/main.go

@@ -1,6 +1,7 @@
 package model
 
 import (
+	"database/sql"
 	"fmt"
 	"github.com/songquanpeng/one-api/common"
 	"github.com/songquanpeng/one-api/common/config"
@@ -29,13 +30,17 @@ func CreateRootAccountIfNeed() error {
 		if err != nil {
 			return err
 		}
+		accessToken := random.GetUUID()
+		if config.InitialRootAccessToken != "" {
+			accessToken = config.InitialRootAccessToken
+		}
 		rootUser := User{
 			Username:    "root",
 			Password:    hashedPassword,
 			Role:        RoleRootUser,
 			Status:      UserStatusEnabled,
 			DisplayName: "Root User",
-			AccessToken: random.GetUUID(),
+			AccessToken: accessToken,
 			Quota:       500000000000000,
 		}
 		DB.Create(&rootUser)
@@ -60,90 +65,156 @@ func CreateRootAccountIfNeed() error {
 }
 
 func chooseDB(envName string) (*gorm.DB, error) {
-	if os.Getenv(envName) != "" {
+	dsn := os.Getenv(envName)
-		dsn := os.Getenv(envName)
+
-		if strings.HasPrefix(dsn, "postgres://") {
+	switch {
-			// Use PostgreSQL
+	case strings.HasPrefix(dsn, "postgres://"):
-			logger.SysLog("using PostgreSQL as database")
+		// Use PostgreSQL
-			common.UsingPostgreSQL = true
+		return openPostgreSQL(dsn)
-			return gorm.Open(postgres.New(postgres.Config{
+	case dsn != "":
-				DSN:                  dsn,
-				PreferSimpleProtocol: true, // disables implicit prepared statement usage
-			}), &gorm.Config{
-				PrepareStmt: true, // precompile SQL
-			})
-		}
 		// Use MySQL
-		logger.SysLog("using MySQL as database")
+		return openMySQL(dsn)
-		common.UsingMySQL = true
+	default:
-		return gorm.Open(mysql.Open(dsn), &gorm.Config{
+		// Use SQLite
-			PrepareStmt: true, // precompile SQL
+		return openSQLite()
-		})
 	}
-	// Use SQLite
+}
-	logger.SysLog("SQL_DSN not set, using SQLite as database")
+
-	common.UsingSQLite = true
+func openPostgreSQL(dsn string) (*gorm.DB, error) {
-	config := fmt.Sprintf("?_busy_timeout=%d", common.SQLiteBusyTimeout)
+	logger.SysLog("using PostgreSQL as database")
-	return gorm.Open(sqlite.Open(common.SQLitePath+config), &gorm.Config{
+	common.UsingPostgreSQL = true
+	return gorm.Open(postgres.New(postgres.Config{
+		DSN:                  dsn,
+		PreferSimpleProtocol: true, // disables implicit prepared statement usage
+	}), &gorm.Config{
 		PrepareStmt: true, // precompile SQL
 	})
 }
 
-func InitDB(envName string) (db *gorm.DB, err error) {
+func openMySQL(dsn string) (*gorm.DB, error) {
-	db, err = chooseDB(envName)
+	logger.SysLog("using MySQL as database")
-	if err == nil {
+	common.UsingMySQL = true
-		if config.DebugSQLEnabled {
+	return gorm.Open(mysql.Open(dsn), &gorm.Config{
-			db = db.Debug()
+		PrepareStmt: true, // precompile SQL
-		}
+	})
-		sqlDB, err := db.DB()
+}
-		if err != nil {
-			return nil, err
-		}
-		sqlDB.SetMaxIdleConns(env.Int("SQL_MAX_IDLE_CONNS", 100))
-		sqlDB.SetMaxOpenConns(env.Int("SQL_MAX_OPEN_CONNS", 1000))
-		sqlDB.SetConnMaxLifetime(time.Second * time.Duration(env.Int("SQL_MAX_LIFETIME", 60)))
 
-		if !config.IsMasterNode {
+func openSQLite() (*gorm.DB, error) {
-			return db, err
+	logger.SysLog("SQL_DSN not set, using SQLite as database")
-		}
+	common.UsingSQLite = true
-		if common.UsingMySQL {
+	dsn := fmt.Sprintf("%s?_busy_timeout=%d", common.SQLitePath, common.SQLiteBusyTimeout)
-			_, _ = sqlDB.Exec("DROP INDEX idx_channels_key ON channels;") // TODO: delete this line when most users have upgraded
+	return gorm.Open(sqlite.Open(dsn), &gorm.Config{
-		}
+		PrepareStmt: true, // precompile SQL
-		logger.SysLog("database migration started")
+	})
-		err = db.AutoMigrate(&Channel{})
+}
-		if err != nil {
+
-			return nil, err
+func InitDB() {
-		}
+	var err error
-		err = db.AutoMigrate(&Token{})
+	DB, err = chooseDB("SQL_DSN")
-		if err != nil {
+	if err != nil {
-			return nil, err
+		logger.FatalLog("failed to initialize database: " + err.Error())
-		}
+		return
-		err = db.AutoMigrate(&User{})
-		if err != nil {
-			return nil, err
-		}
-		err = db.AutoMigrate(&Option{})
-		if err != nil {
-			return nil, err
-		}
-		err = db.AutoMigrate(&Redemption{})
-		if err != nil {
-			return nil, err
-		}
-		err = db.AutoMigrate(&Ability{})
-		if err != nil {
-			return nil, err
-		}
-		err = db.AutoMigrate(&Log{})
-		if err != nil {
-			return nil, err
-		}
-		logger.SysLog("database migrated")
-		return db, err
-	} else {
-		logger.FatalLog(err)
 	}
-	return db, err
+
+	sqlDB := setDBConns(DB)
+
+	if !config.IsMasterNode {
+		return
+	}
+
+	if common.UsingMySQL {
+		_, _ = sqlDB.Exec("DROP INDEX idx_channels_key ON channels;") // TODO: delete this line when most users have upgraded
+	}
+
+	logger.SysLog("database migration started")
+	if err = migrateDB(); err != nil {
+		logger.FatalLog("failed to migrate database: " + err.Error())
+		return
+	}
+	logger.SysLog("database migrated")
+}
+
+func migrateDB() error {
+	var err error
+	if err = DB.AutoMigrate(&Channel{}); err != nil {
+		return err
+	}
+	if err = DB.AutoMigrate(&Token{}); err != nil {
+		return err
+	}
+	if err = DB.AutoMigrate(&User{}); err != nil {
+		return err
+	}
+	if err = DB.AutoMigrate(&Option{}); err != nil {
+		return err
+	}
+	if err = DB.AutoMigrate(&Redemption{}); err != nil {
+		return err
+	}
+	if err = DB.AutoMigrate(&Ability{}); err != nil {
+		return err
+	}
+	if err = DB.AutoMigrate(&Log{}); err != nil {
+		return err
+	}
+	if err = DB.AutoMigrate(&Channel{}); err != nil {
+		return err
+	}
+	return nil
+}
+
+func InitLogDB() {
+	if os.Getenv("LOG_SQL_DSN") == "" {
+		LOG_DB = DB
+		return
+	}
+
+	logger.SysLog("using secondary database for table logs")
+	var err error
+	LOG_DB, err = chooseDB("LOG_SQL_DSN")
+	if err != nil {
+		logger.FatalLog("failed to initialize secondary database: " + err.Error())
+		return
+	}
+
+	setDBConns(LOG_DB)
+
+	if !config.IsMasterNode {
+		return
+	}
+
+	logger.SysLog("secondary database migration started")
+	err = migrateLOGDB()
+	if err != nil {
+		logger.FatalLog("failed to migrate secondary database: " + err.Error())
+		return
+	}
+	logger.SysLog("secondary database migrated")
+}
+
+func migrateLOGDB() error {
+	var err error
+	if err = LOG_DB.AutoMigrate(&Log{}); err != nil {
+		return err
+	}
+	return nil
+}
+
+func setDBConns(db *gorm.DB) *sql.DB {
+	if config.DebugSQLEnabled {
+		db = db.Debug()
+	}
+
+	sqlDB, err := db.DB()
+	if err != nil {
+		logger.FatalLog("failed to connect database: " + err.Error())
+		return nil
+	}
+
+	sqlDB.SetMaxIdleConns(env.Int("SQL_MAX_IDLE_CONNS", 100))
+	sqlDB.SetMaxOpenConns(env.Int("SQL_MAX_OPEN_CONNS", 1000))
+	sqlDB.SetConnMaxLifetime(time.Second * time.Duration(env.Int("SQL_MAX_LIFETIME", 60)))
+	return sqlDB
 }
 
 func closeDB(db *gorm.DB) error {

relay/adaptor.go

@@ -16,6 +16,7 @@ import (
 	"github.com/songquanpeng/one-api/relay/adaptor/openai"
 	"github.com/songquanpeng/one-api/relay/adaptor/palm"
 	"github.com/songquanpeng/one-api/relay/adaptor/tencent"
+	"github.com/songquanpeng/one-api/relay/adaptor/vertexai"
 	"github.com/songquanpeng/one-api/relay/adaptor/xunfei"
 	"github.com/songquanpeng/one-api/relay/adaptor/zhipu"
 	"github.com/songquanpeng/one-api/relay/apitype"
@@ -55,6 +56,8 @@ func GetAdaptor(apiType int) adaptor.Adaptor {
 		return &cloudflare.Adaptor{}
 	case apitype.DeepL:
 		return &deepl.Adaptor{}
+	case apitype.VertexAI:
+		return &vertexai.Adaptor{}
 	}
 	return nil
 }

relay/adaptor/anthropic/adaptor.go

@@ -3,12 +3,14 @@ package anthropic
 import (
 	"errors"
 	"fmt"
+	"io"
+	"net/http"
+	"strings"
+
 	"github.com/gin-gonic/gin"
 	"github.com/songquanpeng/one-api/relay/adaptor"
 	"github.com/songquanpeng/one-api/relay/meta"
 	"github.com/songquanpeng/one-api/relay/model"
-	"io"
-	"net/http"
 )
 
 type Adaptor struct {
@@ -31,6 +33,13 @@ func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *me
 	}
 	req.Header.Set("anthropic-version", anthropicVersion)
 	req.Header.Set("anthropic-beta", "messages-2023-12-15")
+
+	// https://x.com/alexalbert__/status/1812921642143900036
+	// claude-3-5-sonnet can support 8k context
+	if strings.HasPrefix(meta.ActualModelName, "claude-3-5-sonnet") {
+		req.Header.Set("anthropic-beta", "max-tokens-3-5-sonnet-2024-07-15")
+	}
+
 	return nil
 }
 

relay/adaptor/anthropic/main.go

@@ -29,12 +29,30 @@ func stopReasonClaude2OpenAI(reason *string) string {
 		return "stop"
 	case "max_tokens":
 		return "length"
+	case "tool_use":
+		return "tool_calls"
 	default:
 		return *reason
 	}
 }
 
 func ConvertRequest(textRequest model.GeneralOpenAIRequest) *Request {
+	claudeTools := make([]Tool, 0, len(textRequest.Tools))
+
+	for _, tool := range textRequest.Tools {
+		if params, ok := tool.Function.Parameters.(map[string]any); ok {
+			claudeTools = append(claudeTools, Tool{
+				Name:        tool.Function.Name,
+				Description: tool.Function.Description,
+				InputSchema: InputSchema{
+					Type:       params["type"].(string),
+					Properties: params["properties"],
+					Required:   params["required"],
+				},
+			})
+		}
+	}
+
 	claudeRequest := Request{
 		Model:       textRequest.Model,
 		MaxTokens:   textRequest.MaxTokens,
@@ -42,6 +60,24 @@ func ConvertRequest(textRequest model.GeneralOpenAIRequest) *Request {
 		TopP:        textRequest.TopP,
 		TopK:        textRequest.TopK,
 		Stream:      textRequest.Stream,
+		Tools:       claudeTools,
+	}
+	if len(claudeTools) > 0 {
+		claudeToolChoice := struct {
+			Type string `json:"type"`
+			Name string `json:"name,omitempty"`
+		}{Type: "auto"} // default value https://docs.anthropic.com/en/docs/build-with-claude/tool-use#controlling-claudes-output
+		if choice, ok := textRequest.ToolChoice.(map[string]any); ok {
+			if function, ok := choice["function"].(map[string]any); ok {
+				claudeToolChoice.Type = "tool"
+				claudeToolChoice.Name = function["name"].(string)
+			}
+		} else if toolChoiceType, ok := textRequest.ToolChoice.(string); ok {
+			if toolChoiceType == "any" {
+				claudeToolChoice.Type = toolChoiceType
+			}
+		}
+		claudeRequest.ToolChoice = claudeToolChoice
 	}
 	if claudeRequest.MaxTokens == 0 {
 		claudeRequest.MaxTokens = 4096
@@ -64,7 +100,24 @@ func ConvertRequest(textRequest model.GeneralOpenAIRequest) *Request {
 		if message.IsStringContent() {
 			content.Type = "text"
 			content.Text = message.StringContent()
+			if message.Role == "tool" {
+				claudeMessage.Role = "user"
+				content.Type = "tool_result"
+				content.Content = content.Text
+				content.Text = ""
+				content.ToolUseId = message.ToolCallId
+			}
 			claudeMessage.Content = append(claudeMessage.Content, content)
+			for i := range message.ToolCalls {
+				inputParam := make(map[string]any)
+				_ = json.Unmarshal([]byte(message.ToolCalls[i].Function.Arguments.(string)), &inputParam)
+				claudeMessage.Content = append(claudeMessage.Content, Content{
+					Type:  "tool_use",
+					Id:    message.ToolCalls[i].Id,
+					Name:  message.ToolCalls[i].Function.Name,
+					Input: inputParam,
+				})
+			}
 			claudeRequest.Messages = append(claudeRequest.Messages, claudeMessage)
 			continue
 		}
@@ -97,16 +150,35 @@ func StreamResponseClaude2OpenAI(claudeResponse *StreamResponse) (*openai.ChatCo
 	var response *Response
 	var responseText string
 	var stopReason string
+	tools := make([]model.Tool, 0)
+
 	switch claudeResponse.Type {
 	case "message_start":
 		return nil, claudeResponse.Message
 	case "content_block_start":
 		if claudeResponse.ContentBlock != nil {
 			responseText = claudeResponse.ContentBlock.Text
+			if claudeResponse.ContentBlock.Type == "tool_use" {
+				tools = append(tools, model.Tool{
+					Id:   claudeResponse.ContentBlock.Id,
+					Type: "function",
+					Function: model.Function{
+						Name:      claudeResponse.ContentBlock.Name,
+						Arguments: "",
+					},
+				})
+			}
 		}
 	case "content_block_delta":
 		if claudeResponse.Delta != nil {
 			responseText = claudeResponse.Delta.Text
+			if claudeResponse.Delta.Type == "input_json_delta" {
+				tools = append(tools, model.Tool{
+					Function: model.Function{
+						Arguments: claudeResponse.Delta.PartialJson,
+					},
+				})
+			}
 		}
 	case "message_delta":
 		if claudeResponse.Usage != nil {
@@ -120,6 +192,10 @@ func StreamResponseClaude2OpenAI(claudeResponse *StreamResponse) (*openai.ChatCo
 	}
 	var choice openai.ChatCompletionsStreamResponseChoice
 	choice.Delta.Content = responseText
+	if len(tools) > 0 {
+		choice.Delta.Content = nil // compatible with other OpenAI derivative applications, like LobeOpenAICompatibleFactory ...
+		choice.Delta.ToolCalls = tools
+	}
 	choice.Delta.Role = "assistant"
 	finishReason := stopReasonClaude2OpenAI(&stopReason)
 	if finishReason != "null" {
@@ -136,12 +212,27 @@ func ResponseClaude2OpenAI(claudeResponse *Response) *openai.TextResponse {
 	if len(claudeResponse.Content) > 0 {
 		responseText = claudeResponse.Content[0].Text
 	}
+	tools := make([]model.Tool, 0)
+	for _, v := range claudeResponse.Content {
+		if v.Type == "tool_use" {
+			args, _ := json.Marshal(v.Input)
+			tools = append(tools, model.Tool{
+				Id:   v.Id,
+				Type: "function", // compatible with other OpenAI derivative applications
+				Function: model.Function{
+					Name:      v.Name,
+					Arguments: string(args),
+				},
+			})
+		}
+	}
 	choice := openai.TextResponseChoice{
 		Index: 0,
 		Message: model.Message{
-			Role:    "assistant",
+			Role:      "assistant",
-			Content: responseText,
+			Content:   responseText,
-			Name:    nil,
+			Name:      nil,
+			ToolCalls: tools,
 		},
 		FinishReason: stopReasonClaude2OpenAI(claudeResponse.StopReason),
 	}
@@ -176,6 +267,7 @@ func StreamHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusC
 	var usage model.Usage
 	var modelName string
 	var id string
+	var lastToolCallChoice openai.ChatCompletionsStreamResponseChoice
 
 	for scanner.Scan() {
 		data := scanner.Text()
@@ -196,9 +288,20 @@ func StreamHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusC
 		if meta != nil {
 			usage.PromptTokens += meta.Usage.InputTokens
 			usage.CompletionTokens += meta.Usage.OutputTokens
-			modelName = meta.Model
+			if len(meta.Id) > 0 { // only message_start has an id, otherwise it's a finish_reason event.
-			id = fmt.Sprintf("chatcmpl-%s", meta.Id)
+				modelName = meta.Model
-			continue
+				id = fmt.Sprintf("chatcmpl-%s", meta.Id)
+				continue
+			} else { // finish_reason case
+				if len(lastToolCallChoice.Delta.ToolCalls) > 0 {
+					lastArgs := &lastToolCallChoice.Delta.ToolCalls[len(lastToolCallChoice.Delta.ToolCalls)-1].Function
+					if len(lastArgs.Arguments.(string)) == 0 { // compatible with OpenAI sending an empty object `{}` when no arguments.
+						lastArgs.Arguments = "{}"
+						response.Choices[len(response.Choices)-1].Delta.Content = nil
+						response.Choices[len(response.Choices)-1].Delta.ToolCalls = lastToolCallChoice.Delta.ToolCalls
+					}
+				}
+			}
 		}
 		if response == nil {
 			continue
@@ -207,6 +310,12 @@ func StreamHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusC
 		response.Id = id
 		response.Model = modelName
 		response.Created = createdTime
+
+		for _, choice := range response.Choices {
+			if len(choice.Delta.ToolCalls) > 0 {
+				lastToolCallChoice = choice
+			}
+		}
 		err = render.ObjectData(c, response)
 		if err != nil {
 			logger.SysError(err.Error())

relay/adaptor/anthropic/model.go

@@ -16,6 +16,12 @@ type Content struct {
 	Type   string       `json:"type"`
 	Text   string       `json:"text,omitempty"`
 	Source *ImageSource `json:"source,omitempty"`
+	// tool_calls
+	Id        string `json:"id,omitempty"`
+	Name      string `json:"name,omitempty"`
+	Input     any    `json:"input,omitempty"`
+	Content   string `json:"content,omitempty"`
+	ToolUseId string `json:"tool_use_id,omitempty"`
 }
 
 type Message struct {
@@ -23,6 +29,18 @@ type Message struct {
 	Content []Content `json:"content"`
 }
 
+type Tool struct {
+	Name        string      `json:"name"`
+	Description string      `json:"description,omitempty"`
+	InputSchema InputSchema `json:"input_schema"`
+}
+
+type InputSchema struct {
+	Type       string `json:"type"`
+	Properties any    `json:"properties,omitempty"`
+	Required   any    `json:"required,omitempty"`
+}
+
 type Request struct {
 	Model         string    `json:"model"`
 	Messages      []Message `json:"messages"`
@@ -33,6 +51,8 @@ type Request struct {
 	Temperature   float64   `json:"temperature,omitempty"`
 	TopP          float64   `json:"top_p,omitempty"`
 	TopK          int       `json:"top_k,omitempty"`
+	Tools         []Tool    `json:"tools,omitempty"`
+	ToolChoice    any       `json:"tool_choice,omitempty"`
 	//Metadata    `json:"metadata,omitempty"`
 }
 
@@ -61,6 +81,7 @@ type Response struct {
 type Delta struct {
 	Type         string  `json:"type"`
 	Text         string  `json:"text"`
+	PartialJson  string  `json:"partial_json,omitempty"`
 	StopReason   *string `json:"stop_reason"`
 	StopSequence *string `json:"stop_sequence"`
 }

relay/adaptor/aws/adaptor.go

@@ -1,17 +1,16 @@
 package aws
 
 import (
-	"github.com/aws/aws-sdk-go-v2/aws"
+	"errors"
-	"github.com/aws/aws-sdk-go-v2/credentials"
-	"github.com/aws/aws-sdk-go-v2/service/bedrockruntime"
-	"github.com/songquanpeng/one-api/common/ctxkey"
 	"io"
 	"net/http"
 
+	"github.com/aws/aws-sdk-go-v2/aws"
+	"github.com/aws/aws-sdk-go-v2/credentials"
+	"github.com/aws/aws-sdk-go-v2/service/bedrockruntime"
 	"github.com/gin-gonic/gin"
-	"github.com/pkg/errors"
 	"github.com/songquanpeng/one-api/relay/adaptor"
-	"github.com/songquanpeng/one-api/relay/adaptor/anthropic"
+	"github.com/songquanpeng/one-api/relay/adaptor/aws/utils"
 	"github.com/songquanpeng/one-api/relay/meta"
 	"github.com/songquanpeng/one-api/relay/model"
 )
@@ -19,18 +18,52 @@ import (
 var _ adaptor.Adaptor = new(Adaptor)
 
 type Adaptor struct {
-	meta      *meta.Meta
+	awsAdapter utils.AwsAdapter
-	awsClient *bedrockruntime.Client
+
+	Meta      *meta.Meta
+	AwsClient *bedrockruntime.Client
 }
 
 func (a *Adaptor) Init(meta *meta.Meta) {
-	a.meta = meta
+	a.Meta = meta
-	a.awsClient = bedrockruntime.New(bedrockruntime.Options{
+	a.AwsClient = bedrockruntime.New(bedrockruntime.Options{
 		Region:      meta.Config.Region,
 		Credentials: aws.NewCredentialsCache(credentials.NewStaticCredentialsProvider(meta.Config.AK, meta.Config.SK, "")),
 	})
 }
 
+func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+
+	adaptor := GetAdaptor(request.Model)
+	if adaptor == nil {
+		return nil, errors.New("adaptor not found")
+	}
+
+	a.awsAdapter = adaptor
+	return adaptor.ConvertRequest(c, relayMode, request)
+}
+
+func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
+	if a.awsAdapter == nil {
+		return nil, utils.WrapErr(errors.New("awsAdapter is nil"))
+	}
+	return a.awsAdapter.DoResponse(c, a.AwsClient, meta)
+}
+
+func (a *Adaptor) GetModelList() (models []string) {
+	for model := range adaptors {
+		models = append(models, model)
+	}
+	return
+}
+
+func (a *Adaptor) GetChannelName() string {
+	return "aws"
+}
+
 func (a *Adaptor) GetRequestURL(meta *meta.Meta) (string, error) {
 	return "", nil
 }
@@ -39,17 +72,6 @@ func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *me
 	return nil
 }
 
-func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error) {
-	if request == nil {
-		return nil, errors.New("request is nil")
-	}
-
-	claudeReq := anthropic.ConvertRequest(*request)
-	c.Set(ctxkey.RequestModel, request.Model)
-	c.Set(ctxkey.ConvertedRequest, claudeReq)
-	return claudeReq, nil
-}
-
 func (a *Adaptor) ConvertImageRequest(request *model.ImageRequest) (any, error) {
 	if request == nil {
 		return nil, errors.New("request is nil")
@@ -60,23 +82,3 @@ func (a *Adaptor) ConvertImageRequest(request *model.ImageRequest) (any, error)
 func (a *Adaptor) DoRequest(c *gin.Context, meta *meta.Meta, requestBody io.Reader) (*http.Response, error) {
 	return nil, nil
 }
-
-func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
-	if meta.IsStream {
-		err, usage = StreamHandler(c, a.awsClient)
-	} else {
-		err, usage = Handler(c, a.awsClient, meta.ActualModelName)
-	}
-	return
-}
-
-func (a *Adaptor) GetModelList() (models []string) {
-	for n := range awsModelIDMap {
-		models = append(models, n)
-	}
-	return
-}
-
-func (a *Adaptor) GetChannelName() string {
-	return "aws"
-}

relay/adaptor/aws/claude/adapter.go

@@ -0,0 +1,37 @@
+package aws
+
+import (
+	"github.com/aws/aws-sdk-go-v2/service/bedrockruntime"
+	"github.com/gin-gonic/gin"
+	"github.com/pkg/errors"
+	"github.com/songquanpeng/one-api/common/ctxkey"
+	"github.com/songquanpeng/one-api/relay/adaptor/anthropic"
+	"github.com/songquanpeng/one-api/relay/adaptor/aws/utils"
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/model"
+)
+
+var _ utils.AwsAdapter = new(Adaptor)
+
+type Adaptor struct {
+}
+
+func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+
+	claudeReq := anthropic.ConvertRequest(*request)
+	c.Set(ctxkey.RequestModel, request.Model)
+	c.Set(ctxkey.ConvertedRequest, claudeReq)
+	return claudeReq, nil
+}
+
+func (a *Adaptor) DoResponse(c *gin.Context, awsCli *bedrockruntime.Client, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
+	if meta.IsStream {
+		err, usage = StreamHandler(c, awsCli)
+	} else {
+		err, usage = Handler(c, awsCli, meta.ActualModelName)
+	}
+	return
+}

relay/adaptor/aws/claude/main.go

@@ -5,7 +5,6 @@ import (
 	"bytes"
 	"encoding/json"
 	"fmt"
-	"github.com/songquanpeng/one-api/common/ctxkey"
 	"io"
 	"net/http"
 
@@ -16,23 +15,17 @@ import (
 	"github.com/jinzhu/copier"
 	"github.com/pkg/errors"
 	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/ctxkey"
 	"github.com/songquanpeng/one-api/common/helper"
 	"github.com/songquanpeng/one-api/common/logger"
 	"github.com/songquanpeng/one-api/relay/adaptor/anthropic"
+	"github.com/songquanpeng/one-api/relay/adaptor/aws/utils"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
 	relaymodel "github.com/songquanpeng/one-api/relay/model"
 )
 
-func wrapErr(err error) *relaymodel.ErrorWithStatusCode {
-	return &relaymodel.ErrorWithStatusCode{
-		StatusCode: http.StatusInternalServerError,
-		Error: relaymodel.Error{
-			Message: fmt.Sprintf("%s", err.Error()),
-		},
-	}
-}
-
 // https://docs.aws.amazon.com/bedrock/latest/userguide/model-ids.html
-var awsModelIDMap = map[string]string{
+var AwsModelIDMap = map[string]string{
 	"claude-instant-1.2":         "anthropic.claude-instant-v1",
 	"claude-2.0":                 "anthropic.claude-v2",
 	"claude-2.1":                 "anthropic.claude-v2:1",
@@ -43,7 +36,7 @@ var awsModelIDMap = map[string]string{
 }
 
 func awsModelID(requestModel string) (string, error) {
-	if awsModelID, ok := awsModelIDMap[requestModel]; ok {
+	if awsModelID, ok := AwsModelIDMap[requestModel]; ok {
 		return awsModelID, nil
 	}
 
@@ -53,7 +46,7 @@ func awsModelID(requestModel string) (string, error) {
 func Handler(c *gin.Context, awsCli *bedrockruntime.Client, modelName string) (*relaymodel.ErrorWithStatusCode, *relaymodel.Usage) {
 	awsModelId, err := awsModelID(c.GetString(ctxkey.RequestModel))
 	if err != nil {
-		return wrapErr(errors.Wrap(err, "awsModelID")), nil
+		return utils.WrapErr(errors.Wrap(err, "awsModelID")), nil
 	}
 
 	awsReq := &bedrockruntime.InvokeModelInput{
@@ -64,30 +57,30 @@ func Handler(c *gin.Context, awsCli *bedrockruntime.Client, modelName string) (*
 
 	claudeReq_, ok := c.Get(ctxkey.ConvertedRequest)
 	if !ok {
-		return wrapErr(errors.New("request not found")), nil
+		return utils.WrapErr(errors.New("request not found")), nil
 	}
 	claudeReq := claudeReq_.(*anthropic.Request)
 	awsClaudeReq := &Request{
 		AnthropicVersion: "bedrock-2023-05-31",
 	}
 	if err = copier.Copy(awsClaudeReq, claudeReq); err != nil {
-		return wrapErr(errors.Wrap(err, "copy request")), nil
+		return utils.WrapErr(errors.Wrap(err, "copy request")), nil
 	}
 
 	awsReq.Body, err = json.Marshal(awsClaudeReq)
 	if err != nil {
-		return wrapErr(errors.Wrap(err, "marshal request")), nil
+		return utils.WrapErr(errors.Wrap(err, "marshal request")), nil
 	}
 
 	awsResp, err := awsCli.InvokeModel(c.Request.Context(), awsReq)
 	if err != nil {
-		return wrapErr(errors.Wrap(err, "InvokeModel")), nil
+		return utils.WrapErr(errors.Wrap(err, "InvokeModel")), nil
 	}
 
 	claudeResponse := new(anthropic.Response)
 	err = json.Unmarshal(awsResp.Body, claudeResponse)
 	if err != nil {
-		return wrapErr(errors.Wrap(err, "unmarshal response")), nil
+		return utils.WrapErr(errors.Wrap(err, "unmarshal response")), nil
 	}
 
 	openaiResp := anthropic.ResponseClaude2OpenAI(claudeResponse)
@@ -107,7 +100,7 @@ func StreamHandler(c *gin.Context, awsCli *bedrockruntime.Client) (*relaymodel.E
 	createdTime := helper.GetTimestamp()
 	awsModelId, err := awsModelID(c.GetString(ctxkey.RequestModel))
 	if err != nil {
-		return wrapErr(errors.Wrap(err, "awsModelID")), nil
+		return utils.WrapErr(errors.Wrap(err, "awsModelID")), nil
 	}
 
 	awsReq := &bedrockruntime.InvokeModelWithResponseStreamInput{
@@ -118,7 +111,7 @@ func StreamHandler(c *gin.Context, awsCli *bedrockruntime.Client) (*relaymodel.E
 
 	claudeReq_, ok := c.Get(ctxkey.ConvertedRequest)
 	if !ok {
-		return wrapErr(errors.New("request not found")), nil
+		return utils.WrapErr(errors.New("request not found")), nil
 	}
 	claudeReq := claudeReq_.(*anthropic.Request)
 
@@ -126,16 +119,16 @@ func StreamHandler(c *gin.Context, awsCli *bedrockruntime.Client) (*relaymodel.E
 		AnthropicVersion: "bedrock-2023-05-31",
 	}
 	if err = copier.Copy(awsClaudeReq, claudeReq); err != nil {
-		return wrapErr(errors.Wrap(err, "copy request")), nil
+		return utils.WrapErr(errors.Wrap(err, "copy request")), nil
 	}
 	awsReq.Body, err = json.Marshal(awsClaudeReq)
 	if err != nil {
-		return wrapErr(errors.Wrap(err, "marshal request")), nil
+		return utils.WrapErr(errors.Wrap(err, "marshal request")), nil
 	}
 
 	awsResp, err := awsCli.InvokeModelWithResponseStream(c.Request.Context(), awsReq)
 	if err != nil {
-		return wrapErr(errors.Wrap(err, "InvokeModelWithResponseStream")), nil
+		return utils.WrapErr(errors.Wrap(err, "InvokeModelWithResponseStream")), nil
 	}
 	stream := awsResp.GetStream()
 	defer stream.Close()
@@ -143,6 +136,8 @@ func StreamHandler(c *gin.Context, awsCli *bedrockruntime.Client) (*relaymodel.E
 	c.Writer.Header().Set("Content-Type", "text/event-stream")
 	var usage relaymodel.Usage
 	var id string
+	var lastToolCallChoice openai.ChatCompletionsStreamResponseChoice
+
 	c.Stream(func(w io.Writer) bool {
 		event, ok := <-stream.Events()
 		if !ok {
@@ -163,8 +158,19 @@ func StreamHandler(c *gin.Context, awsCli *bedrockruntime.Client) (*relaymodel.E
 			if meta != nil {
 				usage.PromptTokens += meta.Usage.InputTokens
 				usage.CompletionTokens += meta.Usage.OutputTokens
-				id = fmt.Sprintf("chatcmpl-%s", meta.Id)
+				if len(meta.Id) > 0 { // only message_start has an id, otherwise it's a finish_reason event.
-				return true
+					id = fmt.Sprintf("chatcmpl-%s", meta.Id)
+					return true
+				} else { // finish_reason case
+					if len(lastToolCallChoice.Delta.ToolCalls) > 0 {
+						lastArgs := &lastToolCallChoice.Delta.ToolCalls[len(lastToolCallChoice.Delta.ToolCalls)-1].Function
+						if len(lastArgs.Arguments.(string)) == 0 { // compatible with OpenAI sending an empty object `{}` when no arguments.
+							lastArgs.Arguments = "{}"
+							response.Choices[len(response.Choices)-1].Delta.Content = nil
+							response.Choices[len(response.Choices)-1].Delta.ToolCalls = lastToolCallChoice.Delta.ToolCalls
+						}
+					}
+				}
 			}
 			if response == nil {
 				return true
@@ -172,6 +178,12 @@ func StreamHandler(c *gin.Context, awsCli *bedrockruntime.Client) (*relaymodel.E
 			response.Id = id
 			response.Model = c.GetString(ctxkey.OriginalModel)
 			response.Created = createdTime
+
+			for _, choice := range response.Choices {
+				if len(choice.Delta.ToolCalls) > 0 {
+					lastToolCallChoice = choice
+				}
+			}
 			jsonStr, err := json.Marshal(response)
 			if err != nil {
 				logger.SysError("error marshalling stream response: " + err.Error())

relay/adaptor/aws/claude/model.go

@@ -9,9 +9,12 @@ type Request struct {
 	// AnthropicVersion should be "bedrock-2023-05-31"
 	AnthropicVersion string              `json:"anthropic_version"`
 	Messages         []anthropic.Message `json:"messages"`
+	System           string              `json:"system,omitempty"`
 	MaxTokens        int                 `json:"max_tokens,omitempty"`
 	Temperature      float64             `json:"temperature,omitempty"`
 	TopP             float64             `json:"top_p,omitempty"`
 	TopK             int                 `json:"top_k,omitempty"`
 	StopSequences    []string            `json:"stop_sequences,omitempty"`
+	Tools            []anthropic.Tool    `json:"tools,omitempty"`
+	ToolChoice       any                 `json:"tool_choice,omitempty"`
 }

relay/adaptor/aws/llama3/adapter.go

@@ -0,0 +1,37 @@
+package aws
+
+import (
+	"github.com/aws/aws-sdk-go-v2/service/bedrockruntime"
+	"github.com/songquanpeng/one-api/common/ctxkey"
+
+	"github.com/gin-gonic/gin"
+	"github.com/pkg/errors"
+	"github.com/songquanpeng/one-api/relay/adaptor/aws/utils"
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/model"
+)
+
+var _ utils.AwsAdapter = new(Adaptor)
+
+type Adaptor struct {
+}
+
+func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+
+	llamaReq := ConvertRequest(*request)
+	c.Set(ctxkey.RequestModel, request.Model)
+	c.Set(ctxkey.ConvertedRequest, llamaReq)
+	return llamaReq, nil
+}
+
+func (a *Adaptor) DoResponse(c *gin.Context, awsCli *bedrockruntime.Client, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
+	if meta.IsStream {
+		err, usage = StreamHandler(c, awsCli)
+	} else {
+		err, usage = Handler(c, awsCli, meta.ActualModelName)
+	}
+	return
+}

relay/adaptor/aws/llama3/main.go

@@ -0,0 +1,231 @@
+// Package aws provides the AWS adaptor for the relay service.
+package aws
+
+import (
+	"bytes"
+	"encoding/json"
+	"fmt"
+	"io"
+	"net/http"
+	"text/template"
+
+	"github.com/songquanpeng/one-api/common/ctxkey"
+	"github.com/songquanpeng/one-api/common/random"
+
+	"github.com/aws/aws-sdk-go-v2/aws"
+	"github.com/aws/aws-sdk-go-v2/service/bedrockruntime"
+	"github.com/aws/aws-sdk-go-v2/service/bedrockruntime/types"
+	"github.com/gin-gonic/gin"
+	"github.com/pkg/errors"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/helper"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/relay/adaptor/aws/utils"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
+	relaymodel "github.com/songquanpeng/one-api/relay/model"
+)
+
+// Only support llama-3-8b and llama-3-70b instruction models
+// https://docs.aws.amazon.com/bedrock/latest/userguide/model-ids.html
+var AwsModelIDMap = map[string]string{
+	"llama3-8b-8192":  "meta.llama3-8b-instruct-v1:0",
+	"llama3-70b-8192": "meta.llama3-70b-instruct-v1:0",
+}
+
+func awsModelID(requestModel string) (string, error) {
+	if awsModelID, ok := AwsModelIDMap[requestModel]; ok {
+		return awsModelID, nil
+	}
+
+	return "", errors.Errorf("model %s not found", requestModel)
+}
+
+// promptTemplate with range
+const promptTemplate = `<|begin_of_text|>{{range .Messages}}<|start_header_id|>{{.Role}}<|end_header_id|>{{.StringContent}}<|eot_id|>{{end}}<|start_header_id|>assistant<|end_header_id|>
+`
+
+var promptTpl = template.Must(template.New("llama3-chat").Parse(promptTemplate))
+
+func RenderPrompt(messages []relaymodel.Message) string {
+	var buf bytes.Buffer
+	err := promptTpl.Execute(&buf, struct{ Messages []relaymodel.Message }{messages})
+	if err != nil {
+		logger.SysError("error rendering prompt messages: " + err.Error())
+	}
+	return buf.String()
+}
+
+func ConvertRequest(textRequest relaymodel.GeneralOpenAIRequest) *Request {
+	llamaRequest := Request{
+		MaxGenLen:   textRequest.MaxTokens,
+		Temperature: textRequest.Temperature,
+		TopP:        textRequest.TopP,
+	}
+	if llamaRequest.MaxGenLen == 0 {
+		llamaRequest.MaxGenLen = 2048
+	}
+	prompt := RenderPrompt(textRequest.Messages)
+	llamaRequest.Prompt = prompt
+	return &llamaRequest
+}
+
+func Handler(c *gin.Context, awsCli *bedrockruntime.Client, modelName string) (*relaymodel.ErrorWithStatusCode, *relaymodel.Usage) {
+	awsModelId, err := awsModelID(c.GetString(ctxkey.RequestModel))
+	if err != nil {
+		return utils.WrapErr(errors.Wrap(err, "awsModelID")), nil
+	}
+
+	awsReq := &bedrockruntime.InvokeModelInput{
+		ModelId:     aws.String(awsModelId),
+		Accept:      aws.String("application/json"),
+		ContentType: aws.String("application/json"),
+	}
+
+	llamaReq, ok := c.Get(ctxkey.ConvertedRequest)
+	if !ok {
+		return utils.WrapErr(errors.New("request not found")), nil
+	}
+
+	awsReq.Body, err = json.Marshal(llamaReq)
+	if err != nil {
+		return utils.WrapErr(errors.Wrap(err, "marshal request")), nil
+	}
+
+	awsResp, err := awsCli.InvokeModel(c.Request.Context(), awsReq)
+	if err != nil {
+		return utils.WrapErr(errors.Wrap(err, "InvokeModel")), nil
+	}
+
+	var llamaResponse Response
+	err = json.Unmarshal(awsResp.Body, &llamaResponse)
+	if err != nil {
+		return utils.WrapErr(errors.Wrap(err, "unmarshal response")), nil
+	}
+
+	openaiResp := ResponseLlama2OpenAI(&llamaResponse)
+	openaiResp.Model = modelName
+	usage := relaymodel.Usage{
+		PromptTokens:     llamaResponse.PromptTokenCount,
+		CompletionTokens: llamaResponse.GenerationTokenCount,
+		TotalTokens:      llamaResponse.PromptTokenCount + llamaResponse.GenerationTokenCount,
+	}
+	openaiResp.Usage = usage
+
+	c.JSON(http.StatusOK, openaiResp)
+	return nil, &usage
+}
+
+func ResponseLlama2OpenAI(llamaResponse *Response) *openai.TextResponse {
+	var responseText string
+	if len(llamaResponse.Generation) > 0 {
+		responseText = llamaResponse.Generation
+	}
+	choice := openai.TextResponseChoice{
+		Index: 0,
+		Message: relaymodel.Message{
+			Role:    "assistant",
+			Content: responseText,
+			Name:    nil,
+		},
+		FinishReason: llamaResponse.StopReason,
+	}
+	fullTextResponse := openai.TextResponse{
+		Id:      fmt.Sprintf("chatcmpl-%s", random.GetUUID()),
+		Object:  "chat.completion",
+		Created: helper.GetTimestamp(),
+		Choices: []openai.TextResponseChoice{choice},
+	}
+	return &fullTextResponse
+}
+
+func StreamHandler(c *gin.Context, awsCli *bedrockruntime.Client) (*relaymodel.ErrorWithStatusCode, *relaymodel.Usage) {
+	createdTime := helper.GetTimestamp()
+	awsModelId, err := awsModelID(c.GetString(ctxkey.RequestModel))
+	if err != nil {
+		return utils.WrapErr(errors.Wrap(err, "awsModelID")), nil
+	}
+
+	awsReq := &bedrockruntime.InvokeModelWithResponseStreamInput{
+		ModelId:     aws.String(awsModelId),
+		Accept:      aws.String("application/json"),
+		ContentType: aws.String("application/json"),
+	}
+
+	llamaReq, ok := c.Get(ctxkey.ConvertedRequest)
+	if !ok {
+		return utils.WrapErr(errors.New("request not found")), nil
+	}
+
+	awsReq.Body, err = json.Marshal(llamaReq)
+	if err != nil {
+		return utils.WrapErr(errors.Wrap(err, "marshal request")), nil
+	}
+
+	awsResp, err := awsCli.InvokeModelWithResponseStream(c.Request.Context(), awsReq)
+	if err != nil {
+		return utils.WrapErr(errors.Wrap(err, "InvokeModelWithResponseStream")), nil
+	}
+	stream := awsResp.GetStream()
+	defer stream.Close()
+
+	c.Writer.Header().Set("Content-Type", "text/event-stream")
+	var usage relaymodel.Usage
+	c.Stream(func(w io.Writer) bool {
+		event, ok := <-stream.Events()
+		if !ok {
+			c.Render(-1, common.CustomEvent{Data: "data: [DONE]"})
+			return false
+		}
+
+		switch v := event.(type) {
+		case *types.ResponseStreamMemberChunk:
+			var llamaResp StreamResponse
+			err := json.NewDecoder(bytes.NewReader(v.Value.Bytes)).Decode(&llamaResp)
+			if err != nil {
+				logger.SysError("error unmarshalling stream response: " + err.Error())
+				return false
+			}
+
+			if llamaResp.PromptTokenCount > 0 {
+				usage.PromptTokens = llamaResp.PromptTokenCount
+			}
+			if llamaResp.StopReason == "stop" {
+				usage.CompletionTokens = llamaResp.GenerationTokenCount
+				usage.TotalTokens = usage.PromptTokens + usage.CompletionTokens
+			}
+			response := StreamResponseLlama2OpenAI(&llamaResp)
+			response.Id = fmt.Sprintf("chatcmpl-%s", random.GetUUID())
+			response.Model = c.GetString(ctxkey.OriginalModel)
+			response.Created = createdTime
+			jsonStr, err := json.Marshal(response)
+			if err != nil {
+				logger.SysError("error marshalling stream response: " + err.Error())
+				return true
+			}
+			c.Render(-1, common.CustomEvent{Data: "data: " + string(jsonStr)})
+			return true
+		case *types.UnknownUnionMember:
+			fmt.Println("unknown tag:", v.Tag)
+			return false
+		default:
+			fmt.Println("union is nil or unknown type")
+			return false
+		}
+	})
+
+	return nil, &usage
+}
+
+func StreamResponseLlama2OpenAI(llamaResponse *StreamResponse) *openai.ChatCompletionsStreamResponse {
+	var choice openai.ChatCompletionsStreamResponseChoice
+	choice.Delta.Content = llamaResponse.Generation
+	choice.Delta.Role = "assistant"
+	finishReason := llamaResponse.StopReason
+	if finishReason != "null" {
+		choice.FinishReason = &finishReason
+	}
+	var openaiResponse openai.ChatCompletionsStreamResponse
+	openaiResponse.Object = "chat.completion.chunk"
+	openaiResponse.Choices = []openai.ChatCompletionsStreamResponseChoice{choice}
+	return &openaiResponse
+}

relay/adaptor/aws/llama3/main_test.go

@@ -0,0 +1,45 @@
+package aws_test
+
+import (
+	"testing"
+
+	aws "github.com/songquanpeng/one-api/relay/adaptor/aws/llama3"
+	relaymodel "github.com/songquanpeng/one-api/relay/model"
+	"github.com/stretchr/testify/assert"
+)
+
+func TestRenderPrompt(t *testing.T) {
+	messages := []relaymodel.Message{
+		{
+			Role:    "user",
+			Content: "What's your name?",
+		},
+	}
+	prompt := aws.RenderPrompt(messages)
+	expected := `<|begin_of_text|><|start_header_id|>user<|end_header_id|>What's your name?<|eot_id|><|start_header_id|>assistant<|end_header_id|>
+`
+	assert.Equal(t, expected, prompt)
+
+	messages = []relaymodel.Message{
+		{
+			Role:    "system",
+			Content: "Your name is Kat. You are a detective.",
+		},
+		{
+			Role:    "user",
+			Content: "What's your name?",
+		},
+		{
+			Role:    "assistant",
+			Content: "Kat",
+		},
+		{
+			Role:    "user",
+			Content: "What's your job?",
+		},
+	}
+	prompt = aws.RenderPrompt(messages)
+	expected = `<|begin_of_text|><|start_header_id|>system<|end_header_id|>Your name is Kat. You are a detective.<|eot_id|><|start_header_id|>user<|end_header_id|>What's your name?<|eot_id|><|start_header_id|>assistant<|end_header_id|>Kat<|eot_id|><|start_header_id|>user<|end_header_id|>What's your job?<|eot_id|><|start_header_id|>assistant<|end_header_id|>
+`
+	assert.Equal(t, expected, prompt)
+}

relay/adaptor/aws/llama3/model.go

@@ -0,0 +1,29 @@
+package aws
+
+// Request is the request to AWS Llama3
+//
+// https://docs.aws.amazon.com/bedrock/latest/userguide/model-parameters-meta.html
+type Request struct {
+	Prompt      string  `json:"prompt"`
+	MaxGenLen   int     `json:"max_gen_len,omitempty"`
+	Temperature float64 `json:"temperature,omitempty"`
+	TopP        float64 `json:"top_p,omitempty"`
+}
+
+// Response is the response from AWS Llama3
+//
+// https://docs.aws.amazon.com/bedrock/latest/userguide/model-parameters-meta.html
+type Response struct {
+	Generation           string `json:"generation"`
+	PromptTokenCount     int    `json:"prompt_token_count"`
+	GenerationTokenCount int    `json:"generation_token_count"`
+	StopReason           string `json:"stop_reason"`
+}
+
+// {'generation': 'Hi', 'prompt_token_count': 15, 'generation_token_count': 1, 'stop_reason': None}
+type StreamResponse struct {
+	Generation           string `json:"generation"`
+	PromptTokenCount     int    `json:"prompt_token_count"`
+	GenerationTokenCount int    `json:"generation_token_count"`
+	StopReason           string `json:"stop_reason"`
+}

relay/adaptor/aws/registry.go

@@ -0,0 +1,39 @@
+package aws
+
+import (
+	claude "github.com/songquanpeng/one-api/relay/adaptor/aws/claude"
+	llama3 "github.com/songquanpeng/one-api/relay/adaptor/aws/llama3"
+	"github.com/songquanpeng/one-api/relay/adaptor/aws/utils"
+)
+
+type AwsModelType int
+
+const (
+	AwsClaude AwsModelType = iota + 1
+	AwsLlama3
+)
+
+var (
+	adaptors = map[string]AwsModelType{}
+)
+
+func init() {
+	for model := range claude.AwsModelIDMap {
+		adaptors[model] = AwsClaude
+	}
+	for model := range llama3.AwsModelIDMap {
+		adaptors[model] = AwsLlama3
+	}
+}
+
+func GetAdaptor(model string) utils.AwsAdapter {
+	adaptorType := adaptors[model]
+	switch adaptorType {
+	case AwsClaude:
+		return &claude.Adaptor{}
+	case AwsLlama3:
+		return &llama3.Adaptor{}
+	default:
+		return nil
+	}
+}

relay/adaptor/aws/utils/adaptor.go

@@ -0,0 +1,51 @@
+package utils
+
+import (
+	"errors"
+	"io"
+	"net/http"
+
+	"github.com/aws/aws-sdk-go-v2/aws"
+	"github.com/aws/aws-sdk-go-v2/credentials"
+	"github.com/aws/aws-sdk-go-v2/service/bedrockruntime"
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/model"
+)
+
+type AwsAdapter interface {
+	ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error)
+	DoResponse(c *gin.Context, awsCli *bedrockruntime.Client, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode)
+}
+
+type Adaptor struct {
+	Meta      *meta.Meta
+	AwsClient *bedrockruntime.Client
+}
+
+func (a *Adaptor) Init(meta *meta.Meta) {
+	a.Meta = meta
+	a.AwsClient = bedrockruntime.New(bedrockruntime.Options{
+		Region:      meta.Config.Region,
+		Credentials: aws.NewCredentialsCache(credentials.NewStaticCredentialsProvider(meta.Config.AK, meta.Config.SK, "")),
+	})
+}
+
+func (a *Adaptor) GetRequestURL(meta *meta.Meta) (string, error) {
+	return "", nil
+}
+
+func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *meta.Meta) error {
+	return nil
+}
+
+func (a *Adaptor) ConvertImageRequest(request *model.ImageRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+	return request, nil
+}
+
+func (a *Adaptor) DoRequest(c *gin.Context, meta *meta.Meta, requestBody io.Reader) (*http.Response, error) {
+	return nil, nil
+}

relay/adaptor/aws/utils/utils.go

@@ -0,0 +1,16 @@
+package utils
+
+import (
+	"net/http"
+
+	relaymodel "github.com/songquanpeng/one-api/relay/model"
+)
+
+func WrapErr(err error) *relaymodel.ErrorWithStatusCode {
+	return &relaymodel.ErrorWithStatusCode{
+		StatusCode: http.StatusInternalServerError,
+		Error: relaymodel.Error{
+			Message: err.Error(),
+		},
+	}
+}

relay/adaptor/cloudflare/adaptor.go

@@ -5,11 +5,13 @@ import (
 	"fmt"
 	"io"
 	"net/http"
+	"strings"
 
 	"github.com/gin-gonic/gin"
 	"github.com/songquanpeng/one-api/relay/adaptor"
 	"github.com/songquanpeng/one-api/relay/meta"
 	"github.com/songquanpeng/one-api/relay/model"
+	"github.com/songquanpeng/one-api/relay/relaymode"
 )
 
 type Adaptor struct {
@@ -27,8 +29,33 @@ func (a *Adaptor) Init(meta *meta.Meta) {
 	a.meta = meta
 }
 
+// WorkerAI cannot be used across accounts with AIGateWay
+// https://developers.cloudflare.com/ai-gateway/providers/workersai/#openai-compatible-endpoints
+// https://gateway.ai.cloudflare.com/v1/{account_id}/{gateway_id}/workers-ai
+func (a *Adaptor) isAIGateWay(baseURL string) bool {
+	return strings.HasPrefix(baseURL, "https://gateway.ai.cloudflare.com") && strings.HasSuffix(baseURL, "/workers-ai")
+}
+
 func (a *Adaptor) GetRequestURL(meta *meta.Meta) (string, error) {
-	return fmt.Sprintf("%s/client/v4/accounts/%s/ai/run/%s", meta.BaseURL, meta.Config.UserID, meta.ActualModelName), nil
+	isAIGateWay := a.isAIGateWay(meta.BaseURL)
+	var urlPrefix string
+	if isAIGateWay {
+		urlPrefix = meta.BaseURL
+	} else {
+		urlPrefix = fmt.Sprintf("%s/client/v4/accounts/%s/ai", meta.BaseURL, meta.Config.UserID)
+	}
+
+	switch meta.Mode {
+	case relaymode.ChatCompletions:
+		return fmt.Sprintf("%s/v1/chat/completions", urlPrefix), nil
+	case relaymode.Embeddings:
+		return fmt.Sprintf("%s/v1/embeddings", urlPrefix), nil
+	default:
+		if isAIGateWay {
+			return fmt.Sprintf("%s/%s", urlPrefix, meta.ActualModelName), nil
+		}
+		return fmt.Sprintf("%s/run/%s", urlPrefix, meta.ActualModelName), nil
+	}
 }
 
 func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *meta.Meta) error {
@@ -41,7 +68,14 @@ func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.G
 	if request == nil {
 		return nil, errors.New("request is nil")
 	}
-	return ConvertRequest(*request), nil
+	switch relayMode {
+	case relaymode.Completions:
+		return ConvertCompletionsRequest(*request), nil
+	case relaymode.ChatCompletions, relaymode.Embeddings:
+		return request, nil
+	default:
+		return nil, errors.New("not implemented")
+	}
 }
 
 func (a *Adaptor) DoRequest(c *gin.Context, meta *meta.Meta, requestBody io.Reader) (*http.Response, error) {

relay/adaptor/cloudflare/main.go

@@ -3,11 +3,13 @@ package cloudflare
 import (
 	"bufio"
 	"encoding/json"
-	"github.com/songquanpeng/one-api/common/render"
 	"io"
 	"net/http"
 	"strings"
 
+	"github.com/songquanpeng/one-api/common/ctxkey"
+	"github.com/songquanpeng/one-api/common/render"
+
 	"github.com/gin-gonic/gin"
 	"github.com/songquanpeng/one-api/common"
 	"github.com/songquanpeng/one-api/common/helper"
@@ -16,57 +18,23 @@ import (
 	"github.com/songquanpeng/one-api/relay/model"
 )
 
-func ConvertRequest(textRequest model.GeneralOpenAIRequest) *Request {
+func ConvertCompletionsRequest(textRequest model.GeneralOpenAIRequest) *Request {
-	var promptBuilder strings.Builder
+	p, _ := textRequest.Prompt.(string)
-	for _, message := range textRequest.Messages {
-		promptBuilder.WriteString(message.StringContent())
-		promptBuilder.WriteString("\n") // 添加换行符来分隔每个消息
-	}
-
 	return &Request{
+		Prompt:      p,
 		MaxTokens:   textRequest.MaxTokens,
-		Prompt:      promptBuilder.String(),
 		Stream:      textRequest.Stream,
 		Temperature: textRequest.Temperature,
 	}
 }
 
-func ResponseCloudflare2OpenAI(cloudflareResponse *Response) *openai.TextResponse {
-	choice := openai.TextResponseChoice{
-		Index: 0,
-		Message: model.Message{
-			Role:    "assistant",
-			Content: cloudflareResponse.Result.Response,
-		},
-		FinishReason: "stop",
-	}
-	fullTextResponse := openai.TextResponse{
-		Object:  "chat.completion",
-		Created: helper.GetTimestamp(),
-		Choices: []openai.TextResponseChoice{choice},
-	}
-	return &fullTextResponse
-}
-
-func StreamResponseCloudflare2OpenAI(cloudflareResponse *StreamResponse) *openai.ChatCompletionsStreamResponse {
-	var choice openai.ChatCompletionsStreamResponseChoice
-	choice.Delta.Content = cloudflareResponse.Response
-	choice.Delta.Role = "assistant"
-	openaiResponse := openai.ChatCompletionsStreamResponse{
-		Object:  "chat.completion.chunk",
-		Choices: []openai.ChatCompletionsStreamResponseChoice{choice},
-		Created: helper.GetTimestamp(),
-	}
-	return &openaiResponse
-}
-
 func StreamHandler(c *gin.Context, resp *http.Response, promptTokens int, modelName string) (*model.ErrorWithStatusCode, *model.Usage) {
 	scanner := bufio.NewScanner(resp.Body)
 	scanner.Split(bufio.ScanLines)
 
 	common.SetEventStreamHeaders(c)
 	id := helper.GetResponseID(c)
-	responseModel := c.GetString("original_model")
+	responseModel := c.GetString(ctxkey.OriginalModel)
 	var responseText string
 
 	for scanner.Scan() {
@@ -77,22 +45,22 @@ func StreamHandler(c *gin.Context, resp *http.Response, promptTokens int, modelN
 		data = strings.TrimPrefix(data, "data: ")
 		data = strings.TrimSuffix(data, "\r")
 
-		var cloudflareResponse StreamResponse
+		if data == "[DONE]" {
-		err := json.Unmarshal([]byte(data), &cloudflareResponse)
+			break
+		}
+
+		var response openai.ChatCompletionsStreamResponse
+		err := json.Unmarshal([]byte(data), &response)
 		if err != nil {
 			logger.SysError("error unmarshalling stream response: " + err.Error())
 			continue
 		}
-
+		for _, v := range response.Choices {
-		response := StreamResponseCloudflare2OpenAI(&cloudflareResponse)
+			v.Delta.Role = "assistant"
-		if response == nil {
+			responseText += v.Delta.StringContent()
-			continue
 		}
-
-		responseText += cloudflareResponse.Response
 		response.Id = id
-		response.Model = responseModel
+		response.Model = modelName
-
 		err = render.ObjectData(c, response)
 		if err != nil {
 			logger.SysError(err.Error())
@@ -123,22 +91,25 @@ func Handler(c *gin.Context, resp *http.Response, promptTokens int, modelName st
 	if err != nil {
 		return openai.ErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), nil
 	}
-	var cloudflareResponse Response
+	var response openai.TextResponse
-	err = json.Unmarshal(responseBody, &cloudflareResponse)
+	err = json.Unmarshal(responseBody, &response)
 	if err != nil {
 		return openai.ErrorWrapper(err, "unmarshal_response_body_failed", http.StatusInternalServerError), nil
 	}
-	fullTextResponse := ResponseCloudflare2OpenAI(&cloudflareResponse)
+	response.Model = modelName
-	fullTextResponse.Model = modelName
+	var responseText string
-	usage := openai.ResponseText2Usage(cloudflareResponse.Result.Response, modelName, promptTokens)
+	for _, v := range response.Choices {
-	fullTextResponse.Usage = *usage
+		responseText += v.Message.Content.(string)
-	fullTextResponse.Id = helper.GetResponseID(c)
+	}
-	jsonResponse, err := json.Marshal(fullTextResponse)
+	usage := openai.ResponseText2Usage(responseText, modelName, promptTokens)
+	response.Usage = *usage
+	response.Id = helper.GetResponseID(c)
+	jsonResponse, err := json.Marshal(response)
 	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)
+	_, _ = c.Writer.Write(jsonResponse)
 	return nil, usage
 }

relay/adaptor/cloudflare/model.go

@@ -1,25 +1,13 @@
 package cloudflare
 
+import "github.com/songquanpeng/one-api/relay/model"
+
 type Request struct {
-	Lora        string  `json:"lora,omitempty"`
+	Messages    []model.Message `json:"messages,omitempty"`
-	MaxTokens   int     `json:"max_tokens,omitempty"`
+	Lora        string          `json:"lora,omitempty"`
-	Prompt      string  `json:"prompt,omitempty"`
+	MaxTokens   int             `json:"max_tokens,omitempty"`
-	Raw         bool    `json:"raw,omitempty"`
+	Prompt      string          `json:"prompt,omitempty"`
-	Stream      bool    `json:"stream,omitempty"`
+	Raw         bool            `json:"raw,omitempty"`
-	Temperature float64 `json:"temperature,omitempty"`
+	Stream      bool            `json:"stream,omitempty"`
-}
+	Temperature float64         `json:"temperature,omitempty"`
-
-type Result struct {
-	Response string `json:"response"`
-}
-
-type Response struct {
-	Result   Result   `json:"result"`
-	Success  bool     `json:"success"`
-	Errors   []string `json:"errors"`
-	Messages []string `json:"messages"`
-}
-
-type StreamResponse struct {
-	Response string `json:"response"`
 }

relay/adaptor/novita/constants.go

@@ -0,0 +1,19 @@
+package novita
+
+// https://novita.ai/llm-api
+
+var ModelList = []string{
+	"meta-llama/llama-3-8b-instruct",
+	"meta-llama/llama-3-70b-instruct",
+	"nousresearch/hermes-2-pro-llama-3-8b",
+	"nousresearch/nous-hermes-llama2-13b",
+	"mistralai/mistral-7b-instruct",
+	"cognitivecomputations/dolphin-mixtral-8x22b",
+	"sao10k/l3-70b-euryale-v2.1",
+	"sophosympatheia/midnight-rose-70b",
+	"gryphe/mythomax-l2-13b",
+	"Nous-Hermes-2-Mixtral-8x7B-DPO",
+	"lzlv_70b",
+	"teknium/openhermes-2.5-mistral-7b",
+	"microsoft/wizardlm-2-8x22b",
+}

relay/adaptor/novita/main.go

@@ -0,0 +1,15 @@
+package novita
+
+import (
+	"fmt"
+
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/relaymode"
+)
+
+func GetRequestURL(meta *meta.Meta) (string, error) {
+	if meta.Mode == relaymode.ChatCompletions {
+		return fmt.Sprintf("%s/chat/completions", meta.BaseURL), nil
+	}
+	return "", fmt.Errorf("unsupported relay mode %d for novita", meta.Mode)
+}

relay/adaptor/openai/adaptor.go

@@ -3,17 +3,19 @@ package openai
 import (
 	"errors"
 	"fmt"
+	"io"
+	"net/http"
+	"strings"
+
 	"github.com/gin-gonic/gin"
 	"github.com/songquanpeng/one-api/relay/adaptor"
 	"github.com/songquanpeng/one-api/relay/adaptor/doubao"
 	"github.com/songquanpeng/one-api/relay/adaptor/minimax"
+	"github.com/songquanpeng/one-api/relay/adaptor/novita"
 	"github.com/songquanpeng/one-api/relay/channeltype"
 	"github.com/songquanpeng/one-api/relay/meta"
 	"github.com/songquanpeng/one-api/relay/model"
 	"github.com/songquanpeng/one-api/relay/relaymode"
-	"io"
-	"net/http"
-	"strings"
 )
 
 type Adaptor struct {
@@ -48,6 +50,8 @@ func (a *Adaptor) GetRequestURL(meta *meta.Meta) (string, error) {
 		return minimax.GetRequestURL(meta)
 	case channeltype.Doubao:
 		return doubao.GetRequestURL(meta)
+	case channeltype.Novita:
+		return novita.GetRequestURL(meta)
 	default:
 		return GetFullRequestURL(meta.BaseURL, meta.RequestURLPath, meta.ChannelType), nil
 	}

relay/adaptor/openai/compatible.go

@@ -10,6 +10,7 @@ import (
 	"github.com/songquanpeng/one-api/relay/adaptor/minimax"
 	"github.com/songquanpeng/one-api/relay/adaptor/mistral"
 	"github.com/songquanpeng/one-api/relay/adaptor/moonshot"
+	"github.com/songquanpeng/one-api/relay/adaptor/novita"
 	"github.com/songquanpeng/one-api/relay/adaptor/stepfun"
 	"github.com/songquanpeng/one-api/relay/adaptor/togetherai"
 	"github.com/songquanpeng/one-api/relay/channeltype"
@@ -28,6 +29,7 @@ var CompatibleChannels = []int{
 	channeltype.StepFun,
 	channeltype.DeepSeek,
 	channeltype.TogetherAI,
+	channeltype.Novita,
 }
 
 func GetCompatibleChannelMeta(channelType int) (string, []string) {
@@ -56,6 +58,8 @@ func GetCompatibleChannelMeta(channelType int) (string, []string) {
 		return "together.ai", togetherai.ModelList
 	case channeltype.Doubao:
 		return "doubao", doubao.ModelList
+	case channeltype.Novita:
+		return "novita", novita.ModelList
 	default:
 		return "openai", ModelList
 	}

relay/adaptor/openai/main.go

@@ -4,11 +4,12 @@ import (
 	"bufio"
 	"bytes"
 	"encoding/json"
-	"github.com/songquanpeng/one-api/common/render"
 	"io"
 	"net/http"
 	"strings"
 
+	"github.com/songquanpeng/one-api/common/render"
+
 	"github.com/gin-gonic/gin"
 	"github.com/songquanpeng/one-api/common"
 	"github.com/songquanpeng/one-api/common/conv"
@@ -31,6 +32,7 @@ func StreamHandler(c *gin.Context, resp *http.Response, relayMode int) (*model.E
 
 	common.SetEventStreamHeaders(c)
 
+	doneRendered := false
 	for scanner.Scan() {
 		data := scanner.Text()
 		if len(data) < dataPrefixLength { // ignore blank line or wrong format
@@ -41,6 +43,7 @@ func StreamHandler(c *gin.Context, resp *http.Response, relayMode int) (*model.E
 		}
 		if strings.HasPrefix(data[dataPrefixLength:], done) {
 			render.StringData(c, data)
+			doneRendered = true
 			continue
 		}
 		switch relayMode {
@@ -81,7 +84,9 @@ func StreamHandler(c *gin.Context, resp *http.Response, relayMode int) (*model.E
 		logger.SysError("error reading stream: " + err.Error())
 	}
 
-	render.Done(c)
+	if !doneRendered {
+		render.Done(c)
+	}
 
 	err := resp.Body.Close()
 	if err != nil {

relay/adaptor/openai/token.go

@@ -97,7 +97,11 @@ func CountTokenMessages(messages []model.Message, model string) int {
 				m := it.(map[string]any)
 				switch m["type"] {
 				case "text":
-					tokenNum += getTokenNum(tokenEncoder, m["text"].(string))
+					if textValue, ok := m["text"]; ok {
+						if textString, ok := textValue.(string); ok {
+							tokenNum += getTokenNum(tokenEncoder, textString)
+						}
+					}
 				case "image_url":
 					imageUrl, ok := m["image_url"].(map[string]any)
 					if ok {

relay/adaptor/vertexai/adaptor.go

@@ -0,0 +1,117 @@
+package vertexai
+
+import (
+	"errors"
+	"fmt"
+	"io"
+	"net/http"
+	"strings"
+
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/relay/adaptor"
+	channelhelper "github.com/songquanpeng/one-api/relay/adaptor"
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/model"
+	relaymodel "github.com/songquanpeng/one-api/relay/model"
+)
+
+var _ adaptor.Adaptor = new(Adaptor)
+
+const channelName = "vertexai"
+
+type Adaptor struct{}
+
+func (a *Adaptor) Init(meta *meta.Meta) {
+}
+
+func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+
+	adaptor := GetAdaptor(request.Model)
+	if adaptor == nil {
+		return nil, errors.New("adaptor not found")
+	}
+
+	return adaptor.ConvertRequest(c, relayMode, request)
+}
+
+func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
+	adaptor := GetAdaptor(meta.ActualModelName)
+	if adaptor == nil {
+		return nil, &relaymodel.ErrorWithStatusCode{
+			StatusCode: http.StatusInternalServerError,
+			Error: relaymodel.Error{
+				Message: "adaptor not found",
+			},
+		}
+	}
+	return adaptor.DoResponse(c, resp, meta)
+}
+
+func (a *Adaptor) GetModelList() (models []string) {
+	models = modelList
+	return
+}
+
+func (a *Adaptor) GetChannelName() string {
+	return channelName
+}
+
+func (a *Adaptor) GetRequestURL(meta *meta.Meta) (string, error) {
+	suffix := ""
+	if strings.HasPrefix(meta.ActualModelName, "gemini") {
+		if meta.IsStream {
+			suffix = "streamGenerateContent?alt=sse"
+		} else {
+			suffix = "generateContent"
+		}
+	} else {
+		if meta.IsStream {
+			suffix = "streamRawPredict?alt=sse"
+		} else {
+			suffix = "rawPredict"
+		}
+	}
+
+	if meta.BaseURL != "" {
+		return fmt.Sprintf(
+			"%s/v1/projects/%s/locations/%s/publishers/google/models/%s:%s",
+			meta.BaseURL,
+			meta.Config.VertexAIProjectID,
+			meta.Config.Region,
+			meta.ActualModelName,
+			suffix,
+		), nil
+	}
+	return fmt.Sprintf(
+		"https://%s-aiplatform.googleapis.com/v1/projects/%s/locations/%s/publishers/google/models/%s:%s",
+		meta.Config.Region,
+		meta.Config.VertexAIProjectID,
+		meta.Config.Region,
+		meta.ActualModelName,
+		suffix,
+	), nil
+}
+
+func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *meta.Meta) error {
+	adaptor.SetupCommonRequestHeader(c, req, meta)
+	token, err := getToken(c, meta.ChannelId, meta.Config.VertexAIADC)
+	if err != nil {
+		return err
+	}
+	req.Header.Set("Authorization", "Bearer "+token)
+	return nil
+}
+
+func (a *Adaptor) ConvertImageRequest(request *model.ImageRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+	return request, nil
+}
+
+func (a *Adaptor) DoRequest(c *gin.Context, meta *meta.Meta, requestBody io.Reader) (*http.Response, error) {
+	return channelhelper.DoRequestHelper(a, c, meta, requestBody)
+}

relay/adaptor/vertexai/claude/adapter.go

@@ -0,0 +1,55 @@
+package vertexai
+
+import (
+	"net/http"
+
+	"github.com/gin-gonic/gin"
+	"github.com/pkg/errors"
+	"github.com/songquanpeng/one-api/common/ctxkey"
+	"github.com/songquanpeng/one-api/relay/adaptor/anthropic"
+
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/model"
+)
+
+var ModelList = []string{
+	"claude-3-haiku@20240307", "claude-3-opus@20240229", "claude-3-5-sonnet@20240620", "claude-3-sonnet@20240229",
+}
+
+const anthropicVersion = "vertex-2023-10-16"
+
+type Adaptor struct {
+}
+
+func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+
+	claudeReq := anthropic.ConvertRequest(*request)
+	req := Request{
+		AnthropicVersion: anthropicVersion,
+		// Model:            claudeReq.Model,
+		Messages:    claudeReq.Messages,
+		System:      claudeReq.System,
+		MaxTokens:   claudeReq.MaxTokens,
+		Temperature: claudeReq.Temperature,
+		TopP:        claudeReq.TopP,
+		TopK:        claudeReq.TopK,
+		Stream:      claudeReq.Stream,
+		Tools:       claudeReq.Tools,
+	}
+
+	c.Set(ctxkey.RequestModel, request.Model)
+	c.Set(ctxkey.ConvertedRequest, req)
+	return req, nil
+}
+
+func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
+	if meta.IsStream {
+		err, usage = anthropic.StreamHandler(c, resp)
+	} else {
+		err, usage = anthropic.Handler(c, resp, meta.PromptTokens, meta.ActualModelName)
+	}
+	return
+}

relay/adaptor/vertexai/claude/model.go

@@ -0,0 +1,19 @@
+package vertexai
+
+import "github.com/songquanpeng/one-api/relay/adaptor/anthropic"
+
+type Request struct {
+	// AnthropicVersion must be "vertex-2023-10-16"
+	AnthropicVersion string `json:"anthropic_version"`
+	// Model            string              `json:"model"`
+	Messages      []anthropic.Message `json:"messages"`
+	System        string              `json:"system,omitempty"`
+	MaxTokens     int                 `json:"max_tokens,omitempty"`
+	StopSequences []string            `json:"stop_sequences,omitempty"`
+	Stream        bool                `json:"stream,omitempty"`
+	Temperature   float64             `json:"temperature,omitempty"`
+	TopP          float64             `json:"top_p,omitempty"`
+	TopK          int                 `json:"top_k,omitempty"`
+	Tools         []anthropic.Tool    `json:"tools,omitempty"`
+	ToolChoice    any                 `json:"tool_choice,omitempty"`
+}

relay/adaptor/vertexai/gemini/adapter.go

@@ -0,0 +1,49 @@
+package vertexai
+
+import (
+	"net/http"
+
+	"github.com/gin-gonic/gin"
+	"github.com/pkg/errors"
+	"github.com/songquanpeng/one-api/common/ctxkey"
+	"github.com/songquanpeng/one-api/relay/adaptor/gemini"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
+	"github.com/songquanpeng/one-api/relay/relaymode"
+
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/model"
+)
+
+var ModelList = []string{
+	"gemini-1.5-pro-001", "gemini-1.5-flash-001", "gemini-pro", "gemini-pro-vision",
+}
+
+type Adaptor struct {
+}
+
+func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+
+	geminiRequest := gemini.ConvertRequest(*request)
+	c.Set(ctxkey.RequestModel, request.Model)
+	c.Set(ctxkey.ConvertedRequest, geminiRequest)
+	return geminiRequest, nil
+}
+
+func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
+	if meta.IsStream {
+		var responseText string
+		err, responseText = gemini.StreamHandler(c, resp)
+		usage = openai.ResponseText2Usage(responseText, meta.ActualModelName, meta.PromptTokens)
+	} else {
+		switch meta.Mode {
+		case relaymode.Embeddings:
+			err, usage = gemini.EmbeddingHandler(c, resp)
+		default:
+			err, usage = gemini.Handler(c, resp, meta.PromptTokens, meta.ActualModelName)
+		}
+	}
+	return
+}

relay/adaptor/vertexai/registry.go

@@ -0,0 +1,50 @@
+package vertexai
+
+import (
+	"net/http"
+
+	"github.com/gin-gonic/gin"
+	claude "github.com/songquanpeng/one-api/relay/adaptor/vertexai/claude"
+	gemini "github.com/songquanpeng/one-api/relay/adaptor/vertexai/gemini"
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/model"
+)
+
+type VertexAIModelType int
+
+const (
+	VerterAIClaude VertexAIModelType = iota + 1
+	VerterAIGemini
+)
+
+var modelMapping = map[string]VertexAIModelType{}
+var modelList = []string{}
+
+func init() {
+	modelList = append(modelList, claude.ModelList...)
+	for _, model := range claude.ModelList {
+		modelMapping[model] = VerterAIClaude
+	}
+
+	modelList = append(modelList, gemini.ModelList...)
+	for _, model := range gemini.ModelList {
+		modelMapping[model] = VerterAIGemini
+	}
+}
+
+type innerAIAdapter interface {
+	ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error)
+	DoResponse(c *gin.Context, resp *http.Response, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode)
+}
+
+func GetAdaptor(model string) innerAIAdapter {
+	adaptorType := modelMapping[model]
+	switch adaptorType {
+	case VerterAIClaude:
+		return &claude.Adaptor{}
+	case VerterAIGemini:
+		return &gemini.Adaptor{}
+	default:
+		return nil
+	}
+}

relay/adaptor/vertexai/token.go

@@ -0,0 +1,62 @@
+package vertexai
+
+import (
+	"context"
+	"encoding/json"
+	"fmt"
+	"time"
+
+	credentials "cloud.google.com/go/iam/credentials/apiv1"
+	"cloud.google.com/go/iam/credentials/apiv1/credentialspb"
+	"github.com/patrickmn/go-cache"
+	"google.golang.org/api/option"
+)
+
+type ApplicationDefaultCredentials struct {
+	Type                    string `json:"type"`
+	ProjectID               string `json:"project_id"`
+	PrivateKeyID            string `json:"private_key_id"`
+	PrivateKey              string `json:"private_key"`
+	ClientEmail             string `json:"client_email"`
+	ClientID                string `json:"client_id"`
+	AuthURI                 string `json:"auth_uri"`
+	TokenURI                string `json:"token_uri"`
+	AuthProviderX509CertURL string `json:"auth_provider_x509_cert_url"`
+	ClientX509CertURL       string `json:"client_x509_cert_url"`
+	UniverseDomain          string `json:"universe_domain"`
+}
+
+var Cache = cache.New(50*time.Minute, 55*time.Minute)
+
+const defaultScope = "https://www.googleapis.com/auth/cloud-platform"
+
+func getToken(ctx context.Context, channelId int, adcJson string) (string, error) {
+	cacheKey := fmt.Sprintf("vertexai-token-%d", channelId)
+	if token, found := Cache.Get(cacheKey); found {
+		return token.(string), nil
+	}
+	adc := &ApplicationDefaultCredentials{}
+	if err := json.Unmarshal([]byte(adcJson), adc); err != nil {
+		return "", fmt.Errorf("Failed to decode credentials file: %w", err)
+	}
+
+	c, err := credentials.NewIamCredentialsClient(ctx, option.WithCredentialsJSON([]byte(adcJson)))
+	if err != nil {
+		return "", fmt.Errorf("Failed to create client: %w", err)
+	}
+	defer c.Close()
+
+	req := &credentialspb.GenerateAccessTokenRequest{
+		// See https://pkg.go.dev/cloud.google.com/go/iam/credentials/apiv1/credentialspb#GenerateAccessTokenRequest.
+		Name:  fmt.Sprintf("projects/-/serviceAccounts/%s", adc.ClientEmail),
+		Scope: []string{defaultScope},
+	}
+	resp, err := c.GenerateAccessToken(ctx, req)
+	if err != nil {
+		return "", fmt.Errorf("Failed to generate access token: %w", err)
+	}
+	_ = resp
+
+	Cache.Set(cacheKey, resp.AccessToken, cache.DefaultExpiration)
+	return resp.AccessToken, nil
+}

relay/adaptor/xunfei/constants.go

@@ -6,4 +6,5 @@ var ModelList = []string{
 	"SparkDesk-v2.1",
 	"SparkDesk-v3.1",
 	"SparkDesk-v3.5",
+	"SparkDesk-v4.0",
 }

relay/adaptor/xunfei/main.go

@@ -44,7 +44,7 @@ func requestOpenAI2Xunfei(request model.GeneralOpenAIRequest, xunfeiAppId string
 	xunfeiRequest.Parameter.Chat.MaxTokens = request.MaxTokens
 	xunfeiRequest.Payload.Message.Text = messages
 
-	if strings.HasPrefix(domain, "generalv3") {
+	if strings.HasPrefix(domain, "generalv3") || domain == "4.0Ultra" {
 		functions := make([]model.Function, len(request.Tools))
 		for i, tool := range request.Tools {
 			functions[i] = tool.Function
@@ -290,6 +290,8 @@ func apiVersion2domain(apiVersion string) string {
 		return "generalv3"
 	case "v3.5":
 		return "generalv3.5"
+	case "v4.0":
+		return "4.0Ultra"
 	}
 	return "general" + apiVersion
 }

relay/apitype/define.go

@@ -17,6 +17,7 @@ const (
 	Cohere
 	Cloudflare
 	DeepL
+	VertexAI
 
 	Dummy // this one is only for count, do not add any channel after this
 )

relay/billing/ratio/model.go

@@ -2,6 +2,7 @@ package ratio
 
 import (
 	"encoding/json"
+	"fmt"
 	"strings"
 
 	"github.com/songquanpeng/one-api/common/logger"
@@ -125,6 +126,7 @@ var ModelRatio = map[string]float64{
 	"SparkDesk-v2.1":            1.2858, // ￥0.018 / 1k tokens
 	"SparkDesk-v3.1":            1.2858, // ￥0.018 / 1k tokens
 	"SparkDesk-v3.5":            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
 	"embedding_s1_v1":           0.0715, // ¥0.001 / 1k tokens
@@ -168,6 +170,9 @@ var ModelRatio = map[string]float64{
 	"step-1v-32k": 0.024 * RMB,
 	"step-1-32k":  0.024 * RMB,
 	"step-1-200k": 0.15 * RMB,
+	// aws llama3 https://aws.amazon.com/cn/bedrock/pricing/
+	"llama3-8b-8192(33)":  0.0003 / 0.002,  // $0.0003 / 1K tokens
+	"llama3-70b-8192(33)": 0.00265 / 0.002, // $0.00265 / 1K tokens
 	// https://cohere.com/pricing
 	"command":               0.5,
 	"command-nightly":       0.5,
@@ -184,7 +189,11 @@ var ModelRatio = map[string]float64{
 	"deepl-ja": 25.0 / 1000 * USD,
 }
 
-var CompletionRatio = map[string]float64{}
+var CompletionRatio = map[string]float64{
+	// aws llama3
+	"llama3-8b-8192(33)":  0.0006 / 0.0003,
+	"llama3-70b-8192(33)": 0.0035 / 0.00265,
+}
 
 var DefaultModelRatio map[string]float64
 var DefaultCompletionRatio map[string]float64
@@ -233,22 +242,28 @@ func UpdateModelRatioByJSONString(jsonStr string) error {
 	return json.Unmarshal([]byte(jsonStr), &ModelRatio)
 }
 
-func GetModelRatio(name string) float64 {
+func GetModelRatio(name string, channelType int) float64 {
 	if strings.HasPrefix(name, "qwen-") && strings.HasSuffix(name, "-internet") {
 		name = strings.TrimSuffix(name, "-internet")
 	}
 	if strings.HasPrefix(name, "command-") && strings.HasSuffix(name, "-internet") {
 		name = strings.TrimSuffix(name, "-internet")
 	}
-	ratio, ok := ModelRatio[name]
+	model := fmt.Sprintf("%s(%d)", name, channelType)
-	if !ok {
+	if ratio, ok := ModelRatio[model]; ok {
-		ratio, ok = DefaultModelRatio[name]
+		return ratio
 	}
-	if !ok {
+	if ratio, ok := DefaultModelRatio[model]; ok {
-		logger.SysError("model ratio not found: " + name)
+		return ratio
-		return 30
 	}
-	return ratio
+	if ratio, ok := ModelRatio[name]; ok {
+		return ratio
+	}
+	if ratio, ok := DefaultModelRatio[name]; ok {
+		return ratio
+	}
+	logger.SysError("model ratio not found: " + name)
+	return 30
 }
 
 func CompletionRatio2JSONString() string {
@@ -264,7 +279,17 @@ func UpdateCompletionRatioByJSONString(jsonStr string) error {
 	return json.Unmarshal([]byte(jsonStr), &CompletionRatio)
 }
 
-func GetCompletionRatio(name string) float64 {
+func GetCompletionRatio(name string, channelType int) float64 {
+	if strings.HasPrefix(name, "qwen-") && strings.HasSuffix(name, "-internet") {
+		name = strings.TrimSuffix(name, "-internet")
+	}
+	model := fmt.Sprintf("%s(%d)", name, channelType)
+	if ratio, ok := CompletionRatio[model]; ok {
+		return ratio
+	}
+	if ratio, ok := DefaultCompletionRatio[model]; ok {
+		return ratio
+	}
 	if ratio, ok := CompletionRatio[name]; ok {
 		return ratio
 	}

relay/channeltype/define.go

@@ -42,5 +42,7 @@ const (
 	DeepL
 	TogetherAI
 	Doubao
+	Novita
+	VertextAI
 	Dummy
 )

relay/channeltype/helper.go

@@ -35,6 +35,8 @@ func ToAPIType(channelType int) int {
 		apiType = apitype.Cloudflare
 	case DeepL:
 		apiType = apitype.DeepL
+	case VertextAI:
+		apiType = apitype.VertexAI
 	}
 
 	return apiType

relay/channeltype/url.go

@@ -42,6 +42,8 @@ var ChannelBaseURLs = []string{
 	"https://api-free.deepl.com",                // 38
 	"https://api.together.xyz",                  // 39
 	"https://ark.cn-beijing.volces.com",         // 40
+	"https://api.novita.ai/v3/openai",           // 41
+	"",                                          // 42
 }
 
 func init() {

relay/controller/audio.go

@@ -7,6 +7,10 @@ import (
 	"encoding/json"
 	"errors"
 	"fmt"
+	"io"
+	"net/http"
+	"strings"
+
 	"github.com/gin-gonic/gin"
 	"github.com/songquanpeng/one-api/common"
 	"github.com/songquanpeng/one-api/common/client"
@@ -21,9 +25,6 @@ import (
 	"github.com/songquanpeng/one-api/relay/meta"
 	relaymodel "github.com/songquanpeng/one-api/relay/model"
 	"github.com/songquanpeng/one-api/relay/relaymode"
-	"io"
-	"net/http"
-	"strings"
 )
 
 func RelayAudioHelper(c *gin.Context, relayMode int) *relaymodel.ErrorWithStatusCode {
@@ -53,7 +54,7 @@ func RelayAudioHelper(c *gin.Context, relayMode int) *relaymodel.ErrorWithStatus
 		}
 	}
 
-	modelRatio := billingratio.GetModelRatio(audioModel)
+	modelRatio := billingratio.GetModelRatio(audioModel, channelType)
 	groupRatio := billingratio.GetGroupRatio(group)
 	ratio := modelRatio * groupRatio
 	var quota int64

relay/controller/helper.go

@@ -4,6 +4,10 @@ import (
 	"context"
 	"errors"
 	"fmt"
+	"math"
+	"net/http"
+	"strings"
+
 	"github.com/gin-gonic/gin"
 	"github.com/songquanpeng/one-api/common"
 	"github.com/songquanpeng/one-api/common/config"
@@ -16,9 +20,6 @@ import (
 	"github.com/songquanpeng/one-api/relay/meta"
 	relaymodel "github.com/songquanpeng/one-api/relay/model"
 	"github.com/songquanpeng/one-api/relay/relaymode"
-	"math"
-	"net/http"
-	"strings"
 )
 
 func getAndValidateTextRequest(c *gin.Context, relayMode int) (*relaymodel.GeneralOpenAIRequest, error) {
@@ -40,78 +41,6 @@ func getAndValidateTextRequest(c *gin.Context, relayMode int) (*relaymodel.Gener
 	return textRequest, nil
 }
 
-func getImageRequest(c *gin.Context, relayMode int) (*relaymodel.ImageRequest, error) {
-	imageRequest := &relaymodel.ImageRequest{}
-	err := common.UnmarshalBodyReusable(c, imageRequest)
-	if err != nil {
-		return nil, err
-	}
-	if imageRequest.N == 0 {
-		imageRequest.N = 1
-	}
-	if imageRequest.Size == "" {
-		imageRequest.Size = "1024x1024"
-	}
-	if imageRequest.Model == "" {
-		imageRequest.Model = "dall-e-2"
-	}
-	return imageRequest, nil
-}
-
-func isValidImageSize(model string, size string) bool {
-	if model == "cogview-3" {
-		return true
-	}
-	_, ok := billingratio.ImageSizeRatios[model][size]
-	return ok
-}
-
-func getImageSizeRatio(model string, size string) float64 {
-	ratio, ok := billingratio.ImageSizeRatios[model][size]
-	if !ok {
-		return 1
-	}
-	return ratio
-}
-
-func validateImageRequest(imageRequest *relaymodel.ImageRequest, meta *meta.Meta) *relaymodel.ErrorWithStatusCode {
-	// model validation
-	hasValidSize := isValidImageSize(imageRequest.Model, imageRequest.Size)
-	if !hasValidSize {
-		return openai.ErrorWrapper(errors.New("size not supported for this image model"), "size_not_supported", http.StatusBadRequest)
-	}
-	// check prompt length
-	if imageRequest.Prompt == "" {
-		return openai.ErrorWrapper(errors.New("prompt is required"), "prompt_missing", http.StatusBadRequest)
-	}
-	if len(imageRequest.Prompt) > billingratio.ImagePromptLengthLimitations[imageRequest.Model] {
-		return openai.ErrorWrapper(errors.New("prompt is too long"), "prompt_too_long", http.StatusBadRequest)
-	}
-	// Number of generated images validation
-	if !isWithinRange(imageRequest.Model, imageRequest.N) {
-		// channel not azure
-		if meta.ChannelType != channeltype.Azure {
-			return openai.ErrorWrapper(errors.New("invalid value of n"), "n_not_within_range", http.StatusBadRequest)
-		}
-	}
-	return nil
-}
-
-func getImageCostRatio(imageRequest *relaymodel.ImageRequest) (float64, error) {
-	if imageRequest == nil {
-		return 0, errors.New("imageRequest is nil")
-	}
-	imageCostRatio := getImageSizeRatio(imageRequest.Model, imageRequest.Size)
-	if imageRequest.Quality == "hd" && imageRequest.Model == "dall-e-3" {
-		if imageRequest.Size == "1024x1024" {
-			imageCostRatio *= 2
-		} else {
-			imageCostRatio *= 1.5
-		}
-	}
-	return imageCostRatio, nil
-}
-
 func getPromptTokens(textRequest *relaymodel.GeneralOpenAIRequest, relayMode int) int {
 	switch relayMode {
 	case relaymode.ChatCompletions:
@@ -167,7 +96,7 @@ func postConsumeQuota(ctx context.Context, usage *relaymodel.Usage, meta *meta.M
 		return
 	}
 	var quota int64
-	completionRatio := billingratio.GetCompletionRatio(textRequest.Model)
+	completionRatio := billingratio.GetCompletionRatio(textRequest.Model, meta.ChannelType)
 	promptTokens := usage.PromptTokens
 	completionTokens := usage.CompletionTokens
 	quota = int64(math.Ceil((float64(promptTokens) + float64(completionTokens)*completionRatio) * ratio))

relay/controller/image.go

@@ -6,7 +6,11 @@ import (
 	"encoding/json"
 	"errors"
 	"fmt"
+	"io"
+	"net/http"
+
 	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common"
 	"github.com/songquanpeng/one-api/common/ctxkey"
 	"github.com/songquanpeng/one-api/common/logger"
 	"github.com/songquanpeng/one-api/model"
@@ -16,17 +20,86 @@ import (
 	"github.com/songquanpeng/one-api/relay/channeltype"
 	"github.com/songquanpeng/one-api/relay/meta"
 	relaymodel "github.com/songquanpeng/one-api/relay/model"
-	"io"
-	"net/http"
 )
 
-func isWithinRange(element string, value int) bool {
+func getImageRequest(c *gin.Context, relayMode int) (*relaymodel.ImageRequest, error) {
-	if _, ok := billingratio.ImageGenerationAmounts[element]; !ok {
+	imageRequest := &relaymodel.ImageRequest{}
-		return false
+	err := common.UnmarshalBodyReusable(c, imageRequest)
+	if err != nil {
+		return nil, err
 	}
-	min := billingratio.ImageGenerationAmounts[element][0]
+	if imageRequest.N == 0 {
-	max := billingratio.ImageGenerationAmounts[element][1]
+		imageRequest.N = 1
-	return value >= min && value <= max
+	}
+	if imageRequest.Size == "" {
+		imageRequest.Size = "1024x1024"
+	}
+	if imageRequest.Model == "" {
+		imageRequest.Model = "dall-e-2"
+	}
+	return imageRequest, nil
+}
+
+func isValidImageSize(model string, size string) bool {
+	if model == "cogview-3" || billingratio.ImageSizeRatios[model] == nil {
+		return true
+	}
+	_, ok := billingratio.ImageSizeRatios[model][size]
+	return ok
+}
+
+func isValidImagePromptLength(model string, promptLength int) bool {
+	maxPromptLength, ok := billingratio.ImagePromptLengthLimitations[model]
+	return !ok || promptLength <= maxPromptLength
+}
+
+func isWithinRange(element string, value int) bool {
+	amounts, ok := billingratio.ImageGenerationAmounts[element]
+	return !ok || (value >= amounts[0] && value <= amounts[1])
+}
+
+func getImageSizeRatio(model string, size string) float64 {
+	if ratio, ok := billingratio.ImageSizeRatios[model][size]; ok {
+		return ratio
+	}
+	return 1
+}
+
+func validateImageRequest(imageRequest *relaymodel.ImageRequest, meta *meta.Meta) *relaymodel.ErrorWithStatusCode {
+	// check prompt length
+	if imageRequest.Prompt == "" {
+		return openai.ErrorWrapper(errors.New("prompt is required"), "prompt_missing", http.StatusBadRequest)
+	}
+
+	// model validation
+	if !isValidImageSize(imageRequest.Model, imageRequest.Size) {
+		return openai.ErrorWrapper(errors.New("size not supported for this image model"), "size_not_supported", http.StatusBadRequest)
+	}
+
+	if !isValidImagePromptLength(imageRequest.Model, len(imageRequest.Prompt)) {
+		return openai.ErrorWrapper(errors.New("prompt is too long"), "prompt_too_long", http.StatusBadRequest)
+	}
+
+	// Number of generated images validation
+	if !isWithinRange(imageRequest.Model, imageRequest.N) {
+		return openai.ErrorWrapper(errors.New("invalid value of n"), "n_not_within_range", http.StatusBadRequest)
+	}
+	return nil
+}
+
+func getImageCostRatio(imageRequest *relaymodel.ImageRequest) (float64, error) {
+	if imageRequest == nil {
+		return 0, errors.New("imageRequest is nil")
+	}
+	imageCostRatio := getImageSizeRatio(imageRequest.Model, imageRequest.Size)
+	if imageRequest.Quality == "hd" && imageRequest.Model == "dall-e-3" {
+		if imageRequest.Size == "1024x1024" {
+			imageCostRatio *= 2
+		} else {
+			imageCostRatio *= 1.5
+		}
+	}
+	return imageCostRatio, nil
 }
 
 func RelayImageHelper(c *gin.Context, relayMode int) *relaymodel.ErrorWithStatusCode {
@@ -94,7 +167,7 @@ func RelayImageHelper(c *gin.Context, relayMode int) *relaymodel.ErrorWithStatus
 		requestBody = bytes.NewBuffer(jsonStr)
 	}
 
-	modelRatio := billingratio.GetModelRatio(imageModel)
+	modelRatio := billingratio.GetModelRatio(imageModel, meta.ChannelType)
 	groupRatio := billingratio.GetGroupRatio(meta.Group)
 	ratio := modelRatio * groupRatio
 	userQuota, err := model.CacheGetUserQuota(ctx, meta.UserId)

relay/controller/text.go

@@ -4,6 +4,9 @@ import (
 	"bytes"
 	"encoding/json"
 	"fmt"
+	"io"
+	"net/http"
+
 	"github.com/gin-gonic/gin"
 	"github.com/songquanpeng/one-api/common/logger"
 	"github.com/songquanpeng/one-api/relay"
@@ -14,8 +17,6 @@ import (
 	"github.com/songquanpeng/one-api/relay/channeltype"
 	"github.com/songquanpeng/one-api/relay/meta"
 	"github.com/songquanpeng/one-api/relay/model"
-	"io"
-	"net/http"
 )
 
 func RelayTextHelper(c *gin.Context) *model.ErrorWithStatusCode {
@@ -35,7 +36,7 @@ func RelayTextHelper(c *gin.Context) *model.ErrorWithStatusCode {
 	textRequest.Model, isModelMapped = getMappedModelName(textRequest.Model, meta.ModelMapping)
 	meta.ActualModelName = textRequest.Model
 	// get model ratio & group ratio
-	modelRatio := billingratio.GetModelRatio(textRequest.Model)
+	modelRatio := billingratio.GetModelRatio(textRequest.Model, meta.ChannelType)
 	groupRatio := billingratio.GetGroupRatio(meta.Group)
 	ratio := modelRatio * groupRatio
 	// pre-consume quota

relay/meta/relay_meta.go

@@ -10,20 +10,22 @@ import (
 )
 
 type Meta struct {
-	Mode            int
+	Mode         int
-	ChannelType     int
+	ChannelType  int
-	ChannelId       int
+	ChannelId    int
-	TokenId         int
+	TokenId      int
-	TokenName       string
+	TokenName    string
-	UserId          int
+	UserId       int
-	Group           string
+	Group        string
-	ModelMapping    map[string]string
+	ModelMapping map[string]string
-	BaseURL         string
+	BaseURL      string
-	APIKey          string
+	APIKey       string
-	APIType         int
+	APIType      int
-	Config          model.ChannelConfig
+	Config       model.ChannelConfig
-	IsStream        bool
+	IsStream     bool
+	// OriginModelName is the model name from the raw user request
 	OriginModelName string
+	// ActualModelName is the model name after mapping
 	ActualModelName string
 	RequestURLPath  string
 	PromptTokens    int // only for DoResponse

relay/model/general.go

@@ -13,6 +13,7 @@ type GeneralOpenAIRequest struct {
 	PresencePenalty  float64         `json:"presence_penalty,omitempty"`
 	ResponseFormat   *ResponseFormat `json:"response_format,omitempty"`
 	Seed             float64         `json:"seed,omitempty"`
+	Stop             any             `json:"stop,omitempty"`
 	Stream           bool            `json:"stream,omitempty"`
 	Temperature      float64         `json:"temperature,omitempty"`
 	TopP             float64         `json:"top_p,omitempty"`

relay/model/message.go

@@ -1,10 +1,11 @@
 package model
 
 type Message struct {
-	Role      string  `json:"role,omitempty"`
+	Role       string  `json:"role,omitempty"`
-	Content   any     `json:"content,omitempty"`
+	Content    any     `json:"content,omitempty"`
-	Name      *string `json:"name,omitempty"`
+	Name       *string `json:"name,omitempty"`
-	ToolCalls []Tool  `json:"tool_calls,omitempty"`
+	ToolCalls  []Tool  `json:"tool_calls,omitempty"`
+	ToolCallId string  `json:"tool_call_id,omitempty"`
 }
 
 func (m Message) IsStringContent() bool {

relay/model/tool.go

@@ -2,13 +2,13 @@ package model
 
 type Tool struct {
 	Id       string   `json:"id,omitempty"`
-	Type     string   `json:"type"`
+	Type     string   `json:"type,omitempty"` // when splicing claude tools stream messages, it is empty
 	Function Function `json:"function"`
 }
 
 type Function struct {
 	Description string `json:"description,omitempty"`
-	Name        string `json:"name"`
+	Name        string `json:"name,omitempty"`       // when splicing claude tools stream messages, it is empty
 	Parameters  any    `json:"parameters,omitempty"` // request
 	Arguments   any    `json:"arguments,omitempty"`  // response
 }

web/air/src/components/PersonalSetting.js

@@ -47,7 +47,7 @@ const PersonalSetting = () => {
   const [countdown, setCountdown] = useState(30);
   const [affLink, setAffLink] = useState('');
   const [systemToken, setSystemToken] = useState('');
-  // const [models, setModels] = useState([]);
+  const [models, setModels] = useState([]);
   const [openTransfer, setOpenTransfer] = useState(false);
   const [transferAmount, setTransferAmount] = useState(0);
 
@@ -72,7 +72,7 @@ const PersonalSetting = () => {
         console.log(userState);
       }
     );
-    // loadModels().then();
+    loadModels().then();
     getAffLink().then();
     setTransferAmount(getQuotaPerUnit());
   }, []);
@@ -127,16 +127,16 @@ const PersonalSetting = () => {
     }
   };
 
-  // const loadModels = async () => {
+  const loadModels = async () => {
-  //   let res = await API.get(`/api/user/models`);
+    let res = await API.get(`/api/user/available_models`);
-  //   const { success, message, data } = res.data;
+    const { success, message, data } = res.data;
-  //   if (success) {
+    if (success) {
-  //     setModels(data);
+      setModels(data);
-  //     console.log(data);
+      console.log(data);
-  //   } else {
+    } else {
-  //     showError(message);
+      showError(message);
-  //   }
+    }
-  // };
+  };
 
   const handleAffLinkClick = async (e) => {
     e.target.select();
@@ -344,7 +344,7 @@ const PersonalSetting = () => {
               }
             >
               <Typography.Title heading={6}>调用信息</Typography.Title>
-              {/* <Typography.Title heading={6}>可用模型</Typography.Title>
+              <p>可用模型（可点击复制）</p>
               <div style={{ marginTop: 10 }}>
                 <Space wrap>
                   {models.map((model) => (
@@ -355,7 +355,7 @@ const PersonalSetting = () => {
                     </Tag>
                   ))}
                 </Space>
-              </div> */}
+              </div>
             </Card>
             {/* <Card
               footer={

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'];
+                    localModels = ['SparkDesk', 'SparkDesk-v1.1', 'SparkDesk-v2.1', 'SparkDesk-v3.1', 'SparkDesk-v3.5', 'SparkDesk-v4.0'];
                     break;
                 case 19:
                     localModels = ['360GPT_S2_V9', 'embedding-bert-512-v1', 'embedding_s1_v1', 'semantic_similarity_s1_v1'];

web/berry/src/constants/ChannelConstants.js

@@ -13,7 +13,7 @@ export const CHANNEL_OPTIONS = {
   },
   33: {
     key: 33,
-    text: 'AWS Claude',
+    text: 'AWS',
     value: 33,
     color: 'primary'
   },
@@ -161,6 +161,18 @@ export const CHANNEL_OPTIONS = {
     value: 39,
     color: 'primary'
   },
+  42: {
+    key: 42,
+    text: 'VertexAI',
+    value: 42,
+    color: 'primary'
+  },
+  41: {
+    key: 41,
+    text: 'Novita',
+    value: 41,
+    color: 'purple'
+  },
   8: {
     key: 8,
     text: '自定义渠道',

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

@@ -165,6 +165,8 @@ const EditModal = ({ open, channelId, onCancel, onOk }) => {
     if (values.key === '') {
       if (values.config.ak && values.config.sk && values.config.region) {
         values.key = `${values.config.ak}|${values.config.sk}|${values.config.region}`;
+      } else if (values.config.region && values.config.vertex_ai_project_id && values.config.vertex_ai_adc) {
+        values.key = `${values.config.region}|${values.config.vertex_ai_project_id}|${values.config.vertex_ai_adc}`;
       }
     }
 

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']
+      models: ['SparkDesk', 'SparkDesk-v1.1', 'SparkDesk-v2.1', 'SparkDesk-v3.1', 'SparkDesk-v3.5', 'SparkDesk-v4.0']
     },
     prompt: {
       key: '按照如下格式输入：APPID|APISecret|APIKey',
@@ -203,7 +203,26 @@ const typeConfig = {
       }
     },
     modelGroup: 'Coze'
-  }
+  },
+  42: {
+    inputLabel: {
+      key: '',
+      config: {
+        region: 'Vertex AI Region',
+        vertex_ai_project_id: 'Vertex AI Project ID',
+        vertex_ai_adc: 'Google Cloud Application Default Credentials JSON'
+      }
+    },
+    prompt: {
+      key: '',
+      config: {
+        region: 'Vertex AI Region.g. us-east5',
+        vertex_ai_project_id: 'Vertex AI Project ID',
+        vertex_ai_adc: 'Google Cloud Application Default Credentials JSON: https://cloud.google.com/docs/authentication/application-default-credentials'
+      }
+    },
+    modelGroup: 'anthropic'
+  },
 };
 
 export { defaultConfig, typeConfig };

web/default/src/components/ChannelsTable.js

@@ -1,5 +1,5 @@
 import React, { useEffect, useState } from 'react';
-import { Button, Form, Input, Label, Message, Pagination, Popup, Table } from 'semantic-ui-react';
+import { Button, Dropdown, Form, Input, Label, Message, Pagination, Popup, Table } from 'semantic-ui-react';
 import { Link } from 'react-router-dom';
 import {
   API,
@@ -70,13 +70,33 @@ const ChannelsTable = () => {
     const res = await API.get(`/api/channel/?p=${startIdx}`);
     const { success, message, data } = res.data;
     if (success) {
-      if (startIdx === 0) {
+        let localChannels = data.map((channel) => {
-        setChannels(data);
+            if (channel.models === '') {
-      } else {
+                channel.models = [];
-        let newChannels = [...channels];
+                channel.test_model = "";
-        newChannels.splice(startIdx * ITEMS_PER_PAGE, data.length, ...data);
+            } else {
-        setChannels(newChannels);
+                channel.models = channel.models.split(',');
-      }
+                if (channel.models.length > 0) {
+                    channel.test_model = channel.models[0];
+                }
+                channel.model_options = channel.models.map((model) => {
+                    return {
+                        key: model,
+                        text: model,
+                        value: model,
+                    }
+                })
+                console.log('channel', channel)
+            }
+            return channel;
+        });
+        if (startIdx === 0) {
+            setChannels(localChannels);
+        } else {
+            let newChannels = [...channels];
+            newChannels.splice(startIdx * ITEMS_PER_PAGE, data.length, ...localChannels);
+            setChannels(newChannels);
+        }
     } else {
       showError(message);
     }
@@ -225,19 +245,31 @@ const ChannelsTable = () => {
     setSearching(false);
   };
 
-  const testChannel = async (id, name, idx) => {
+  const switchTestModel = async (idx, model) => {
-    const res = await API.get(`/api/channel/test/${id}/`);
+    let newChannels = [...channels];
-    const { success, message, time } = res.data;
+    let realIdx = (activePage - 1) * ITEMS_PER_PAGE + idx;
+    newChannels[realIdx].test_model = model;
+    setChannels(newChannels);
+  };
+
+  const testChannel = async (id, name, idx, m) => {
+    const res = await API.get(`/api/channel/test/${id}?model=${m}`);
+    const { success, message, time, model } = res.data;
     if (success) {
       let newChannels = [...channels];
       let realIdx = (activePage - 1) * ITEMS_PER_PAGE + idx;
       newChannels[realIdx].response_time = time * 1000;
       newChannels[realIdx].test_time = Date.now() / 1000;
       setChannels(newChannels);
-      showInfo(`渠道 ${name} 测试成功，耗时 ${time.toFixed(2)} 秒。`);
+      showInfo(`渠道 ${name} 测试成功，模型 ${model}，耗时 ${time.toFixed(2)} 秒。`);
     } else {
       showError(message);
     }
+    let newChannels = [...channels];
+    let realIdx = (activePage - 1) * ITEMS_PER_PAGE + idx;
+    newChannels[realIdx].response_time = time * 1000;
+    newChannels[realIdx].test_time = Date.now() / 1000;
+    setChannels(newChannels);
   };
 
   const testChannels = async (scope) => {
@@ -405,6 +437,7 @@ const ChannelsTable = () => {
             >
               优先级
             </Table.HeaderCell>
+            <Table.HeaderCell>测试模型</Table.HeaderCell>
             <Table.HeaderCell>操作</Table.HeaderCell>
           </Table.Row>
         </Table.Header>
@@ -459,13 +492,24 @@ const ChannelsTable = () => {
                       basic
                     />
                   </Table.Cell>
+                  <Table.Cell>
+                    <Dropdown
+                      placeholder='请选择测试模型'
+                      selection
+                      options={channel.model_options}
+                      defaultValue={channel.test_model}
+                      onChange={(event, data) => {
+                        switchTestModel(idx, data.value);
+                      }}
+                    />
+                  </Table.Cell>
                   <Table.Cell>
                     <div>
                       <Button
                         size={'small'}
                         positive
                         onClick={() => {
-                          testChannel(channel.id, channel.name, idx);
+                          testChannel(channel.id, channel.name, idx, channel.test_model);
                         }}
                       >
                         测试

web/default/src/constants/channel.constants.js

@@ -1,11 +1,12 @@
 export const CHANNEL_OPTIONS = [
     {key: 1, text: 'OpenAI', value: 1, color: 'green'},
     {key: 14, text: 'Anthropic Claude', value: 14, color: 'black'},
-    {key: 33, text: 'AWS Claude', value: 33, color: 'black'},
+    {key: 33, text: 'AWS', value: 33, color: 'black'},
     {key: 3, text: 'Azure OpenAI', value: 3, color: 'olive'},
     {key: 11, text: 'Google PaLM2', value: 11, color: 'orange'},
     {key: 24, text: 'Google Gemini', value: 24, color: 'orange'},
     {key: 28, text: 'Mistral AI', value: 28, color: 'orange'},
+    {key: 41, text: 'Novita', value: 41, color: 'purple'},
     {key: 40, text: '字节跳动豆包', value: 40, color: 'blue'},
     {key: 15, text: '百度文心千帆', value: 15, color: 'blue'},
     {key: 17, text: '阿里通义千问', value: 17, color: 'orange'},
@@ -26,6 +27,7 @@ export const CHANNEL_OPTIONS = [
     {key: 37, text: 'Cloudflare', value: 37, color: 'orange'},
     {key: 38, text: 'DeepL', value: 38, color: 'black'},
     {key: 39, text: 'together.ai', value: 39, color: 'blue'},
+    {key: 42, text: 'VertexAI', value: 42, color: 'blue'},
     {key: 8, text: '自定义渠道', value: 8, color: 'pink'},
     {key: 22, text: '知识库：FastGPT', value: 22, color: 'blue'},
     {key: 21, text: '知识库：AI Proxy', value: 21, color: 'purple'},

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

@@ -58,7 +58,9 @@ const EditChannel = () => {
     region: '',
     sk: '',
     ak: '',
-    user_id: ''
+    user_id: '',
+    vertex_ai_project_id: '',
+    vertex_ai_adc: ''
   });
   const handleInputChange = (e, { name, value }) => {
     setInputs((inputs) => ({ ...inputs, [name]: value }));
@@ -160,6 +162,8 @@ const EditChannel = () => {
     if (inputs.key === '') {
       if (config.ak !== '' && config.sk !== '' && config.region !== '') {
         inputs.key = `${config.ak}|${config.sk}|${config.region}`;
+      } else if (config.region !== '' && config.vertex_ai_project_id !== '' && config.vertex_ai_adc !== '') {
+        inputs.key = `${config.region}|${config.vertex_ai_project_id}|${config.vertex_ai_adc}`;
       }
     }
     if (!isEdit && (inputs.name === '' || inputs.key === '')) {
@@ -456,6 +460,39 @@ const EditChannel = () => {
               </Form.Field>
             )
           }
+          {
+            inputs.type === 42 && (
+              <Form.Field>
+                <Form.Input
+                  label='Region'
+                  name='region'
+                  required
+                  placeholder={'Vertex AI Region.g. us-east5'}
+                  onChange={handleConfigChange}
+                  value={config.region}
+                  autoComplete=''
+                />
+                <Form.Input
+                  label='Vertex AI Project ID'
+                  name='vertex_ai_project_id'
+                  required
+                  placeholder={'Vertex AI Project ID'}
+                  onChange={handleConfigChange}
+                  value={config.vertex_ai_project_id}
+                  autoComplete=''
+                />
+                <Form.Input
+                  label='Google Cloud Application Default Credentials JSON'
+                  name='vertex_ai_adc'
+                  required
+                  placeholder={'Google Cloud Application Default Credentials JSON'}
+                  onChange={handleConfigChange}
+                  value={config.vertex_ai_adc}
+                  autoComplete=''
+                />
+              </Form.Field>
+            )
+          }
           {
             inputs.type === 34 && (
               <Form.Input
@@ -469,7 +506,7 @@ const EditChannel = () => {
               />)
           }
           {
-            inputs.type !== 33 && (batch ? <Form.Field>
+            inputs.type !== 33 && inputs.type !== 42 && (batch ? <Form.Field>
               <Form.TextArea
                 label='密钥'
                 name='key'