chore: update ratio for baidu embedding

添加百度的另外3个向量模型【"bge-large-zh",
	"bge-large-en",
	"tao-8k",
】

.github/workflows/docker-image-amd64-en.yml

@@ -20,6 +20,13 @@ jobs:
       - name: Check out the repo
         uses: actions/checkout@v3
 
+      - name: Check repository URL
+        run: |
+          REPO_URL=$(git config --get remote.origin.url)
+          if [[ $REPO_URL == *"pro" ]]; then
+            exit 1
+          fi      
+
       - name: Save version info
         run: |
           git describe --tags > VERSION 

.github/workflows/docker-image-amd64.yml

@@ -20,6 +20,13 @@ jobs:
       - name: Check out the repo
         uses: actions/checkout@v3
 
+      - name: Check repository URL
+        run: |
+          REPO_URL=$(git config --get remote.origin.url)
+          if [[ $REPO_URL == *"pro" ]]; then
+            exit 1
+          fi        
+
       - name: Save version info
         run: |
           git describe --tags > VERSION 

.github/workflows/docker-image-arm64.yml

@@ -21,6 +21,13 @@ jobs:
       - name: Check out the repo
         uses: actions/checkout@v3
 
+      - name: Check repository URL
+        run: |
+          REPO_URL=$(git config --get remote.origin.url)
+          if [[ $REPO_URL == *"pro" ]]; then
+            exit 1
+          fi
+
       - name: Save version info
         run: |
           git describe --tags > VERSION 

.github/workflows/linux-release.yml

@@ -20,10 +20,16 @@ jobs:
         uses: actions/checkout@v3
         with:
           fetch-depth: 0
+      - name: Check repository URL
+        run: |
+          REPO_URL=$(git config --get remote.origin.url)
+          if [[ $REPO_URL == *"pro" ]]; then
+            exit 1
+          fi
       - uses: actions/setup-node@v3
         with:
           node-version: 16
-      - name: Build Frontend (theme default)
+      - name: Build Frontend
         env:
           CI: ""
         run: |
@@ -38,7 +44,7 @@ jobs:
       - name: Build Backend (amd64)
         run: |
           go mod download
-          go build -ldflags "-s -w -X 'one-api/common.Version=$(git describe --tags)' -extldflags '-static'" -o one-api
+          go build -ldflags "-s -w -X 'github.com/songquanpeng/one-api/common.Version=$(git describe --tags)' -extldflags '-static'" -o one-api
 
       - name: Build Backend (arm64)
         run: |

.github/workflows/macos-release.yml

@@ -20,10 +20,16 @@ jobs:
         uses: actions/checkout@v3
         with:
           fetch-depth: 0
+      - name: Check repository URL
+        run: |
+          REPO_URL=$(git config --get remote.origin.url)
+          if [[ $REPO_URL == *"pro" ]]; then
+            exit 1
+          fi
       - uses: actions/setup-node@v3
         with:
           node-version: 16
-      - name: Build Frontend (theme default)
+      - name: Build Frontend
         env:
           CI: ""
         run: |
@@ -38,7 +44,7 @@ jobs:
       - name: Build Backend
         run: |
           go mod download
-          go build -ldflags "-X 'one-api/common.Version=$(git describe --tags)'" -o one-api-macos
+          go build -ldflags "-X 'github.com/songquanpeng/one-api/common.Version=$(git describe --tags)'" -o one-api-macos
       - name: Release
         uses: softprops/action-gh-release@v1
         if: startsWith(github.ref, 'refs/tags/')

.github/workflows/windows-release.yml

@@ -23,10 +23,16 @@ jobs:
         uses: actions/checkout@v3
         with:
           fetch-depth: 0
+      - name: Check repository URL
+        run: |
+          REPO_URL=$(git config --get remote.origin.url)
+          if [[ $REPO_URL == *"pro" ]]; then
+            exit 1
+          fi
       - uses: actions/setup-node@v3
         with:
           node-version: 16
-      - name: Build Frontend (theme default)
+      - name: Build Frontend
         env:
           CI: ""
         run: |
@@ -41,7 +47,7 @@ jobs:
       - name: Build Backend
         run: |
           go mod download
-          go build -ldflags "-s -w -X 'one-api/common.Version=$(git describe --tags)'" -o one-api.exe
+          go build -ldflags "-s -w -X 'github.com/songquanpeng/one-api/common.Version=$(git describe --tags)'" -o one-api.exe
       - name: Release
         uses: softprops/action-gh-release@v1
         if: startsWith(github.ref, 'refs/tags/')

.gitignore

@@ -6,4 +6,5 @@ upload
 build
 *.db-journal
 logs
-data
+data
+/web/node_modules

Dockerfile

@@ -23,7 +23,7 @@ ADD go.mod go.sum ./
 RUN go mod download
 COPY . .
 COPY --from=builder /web/build ./web/build
-RUN go build -ldflags "-s -w -X 'one-api/common.Version=$(cat VERSION)' -extldflags '-static'" -o one-api
+RUN go build -ldflags "-s -w -X 'github.com/songquanpeng/one-api/common.Version=$(cat VERSION)' -extldflags '-static'" -o one-api
 
 FROM alpine
 

README.md

@@ -67,12 +67,20 @@ _✨ 通过标准的 OpenAI API 格式访问所有的大模型，开箱即用
    + [x] [OpenAI ChatGPT 系列模型](https://platform.openai.com/docs/guides/gpt/chat-completions-api)（支持 [Azure OpenAI API](https://learn.microsoft.com/en-us/azure/ai-services/openai/reference)）
    + [x] [Anthropic Claude 系列模型](https://anthropic.com)
    + [x] [Google PaLM2/Gemini 系列模型](https://developers.generativeai.google)
+   + [x] [Mistral 系列模型](https://mistral.ai/)
    + [x] [百度文心一言系列模型](https://cloud.baidu.com/doc/WENXINWORKSHOP/index.html)
    + [x] [阿里通义千问系列模型](https://help.aliyun.com/document_detail/2400395.html)
    + [x] [讯飞星火认知大模型](https://www.xfyun.cn/doc/spark/Web.html)
    + [x] [智谱 ChatGLM 系列模型](https://bigmodel.cn)
    + [x] [360 智脑](https://ai.360.cn)
    + [x] [腾讯混元大模型](https://cloud.tencent.com/document/product/1729)
+   + [x] [Moonshot AI](https://platform.moonshot.cn/)
+   + [x] [百川大模型](https://platform.baichuan-ai.com)
+   + [ ] [字节云雀大模型](https://www.volcengine.com/product/ark) (WIP)
+   + [x] [MINIMAX](https://api.minimax.chat/)
+   + [x] [Groq](https://wow.groq.com/)
+   + [x] [Ollama](https://github.com/ollama/ollama)
+   + [x] [零一万物](https://platform.lingyiwanwu.com/)
 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。
@@ -100,6 +108,7 @@ _✨ 通过标准的 OpenAI API 格式访问所有的大模型，开箱即用
     + [GitHub 开放授权](https://github.com/settings/applications/new)。
     + 微信公众号授权（需要额外部署 [WeChat Server](https://github.com/songquanpeng/wechat-server)）。
 23. 支持主题切换，设置环境变量 `THEME` 即可，默认为 `default`，欢迎 PR 更多主题，具体参考[此处](./web/README.md)。
+24. 配合 [Message Pusher](https://github.com/songquanpeng/message-pusher) 可将报警信息推送到多种 App 上。
 
 ## 部署
 ### 基于 Docker 进行部署
@@ -369,6 +378,9 @@ graph LR
 16. `SQLITE_BUSY_TIMEOUT`：SQLite 锁等待超时设置，单位为毫秒，默认 `3000`。
 17. `GEMINI_SAFETY_SETTING`：Gemini 的安全设置，默认 `BLOCK_NONE`。
 18. `THEME`：系统的主题设置，默认为 `default`，具体可选值参考[此处](./web/README.md)。
+19. `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`。
+20. `METRIC_QUEUE_SIZE`：请求成功率统计队列大小，默认为 `10`。
+21. `METRIC_SUCCESS_RATE_THRESHOLD`：请求成功率阈值，默认为 `0.8`。
 
 ### 命令行参数
 1. `--port <port_number>`: 指定服务器监听的端口号，默认为 `3000`。

common/blacklist/main.go

@@ -0,0 +1,29 @@
+package blacklist
+
+import (
+	"fmt"
+	"sync"
+)
+
+var blackList sync.Map
+
+func init() {
+	blackList = sync.Map{}
+}
+
+func userId2Key(id int) string {
+	return fmt.Sprintf("userid_%d", id)
+}
+
+func BanUser(id int) {
+	blackList.Store(userId2Key(id), true)
+}
+
+func UnbanUser(id int) {
+	blackList.Delete(userId2Key(id))
+}
+
+func IsUserBanned(id int) bool {
+	_, ok := blackList.Load(userId2Key(id))
+	return ok
+}

common/config/config.go

@@ -1,7 +1,7 @@
 package config
 
 import (
-	"one-api/common/helper"
+	"github.com/songquanpeng/one-api/common/env"
 	"os"
 	"strconv"
 	"sync"
@@ -52,6 +52,7 @@ var EmailDomainWhitelist = []string{
 }
 
 var DebugEnabled = os.Getenv("DEBUG") == "true"
+var DebugSQLEnabled = os.Getenv("DEBUG_SQL") == "true"
 var MemoryCacheEnabled = os.Getenv("MEMORY_CACHE_ENABLED") == "true"
 
 var LogConsumeEnabled = true
@@ -69,17 +70,20 @@ var WeChatServerAddress = ""
 var WeChatServerToken = ""
 var WeChatAccountQRCodeImageURL = ""
 
+var MessagePusherAddress = ""
+var MessagePusherToken = ""
+
 var TurnstileSiteKey = ""
 var TurnstileSecretKey = ""
 
-var QuotaForNewUser = 0
-var QuotaForInviter = 0
-var QuotaForInvitee = 0
+var QuotaForNewUser int64 = 0
+var QuotaForInviter int64 = 0
+var QuotaForInvitee int64 = 0
 var ChannelDisableThreshold = 5.0
 var AutomaticDisableChannelEnabled = false
 var AutomaticEnableChannelEnabled = false
-var QuotaRemindThreshold = 1000
-var PreConsumedQuota = 500
+var QuotaRemindThreshold int64 = 1000
+var PreConsumedQuota int64 = 500
 var ApproximateTokenEnabled = false
 var RetryTimes = 0
 
@@ -90,16 +94,16 @@ var IsMasterNode = os.Getenv("NODE_TYPE") != "slave"
 var requestInterval, _ = strconv.Atoi(os.Getenv("POLLING_INTERVAL"))
 var RequestInterval = time.Duration(requestInterval) * time.Second
 
-var SyncFrequency = helper.GetOrDefaultEnvInt("SYNC_FREQUENCY", 10*60) // unit is second
+var SyncFrequency = env.Int("SYNC_FREQUENCY", 10*60) // unit is second
 
 var BatchUpdateEnabled = false
-var BatchUpdateInterval = helper.GetOrDefaultEnvInt("BATCH_UPDATE_INTERVAL", 5)
+var BatchUpdateInterval = env.Int("BATCH_UPDATE_INTERVAL", 5)
 
-var RelayTimeout = helper.GetOrDefaultEnvInt("RELAY_TIMEOUT", 0) // unit is second
+var RelayTimeout = env.Int("RELAY_TIMEOUT", 0) // unit is second
 
-var GeminiSafetySetting = helper.GetOrDefaultEnvString("GEMINI_SAFETY_SETTING", "BLOCK_NONE")
+var GeminiSafetySetting = env.String("GEMINI_SAFETY_SETTING", "BLOCK_NONE")
 
-var Theme = helper.GetOrDefaultEnvString("THEME", "default")
+var Theme = env.String("THEME", "default")
 var ValidThemes = map[string]bool{
 	"default": true,
 	"berry":   true,
@@ -108,10 +112,10 @@ var ValidThemes = map[string]bool{
 // All duration's unit is seconds
 // Shouldn't larger then RateLimitKeyExpirationDuration
 var (
-	GlobalApiRateLimitNum            = helper.GetOrDefaultEnvInt("GLOBAL_API_RATE_LIMIT", 180)
+	GlobalApiRateLimitNum            = env.Int("GLOBAL_API_RATE_LIMIT", 180)
 	GlobalApiRateLimitDuration int64 = 3 * 60
 
-	GlobalWebRateLimitNum            = helper.GetOrDefaultEnvInt("GLOBAL_WEB_RATE_LIMIT", 60)
+	GlobalWebRateLimitNum            = env.Int("GLOBAL_WEB_RATE_LIMIT", 60)
 	GlobalWebRateLimitDuration int64 = 3 * 60
 
 	UploadRateLimitNum            = 10
@@ -125,3 +129,9 @@ var (
 )
 
 var RateLimitKeyExpirationDuration = 20 * time.Minute
+
+var EnableMetric = env.Bool("ENABLE_METRIC", false)
+var MetricQueueSize = env.Int("METRIC_QUEUE_SIZE", 10)
+var MetricSuccessRateThreshold = env.Float64("METRIC_SUCCESS_RATE_THRESHOLD", 0.8)
+var MetricSuccessChanSize = env.Int("METRIC_SUCCESS_CHAN_SIZE", 1024)
+var MetricFailChanSize = env.Int("METRIC_FAIL_CHAN_SIZE", 128)

common/constants.go

@@ -15,6 +15,7 @@ const (
 const (
 	UserStatusEnabled  = 1 // don't use 0, 0 is the default value!
 	UserStatusDisabled = 2 // also don't use 0
+	UserStatusDeleted  = 3
 )
 
 const (
@@ -38,31 +39,40 @@ const (
 )
 
 const (
-	ChannelTypeUnknown        = 0
-	ChannelTypeOpenAI         = 1
-	ChannelTypeAPI2D          = 2
-	ChannelTypeAzure          = 3
-	ChannelTypeCloseAI        = 4
-	ChannelTypeOpenAISB       = 5
-	ChannelTypeOpenAIMax      = 6
-	ChannelTypeOhMyGPT        = 7
-	ChannelTypeCustom         = 8
-	ChannelTypeAILS           = 9
-	ChannelTypeAIProxy        = 10
-	ChannelTypePaLM           = 11
-	ChannelTypeAPI2GPT        = 12
-	ChannelTypeAIGC2D         = 13
-	ChannelTypeAnthropic      = 14
-	ChannelTypeBaidu          = 15
-	ChannelTypeZhipu          = 16
-	ChannelTypeAli            = 17
-	ChannelTypeXunfei         = 18
-	ChannelType360            = 19
-	ChannelTypeOpenRouter     = 20
-	ChannelTypeAIProxyLibrary = 21
-	ChannelTypeFastGPT        = 22
-	ChannelTypeTencent        = 23
-	ChannelTypeGemini         = 24
+	ChannelTypeUnknown = iota
+	ChannelTypeOpenAI
+	ChannelTypeAPI2D
+	ChannelTypeAzure
+	ChannelTypeCloseAI
+	ChannelTypeOpenAISB
+	ChannelTypeOpenAIMax
+	ChannelTypeOhMyGPT
+	ChannelTypeCustom
+	ChannelTypeAILS
+	ChannelTypeAIProxy
+	ChannelTypePaLM
+	ChannelTypeAPI2GPT
+	ChannelTypeAIGC2D
+	ChannelTypeAnthropic
+	ChannelTypeBaidu
+	ChannelTypeZhipu
+	ChannelTypeAli
+	ChannelTypeXunfei
+	ChannelType360
+	ChannelTypeOpenRouter
+	ChannelTypeAIProxyLibrary
+	ChannelTypeFastGPT
+	ChannelTypeTencent
+	ChannelTypeGemini
+	ChannelTypeMoonshot
+	ChannelTypeBaichuan
+	ChannelTypeMinimax
+	ChannelTypeMistral
+	ChannelTypeGroq
+	ChannelTypeOllama
+	ChannelTypeLingYiWanWu
+
+	ChannelTypeDummy
 )
 
 var ChannelBaseURLs = []string{
@@ -91,4 +101,19 @@ var ChannelBaseURLs = []string{
 	"https://fastgpt.run/api/openapi",           // 22
 	"https://hunyuan.cloud.tencent.com",         // 23
 	"https://generativelanguage.googleapis.com", // 24
+	"https://api.moonshot.cn",                   // 25
+	"https://api.baichuan-ai.com",               // 26
+	"https://api.minimax.chat",                  // 27
+	"https://api.mistral.ai",                    // 28
+	"https://api.groq.com/openai",               // 29
+	"http://localhost:11434",                    // 30
+	"https://api.lingyiwanwu.com",               // 31
 }
+
+const (
+	ConfigKeyPrefix = "cfg_"
+
+	ConfigKeyAPIVersion = ConfigKeyPrefix + "api_version"
+	ConfigKeyLibraryID  = ConfigKeyPrefix + "library_id"
+	ConfigKeyPlugin     = ConfigKeyPrefix + "plugin"
+)

common/database.go

@@ -1,9 +1,12 @@
 package common
 
-import "one-api/common/helper"
+import (
+	"github.com/songquanpeng/one-api/common/env"
+)
 
 var UsingSQLite = false
 var UsingPostgreSQL = false
+var UsingMySQL = false
 
 var SQLitePath = "one-api.db"
-var SQLiteBusyTimeout = helper.GetOrDefaultEnvInt("SQLITE_BUSY_TIMEOUT", 3000)
+var SQLiteBusyTimeout = env.Int("SQLITE_BUSY_TIMEOUT", 3000)

common/embed-file-system.go

@@ -15,10 +15,7 @@ type embedFileSystem struct {
 
 func (e embedFileSystem) Exists(prefix string, path string) bool {
 	_, err := e.Open(path)
-	if err != nil {
-		return false
-	}
-	return true
+	return err == nil
 }
 
 func EmbedFolder(fsEmbed embed.FS, targetPath string) static.ServeFileSystem {

common/env/helper.go

@@ -0,0 +1,42 @@
+package env
+
+import (
+	"os"
+	"strconv"
+)
+
+func Bool(env string, defaultValue bool) bool {
+	if env == "" || os.Getenv(env) == "" {
+		return defaultValue
+	}
+	return os.Getenv(env) == "true"
+}
+
+func Int(env string, defaultValue int) int {
+	if env == "" || os.Getenv(env) == "" {
+		return defaultValue
+	}
+	num, err := strconv.Atoi(os.Getenv(env))
+	if err != nil {
+		return defaultValue
+	}
+	return num
+}
+
+func Float64(env string, defaultValue float64) float64 {
+	if env == "" || os.Getenv(env) == "" {
+		return defaultValue
+	}
+	num, err := strconv.ParseFloat(os.Getenv(env), 64)
+	if err != nil {
+		return defaultValue
+	}
+	return num
+}
+
+func String(env string, defaultValue string) string {
+	if env == "" || os.Getenv(env) == "" {
+		return defaultValue
+	}
+	return os.Getenv(env)
+}

common/gin.go

@@ -8,12 +8,24 @@ import (
 	"strings"
 )
 
-func UnmarshalBodyReusable(c *gin.Context, v any) error {
+const KeyRequestBody = "key_request_body"
+
+func GetRequestBody(c *gin.Context) ([]byte, error) {
+	requestBody, _ := c.Get(KeyRequestBody)
+	if requestBody != nil {
+		return requestBody.([]byte), nil
+	}
 	requestBody, err := io.ReadAll(c.Request.Body)
 	if err != nil {
-		return err
+		return nil, err
 	}
-	err = c.Request.Body.Close()
+	_ = c.Request.Body.Close()
+	c.Set(KeyRequestBody, requestBody)
+	return requestBody.([]byte), nil
+}
+
+func UnmarshalBodyReusable(c *gin.Context, v any) error {
+	requestBody, err := GetRequestBody(c)
 	if err != nil {
 		return err
 	}

common/group-ratio.go

@@ -2,7 +2,7 @@ package common
 
 import (
 	"encoding/json"
-	"one-api/common/logger"
+	"github.com/songquanpeng/one-api/common/logger"
 )
 
 var GroupRatio = map[string]float64{

common/helper/helper.go

@@ -7,8 +7,6 @@ import (
 	"log"
 	"math/rand"
 	"net"
-	"one-api/common/logger"
-	"os"
 	"os/exec"
 	"runtime"
 	"strconv"
@@ -107,13 +105,13 @@ func Seconds2Time(num int) (time string) {
 }
 
 func Interface2String(inter interface{}) string {
-	switch inter.(type) {
+	switch inter := inter.(type) {
 	case string:
-		return inter.(string)
+		return inter
 	case int:
-		return fmt.Sprintf("%d", inter.(int))
+		return fmt.Sprintf("%d", inter)
 	case float64:
-		return fmt.Sprintf("%f", inter.(float64))
+		return fmt.Sprintf("%f", inter)
 	}
 	return "Not Implemented"
 }
@@ -137,6 +135,7 @@ func GetUUID() string {
 }
 
 const keyChars = "0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ"
+const keyNumbers = "0123456789"
 
 func init() {
 	rand.Seed(time.Now().UnixNano())
@@ -168,6 +167,15 @@ func GetRandomString(length int) string {
 	return string(key)
 }
 
+func GetRandomNumberString(length int) string {
+	rand.Seed(time.Now().UnixNano())
+	key := make([]byte, length)
+	for i := 0; i < length; i++ {
+		key[i] = keyNumbers[rand.Intn(len(keyNumbers))]
+	}
+	return string(key)
+}
+
 func GetTimestamp() int64 {
 	return time.Now().Unix()
 }
@@ -177,6 +185,10 @@ func GetTimeString() string {
 	return fmt.Sprintf("%s%d", now.Format("20060102150405"), now.UnixNano()%1e9)
 }
 
+func GenRequestID() string {
+	return GetTimeString() + GetRandomNumberString(8)
+}
+
 func Max(a int, b int) int {
 	if a >= b {
 		return a
@@ -185,25 +197,6 @@ func Max(a int, b int) int {
 	}
 }
 
-func GetOrDefaultEnvInt(env string, defaultValue int) int {
-	if env == "" || os.Getenv(env) == "" {
-		return defaultValue
-	}
-	num, err := strconv.Atoi(os.Getenv(env))
-	if err != nil {
-		logger.SysError(fmt.Sprintf("failed to parse %s: %s, using default value: %d", env, err.Error(), defaultValue))
-		return defaultValue
-	}
-	return num
-}
-
-func GetOrDefaultEnvString(env string, defaultValue string) string {
-	if env == "" || os.Getenv(env) == "" {
-		return defaultValue
-	}
-	return os.Getenv(env)
-}
-
 func AssignOrDefault(value string, defaultValue string) string {
 	if len(value) != 0 {
 		return value

common/image/image_test.go

@@ -12,7 +12,7 @@ import (
 	"strings"
 	"testing"
 
-	img "one-api/common/image"
+	img "github.com/songquanpeng/one-api/common/image"
 
 	"github.com/stretchr/testify/assert"
 	_ "golang.org/x/image/webp"

common/init.go

@@ -3,9 +3,9 @@ package common
 import (
 	"flag"
 	"fmt"
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/logger"
 	"log"
-	"one-api/common/config"
-	"one-api/common/logger"
 	"os"
 	"path/filepath"
 )

common/logger/logger.go

@@ -4,6 +4,8 @@ import (
 	"context"
 	"fmt"
 	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/helper"
 	"io"
 	"log"
 	"os"
@@ -13,14 +15,12 @@ import (
 )
 
 const (
+	loggerDEBUG = "DEBUG"
 	loggerINFO  = "INFO"
 	loggerWarn  = "WARN"
 	loggerError = "ERR"
 )
 
-const maxLogCount = 1000000
-
-var logCount int
 var setupLogLock sync.Mutex
 var setupLogWorking bool
 
@@ -55,6 +55,12 @@ func SysError(s string) {
 	_, _ = fmt.Fprintf(gin.DefaultErrorWriter, "[SYS] %v | %s \n", t.Format("2006/01/02 - 15:04:05"), s)
 }
 
+func Debug(ctx context.Context, msg string) {
+	if config.DebugEnabled {
+		logHelper(ctx, loggerDEBUG, msg)
+	}
+}
+
 func Info(ctx context.Context, msg string) {
 	logHelper(ctx, loggerINFO, msg)
 }
@@ -67,16 +73,20 @@ func Error(ctx context.Context, msg string) {
 	logHelper(ctx, loggerError, msg)
 }
 
+func Debugf(ctx context.Context, format string, a ...any) {
+	Debug(ctx, fmt.Sprintf(format, a...))
+}
+
 func Infof(ctx context.Context, format string, a ...any) {
-	Info(ctx, fmt.Sprintf(format, a))
+	Info(ctx, fmt.Sprintf(format, a...))
 }
 
 func Warnf(ctx context.Context, format string, a ...any) {
-	Warn(ctx, fmt.Sprintf(format, a))
+	Warn(ctx, fmt.Sprintf(format, a...))
 }
 
 func Errorf(ctx context.Context, format string, a ...any) {
-	Error(ctx, fmt.Sprintf(format, a))
+	Error(ctx, fmt.Sprintf(format, a...))
 }
 
 func logHelper(ctx context.Context, level string, msg string) {
@@ -85,11 +95,12 @@ func logHelper(ctx context.Context, level string, msg string) {
 		writer = gin.DefaultWriter
 	}
 	id := ctx.Value(RequestIdKey)
+	if id == nil {
+		id = helper.GenRequestID()
+	}
 	now := time.Now()
 	_, _ = fmt.Fprintf(writer, "[%s] %v | %s | %s \n", level, now.Format("2006/01/02 - 15:04:05"), id, msg)
-	logCount++ // we don't need accurate count, so no lock here
-	if logCount > maxLogCount && !setupLogWorking {
-		logCount = 0
+	if !setupLogWorking {
 		setupLogWorking = true
 		go func() {
 			SetupLogger()

common/message/email.go

@@ -1,17 +1,20 @@
-package common
+package message
 
 import (
 	"crypto/rand"
 	"crypto/tls"
 	"encoding/base64"
 	"fmt"
+	"github.com/songquanpeng/one-api/common/config"
 	"net/smtp"
-	"one-api/common/config"
 	"strings"
 	"time"
 )
 
 func SendEmail(subject string, receiver string, content string) error {
+	if receiver == "" {
+		return fmt.Errorf("receiver is empty")
+	}
 	if config.SMTPFrom == "" { // for compatibility
 		config.SMTPFrom = config.SMTPAccount
 	}

common/message/main.go

@@ -0,0 +1,22 @@
+package message
+
+import (
+	"fmt"
+	"github.com/songquanpeng/one-api/common/config"
+)
+
+const (
+	ByAll           = "all"
+	ByEmail         = "email"
+	ByMessagePusher = "message_pusher"
+)
+
+func Notify(by string, title string, description string, content string) error {
+	if by == ByEmail {
+		return SendEmail(title, config.RootUserEmail, content)
+	}
+	if by == ByMessagePusher {
+		return SendMessage(title, description, content)
+	}
+	return fmt.Errorf("unknown notify method: %s", by)
+}

common/message/message-pusher.go

@@ -0,0 +1,53 @@
+package message
+
+import (
+	"bytes"
+	"encoding/json"
+	"errors"
+	"github.com/songquanpeng/one-api/common/config"
+	"net/http"
+)
+
+type request struct {
+	Title       string `json:"title"`
+	Description string `json:"description"`
+	Content     string `json:"content"`
+	URL         string `json:"url"`
+	Channel     string `json:"channel"`
+	Token       string `json:"token"`
+}
+
+type response struct {
+	Success bool   `json:"success"`
+	Message string `json:"message"`
+}
+
+func SendMessage(title string, description string, content string) error {
+	if config.MessagePusherAddress == "" {
+		return errors.New("message pusher address is not set")
+	}
+	req := request{
+		Title:       title,
+		Description: description,
+		Content:     content,
+		Token:       config.MessagePusherToken,
+	}
+	data, err := json.Marshal(req)
+	if err != nil {
+		return err
+	}
+	resp, err := http.Post(config.MessagePusherAddress,
+		"application/json", bytes.NewBuffer(data))
+	if err != nil {
+		return err
+	}
+	var res response
+	err = json.NewDecoder(resp.Body).Decode(&res)
+	if err != nil {
+		return err
+	}
+	if !res.Success {
+		return errors.New(res.Message)
+	}
+	return nil
+}

common/model-ratio.go

@@ -2,97 +2,92 @@ package common
 
 import (
 	"encoding/json"
-	"one-api/common/logger"
+	"github.com/songquanpeng/one-api/common/logger"
 	"strings"
 	"time"
 )
 
-var DalleSizeRatios = map[string]map[string]float64{
-	"dall-e-2": {
-		"256x256":   1,
-		"512x512":   1.125,
-		"1024x1024": 1.25,
-	},
-	"dall-e-3": {
-		"1024x1024": 1,
-		"1024x1792": 2,
-		"1792x1024": 2,
-	},
-}
-
-var DalleGenerationImageAmounts = map[string][2]int{
-	"dall-e-2": {1, 10},
-	"dall-e-3": {1, 1}, // OpenAI allows n=1 currently.
-}
-
-var DalleImagePromptLengthLimitations = map[string]int{
-	"dall-e-2": 1000,
-	"dall-e-3": 4000,
-}
+const (
+	USD2RMB = 7
+	USD     = 500 // $0.002 = 1 -> $1 = 500
+	RMB     = USD / USD2RMB
+)
 
 // ModelRatio
 // https://platform.openai.com/docs/models/model-endpoint-compatibility
 // https://cloud.baidu.com/doc/WENXINWORKSHOP/s/Blfmc9dlf
 // https://openai.com/pricing
-// TODO: when a new api is enabled, check the pricing here
 // 1 === $0.002 / 1K tokens
 // 1 === ￥0.014 / 1k tokens
 var ModelRatio = map[string]float64{
-	"gpt-4":                     15,
-	"gpt-4-0314":                15,
-	"gpt-4-0613":                15,
-	"gpt-4-32k":                 30,
-	"gpt-4-32k-0314":            30,
-	"gpt-4-32k-0613":            30,
-	"gpt-4-1106-preview":        5,    // $0.01 / 1K tokens
-	"gpt-4-0125-preview":        5,    // $0.01 / 1K tokens
-	"gpt-4-turbo-preview":       5,    // $0.01 / 1K tokens
-	"gpt-4-vision-preview":      5,    // $0.01 / 1K tokens
-	"gpt-3.5-turbo":             0.75, // $0.0015 / 1K tokens
-	"gpt-3.5-turbo-0301":        0.75,
-	"gpt-3.5-turbo-0613":        0.75,
-	"gpt-3.5-turbo-16k":         1.5, // $0.003 / 1K tokens
-	"gpt-3.5-turbo-16k-0613":    1.5,
-	"gpt-3.5-turbo-instruct":    0.75, // $0.0015 / 1K tokens
-	"gpt-3.5-turbo-1106":        0.5,  // $0.001 / 1K tokens
-	"gpt-3.5-turbo-0125":        0.25, // $0.0005 / 1K tokens
-	"davinci-002":               1,    // $0.002 / 1K tokens
-	"babbage-002":               0.2,  // $0.0004 / 1K tokens
-	"text-ada-001":              0.2,
-	"text-babbage-001":          0.25,
-	"text-curie-001":            1,
-	"text-davinci-002":          10,
-	"text-davinci-003":          10,
-	"text-davinci-edit-001":     10,
-	"code-davinci-edit-001":     10,
-	"whisper-1":                 15,  // $0.006 / minute -> $0.006 / 150 words -> $0.006 / 200 tokens -> $0.03 / 1k tokens
-	"tts-1":                     7.5, // $0.015 / 1K characters
-	"tts-1-1106":                7.5,
-	"tts-1-hd":                  15, // $0.030 / 1K characters
-	"tts-1-hd-1106":             15,
-	"davinci":                   10,
-	"curie":                     10,
-	"babbage":                   10,
-	"ada":                       10,
-	"text-embedding-ada-002":    0.05,
-	"text-embedding-3-small":    0.01,
-	"text-embedding-3-large":    0.065,
-	"text-search-ada-doc-001":   10,
-	"text-moderation-stable":    0.1,
-	"text-moderation-latest":    0.1,
-	"dall-e-2":                  8,      // $0.016 - $0.020 / image
-	"dall-e-3":                  20,     // $0.040 - $0.120 / image
-	"claude-instant-1":          0.815,  // $1.63 / 1M tokens
-	"claude-2":                  5.51,   // $11.02 / 1M tokens
-	"claude-2.0":                5.51,   // $11.02 / 1M tokens
-	"claude-2.1":                5.51,   // $11.02 / 1M tokens
-	"ERNIE-Bot":                 0.8572, // ￥0.012 / 1k tokens
-	"ERNIE-Bot-turbo":           0.5715, // ￥0.008 / 1k tokens
-	"ERNIE-Bot-4":               8.572,  // ￥0.12 / 1k tokens
-	"Embedding-V1":              0.1429, // ￥0.002 / 1k tokens
-	"PaLM-2":                    1,
-	"gemini-pro":                1,      // $0.00025 / 1k characters -> $0.001 / 1k tokens
-	"gemini-pro-vision":         1,      // $0.00025 / 1k characters -> $0.001 / 1k tokens
+	// https://openai.com/pricing
+	"gpt-4":                   15,
+	"gpt-4-0314":              15,
+	"gpt-4-0613":              15,
+	"gpt-4-32k":               30,
+	"gpt-4-32k-0314":          30,
+	"gpt-4-32k-0613":          30,
+	"gpt-4-1106-preview":      5,    // $0.01 / 1K tokens
+	"gpt-4-0125-preview":      5,    // $0.01 / 1K tokens
+	"gpt-4-turbo-preview":     5,    // $0.01 / 1K tokens
+	"gpt-4-vision-preview":    5,    // $0.01 / 1K tokens
+	"gpt-3.5-turbo":           0.75, // $0.0015 / 1K tokens
+	"gpt-3.5-turbo-0301":      0.75,
+	"gpt-3.5-turbo-0613":      0.75,
+	"gpt-3.5-turbo-16k":       1.5, // $0.003 / 1K tokens
+	"gpt-3.5-turbo-16k-0613":  1.5,
+	"gpt-3.5-turbo-instruct":  0.75, // $0.0015 / 1K tokens
+	"gpt-3.5-turbo-1106":      0.5,  // $0.001 / 1K tokens
+	"gpt-3.5-turbo-0125":      0.25, // $0.0005 / 1K tokens
+	"davinci-002":             1,    // $0.002 / 1K tokens
+	"babbage-002":             0.2,  // $0.0004 / 1K tokens
+	"text-ada-001":            0.2,
+	"text-babbage-001":        0.25,
+	"text-curie-001":          1,
+	"text-davinci-002":        10,
+	"text-davinci-003":        10,
+	"text-davinci-edit-001":   10,
+	"code-davinci-edit-001":   10,
+	"whisper-1":               15,  // $0.006 / minute -> $0.006 / 150 words -> $0.006 / 200 tokens -> $0.03 / 1k tokens
+	"tts-1":                   7.5, // $0.015 / 1K characters
+	"tts-1-1106":              7.5,
+	"tts-1-hd":                15, // $0.030 / 1K characters
+	"tts-1-hd-1106":           15,
+	"davinci":                 10,
+	"curie":                   10,
+	"babbage":                 10,
+	"ada":                     10,
+	"text-embedding-ada-002":  0.05,
+	"text-embedding-3-small":  0.01,
+	"text-embedding-3-large":  0.065,
+	"text-search-ada-doc-001": 10,
+	"text-moderation-stable":  0.1,
+	"text-moderation-latest":  0.1,
+	"dall-e-2":                8,  // $0.016 - $0.020 / image
+	"dall-e-3":                20, // $0.040 - $0.120 / image
+	// https://www.anthropic.com/api#pricing
+	"claude-instant-1.2":       0.8 / 1000 * USD,
+	"claude-2.0":               8.0 / 1000 * USD,
+	"claude-2.1":               8.0 / 1000 * USD,
+	"claude-3-haiku-20240307":  0.25 / 1000 * USD,
+	"claude-3-sonnet-20240229": 3.0 / 1000 * USD,
+	"claude-3-opus-20240229":   15.0 / 1000 * USD,
+	// https://cloud.baidu.com/doc/WENXINWORKSHOP/s/hlrk4akp7
+	"ERNIE-Bot":         0.8572,     // ￥0.012 / 1k tokens
+	"ERNIE-Bot-turbo":   0.5715,     // ￥0.008 / 1k tokens
+	"ERNIE-Bot-4":       0.12 * RMB, // ￥0.12 / 1k tokens
+	"ERNIE-Bot-8k":      0.024 * RMB,
+	"Embedding-V1":      0.1429, // ￥0.002 / 1k tokens
+	"bge-large-zh":      0.002 * RMB,
+	"bge-large-en":      0.002 * RMB,
+	"bge-large-8k":      0.002 * RMB,
+	"PaLM-2":            1,
+	"gemini-pro":        1, // $0.00025 / 1k characters -> $0.001 / 1k tokens
+	"gemini-pro-vision": 1, // $0.00025 / 1k characters -> $0.001 / 1k tokens
+	// https://open.bigmodel.cn/pricing
+	"glm-4":                     0.1 * RMB,
+	"glm-4v":                    0.1 * RMB,
+	"glm-3-turbo":               0.005 * RMB,
 	"chatglm_turbo":             0.3572, // ￥0.005 / 1k tokens
 	"chatglm_pro":               0.7143, // ￥0.01 / 1k tokens
 	"chatglm_std":               0.3572, // ￥0.005 / 1k tokens
@@ -103,11 +98,81 @@ var ModelRatio = map[string]float64{
 	"qwen-max-longcontext":      1.4286, // ￥0.02 / 1k tokens
 	"text-embedding-v1":         0.05,   // ￥0.0007 / 1k tokens
 	"SparkDesk":                 1.2858, // ￥0.018 / 1k tokens
+	"SparkDesk-v1.1":            1.2858, // ￥0.018 / 1k tokens
+	"SparkDesk-v2.1":            1.2858, // ￥0.018 / 1k tokens
+	"SparkDesk-v3.1":            1.2858, // ￥0.018 / 1k tokens
+	"SparkDesk-v3.5":            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
 	"semantic_similarity_s1_v1": 0.0715, // ¥0.001 / 1k tokens
 	"hunyuan":                   7.143,  // ¥0.1 / 1k tokens  // https://cloud.tencent.com/document/product/1729/97731#e0e6be58-60c8-469f-bdeb-6c264ce3b4d0
+	"ChatStd":                   0.01 * RMB,
+	"ChatPro":                   0.1 * RMB,
+	// https://platform.moonshot.cn/pricing
+	"moonshot-v1-8k":   0.012 * RMB,
+	"moonshot-v1-32k":  0.024 * RMB,
+	"moonshot-v1-128k": 0.06 * RMB,
+	// https://platform.baichuan-ai.com/price
+	"Baichuan2-Turbo":      0.008 * RMB,
+	"Baichuan2-Turbo-192k": 0.016 * RMB,
+	"Baichuan2-53B":        0.02 * RMB,
+	// https://api.minimax.chat/document/price
+	"abab6-chat":    0.1 * RMB,
+	"abab5.5-chat":  0.015 * RMB,
+	"abab5.5s-chat": 0.005 * RMB,
+	// https://docs.mistral.ai/platform/pricing/
+	"open-mistral-7b":       0.25 / 1000 * USD,
+	"open-mixtral-8x7b":     0.7 / 1000 * USD,
+	"mistral-small-latest":  2.0 / 1000 * USD,
+	"mistral-medium-latest": 2.7 / 1000 * USD,
+	"mistral-large-latest":  8.0 / 1000 * USD,
+	"mistral-embed":         0.1 / 1000 * USD,
+	// https://wow.groq.com/
+	"llama2-70b-4096":    0.7 / 1000 * USD,
+	"llama2-7b-2048":     0.1 / 1000 * USD,
+	"mixtral-8x7b-32768": 0.27 / 1000 * USD,
+	"gemma-7b-it":        0.1 / 1000 * USD,
+	// https://platform.lingyiwanwu.com/docs#-计费单元
+	"yi-34b-chat-0205": 2.5 / 1000000 * RMB,
+	"yi-34b-chat-200k": 12.0 / 1000000 * RMB,
+	"yi-vl-plus":       6.0 / 1000000 * RMB,
+}
+
+var CompletionRatio = map[string]float64{}
+
+var DefaultModelRatio map[string]float64
+var DefaultCompletionRatio map[string]float64
+
+func init() {
+	DefaultModelRatio = make(map[string]float64)
+	for k, v := range ModelRatio {
+		DefaultModelRatio[k] = v
+	}
+	DefaultCompletionRatio = make(map[string]float64)
+	for k, v := range CompletionRatio {
+		DefaultCompletionRatio[k] = v
+	}
+}
+
+func AddNewMissingRatio(oldRatio string) string {
+	newRatio := make(map[string]float64)
+	err := json.Unmarshal([]byte(oldRatio), &newRatio)
+	if err != nil {
+		logger.SysError("error unmarshalling old ratio: " + err.Error())
+		return oldRatio
+	}
+	for k, v := range DefaultModelRatio {
+		if _, ok := newRatio[k]; !ok {
+			newRatio[k] = v
+		}
+	}
+	jsonBytes, err := json.Marshal(newRatio)
+	if err != nil {
+		logger.SysError("error marshalling new ratio: " + err.Error())
+		return oldRatio
+	}
+	return string(jsonBytes)
 }
 
 func ModelRatio2JSONString() string {
@@ -128,6 +193,9 @@ func GetModelRatio(name string) float64 {
 		name = strings.TrimSuffix(name, "-internet")
 	}
 	ratio, ok := ModelRatio[name]
+	if !ok {
+		ratio, ok = DefaultModelRatio[name]
+	}
 	if !ok {
 		logger.SysError("model ratio not found: " + name)
 		return 30
@@ -135,8 +203,6 @@ func GetModelRatio(name string) float64 {
 	return ratio
 }
 
-var CompletionRatio = map[string]float64{}
-
 func CompletionRatio2JSONString() string {
 	jsonBytes, err := json.Marshal(CompletionRatio)
 	if err != nil {
@@ -154,6 +220,9 @@ func GetCompletionRatio(name string) float64 {
 	if ratio, ok := CompletionRatio[name]; ok {
 		return ratio
 	}
+	if ratio, ok := DefaultCompletionRatio[name]; ok {
+		return ratio
+	}
 	if strings.HasPrefix(name, "gpt-3.5") {
 		if strings.HasSuffix(name, "0125") {
 			// https://openai.com/blog/new-embedding-models-and-api-updates
@@ -172,7 +241,7 @@ func GetCompletionRatio(name string) float64 {
 				return 2
 			}
 		}
-		return 1.333333
+		return 4.0 / 3.0
 	}
 	if strings.HasPrefix(name, "gpt-4") {
 		if strings.HasSuffix(name, "preview") {
@@ -180,11 +249,18 @@ func GetCompletionRatio(name string) float64 {
 		}
 		return 2
 	}
-	if strings.HasPrefix(name, "claude-instant-1") {
-		return 3.38
+	if strings.HasPrefix(name, "claude-3") {
+		return 5
 	}
-	if strings.HasPrefix(name, "claude-2") {
-		return 2.965517
+	if strings.HasPrefix(name, "claude-") {
+		return 3
+	}
+	if strings.HasPrefix(name, "mistral-") {
+		return 3
+	}
+	switch name {
+	case "llama2-70b-4096":
+		return 0.8 / 0.7
 	}
 	return 1
 }

common/random.go

@@ -0,0 +1,8 @@
+package common
+
+import "math/rand"
+
+// RandRange returns a random number between min and max (max is not included)
+func RandRange(min, max int) int {
+	return min + rand.Intn(max-min)
+}

common/redis.go

@@ -3,7 +3,7 @@ package common
 import (
 	"context"
 	"github.com/go-redis/redis/v8"
-	"one-api/common/logger"
+	"github.com/songquanpeng/one-api/common/logger"
 	"os"
 	"time"
 )

common/utils.go

@@ -2,10 +2,10 @@ package common
 
 import (
 	"fmt"
-	"one-api/common/config"
+	"github.com/songquanpeng/one-api/common/config"
 )
 
-func LogQuota(quota int) string {
+func LogQuota(quota int64) string {
 	if config.DisplayInCurrencyEnabled {
 		return fmt.Sprintf("＄%.6f 额度", float64(quota)/config.QuotaPerUnit)
 	} else {

controller/billing.go

@@ -2,14 +2,14 @@ package controller
 
 import (
 	"github.com/gin-gonic/gin"
-	"one-api/common/config"
-	"one-api/model"
-	"one-api/relay/channel/openai"
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/model"
+	relaymodel "github.com/songquanpeng/one-api/relay/model"
 )
 
 func GetSubscription(c *gin.Context) {
-	var remainQuota int
-	var usedQuota int
+	var remainQuota int64
+	var usedQuota int64
 	var err error
 	var token *model.Token
 	var expiredTime int64
@@ -22,13 +22,15 @@ func GetSubscription(c *gin.Context) {
 	} else {
 		userId := c.GetInt("id")
 		remainQuota, err = model.GetUserQuota(userId)
-		usedQuota, err = model.GetUserUsedQuota(userId)
+		if err != nil {
+			usedQuota, err = model.GetUserUsedQuota(userId)
+		}
 	}
 	if expiredTime <= 0 {
 		expiredTime = 0
 	}
 	if err != nil {
-		Error := openai.Error{
+		Error := relaymodel.Error{
 			Message: err.Error(),
 			Type:    "upstream_error",
 		}
@@ -58,7 +60,7 @@ func GetSubscription(c *gin.Context) {
 }
 
 func GetUsage(c *gin.Context) {
-	var quota int
+	var quota int64
 	var err error
 	var token *model.Token
 	if config.DisplayTokenStatEnabled {
@@ -70,7 +72,7 @@ func GetUsage(c *gin.Context) {
 		quota, err = model.GetUserUsedQuota(userId)
 	}
 	if err != nil {
-		Error := openai.Error{
+		Error := relaymodel.Error{
 			Message: err.Error(),
 			Type:    "one_api_error",
 		}

controller/channel-billing.go

@@ -4,13 +4,14 @@ import (
 	"encoding/json"
 	"errors"
 	"fmt"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/model"
+	"github.com/songquanpeng/one-api/monitor"
+	"github.com/songquanpeng/one-api/relay/util"
 	"io"
 	"net/http"
-	"one-api/common"
-	"one-api/common/config"
-	"one-api/common/logger"
-	"one-api/model"
-	"one-api/relay/util"
 	"strconv"
 	"time"
 
@@ -295,7 +296,7 @@ func UpdateChannelBalance(c *gin.Context) {
 }
 
 func updateAllChannelsBalance() error {
-	channels, err := model.GetAllChannels(0, 0, true)
+	channels, err := model.GetAllChannels(0, 0, "all")
 	if err != nil {
 		return err
 	}
@@ -313,7 +314,7 @@ func updateAllChannelsBalance() error {
 		} else {
 			// err is nil & balance <= 0 means quota is used up
 			if balance <= 0 {
-				disableChannel(channel.Id, channel.Name, "余额不足")
+				monitor.DisableChannel(channel.Id, channel.Name, "余额不足")
 			}
 		}
 		time.Sleep(config.RequestInterval)
@@ -322,15 +323,14 @@ func updateAllChannelsBalance() error {
 }
 
 func UpdateAllChannelsBalance(c *gin.Context) {
-	// TODO: make it async
-	err := updateAllChannelsBalance()
-	if err != nil {
-		c.JSON(http.StatusOK, gin.H{
-			"success": false,
-			"message": err.Error(),
-		})
-		return
-	}
+	//err := updateAllChannelsBalance()
+	//if err != nil {
+	//	c.JSON(http.StatusOK, gin.H{
+	//		"success": false,
+	//		"message": err.Error(),
+	//	})
+	//	return
+	//}
 	c.JSON(http.StatusOK, gin.H{
 		"success": true,
 		"message": "",

controller/channel-test.go

@@ -5,102 +5,36 @@ import (
 	"encoding/json"
 	"errors"
 	"fmt"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/common/message"
+	"github.com/songquanpeng/one-api/middleware"
+	"github.com/songquanpeng/one-api/model"
+	"github.com/songquanpeng/one-api/monitor"
+	"github.com/songquanpeng/one-api/relay/constant"
+	"github.com/songquanpeng/one-api/relay/helper"
+	relaymodel "github.com/songquanpeng/one-api/relay/model"
+	"github.com/songquanpeng/one-api/relay/util"
 	"io"
 	"net/http"
-	"one-api/common"
-	"one-api/common/config"
-	"one-api/common/logger"
-	"one-api/model"
-	"one-api/relay/channel/openai"
-	"one-api/relay/util"
+	"net/http/httptest"
+	"net/url"
 	"strconv"
+	"strings"
 	"sync"
 	"time"
 
 	"github.com/gin-gonic/gin"
 )
 
-func testChannel(channel *model.Channel, request openai.ChatRequest) (err error, openaiErr *openai.Error) {
-	switch channel.Type {
-	case common.ChannelTypePaLM:
-		fallthrough
-	case common.ChannelTypeGemini:
-		fallthrough
-	case common.ChannelTypeAnthropic:
-		fallthrough
-	case common.ChannelTypeBaidu:
-		fallthrough
-	case common.ChannelTypeZhipu:
-		fallthrough
-	case common.ChannelTypeAli:
-		fallthrough
-	case common.ChannelType360:
-		fallthrough
-	case common.ChannelTypeXunfei:
-		return errors.New("该渠道类型当前版本不支持测试，请手动测试"), nil
-	case common.ChannelTypeAzure:
-		request.Model = "gpt-35-turbo"
-		defer func() {
-			if err != nil {
-				err = errors.New("请确保已在 Azure 上创建了 gpt-35-turbo 模型，并且 apiVersion 已正确填写！")
-			}
-		}()
-	default:
-		request.Model = "gpt-3.5-turbo"
+func buildTestRequest() *relaymodel.GeneralOpenAIRequest {
+	testRequest := &relaymodel.GeneralOpenAIRequest{
+		MaxTokens: 2,
+		Stream:    false,
+		Model:     "gpt-3.5-turbo",
 	}
-	requestURL := common.ChannelBaseURLs[channel.Type]
-	if channel.Type == common.ChannelTypeAzure {
-		requestURL = util.GetFullRequestURL(channel.GetBaseURL(), fmt.Sprintf("/openai/deployments/%s/chat/completions?api-version=2023-03-15-preview", request.Model), channel.Type)
-	} else {
-		if baseURL := channel.GetBaseURL(); len(baseURL) > 0 {
-			requestURL = baseURL
-		}
-
-		requestURL = util.GetFullRequestURL(requestURL, "/v1/chat/completions", channel.Type)
-	}
-	jsonData, err := json.Marshal(request)
-	if err != nil {
-		return err, nil
-	}
-	req, err := http.NewRequest("POST", requestURL, bytes.NewBuffer(jsonData))
-	if err != nil {
-		return err, nil
-	}
-	if channel.Type == common.ChannelTypeAzure {
-		req.Header.Set("api-key", channel.Key)
-	} else {
-		req.Header.Set("Authorization", "Bearer "+channel.Key)
-	}
-	req.Header.Set("Content-Type", "application/json")
-	resp, err := util.HTTPClient.Do(req)
-	if err != nil {
-		return err, nil
-	}
-	defer resp.Body.Close()
-	var response openai.SlimTextResponse
-	body, err := io.ReadAll(resp.Body)
-	if err != nil {
-		return err, nil
-	}
-	err = json.Unmarshal(body, &response)
-	if err != nil {
-		return fmt.Errorf("Error: %s\nResp body: %s", err, body), nil
-	}
-	if response.Usage.CompletionTokens == 0 {
-		if response.Error.Message == "" {
-			response.Error.Message = "补全 tokens 非预期返回 0"
-		}
-		return errors.New(fmt.Sprintf("type %s, code %v, message %s", response.Error.Type, response.Error.Code, response.Error.Message)), &response.Error
-	}
-	return nil, nil
-}
-
-func buildTestRequest() *openai.ChatRequest {
-	testRequest := &openai.ChatRequest{
-		Model:     "", // this will be set later
-		MaxTokens: 1,
-	}
-	testMessage := openai.Message{
+	testMessage := relaymodel.Message{
 		Role:    "user",
 		Content: "hi",
 	}
@@ -108,6 +42,72 @@ func buildTestRequest() *openai.ChatRequest {
 	return testRequest
 }
 
+func testChannel(channel *model.Channel) (err error, openaiErr *relaymodel.Error) {
+	w := httptest.NewRecorder()
+	c, _ := gin.CreateTestContext(w)
+	c.Request = &http.Request{
+		Method: "POST",
+		URL:    &url.URL{Path: "/v1/chat/completions"},
+		Body:   nil,
+		Header: make(http.Header),
+	}
+	c.Request.Header.Set("Authorization", "Bearer "+channel.Key)
+	c.Request.Header.Set("Content-Type", "application/json")
+	c.Set("channel", channel.Type)
+	c.Set("base_url", channel.GetBaseURL())
+	middleware.SetupContextForSelectedChannel(c, channel, "")
+	meta := util.GetRelayMeta(c)
+	apiType := constant.ChannelType2APIType(channel.Type)
+	adaptor := helper.GetAdaptor(apiType)
+	if adaptor == nil {
+		return fmt.Errorf("invalid api type: %d, adaptor is nil", apiType), nil
+	}
+	adaptor.Init(meta)
+	modelName := adaptor.GetModelList()[0]
+	if !strings.Contains(channel.Models, modelName) {
+		modelNames := strings.Split(channel.Models, ",")
+		if len(modelNames) > 0 {
+			modelName = modelNames[0]
+		}
+	}
+	request := buildTestRequest()
+	request.Model = modelName
+	meta.OriginModelName, meta.ActualModelName = modelName, modelName
+	convertedRequest, err := adaptor.ConvertRequest(c, constant.RelayModeChatCompletions, request)
+	if err != nil {
+		return err, nil
+	}
+	jsonData, err := json.Marshal(convertedRequest)
+	if err != nil {
+		return err, nil
+	}
+	requestBody := bytes.NewBuffer(jsonData)
+	c.Request.Body = io.NopCloser(requestBody)
+	resp, err := adaptor.DoRequest(c, meta, requestBody)
+	if err != nil {
+		return err, nil
+	}
+	if resp.StatusCode != http.StatusOK {
+		err := util.RelayErrorHandler(resp)
+		return fmt.Errorf("status code %d: %s", resp.StatusCode, err.Error.Message), &err.Error
+	}
+	usage, respErr := adaptor.DoResponse(c, resp, meta)
+	if respErr != nil {
+		return fmt.Errorf("%s", respErr.Error.Message), &respErr.Error
+	}
+	if usage == nil {
+		return errors.New("usage is nil"), nil
+	}
+	result := w.Result()
+	// print result.Body
+	respBody, err := io.ReadAll(result.Body)
+	if err != nil {
+		return err, nil
+	}
+	logger.SysLog(fmt.Sprintf("testing channel #%d, response: \n%s", channel.Id, string(respBody)))
+	return nil, nil
+}
+
 func TestChannel(c *gin.Context) {
 	id, err := strconv.Atoi(c.Param("id"))
 	if err != nil {
@@ -125,9 +125,8 @@ func TestChannel(c *gin.Context) {
 		})
 		return
 	}
-	testRequest := buildTestRequest()
 	tik := time.Now()
-	err, _ = testChannel(channel, *testRequest)
+	err, _ = testChannel(channel)
 	tok := time.Now()
 	milliseconds := tok.Sub(tik).Milliseconds()
 	go channel.UpdateResponseTime(milliseconds)
@@ -151,33 +150,7 @@ func TestChannel(c *gin.Context) {
 var testAllChannelsLock sync.Mutex
 var testAllChannelsRunning bool = false
 
-func notifyRootUser(subject string, content string) {
-	if config.RootUserEmail == "" {
-		config.RootUserEmail = model.GetRootUserEmail()
-	}
-	err := common.SendEmail(subject, config.RootUserEmail, content)
-	if err != nil {
-		logger.SysError(fmt.Sprintf("failed to send email: %s", err.Error()))
-	}
-}
-
-// disable & notify
-func disableChannel(channelId int, channelName string, reason string) {
-	model.UpdateChannelStatusById(channelId, common.ChannelStatusAutoDisabled)
-	subject := fmt.Sprintf("通道「%s」（#%d）已被禁用", channelName, channelId)
-	content := fmt.Sprintf("通道「%s」（#%d）已被禁用，原因：%s", channelName, channelId, reason)
-	notifyRootUser(subject, content)
-}
-
-// enable & notify
-func enableChannel(channelId int, channelName string) {
-	model.UpdateChannelStatusById(channelId, common.ChannelStatusEnabled)
-	subject := fmt.Sprintf("通道「%s」（#%d）已被启用", channelName, channelId)
-	content := fmt.Sprintf("通道「%s」（#%d）已被启用", channelName, channelId)
-	notifyRootUser(subject, content)
-}
-
-func testAllChannels(notify bool) error {
+func testChannels(notify bool, scope string) error {
 	if config.RootUserEmail == "" {
 		config.RootUserEmail = model.GetRootUserEmail()
 	}
@@ -188,11 +161,10 @@ func testAllChannels(notify bool) error {
 	}
 	testAllChannelsRunning = true
 	testAllChannelsLock.Unlock()
-	channels, err := model.GetAllChannels(0, 0, true)
+	channels, err := model.GetAllChannels(0, 0, scope)
 	if err != nil {
 		return err
 	}
-	testRequest := buildTestRequest()
 	var disableThreshold = int64(config.ChannelDisableThreshold * 1000)
 	if disableThreshold == 0 {
 		disableThreshold = 10000000 // a impossible value
@@ -201,18 +173,22 @@ func testAllChannels(notify bool) error {
 		for _, channel := range channels {
 			isChannelEnabled := channel.Status == common.ChannelStatusEnabled
 			tik := time.Now()
-			err, openaiErr := testChannel(channel, *testRequest)
+			err, openaiErr := testChannel(channel)
 			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))
-				disableChannel(channel.Id, channel.Name, err.Error())
+				if config.AutomaticDisableChannelEnabled {
+					monitor.DisableChannel(channel.Id, channel.Name, err.Error())
+				} else {
+					_ = message.Notify(message.ByAll, fmt.Sprintf("渠道 %s （%d）测试超时", channel.Name, channel.Id), "", err.Error())
+				}
 			}
 			if isChannelEnabled && util.ShouldDisableChannel(openaiErr, -1) {
-				disableChannel(channel.Id, channel.Name, err.Error())
+				monitor.DisableChannel(channel.Id, channel.Name, err.Error())
 			}
 			if !isChannelEnabled && util.ShouldEnableChannel(err, openaiErr) {
-				enableChannel(channel.Id, channel.Name)
+				monitor.EnableChannel(channel.Id, channel.Name)
 			}
 			channel.UpdateResponseTime(milliseconds)
 			time.Sleep(config.RequestInterval)
@@ -221,7 +197,7 @@ func testAllChannels(notify bool) error {
 		testAllChannelsRunning = false
 		testAllChannelsLock.Unlock()
 		if notify {
-			err := common.SendEmail("通道测试完成", config.RootUserEmail, "通道测试完成，如果没有收到禁用通知，说明所有通道都正常")
+			err := message.Notify(message.ByAll, "通道测试完成", "", "通道测试完成，如果没有收到禁用通知，说明所有通道都正常")
 			if err != nil {
 				logger.SysError(fmt.Sprintf("failed to send email: %s", err.Error()))
 			}
@@ -230,8 +206,12 @@ func testAllChannels(notify bool) error {
 	return nil
 }
 
-func TestAllChannels(c *gin.Context) {
-	err := testAllChannels(true)
+func TestChannels(c *gin.Context) {
+	scope := c.Query("scope")
+	if scope == "" {
+		scope = "all"
+	}
+	err := testChannels(true, scope)
 	if err != nil {
 		c.JSON(http.StatusOK, gin.H{
 			"success": false,
@@ -250,7 +230,7 @@ func AutomaticallyTestChannels(frequency int) {
 	for {
 		time.Sleep(time.Duration(frequency) * time.Minute)
 		logger.SysLog("testing all channels")
-		_ = testAllChannels(false)
+		_ = testChannels(false, "all")
 		logger.SysLog("channel test finished")
 	}
 }

controller/channel.go

@@ -2,10 +2,10 @@ package controller
 
 import (
 	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/helper"
+	"github.com/songquanpeng/one-api/model"
 	"net/http"
-	"one-api/common/config"
-	"one-api/common/helper"
-	"one-api/model"
 	"strconv"
 	"strings"
 )
@@ -15,7 +15,7 @@ func GetAllChannels(c *gin.Context) {
 	if p < 0 {
 		p = 0
 	}
-	channels, err := model.GetAllChannels(p*config.ItemsPerPage, config.ItemsPerPage, false)
+	channels, err := model.GetAllChannels(p*config.ItemsPerPage, config.ItemsPerPage, "limited")
 	if err != nil {
 		c.JSON(http.StatusOK, gin.H{
 			"success": false,

controller/github.go

@@ -7,12 +7,12 @@ import (
 	"fmt"
 	"github.com/gin-contrib/sessions"
 	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/helper"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/model"
 	"net/http"
-	"one-api/common"
-	"one-api/common/config"
-	"one-api/common/helper"
-	"one-api/common/logger"
-	"one-api/model"
 	"strconv"
 	"time"
 )

controller/group.go

@@ -2,13 +2,13 @@ package controller
 
 import (
 	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common"
 	"net/http"
-	"one-api/common"
 )
 
 func GetGroups(c *gin.Context) {
 	groupNames := make([]string, 0)
-	for groupName, _ := range common.GroupRatio {
+	for groupName := range common.GroupRatio {
 		groupNames = append(groupNames, groupName)
 	}
 	c.JSON(http.StatusOK, gin.H{

controller/log.go

@@ -2,9 +2,9 @@ package controller
 
 import (
 	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/model"
 	"net/http"
-	"one-api/common/config"
-	"one-api/model"
 	"strconv"
 )
 

controller/misc.go

@@ -3,10 +3,11 @@ package controller
 import (
 	"encoding/json"
 	"fmt"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/message"
+	"github.com/songquanpeng/one-api/model"
 	"net/http"
-	"one-api/common"
-	"one-api/common/config"
-	"one-api/model"
 	"strings"
 
 	"github.com/gin-gonic/gin"
@@ -110,7 +111,7 @@ func SendEmailVerification(c *gin.Context) {
 	content := fmt.Sprintf("<p>您好，你正在进行%s邮箱验证。</p>"+
 		"<p>您的验证码为: <strong>%s</strong></p>"+
 		"<p>验证码 %d 分钟内有效，如果不是本人操作，请忽略。</p>", config.SystemName, code, common.VerificationValidMinutes)
-	err := common.SendEmail(subject, email, content)
+	err := message.SendEmail(subject, email, content)
 	if err != nil {
 		c.JSON(http.StatusOK, gin.H{
 			"success": false,
@@ -149,7 +150,7 @@ func SendPasswordResetEmail(c *gin.Context) {
 		"<p>点击 <a href='%s'>此处</a> 进行密码重置。</p>"+
 		"<p>如果链接无法点击，请尝试点击下面的链接或将其复制到浏览器中打开：<br> %s </p>"+
 		"<p>重置链接 %d 分钟内有效，如果不是本人操作，请忽略。</p>", config.SystemName, link, link, common.VerificationValidMinutes)
-	err := common.SendEmail(subject, email, content)
+	err := message.SendEmail(subject, email, content)
 	if err != nil {
 		c.JSON(http.StatusOK, gin.H{
 			"success": false,

controller/model.go

@@ -3,7 +3,13 @@ package controller
 import (
 	"fmt"
 	"github.com/gin-gonic/gin"
-	"one-api/relay/channel/openai"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/relay/channel/openai"
+	"github.com/songquanpeng/one-api/relay/constant"
+	"github.com/songquanpeng/one-api/relay/helper"
+	relaymodel "github.com/songquanpeng/one-api/relay/model"
+	"github.com/songquanpeng/one-api/relay/util"
+	"net/http"
 )
 
 // https://platform.openai.com/docs/api-reference/models/list
@@ -35,6 +41,7 @@ type OpenAIModels struct {
 
 var openAIModels []OpenAIModels
 var openAIModelsMap map[string]OpenAIModels
+var channelId2Models map[int][]string
 
 func init() {
 	var permission []OpenAIModelPermission
@@ -53,597 +60,63 @@ func init() {
 		IsBlocking:         false,
 	})
 	// https://platform.openai.com/docs/models/model-endpoint-compatibility
-	openAIModels = []OpenAIModels{
-		{
-			Id:         "dall-e-2",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "dall-e-2",
-			Parent:     nil,
-		},
-		{
-			Id:         "dall-e-3",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "dall-e-3",
-			Parent:     nil,
-		},
-		{
-			Id:         "whisper-1",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "whisper-1",
-			Parent:     nil,
-		},
-		{
-			Id:         "tts-1",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "tts-1",
-			Parent:     nil,
-		},
-		{
-			Id:         "tts-1-1106",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "tts-1-1106",
-			Parent:     nil,
-		},
-		{
-			Id:         "tts-1-hd",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "tts-1-hd",
-			Parent:     nil,
-		},
-		{
-			Id:         "tts-1-hd-1106",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "tts-1-hd-1106",
-			Parent:     nil,
-		},
-		{
-			Id:         "gpt-3.5-turbo",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "gpt-3.5-turbo",
-			Parent:     nil,
-		},
-		{
-			Id:         "gpt-3.5-turbo-0301",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "gpt-3.5-turbo-0301",
-			Parent:     nil,
-		},
-		{
-			Id:         "gpt-3.5-turbo-0613",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "gpt-3.5-turbo-0613",
-			Parent:     nil,
-		},
-		{
-			Id:         "gpt-3.5-turbo-16k",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "gpt-3.5-turbo-16k",
-			Parent:     nil,
-		},
-		{
-			Id:         "gpt-3.5-turbo-16k-0613",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "gpt-3.5-turbo-16k-0613",
-			Parent:     nil,
-		},
-		{
-			Id:         "gpt-3.5-turbo-1106",
-			Object:     "model",
-			Created:    1699593571,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "gpt-3.5-turbo-1106",
-			Parent:     nil,
-		},
-		{
-			Id:         "gpt-3.5-turbo-0125",
-			Object:     "model",
-			Created:    1706232090,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "gpt-3.5-turbo-0125",
-			Parent:     nil,
-		},
-		{
-			Id:         "gpt-3.5-turbo-instruct",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "gpt-3.5-turbo-instruct",
-			Parent:     nil,
-		},
-		{
-			Id:         "gpt-4",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "gpt-4",
-			Parent:     nil,
-		},
-		{
-			Id:         "gpt-4-0314",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "gpt-4-0314",
-			Parent:     nil,
-		},
-		{
-			Id:         "gpt-4-0613",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "gpt-4-0613",
-			Parent:     nil,
-		},
-		{
-			Id:         "gpt-4-32k",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "gpt-4-32k",
-			Parent:     nil,
-		},
-		{
-			Id:         "gpt-4-32k-0314",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "gpt-4-32k-0314",
-			Parent:     nil,
-		},
-		{
-			Id:         "gpt-4-32k-0613",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "gpt-4-32k-0613",
-			Parent:     nil,
-		},
-		{
-			Id:         "gpt-4-1106-preview",
-			Object:     "model",
-			Created:    1699593571,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "gpt-4-1106-preview",
-			Parent:     nil,
-		},
-		{
-			Id:         "gpt-4-0125-preview",
-			Object:     "model",
-			Created:    1706232090,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "gpt-4-0125-preview",
-			Parent:     nil,
-		},
-		{
-			Id:         "gpt-4-turbo-preview",
-			Object:     "model",
-			Created:    1706232090,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "gpt-4-turbo-preview",
-			Parent:     nil,
-		},
-		{
-			Id:         "gpt-4-vision-preview",
-			Object:     "model",
-			Created:    1699593571,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "gpt-4-vision-preview",
-			Parent:     nil,
-		},
-		{
-			Id:         "text-embedding-ada-002",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "text-embedding-ada-002",
-			Parent:     nil,
-		},
-		{
-			Id:         "text-embedding-3-small",
-			Object:     "model",
-			Created:    1706232090,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "text-embedding-3-small",
-			Parent:     nil,
-		},
-		{
-			Id:         "text-embedding-3-large",
-			Object:     "model",
-			Created:    1706232090,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "text-embedding-3-large",
-			Parent:     nil,
-		},
-		{
-			Id:         "text-davinci-003",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "text-davinci-003",
-			Parent:     nil,
-		},
-		{
-			Id:         "text-davinci-002",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "text-davinci-002",
-			Parent:     nil,
-		},
-		{
-			Id:         "text-curie-001",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "text-curie-001",
-			Parent:     nil,
-		},
-		{
-			Id:         "text-babbage-001",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "text-babbage-001",
-			Parent:     nil,
-		},
-		{
-			Id:         "text-ada-001",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "text-ada-001",
-			Parent:     nil,
-		},
-		{
-			Id:         "text-moderation-latest",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "text-moderation-latest",
-			Parent:     nil,
-		},
-		{
-			Id:         "text-moderation-stable",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "text-moderation-stable",
-			Parent:     nil,
-		},
-		{
-			Id:         "text-davinci-edit-001",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "text-davinci-edit-001",
-			Parent:     nil,
-		},
-		{
-			Id:         "code-davinci-edit-001",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "code-davinci-edit-001",
-			Parent:     nil,
-		},
-		{
-			Id:         "davinci-002",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "davinci-002",
-			Parent:     nil,
-		},
-		{
-			Id:         "babbage-002",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "openai",
-			Permission: permission,
-			Root:       "babbage-002",
-			Parent:     nil,
-		},
-		{
-			Id:         "claude-instant-1",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "anthropic",
-			Permission: permission,
-			Root:       "claude-instant-1",
-			Parent:     nil,
-		},
-		{
-			Id:         "claude-2",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "anthropic",
-			Permission: permission,
-			Root:       "claude-2",
-			Parent:     nil,
-		},
-		{
-			Id:         "claude-2.1",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "anthropic",
-			Permission: permission,
-			Root:       "claude-2.1",
-			Parent:     nil,
-		},
-		{
-			Id:         "claude-2.0",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "anthropic",
-			Permission: permission,
-			Root:       "claude-2.0",
-			Parent:     nil,
-		},
-		{
-			Id:         "ERNIE-Bot",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "baidu",
-			Permission: permission,
-			Root:       "ERNIE-Bot",
-			Parent:     nil,
-		},
-		{
-			Id:         "ERNIE-Bot-turbo",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "baidu",
-			Permission: permission,
-			Root:       "ERNIE-Bot-turbo",
-			Parent:     nil,
-		},
-		{
-			Id:         "ERNIE-Bot-4",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "baidu",
-			Permission: permission,
-			Root:       "ERNIE-Bot-4",
-			Parent:     nil,
-		},
-		{
-			Id:         "Embedding-V1",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "baidu",
-			Permission: permission,
-			Root:       "Embedding-V1",
-			Parent:     nil,
-		},
-		{
-			Id:         "PaLM-2",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "google palm",
-			Permission: permission,
-			Root:       "PaLM-2",
-			Parent:     nil,
-		},
-		{
-			Id:         "gemini-pro",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "google gemini",
-			Permission: permission,
-			Root:       "gemini-pro",
-			Parent:     nil,
-		},
-		{
-			Id:         "gemini-pro-vision",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "google gemini",
-			Permission: permission,
-			Root:       "gemini-pro-vision",
-			Parent:     nil,
-		},
-		{
-			Id:         "chatglm_turbo",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "zhipu",
-			Permission: permission,
-			Root:       "chatglm_turbo",
-			Parent:     nil,
-		},
-		{
-			Id:         "chatglm_pro",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "zhipu",
-			Permission: permission,
-			Root:       "chatglm_pro",
-			Parent:     nil,
-		},
-		{
-			Id:         "chatglm_std",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "zhipu",
-			Permission: permission,
-			Root:       "chatglm_std",
-			Parent:     nil,
-		},
-		{
-			Id:         "chatglm_lite",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "zhipu",
-			Permission: permission,
-			Root:       "chatglm_lite",
-			Parent:     nil,
-		},
-		{
-			Id:         "qwen-turbo",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "ali",
-			Permission: permission,
-			Root:       "qwen-turbo",
-			Parent:     nil,
-		},
-		{
-			Id:         "qwen-plus",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "ali",
-			Permission: permission,
-			Root:       "qwen-plus",
-			Parent:     nil,
-		},
-		{
-			Id:         "qwen-max",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "ali",
-			Permission: permission,
-			Root:       "qwen-max",
-			Parent:     nil,
-		},
-		{
-			Id:         "qwen-max-longcontext",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "ali",
-			Permission: permission,
-			Root:       "qwen-max-longcontext",
-			Parent:     nil,
-		},
-		{
-			Id:         "text-embedding-v1",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "ali",
-			Permission: permission,
-			Root:       "text-embedding-v1",
-			Parent:     nil,
-		},
-		{
-			Id:         "SparkDesk",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "xunfei",
-			Permission: permission,
-			Root:       "SparkDesk",
-			Parent:     nil,
-		},
-		{
-			Id:         "360GPT_S2_V9",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "360",
-			Permission: permission,
-			Root:       "360GPT_S2_V9",
-			Parent:     nil,
-		},
-		{
-			Id:         "embedding-bert-512-v1",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "360",
-			Permission: permission,
-			Root:       "embedding-bert-512-v1",
-			Parent:     nil,
-		},
-		{
-			Id:         "embedding_s1_v1",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "360",
-			Permission: permission,
-			Root:       "embedding_s1_v1",
-			Parent:     nil,
-		},
-		{
-			Id:         "semantic_similarity_s1_v1",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "360",
-			Permission: permission,
-			Root:       "semantic_similarity_s1_v1",
-			Parent:     nil,
-		},
-		{
-			Id:         "hunyuan",
-			Object:     "model",
-			Created:    1677649963,
-			OwnedBy:    "tencent",
-			Permission: permission,
-			Root:       "hunyuan",
-			Parent:     nil,
-		},
+	for i := 0; i < constant.APITypeDummy; i++ {
+		if i == constant.APITypeAIProxyLibrary {
+			continue
+		}
+		adaptor := helper.GetAdaptor(i)
+		channelName := adaptor.GetChannelName()
+		modelNames := adaptor.GetModelList()
+		for _, modelName := range modelNames {
+			openAIModels = append(openAIModels, OpenAIModels{
+				Id:         modelName,
+				Object:     "model",
+				Created:    1626777600,
+				OwnedBy:    channelName,
+				Permission: permission,
+				Root:       modelName,
+				Parent:     nil,
+			})
+		}
+	}
+	for _, channelType := range openai.CompatibleChannels {
+		if channelType == common.ChannelTypeAzure {
+			continue
+		}
+		channelName, channelModelList := openai.GetCompatibleChannelMeta(channelType)
+		for _, modelName := range channelModelList {
+			openAIModels = append(openAIModels, OpenAIModels{
+				Id:         modelName,
+				Object:     "model",
+				Created:    1626777600,
+				OwnedBy:    channelName,
+				Permission: permission,
+				Root:       modelName,
+				Parent:     nil,
+			})
+		}
 	}
 	openAIModelsMap = make(map[string]OpenAIModels)
 	for _, model := range openAIModels {
 		openAIModelsMap[model.Id] = model
 	}
+	channelId2Models = make(map[int][]string)
+	for i := 1; i < common.ChannelTypeDummy; i++ {
+		adaptor := helper.GetAdaptor(constant.ChannelType2APIType(i))
+		meta := &util.RelayMeta{
+			ChannelType: i,
+		}
+		adaptor.Init(meta)
+		channelId2Models[i] = adaptor.GetModelList()
+	}
+}
+
+func DashboardListModels(c *gin.Context) {
+	c.JSON(http.StatusOK, gin.H{
+		"success": true,
+		"message": "",
+		"data":    channelId2Models,
+	})
 }
 
 func ListModels(c *gin.Context) {
@@ -658,7 +131,7 @@ func RetrieveModel(c *gin.Context) {
 	if model, ok := openAIModelsMap[modelId]; ok {
 		c.JSON(200, model)
 	} else {
-		Error := openai.Error{
+		Error := relaymodel.Error{
 			Message: fmt.Sprintf("The model '%s' does not exist", modelId),
 			Type:    "invalid_request_error",
 			Param:   "model",

controller/option.go

@@ -2,10 +2,10 @@ package controller
 
 import (
 	"encoding/json"
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/helper"
+	"github.com/songquanpeng/one-api/model"
 	"net/http"
-	"one-api/common/config"
-	"one-api/common/helper"
-	"one-api/model"
 	"strings"
 
 	"github.com/gin-gonic/gin"

controller/redemption.go

@@ -2,10 +2,10 @@ package controller
 
 import (
 	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/helper"
+	"github.com/songquanpeng/one-api/model"
 	"net/http"
-	"one-api/common/config"
-	"one-api/common/helper"
-	"one-api/model"
 	"strconv"
 )
 

controller/relay.go

@@ -1,24 +1,29 @@
 package controller
 
 import (
+	"bytes"
+	"context"
 	"fmt"
 	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/helper"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/middleware"
+	dbmodel "github.com/songquanpeng/one-api/model"
+	"github.com/songquanpeng/one-api/monitor"
+	"github.com/songquanpeng/one-api/relay/constant"
+	"github.com/songquanpeng/one-api/relay/controller"
+	"github.com/songquanpeng/one-api/relay/model"
+	"github.com/songquanpeng/one-api/relay/util"
+	"io"
 	"net/http"
-	"one-api/common/config"
-	"one-api/common/helper"
-	"one-api/common/logger"
-	"one-api/relay/channel/openai"
-	"one-api/relay/constant"
-	"one-api/relay/controller"
-	"one-api/relay/util"
-	"strconv"
 )
 
 // https://platform.openai.com/docs/api-reference/chat
 
-func Relay(c *gin.Context) {
-	relayMode := constant.Path2RelayMode(c.Request.URL.Path)
-	var err *openai.ErrorWithStatusCode
+func relay(c *gin.Context, relayMode int) *model.ErrorWithStatusCode {
+	var err *model.ErrorWithStatusCode
 	switch relayMode {
 	case constant.RelayModeImagesGenerations:
 		err = controller.RelayImageHelper(c, relayMode)
@@ -29,39 +34,99 @@ func Relay(c *gin.Context) {
 	case constant.RelayModeAudioTranscription:
 		err = controller.RelayAudioHelper(c, relayMode)
 	default:
-		err = controller.RelayTextHelper(c, relayMode)
+		err = controller.RelayTextHelper(c)
 	}
-	if err != nil {
-		requestId := c.GetString(logger.RequestIdKey)
-		retryTimesStr := c.Query("retry")
-		retryTimes, _ := strconv.Atoi(retryTimesStr)
-		if retryTimesStr == "" {
-			retryTimes = config.RetryTimes
+	return err
+}
+
+func Relay(c *gin.Context) {
+	ctx := c.Request.Context()
+	relayMode := constant.Path2RelayMode(c.Request.URL.Path)
+	if config.DebugEnabled {
+		requestBody, _ := common.GetRequestBody(c)
+		logger.Debugf(ctx, "request body: %s", string(requestBody))
+	}
+	channelId := c.GetInt("channel_id")
+	bizErr := relay(c, relayMode)
+	if bizErr == nil {
+		monitor.Emit(channelId, true)
+		return
+	}
+	lastFailedChannelId := channelId
+	channelName := c.GetString("channel_name")
+	group := c.GetString("group")
+	originalModel := c.GetString("original_model")
+	go processChannelRelayError(ctx, channelId, channelName, bizErr)
+	requestId := c.GetString(logger.RequestIdKey)
+	retryTimes := config.RetryTimes
+	if !shouldRetry(c, bizErr.StatusCode) {
+		logger.Errorf(ctx, "relay error happen, status code is %d, won't retry in this case", bizErr.StatusCode)
+		retryTimes = 0
+	}
+	for i := retryTimes; i > 0; i-- {
+		channel, err := dbmodel.CacheGetRandomSatisfiedChannel(group, originalModel, i != retryTimes)
+		if err != nil {
+			logger.Errorf(ctx, "CacheGetRandomSatisfiedChannel failed: %w", err)
+			break
 		}
-		if retryTimes > 0 {
-			c.Redirect(http.StatusTemporaryRedirect, fmt.Sprintf("%s?retry=%d", c.Request.URL.Path, retryTimes-1))
-		} else {
-			if err.StatusCode == http.StatusTooManyRequests {
-				err.Error.Message = "当前分组上游负载已饱和，请稍后再试"
-			}
-			err.Error.Message = helper.MessageWithRequestId(err.Error.Message, requestId)
-			c.JSON(err.StatusCode, gin.H{
-				"error": err.Error,
-			})
+		logger.Infof(ctx, "using channel #%d to retry (remain times %d)", channel.Id, i)
+		if channel.Id == lastFailedChannelId {
+			continue
+		}
+		middleware.SetupContextForSelectedChannel(c, channel, originalModel)
+		requestBody, err := common.GetRequestBody(c)
+		c.Request.Body = io.NopCloser(bytes.NewBuffer(requestBody))
+		bizErr = relay(c, relayMode)
+		if bizErr == nil {
+			return
 		}
 		channelId := c.GetInt("channel_id")
-		logger.Error(c.Request.Context(), fmt.Sprintf("relay error (channel #%d): %s", channelId, err.Message))
-		// https://platform.openai.com/docs/guides/error-codes/api-errors
-		if util.ShouldDisableChannel(&err.Error, err.StatusCode) {
-			channelId := c.GetInt("channel_id")
-			channelName := c.GetString("channel_name")
-			disableChannel(channelId, channelName, err.Message)
+		lastFailedChannelId = channelId
+		channelName := c.GetString("channel_name")
+		go processChannelRelayError(ctx, channelId, channelName, bizErr)
+	}
+	if bizErr != nil {
+		if bizErr.StatusCode == http.StatusTooManyRequests {
+			bizErr.Error.Message = "当前分组上游负载已饱和，请稍后再试"
 		}
+		bizErr.Error.Message = helper.MessageWithRequestId(bizErr.Error.Message, requestId)
+		c.JSON(bizErr.StatusCode, gin.H{
+			"error": bizErr.Error,
+		})
+	}
+}
+
+func shouldRetry(c *gin.Context, statusCode int) bool {
+	if _, ok := c.Get("specific_channel_id"); ok {
+		return false
+	}
+	if statusCode == http.StatusTooManyRequests {
+		return true
+	}
+	if statusCode/100 == 5 {
+		return true
+	}
+	if statusCode == http.StatusBadRequest {
+		return false
+	}
+	if statusCode/100 == 2 {
+		return false
+	}
+	return true
+}
+
+func processChannelRelayError(ctx context.Context, channelId int, channelName string, err *model.ErrorWithStatusCode) {
+	logger.Errorf(ctx, "relay error (channel #%d): %s", channelId, err.Message)
+	// https://platform.openai.com/docs/guides/error-codes/api-errors
+	if util.ShouldDisableChannel(&err.Error, err.StatusCode) {
+		monitor.DisableChannel(channelId, channelName, err.Message)
+	} else {
+		monitor.Emit(channelId, false)
 	}
 }
 
 func RelayNotImplemented(c *gin.Context) {
-	err := openai.Error{
+	err := model.Error{
 		Message: "API not implemented",
 		Type:    "one_api_error",
 		Param:   "",
@@ -73,7 +138,7 @@ func RelayNotImplemented(c *gin.Context) {
 }
 
 func RelayNotFound(c *gin.Context) {
-	err := openai.Error{
+	err := model.Error{
 		Message: fmt.Sprintf("Invalid URL (%s %s)", c.Request.Method, c.Request.URL.Path),
 		Type:    "invalid_request_error",
 		Param:   "",

controller/token.go

@@ -2,11 +2,11 @@ package controller
 
 import (
 	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/helper"
+	"github.com/songquanpeng/one-api/model"
 	"net/http"
-	"one-api/common"
-	"one-api/common/config"
-	"one-api/common/helper"
-	"one-api/model"
 	"strconv"
 )
 

controller/user.go

@@ -3,11 +3,11 @@ package controller
 import (
 	"encoding/json"
 	"fmt"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/helper"
+	"github.com/songquanpeng/one-api/model"
 	"net/http"
-	"one-api/common"
-	"one-api/common/config"
-	"one-api/common/helper"
-	"one-api/model"
 	"strconv"
 	"time"
 

controller/wechat.go

@@ -5,10 +5,10 @@ import (
 	"errors"
 	"fmt"
 	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/model"
 	"net/http"
-	"one-api/common"
-	"one-api/common/config"
-	"one-api/model"
 	"strconv"
 	"time"
 )

go.mod

@@ -1,4 +1,4 @@
-module one-api
+module github.com/songquanpeng/one-api
 
 // +heroku goVersion go1.18
 go 1.18

i18n/en.json

@@ -456,6 +456,7 @@
   "已绑定的邮箱账户": "Email Account Bound",
   "用户信息更新成功！": "User information updated successfully!",
   "模型倍率 %.2f，分组倍率 %.2f": "model rate %.2f, group rate %.2f",
+  "模型倍率 %.2f，分组倍率 %.2f，补全倍率 %.2f": "model rate %.2f, group rate %.2f, completion rate %.2f",
   "使用明细（总消耗额度：{renderQuota(stat.quota)}）": "Usage Details (Total Consumption Quota: {renderQuota(stat.quota)})",
   "用户名称": "User Name",
   "令牌名称": "Token Name",

main.go

@@ -6,14 +6,14 @@ import (
 	"github.com/gin-contrib/sessions"
 	"github.com/gin-contrib/sessions/cookie"
 	"github.com/gin-gonic/gin"
-	"one-api/common"
-	"one-api/common/config"
-	"one-api/common/logger"
-	"one-api/controller"
-	"one-api/middleware"
-	"one-api/model"
-	"one-api/relay/channel/openai"
-	"one-api/router"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/controller"
+	"github.com/songquanpeng/one-api/middleware"
+	"github.com/songquanpeng/one-api/model"
+	"github.com/songquanpeng/one-api/relay/channel/openai"
+	"github.com/songquanpeng/one-api/router"
 	"os"
 	"strconv"
 )
@@ -64,13 +64,6 @@ func main() {
 		go model.SyncOptions(config.SyncFrequency)
 		go model.SyncChannelCache(config.SyncFrequency)
 	}
-	if os.Getenv("CHANNEL_UPDATE_FREQUENCY") != "" {
-		frequency, err := strconv.Atoi(os.Getenv("CHANNEL_UPDATE_FREQUENCY"))
-		if err != nil {
-			logger.FatalLog("failed to parse CHANNEL_UPDATE_FREQUENCY: " + err.Error())
-		}
-		go controller.AutomaticallyUpdateChannels(frequency)
-	}
 	if os.Getenv("CHANNEL_TEST_FREQUENCY") != "" {
 		frequency, err := strconv.Atoi(os.Getenv("CHANNEL_TEST_FREQUENCY"))
 		if err != nil {
@@ -83,6 +76,9 @@ func main() {
 		logger.SysLog("batch update enabled with interval " + strconv.Itoa(config.BatchUpdateInterval) + "s")
 		model.InitBatchUpdater()
 	}
+	if config.EnableMetric {
+		logger.SysLog("metric enabled, will disable channel if too much request failed")
+	}
 	openai.InitTokenEncoders()
 
 	// Initialize HTTP server

middleware/auth.go

@@ -3,9 +3,10 @@ package middleware
 import (
 	"github.com/gin-contrib/sessions"
 	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/blacklist"
+	"github.com/songquanpeng/one-api/model"
 	"net/http"
-	"one-api/common"
-	"one-api/model"
 	"strings"
 )
 
@@ -42,11 +43,14 @@ func authHelper(c *gin.Context, minRole int) {
 			return
 		}
 	}
-	if status.(int) == common.UserStatusDisabled {
+	if status.(int) == common.UserStatusDisabled || blacklist.IsUserBanned(id.(int)) {
 		c.JSON(http.StatusOK, gin.H{
 			"success": false,
 			"message": "用户已被封禁",
 		})
+		session := sessions.Default(c)
+		session.Clear()
+		_ = session.Save()
 		c.Abort()
 		return
 	}
@@ -99,7 +103,7 @@ func TokenAuth() func(c *gin.Context) {
 			abortWithMessage(c, http.StatusInternalServerError, err.Error())
 			return
 		}
-		if !userEnabled {
+		if !userEnabled || blacklist.IsUserBanned(token.UserId) {
 			abortWithMessage(c, http.StatusForbidden, "用户已被封禁")
 			return
 		}
@@ -108,7 +112,7 @@ func TokenAuth() func(c *gin.Context) {
 		c.Set("token_name", token.Name)
 		if len(parts) > 1 {
 			if model.IsAdmin(token.UserId) {
-				c.Set("channelId", parts[1])
+				c.Set("specific_channel_id", parts[1])
 			} else {
 				abortWithMessage(c, http.StatusForbidden, "普通用户不支持指定渠道")
 				return

middleware/distributor.go

@@ -2,10 +2,10 @@ package middleware
 
 import (
 	"fmt"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/model"
 	"net/http"
-	"one-api/common"
-	"one-api/common/logger"
-	"one-api/model"
 	"strconv"
 	"strings"
 
@@ -21,8 +21,9 @@ func Distribute() func(c *gin.Context) {
 		userId := c.GetInt("id")
 		userGroup, _ := model.CacheGetUserGroup(userId)
 		c.Set("group", userGroup)
+		var requestModel string
 		var channel *model.Channel
-		channelId, ok := c.Get("channelId")
+		channelId, ok := c.Get("specific_channel_id")
 		if ok {
 			id, err := strconv.Atoi(channelId.(string))
 			if err != nil {
@@ -66,7 +67,8 @@ func Distribute() func(c *gin.Context) {
 					modelRequest.Model = "whisper-1"
 				}
 			}
-			channel, err = model.CacheGetRandomSatisfiedChannel(userGroup, modelRequest.Model)
+			requestModel = modelRequest.Model
+			channel, err = model.CacheGetRandomSatisfiedChannel(userGroup, modelRequest.Model, false)
 			if err != nil {
 				message := fmt.Sprintf("当前分组 %s 下对于模型 %s 无可用渠道", userGroup, modelRequest.Model)
 				if channel != nil {
@@ -77,24 +79,34 @@ func Distribute() func(c *gin.Context) {
 				return
 			}
 		}
-		c.Set("channel", channel.Type)
-		c.Set("channel_id", channel.Id)
-		c.Set("channel_name", channel.Name)
-		c.Set("model_mapping", channel.GetModelMapping())
-		c.Request.Header.Set("Authorization", fmt.Sprintf("Bearer %s", channel.Key))
-		c.Set("base_url", channel.GetBaseURL())
-		switch channel.Type {
-		case common.ChannelTypeAzure:
-			c.Set("api_version", channel.Other)
-		case common.ChannelTypeXunfei:
-			c.Set("api_version", channel.Other)
-		case common.ChannelTypeGemini:
-			c.Set("api_version", channel.Other)
-		case common.ChannelTypeAIProxyLibrary:
-			c.Set("library_id", channel.Other)
-		case common.ChannelTypeAli:
-			c.Set("plugin", channel.Other)
-		}
+		SetupContextForSelectedChannel(c, channel, requestModel)
 		c.Next()
 	}
 }
+
+func SetupContextForSelectedChannel(c *gin.Context, channel *model.Channel, modelName string) {
+	c.Set("channel", channel.Type)
+	c.Set("channel_id", channel.Id)
+	c.Set("channel_name", channel.Name)
+	c.Set("model_mapping", channel.GetModelMapping())
+	c.Set("original_model", modelName) // for retry
+	c.Request.Header.Set("Authorization", fmt.Sprintf("Bearer %s", channel.Key))
+	c.Set("base_url", channel.GetBaseURL())
+	// this is for backward compatibility
+	switch channel.Type {
+	case common.ChannelTypeAzure:
+		c.Set(common.ConfigKeyAPIVersion, channel.Other)
+	case common.ChannelTypeXunfei:
+		c.Set(common.ConfigKeyAPIVersion, channel.Other)
+	case common.ChannelTypeGemini:
+		c.Set(common.ConfigKeyAPIVersion, channel.Other)
+	case common.ChannelTypeAIProxyLibrary:
+		c.Set(common.ConfigKeyLibraryID, channel.Other)
+	case common.ChannelTypeAli:
+		c.Set(common.ConfigKeyPlugin, channel.Other)
+	}
+	cfg, _ := channel.LoadConfig()
+	for k, v := range cfg {
+		c.Set(common.ConfigKeyPrefix+k, v)
+	}
+}

middleware/logger.go

@@ -3,7 +3,7 @@ package middleware
 import (
 	"fmt"
 	"github.com/gin-gonic/gin"
-	"one-api/common/logger"
+	"github.com/songquanpeng/one-api/common/logger"
 )
 
 func SetUpLogger(server *gin.Engine) {

middleware/rate-limit.go

@@ -4,9 +4,9 @@ import (
 	"context"
 	"fmt"
 	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/config"
 	"net/http"
-	"one-api/common"
-	"one-api/common/config"
 	"time"
 )
 

middleware/recover.go

@@ -3,8 +3,9 @@ package middleware
 import (
 	"fmt"
 	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/logger"
 	"net/http"
-	"one-api/common/logger"
 	"runtime/debug"
 )
 
@@ -12,11 +13,15 @@ func RelayPanicRecover() gin.HandlerFunc {
 	return func(c *gin.Context) {
 		defer func() {
 			if err := recover(); err != nil {
-				logger.SysError(fmt.Sprintf("panic detected: %v", err))
-				logger.SysError(fmt.Sprintf("stacktrace from panic: %s", string(debug.Stack())))
+				ctx := c.Request.Context()
+				logger.Errorf(ctx, fmt.Sprintf("panic detected: %v", err))
+				logger.Errorf(ctx, fmt.Sprintf("stacktrace from panic: %s", string(debug.Stack())))
+				logger.Errorf(ctx, fmt.Sprintf("request: %s %s", c.Request.Method, c.Request.URL.Path))
+				body, _ := common.GetRequestBody(c)
+				logger.Errorf(ctx, fmt.Sprintf("request body: %s", string(body)))
 				c.JSON(http.StatusInternalServerError, gin.H{
 					"error": gin.H{
-						"message": fmt.Sprintf("Panic detected, error: %v. Please submit a issue here: https://github.com/songquanpeng/one-api", err),
+						"message": fmt.Sprintf("Panic detected, error: %v. Please submit an issue with the related log here: https://github.com/songquanpeng/one-api", err),
 						"type":    "one_api_panic",
 					},
 				})

middleware/request-id.go

@@ -3,13 +3,13 @@ package middleware
 import (
 	"context"
 	"github.com/gin-gonic/gin"
-	"one-api/common/helper"
-	"one-api/common/logger"
+	"github.com/songquanpeng/one-api/common/helper"
+	"github.com/songquanpeng/one-api/common/logger"
 )
 
 func RequestId() func(c *gin.Context) {
 	return func(c *gin.Context) {
-		id := helper.GetTimeString() + helper.GetRandomString(8)
+		id := helper.GenRequestID()
 		c.Set(logger.RequestIdKey, id)
 		ctx := context.WithValue(c.Request.Context(), logger.RequestIdKey, id)
 		c.Request = c.Request.WithContext(ctx)

middleware/turnstile-check.go

@@ -4,10 +4,10 @@ import (
 	"encoding/json"
 	"github.com/gin-contrib/sessions"
 	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/logger"
 	"net/http"
 	"net/url"
-	"one-api/common/config"
-	"one-api/common/logger"
 )
 
 type turnstileCheckResponse struct {

middleware/utils.go

@@ -2,8 +2,8 @@ package middleware
 
 import (
 	"github.com/gin-gonic/gin"
-	"one-api/common/helper"
-	"one-api/common/logger"
+	"github.com/songquanpeng/one-api/common/helper"
+	"github.com/songquanpeng/one-api/common/logger"
 )
 
 func abortWithMessage(c *gin.Context, statusCode int, message string) {

model/ability.go

@@ -1,7 +1,7 @@
 package model
 
 import (
-	"one-api/common"
+	"github.com/songquanpeng/one-api/common"
 	"strings"
 )
 

model/cache.go

@@ -1,13 +1,14 @@
 package model
 
 import (
+	"context"
 	"encoding/json"
 	"errors"
 	"fmt"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/logger"
 	"math/rand"
-	"one-api/common"
-	"one-api/common/config"
-	"one-api/common/logger"
 	"sort"
 	"strconv"
 	"strings"
@@ -70,31 +71,42 @@ func CacheGetUserGroup(id int) (group string, err error) {
 	return group, err
 }
 
-func CacheGetUserQuota(id int) (quota int, err error) {
+func fetchAndUpdateUserQuota(ctx context.Context, id int) (quota int64, err error) {
+	quota, err = GetUserQuota(id)
+	if err != nil {
+		return 0, err
+	}
+	err = common.RedisSet(fmt.Sprintf("user_quota:%d", id), fmt.Sprintf("%d", quota), time.Duration(UserId2QuotaCacheSeconds)*time.Second)
+	if err != nil {
+		logger.Error(ctx, "Redis set user quota error: "+err.Error())
+	}
+	return
+}
+
+func CacheGetUserQuota(ctx context.Context, id int) (quota int64, err error) {
 	if !common.RedisEnabled {
 		return GetUserQuota(id)
 	}
 	quotaString, err := common.RedisGet(fmt.Sprintf("user_quota:%d", id))
 	if err != nil {
-		quota, err = GetUserQuota(id)
-		if err != nil {
-			return 0, err
-		}
-		err = common.RedisSet(fmt.Sprintf("user_quota:%d", id), fmt.Sprintf("%d", quota), time.Duration(UserId2QuotaCacheSeconds)*time.Second)
-		if err != nil {
-			logger.SysError("Redis set user quota error: " + err.Error())
-		}
-		return quota, err
+		return fetchAndUpdateUserQuota(ctx, id)
 	}
-	quota, err = strconv.Atoi(quotaString)
-	return quota, err
+	quota, err = strconv.ParseInt(quotaString, 10, 64)
+	if err != nil {
+		return 0, nil
+	}
+	if quota <= config.PreConsumedQuota { // when user's quota is less than pre-consumed quota, we need to fetch from db
+		logger.Infof(ctx, "user %d's cached quota is too low: %d, refreshing from db", quota, id)
+		return fetchAndUpdateUserQuota(ctx, id)
+	}
+	return quota, nil
 }
 
-func CacheUpdateUserQuota(id int) error {
+func CacheUpdateUserQuota(ctx context.Context, id int) error {
 	if !common.RedisEnabled {
 		return nil
 	}
-	quota, err := GetUserQuota(id)
+	quota, err := CacheGetUserQuota(ctx, id)
 	if err != nil {
 		return err
 	}
@@ -102,7 +114,7 @@ func CacheUpdateUserQuota(id int) error {
 	return err
 }
 
-func CacheDecreaseUserQuota(id int, quota int) error {
+func CacheDecreaseUserQuota(id int, quota int64) error {
 	if !common.RedisEnabled {
 		return nil
 	}
@@ -191,7 +203,7 @@ func SyncChannelCache(frequency int) {
 	}
 }
 
-func CacheGetRandomSatisfiedChannel(group string, model string) (*Channel, error) {
+func CacheGetRandomSatisfiedChannel(group string, model string, ignoreFirstPriority bool) (*Channel, error) {
 	if !config.MemoryCacheEnabled {
 		return GetRandomSatisfiedChannel(group, model)
 	}
@@ -213,5 +225,10 @@ func CacheGetRandomSatisfiedChannel(group string, model string) (*Channel, error
 		}
 	}
 	idx := rand.Intn(endIdx)
+	if ignoreFirstPriority {
+		if endIdx < len(channels) { // which means there are more than one priority
+			idx = common.RandRange(endIdx, len(channels))
+		}
+	}
 	return channels[idx], nil
 }

model/channel.go

@@ -3,17 +3,17 @@ package model
 import (
 	"encoding/json"
 	"fmt"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/helper"
+	"github.com/songquanpeng/one-api/common/logger"
 	"gorm.io/gorm"
-	"one-api/common"
-	"one-api/common/config"
-	"one-api/common/helper"
-	"one-api/common/logger"
 )
 
 type Channel struct {
 	Id                 int     `json:"id"`
 	Type               int     `json:"type" gorm:"default:0"`
-	Key                string  `json:"key" gorm:"not null;index"`
+	Key                string  `json:"key" gorm:"type:text"`
 	Status             int     `json:"status" gorm:"default:1"`
 	Name               string  `json:"name" gorm:"index"`
 	Weight             *uint   `json:"weight" gorm:"default:0"`
@@ -21,7 +21,7 @@ type Channel struct {
 	TestTime           int64   `json:"test_time" gorm:"bigint"`
 	ResponseTime       int     `json:"response_time"` // in milliseconds
 	BaseURL            *string `json:"base_url" gorm:"column:base_url;default:''"`
-	Other              string  `json:"other"`
+	Other              string  `json:"other"`   // DEPRECATED: please save config to field Config
 	Balance            float64 `json:"balance"` // in USD
 	BalanceUpdatedTime int64   `json:"balance_updated_time" gorm:"bigint"`
 	Models             string  `json:"models"`
@@ -29,25 +29,25 @@ type Channel struct {
 	UsedQuota          int64   `json:"used_quota" gorm:"bigint;default:0"`
 	ModelMapping       *string `json:"model_mapping" gorm:"type:varchar(1024);default:''"`
 	Priority           *int64  `json:"priority" gorm:"bigint;default:0"`
+	Config             string  `json:"config"`
 }
 
-func GetAllChannels(startIdx int, num int, selectAll bool) ([]*Channel, error) {
+func GetAllChannels(startIdx int, num int, scope string) ([]*Channel, error) {
 	var channels []*Channel
 	var err error
-	if selectAll {
+	switch scope {
+	case "all":
 		err = DB.Order("id desc").Find(&channels).Error
-	} else {
+	case "disabled":
+		err = DB.Order("id desc").Where("status = ? or status = ?", common.ChannelStatusAutoDisabled, common.ChannelStatusManuallyDisabled).Find(&channels).Error
+	default:
 		err = DB.Order("id desc").Limit(num).Offset(startIdx).Omit("key").Find(&channels).Error
 	}
 	return channels, err
 }
 
 func SearchChannels(keyword string) (channels []*Channel, err error) {
-	keyCol := "`key`"
-	if common.UsingPostgreSQL {
-		keyCol = `"key"`
-	}
-	err = DB.Omit("key").Where("id = ? or name LIKE ? or "+keyCol+" = ?", helper.String2Int(keyword), keyword+"%", keyword).Find(&channels).Error
+	err = DB.Omit("key").Where("id = ? or name LIKE ?", helper.String2Int(keyword), keyword+"%").Find(&channels).Error
 	return channels, err
 }
 
@@ -155,6 +155,18 @@ func (channel *Channel) Delete() error {
 	return err
 }
 
+func (channel *Channel) LoadConfig() (map[string]string, error) {
+	if channel.Config == "" {
+		return nil, nil
+	}
+	cfg := make(map[string]string)
+	err := json.Unmarshal([]byte(channel.Config), &cfg)
+	if err != nil {
+		return nil, err
+	}
+	return cfg, nil
+}
+
 func UpdateChannelStatusById(id int, status int) {
 	err := UpdateAbilityStatus(id, status == common.ChannelStatusEnabled)
 	if err != nil {
@@ -166,7 +178,7 @@ func UpdateChannelStatusById(id int, status int) {
 	}
 }
 
-func UpdateChannelUsedQuota(id int, quota int) {
+func UpdateChannelUsedQuota(id int, quota int64) {
 	if config.BatchUpdateEnabled {
 		addNewRecord(BatchUpdateTypeChannelUsedQuota, id, quota)
 		return
@@ -174,7 +186,7 @@ func UpdateChannelUsedQuota(id int, quota int) {
 	updateChannelUsedQuota(id, quota)
 }
 
-func updateChannelUsedQuota(id int, quota int) {
+func updateChannelUsedQuota(id int, quota int64) {
 	err := DB.Model(&Channel{}).Where("id = ?", id).Update("used_quota", gorm.Expr("used_quota + ?", quota)).Error
 	if err != nil {
 		logger.SysError("failed to update channel used quota: " + err.Error())

model/log.go

@@ -3,10 +3,10 @@ package model
 import (
 	"context"
 	"fmt"
-	"one-api/common"
-	"one-api/common/config"
-	"one-api/common/helper"
-	"one-api/common/logger"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/helper"
+	"github.com/songquanpeng/one-api/common/logger"
 
 	"gorm.io/gorm"
 )
@@ -51,7 +51,7 @@ func RecordLog(userId int, logType int, content string) {
 	}
 }
 
-func RecordConsumeLog(ctx context.Context, userId int, channelId int, promptTokens int, completionTokens int, modelName string, tokenName string, quota int, content string) {
+func RecordConsumeLog(ctx context.Context, userId int, channelId int, promptTokens int, completionTokens int, modelName string, tokenName string, quota int64, content string) {
 	logger.Info(ctx, fmt.Sprintf("record consume log: userId=%d, channelId=%d, promptTokens=%d, completionTokens=%d, modelName=%s, tokenName=%s, quota=%d, content=%s", userId, channelId, promptTokens, completionTokens, modelName, tokenName, quota, content))
 	if !config.LogConsumeEnabled {
 		return
@@ -66,7 +66,7 @@ func RecordConsumeLog(ctx context.Context, userId int, channelId int, promptToke
 		CompletionTokens: completionTokens,
 		TokenName:        tokenName,
 		ModelName:        modelName,
-		Quota:            quota,
+		Quota:            int(quota),
 		ChannelId:        channelId,
 	}
 	err := DB.Create(log).Error
@@ -137,7 +137,7 @@ func SearchUserLogs(userId int, keyword string) (logs []*Log, err error) {
 	return logs, err
 }
 
-func SumUsedQuota(logType int, startTimestamp int64, endTimestamp int64, modelName string, username string, tokenName string, channel int) (quota int) {
+func SumUsedQuota(logType int, startTimestamp int64, endTimestamp int64, modelName string, username string, tokenName string, channel int) (quota int64) {
 	tx := DB.Table("logs").Select("ifnull(sum(quota),0)")
 	if username != "" {
 		tx = tx.Where("username = ?", username)

model/main.go

@@ -2,14 +2,15 @@ package model
 
 import (
 	"fmt"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/env"
+	"github.com/songquanpeng/one-api/common/helper"
+	"github.com/songquanpeng/one-api/common/logger"
 	"gorm.io/driver/mysql"
 	"gorm.io/driver/postgres"
 	"gorm.io/driver/sqlite"
 	"gorm.io/gorm"
-	"one-api/common"
-	"one-api/common/config"
-	"one-api/common/helper"
-	"one-api/common/logger"
 	"os"
 	"strings"
 	"time"
@@ -56,6 +57,7 @@ func chooseDB() (*gorm.DB, error) {
 		}
 		// Use MySQL
 		logger.SysLog("using MySQL as database")
+		common.UsingMySQL = true
 		return gorm.Open(mysql.Open(dsn), &gorm.Config{
 			PrepareStmt: true, // precompile SQL
 		})
@@ -72,7 +74,7 @@ func chooseDB() (*gorm.DB, error) {
 func InitDB() (err error) {
 	db, err := chooseDB()
 	if err == nil {
-		if config.DebugEnabled {
+		if config.DebugSQLEnabled {
 			db = db.Debug()
 		}
 		DB = db
@@ -80,13 +82,16 @@ func InitDB() (err error) {
 		if err != nil {
 			return err
 		}
-		sqlDB.SetMaxIdleConns(helper.GetOrDefaultEnvInt("SQL_MAX_IDLE_CONNS", 100))
-		sqlDB.SetMaxOpenConns(helper.GetOrDefaultEnvInt("SQL_MAX_OPEN_CONNS", 1000))
-		sqlDB.SetConnMaxLifetime(time.Second * time.Duration(helper.GetOrDefaultEnvInt("SQL_MAX_LIFETIME", 60)))
+		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 {
 			return nil
 		}
+		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")
 		err = db.AutoMigrate(&Channel{})
 		if err != nil {

model/option.go

@@ -1,9 +1,9 @@
 package model
 
 import (
-	"one-api/common"
-	"one-api/common/config"
-	"one-api/common/logger"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/logger"
 	"strconv"
 	"strings"
 	"time"
@@ -57,13 +57,15 @@ func InitOptionMap() {
 	config.OptionMap["WeChatServerAddress"] = ""
 	config.OptionMap["WeChatServerToken"] = ""
 	config.OptionMap["WeChatAccountQRCodeImageURL"] = ""
+	config.OptionMap["MessagePusherAddress"] = ""
+	config.OptionMap["MessagePusherToken"] = ""
 	config.OptionMap["TurnstileSiteKey"] = ""
 	config.OptionMap["TurnstileSecretKey"] = ""
-	config.OptionMap["QuotaForNewUser"] = strconv.Itoa(config.QuotaForNewUser)
-	config.OptionMap["QuotaForInviter"] = strconv.Itoa(config.QuotaForInviter)
-	config.OptionMap["QuotaForInvitee"] = strconv.Itoa(config.QuotaForInvitee)
-	config.OptionMap["QuotaRemindThreshold"] = strconv.Itoa(config.QuotaRemindThreshold)
-	config.OptionMap["PreConsumedQuota"] = strconv.Itoa(config.PreConsumedQuota)
+	config.OptionMap["QuotaForNewUser"] = strconv.FormatInt(config.QuotaForNewUser, 10)
+	config.OptionMap["QuotaForInviter"] = strconv.FormatInt(config.QuotaForInviter, 10)
+	config.OptionMap["QuotaForInvitee"] = strconv.FormatInt(config.QuotaForInvitee, 10)
+	config.OptionMap["QuotaRemindThreshold"] = strconv.FormatInt(config.QuotaRemindThreshold, 10)
+	config.OptionMap["PreConsumedQuota"] = strconv.FormatInt(config.PreConsumedQuota, 10)
 	config.OptionMap["ModelRatio"] = common.ModelRatio2JSONString()
 	config.OptionMap["GroupRatio"] = common.GroupRatio2JSONString()
 	config.OptionMap["CompletionRatio"] = common.CompletionRatio2JSONString()
@@ -79,6 +81,9 @@ func InitOptionMap() {
 func loadOptionsFromDatabase() {
 	options, _ := AllOption()
 	for _, option := range options {
+		if option.Key == "ModelRatio" {
+			option.Value = common.AddNewMissingRatio(option.Value)
+		}
 		err := updateOptionMap(option.Key, option.Value)
 		if err != nil {
 			logger.SysError("failed to update option map: " + err.Error())
@@ -179,20 +184,24 @@ func updateOptionMap(key string, value string) (err error) {
 		config.WeChatServerToken = value
 	case "WeChatAccountQRCodeImageURL":
 		config.WeChatAccountQRCodeImageURL = value
+	case "MessagePusherAddress":
+		config.MessagePusherAddress = value
+	case "MessagePusherToken":
+		config.MessagePusherToken = value
 	case "TurnstileSiteKey":
 		config.TurnstileSiteKey = value
 	case "TurnstileSecretKey":
 		config.TurnstileSecretKey = value
 	case "QuotaForNewUser":
-		config.QuotaForNewUser, _ = strconv.Atoi(value)
+		config.QuotaForNewUser, _ = strconv.ParseInt(value, 10, 64)
 	case "QuotaForInviter":
-		config.QuotaForInviter, _ = strconv.Atoi(value)
+		config.QuotaForInviter, _ = strconv.ParseInt(value, 10, 64)
 	case "QuotaForInvitee":
-		config.QuotaForInvitee, _ = strconv.Atoi(value)
+		config.QuotaForInvitee, _ = strconv.ParseInt(value, 10, 64)
 	case "QuotaRemindThreshold":
-		config.QuotaRemindThreshold, _ = strconv.Atoi(value)
+		config.QuotaRemindThreshold, _ = strconv.ParseInt(value, 10, 64)
 	case "PreConsumedQuota":
-		config.PreConsumedQuota, _ = strconv.Atoi(value)
+		config.PreConsumedQuota, _ = strconv.ParseInt(value, 10, 64)
 	case "RetryTimes":
 		config.RetryTimes, _ = strconv.Atoi(value)
 	case "ModelRatio":

model/redemption.go

@@ -3,9 +3,9 @@ package model
 import (
 	"errors"
 	"fmt"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/helper"
 	"gorm.io/gorm"
-	"one-api/common"
-	"one-api/common/helper"
 )
 
 type Redemption struct {
@@ -14,7 +14,7 @@ type Redemption struct {
 	Key          string `json:"key" gorm:"type:char(32);uniqueIndex"`
 	Status       int    `json:"status" gorm:"default:1"`
 	Name         string `json:"name" gorm:"index"`
-	Quota        int    `json:"quota" gorm:"default:100"`
+	Quota        int64  `json:"quota" gorm:"default:100"`
 	CreatedTime  int64  `json:"created_time" gorm:"bigint"`
 	RedeemedTime int64  `json:"redeemed_time" gorm:"bigint"`
 	Count        int    `json:"count" gorm:"-:all"` // only for api request
@@ -42,7 +42,7 @@ func GetRedemptionById(id int) (*Redemption, error) {
 	return &redemption, err
 }
 
-func Redeem(key string, userId int) (quota int, err error) {
+func Redeem(key string, userId int) (quota int64, err error) {
 	if key == "" {
 		return 0, errors.New("未提供兑换码")
 	}

model/token.go

@@ -3,11 +3,12 @@ package model
 import (
 	"errors"
 	"fmt"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/helper"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/common/message"
 	"gorm.io/gorm"
-	"one-api/common"
-	"one-api/common/config"
-	"one-api/common/helper"
-	"one-api/common/logger"
 )
 
 type Token struct {
@@ -19,9 +20,9 @@ type Token struct {
 	CreatedTime    int64  `json:"created_time" gorm:"bigint"`
 	AccessedTime   int64  `json:"accessed_time" gorm:"bigint"`
 	ExpiredTime    int64  `json:"expired_time" gorm:"bigint;default:-1"` // -1 means never expired
-	RemainQuota    int    `json:"remain_quota" gorm:"default:0"`
+	RemainQuota    int64  `json:"remain_quota" gorm:"default:0"`
 	UnlimitedQuota bool   `json:"unlimited_quota" gorm:"default:false"`
-	UsedQuota      int    `json:"used_quota" gorm:"default:0"` // used quota
+	UsedQuota      int64  `json:"used_quota" gorm:"default:0"` // used quota
 }
 
 func GetAllUserTokens(userId int, startIdx int, num int) ([]*Token, error) {
@@ -137,7 +138,7 @@ func DeleteTokenById(id int, userId int) (err error) {
 	return token.Delete()
 }
 
-func IncreaseTokenQuota(id int, quota int) (err error) {
+func IncreaseTokenQuota(id int, quota int64) (err error) {
 	if quota < 0 {
 		return errors.New("quota 不能为负数！")
 	}
@@ -148,7 +149,7 @@ func IncreaseTokenQuota(id int, quota int) (err error) {
 	return increaseTokenQuota(id, quota)
 }
 
-func increaseTokenQuota(id int, quota int) (err error) {
+func increaseTokenQuota(id int, quota int64) (err error) {
 	err = DB.Model(&Token{}).Where("id = ?", id).Updates(
 		map[string]interface{}{
 			"remain_quota":  gorm.Expr("remain_quota + ?", quota),
@@ -159,7 +160,7 @@ func increaseTokenQuota(id int, quota int) (err error) {
 	return err
 }
 
-func DecreaseTokenQuota(id int, quota int) (err error) {
+func DecreaseTokenQuota(id int, quota int64) (err error) {
 	if quota < 0 {
 		return errors.New("quota 不能为负数！")
 	}
@@ -170,7 +171,7 @@ func DecreaseTokenQuota(id int, quota int) (err error) {
 	return decreaseTokenQuota(id, quota)
 }
 
-func decreaseTokenQuota(id int, quota int) (err error) {
+func decreaseTokenQuota(id int, quota int64) (err error) {
 	err = DB.Model(&Token{}).Where("id = ?", id).Updates(
 		map[string]interface{}{
 			"remain_quota":  gorm.Expr("remain_quota - ?", quota),
@@ -181,7 +182,7 @@ func decreaseTokenQuota(id int, quota int) (err error) {
 	return err
 }
 
-func PreConsumeTokenQuota(tokenId int, quota int) (err error) {
+func PreConsumeTokenQuota(tokenId int, quota int64) (err error) {
 	if quota < 0 {
 		return errors.New("quota 不能为负数！")
 	}
@@ -213,7 +214,7 @@ func PreConsumeTokenQuota(tokenId int, quota int) (err error) {
 			}
 			if email != "" {
 				topUpLink := fmt.Sprintf("%s/topup", config.ServerAddress)
-				err = common.SendEmail(prompt, email,
+				err = message.SendEmail(prompt, email,
 					fmt.Sprintf("%s，当前剩余额度为 %d，为了不影响您的使用，请及时充值。<br/>充值链接：<a href='%s'>%s</a>", prompt, userQuota, topUpLink, topUpLink))
 				if err != nil {
 					logger.SysError("failed to send email" + err.Error())
@@ -231,7 +232,7 @@ func PreConsumeTokenQuota(tokenId int, quota int) (err error) {
 	return err
 }
 
-func PostConsumeTokenQuota(tokenId int, quota int) (err error) {
+func PostConsumeTokenQuota(tokenId int, quota int64) (err error) {
 	token, err := GetTokenById(tokenId)
 	if quota > 0 {
 		err = DecreaseUserQuota(token.UserId, quota)

model/user.go

@@ -3,11 +3,12 @@ package model
 import (
 	"errors"
 	"fmt"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/blacklist"
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/helper"
+	"github.com/songquanpeng/one-api/common/logger"
 	"gorm.io/gorm"
-	"one-api/common"
-	"one-api/common/config"
-	"one-api/common/helper"
-	"one-api/common/logger"
 	"strings"
 )
 
@@ -25,8 +26,8 @@ type User struct {
 	WeChatId         string `json:"wechat_id" gorm:"column:wechat_id;index"`
 	VerificationCode string `json:"verification_code" gorm:"-:all"`                                    // this field is only for Email verification, don't save it to database!
 	AccessToken      string `json:"access_token" gorm:"type:char(32);column:access_token;uniqueIndex"` // this token is for system management
-	Quota            int    `json:"quota" gorm:"type:int;default:0"`
-	UsedQuota        int    `json:"used_quota" gorm:"type:int;default:0;column:used_quota"` // used quota
+	Quota            int64  `json:"quota" gorm:"type:int;default:0"`
+	UsedQuota        int64  `json:"used_quota" gorm:"type:int;default:0;column:used_quota"` // used quota
 	RequestCount     int    `json:"request_count" gorm:"type:int;default:0;"`               // request number
 	Group            string `json:"group" gorm:"type:varchar(32);default:'default'"`
 	AffCode          string `json:"aff_code" gorm:"type:varchar(32);column:aff_code;uniqueIndex"`
@@ -40,7 +41,7 @@ func GetMaxUserId() int {
 }
 
 func GetAllUsers(startIdx int, num int) (users []*User, err error) {
-	err = DB.Order("id desc").Limit(num).Offset(startIdx).Omit("password").Find(&users).Error
+	err = DB.Order("id desc").Limit(num).Offset(startIdx).Omit("password").Where("status != ?", common.UserStatusDeleted).Find(&users).Error
 	return users, err
 }
 
@@ -123,6 +124,11 @@ func (user *User) Update(updatePassword bool) error {
 			return err
 		}
 	}
+	if user.Status == common.UserStatusDisabled {
+		blacklist.BanUser(user.Id)
+	} else if user.Status == common.UserStatusEnabled {
+		blacklist.UnbanUser(user.Id)
+	}
 	err = DB.Model(user).Updates(user).Error
 	return err
 }
@@ -131,7 +137,10 @@ func (user *User) Delete() error {
 	if user.Id == 0 {
 		return errors.New("id 为空！")
 	}
-	err := DB.Delete(user).Error
+	blacklist.BanUser(user.Id)
+	user.Username = fmt.Sprintf("deleted_%s", helper.GetUUID())
+	user.Status = common.UserStatusDeleted
+	err := DB.Model(user).Updates(user).Error
 	return err
 }
 
@@ -265,12 +274,12 @@ func ValidateAccessToken(token string) (user *User) {
 	return nil
 }
 
-func GetUserQuota(id int) (quota int, err error) {
+func GetUserQuota(id int) (quota int64, err error) {
 	err = DB.Model(&User{}).Where("id = ?", id).Select("quota").Find(&quota).Error
 	return quota, err
 }
 
-func GetUserUsedQuota(id int) (quota int, err error) {
+func GetUserUsedQuota(id int) (quota int64, err error) {
 	err = DB.Model(&User{}).Where("id = ?", id).Select("used_quota").Find(&quota).Error
 	return quota, err
 }
@@ -290,7 +299,7 @@ func GetUserGroup(id int) (group string, err error) {
 	return group, err
 }
 
-func IncreaseUserQuota(id int, quota int) (err error) {
+func IncreaseUserQuota(id int, quota int64) (err error) {
 	if quota < 0 {
 		return errors.New("quota 不能为负数！")
 	}
@@ -301,12 +310,12 @@ func IncreaseUserQuota(id int, quota int) (err error) {
 	return increaseUserQuota(id, quota)
 }
 
-func increaseUserQuota(id int, quota int) (err error) {
+func increaseUserQuota(id int, quota int64) (err error) {
 	err = DB.Model(&User{}).Where("id = ?", id).Update("quota", gorm.Expr("quota + ?", quota)).Error
 	return err
 }
 
-func DecreaseUserQuota(id int, quota int) (err error) {
+func DecreaseUserQuota(id int, quota int64) (err error) {
 	if quota < 0 {
 		return errors.New("quota 不能为负数！")
 	}
@@ -317,7 +326,7 @@ func DecreaseUserQuota(id int, quota int) (err error) {
 	return decreaseUserQuota(id, quota)
 }
 
-func decreaseUserQuota(id int, quota int) (err error) {
+func decreaseUserQuota(id int, quota int64) (err error) {
 	err = DB.Model(&User{}).Where("id = ?", id).Update("quota", gorm.Expr("quota - ?", quota)).Error
 	return err
 }
@@ -327,7 +336,7 @@ func GetRootUserEmail() (email string) {
 	return email
 }
 
-func UpdateUserUsedQuotaAndRequestCount(id int, quota int) {
+func UpdateUserUsedQuotaAndRequestCount(id int, quota int64) {
 	if config.BatchUpdateEnabled {
 		addNewRecord(BatchUpdateTypeUsedQuota, id, quota)
 		addNewRecord(BatchUpdateTypeRequestCount, id, 1)
@@ -336,7 +345,7 @@ func UpdateUserUsedQuotaAndRequestCount(id int, quota int) {
 	updateUserUsedQuotaAndRequestCount(id, quota, 1)
 }
 
-func updateUserUsedQuotaAndRequestCount(id int, quota int, count int) {
+func updateUserUsedQuotaAndRequestCount(id int, quota int64, count int) {
 	err := DB.Model(&User{}).Where("id = ?", id).Updates(
 		map[string]interface{}{
 			"used_quota":    gorm.Expr("used_quota + ?", quota),
@@ -348,7 +357,7 @@ func updateUserUsedQuotaAndRequestCount(id int, quota int, count int) {
 	}
 }
 
-func updateUserUsedQuota(id int, quota int) {
+func updateUserUsedQuota(id int, quota int64) {
 	err := DB.Model(&User{}).Where("id = ?", id).Updates(
 		map[string]interface{}{
 			"used_quota": gorm.Expr("used_quota + ?", quota),

model/utils.go

@@ -1,8 +1,8 @@
 package model
 
 import (
-	"one-api/common/config"
-	"one-api/common/logger"
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/logger"
 	"sync"
 	"time"
 )
@@ -16,12 +16,12 @@ const (
 	BatchUpdateTypeCount // if you add a new type, you need to add a new map and a new lock
 )
 
-var batchUpdateStores []map[int]int
+var batchUpdateStores []map[int]int64
 var batchUpdateLocks []sync.Mutex
 
 func init() {
 	for i := 0; i < BatchUpdateTypeCount; i++ {
-		batchUpdateStores = append(batchUpdateStores, make(map[int]int))
+		batchUpdateStores = append(batchUpdateStores, make(map[int]int64))
 		batchUpdateLocks = append(batchUpdateLocks, sync.Mutex{})
 	}
 }
@@ -35,7 +35,7 @@ func InitBatchUpdater() {
 	}()
 }
 
-func addNewRecord(type_ int, id int, value int) {
+func addNewRecord(type_ int, id int, value int64) {
 	batchUpdateLocks[type_].Lock()
 	defer batchUpdateLocks[type_].Unlock()
 	if _, ok := batchUpdateStores[type_][id]; !ok {
@@ -50,7 +50,7 @@ func batchUpdate() {
 	for i := 0; i < BatchUpdateTypeCount; i++ {
 		batchUpdateLocks[i].Lock()
 		store := batchUpdateStores[i]
-		batchUpdateStores[i] = make(map[int]int)
+		batchUpdateStores[i] = make(map[int]int64)
 		batchUpdateLocks[i].Unlock()
 		// TODO: maybe we can combine updates with same key?
 		for key, value := range store {
@@ -68,7 +68,7 @@ func batchUpdate() {
 			case BatchUpdateTypeUsedQuota:
 				updateUserUsedQuota(key, value)
 			case BatchUpdateTypeRequestCount:
-				updateUserRequestCount(key, value)
+				updateUserRequestCount(key, int(value))
 			case BatchUpdateTypeChannelUsedQuota:
 				updateChannelUsedQuota(key, value)
 			}

monitor/channel.go

@@ -0,0 +1,55 @@
+package monitor
+
+import (
+	"fmt"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/common/message"
+	"github.com/songquanpeng/one-api/model"
+)
+
+func notifyRootUser(subject string, content string) {
+	if config.MessagePusherAddress != "" {
+		err := message.SendMessage(subject, content, content)
+		if err != nil {
+			logger.SysError(fmt.Sprintf("failed to send message: %s", err.Error()))
+		} else {
+			return
+		}
+	}
+	if config.RootUserEmail == "" {
+		config.RootUserEmail = model.GetRootUserEmail()
+	}
+	err := message.SendEmail(subject, config.RootUserEmail, content)
+	if err != nil {
+		logger.SysError(fmt.Sprintf("failed to send email: %s", err.Error()))
+	}
+}
+
+// DisableChannel disable & notify
+func DisableChannel(channelId int, channelName string, reason string) {
+	model.UpdateChannelStatusById(channelId, common.ChannelStatusAutoDisabled)
+	logger.SysLog(fmt.Sprintf("channel #%d has been disabled: %s", channelId, reason))
+	subject := fmt.Sprintf("通道「%s」（#%d）已被禁用", channelName, channelId)
+	content := fmt.Sprintf("通道「%s」（#%d）已被禁用，原因：%s", channelName, channelId, reason)
+	notifyRootUser(subject, content)
+}
+
+func MetricDisableChannel(channelId int, successRate float64) {
+	model.UpdateChannelStatusById(channelId, common.ChannelStatusAutoDisabled)
+	logger.SysLog(fmt.Sprintf("channel #%d has been disabled due to low success rate: %.2f", channelId, successRate*100))
+	subject := fmt.Sprintf("通道 #%d 已被禁用", channelId)
+	content := fmt.Sprintf("该渠道在最近 %d 次调用中成功率为 %.2f%%，低于阈值 %.2f%%，因此被系统自动禁用。",
+		config.MetricQueueSize, successRate*100, config.MetricSuccessRateThreshold*100)
+	notifyRootUser(subject, content)
+}
+
+// EnableChannel enable & notify
+func EnableChannel(channelId int, channelName string) {
+	model.UpdateChannelStatusById(channelId, common.ChannelStatusEnabled)
+	logger.SysLog(fmt.Sprintf("channel #%d has been enabled", channelId))
+	subject := fmt.Sprintf("通道「%s」（#%d）已被启用", channelName, channelId)
+	content := fmt.Sprintf("通道「%s」（#%d）已被启用", channelName, channelId)
+	notifyRootUser(subject, content)
+}

monitor/metric.go

@@ -0,0 +1,79 @@
+package monitor
+
+import (
+	"github.com/songquanpeng/one-api/common/config"
+)
+
+var store = make(map[int][]bool)
+var metricSuccessChan = make(chan int, config.MetricSuccessChanSize)
+var metricFailChan = make(chan int, config.MetricFailChanSize)
+
+func consumeSuccess(channelId int) {
+	if len(store[channelId]) > config.MetricQueueSize {
+		store[channelId] = store[channelId][1:]
+	}
+	store[channelId] = append(store[channelId], true)
+}
+
+func consumeFail(channelId int) (bool, float64) {
+	if len(store[channelId]) > config.MetricQueueSize {
+		store[channelId] = store[channelId][1:]
+	}
+	store[channelId] = append(store[channelId], false)
+	successCount := 0
+	for _, success := range store[channelId] {
+		if success {
+			successCount++
+		}
+	}
+	successRate := float64(successCount) / float64(len(store[channelId]))
+	if len(store[channelId]) < config.MetricQueueSize {
+		return false, successRate
+	}
+	if successRate < config.MetricSuccessRateThreshold {
+		store[channelId] = make([]bool, 0)
+		return true, successRate
+	}
+	return false, successRate
+}
+
+func metricSuccessConsumer() {
+	for {
+		select {
+		case channelId := <-metricSuccessChan:
+			consumeSuccess(channelId)
+		}
+	}
+}
+
+func metricFailConsumer() {
+	for {
+		select {
+		case channelId := <-metricFailChan:
+			disable, successRate := consumeFail(channelId)
+			if disable {
+				go MetricDisableChannel(channelId, successRate)
+			}
+		}
+	}
+}
+
+func init() {
+	if config.EnableMetric {
+		go metricSuccessConsumer()
+		go metricFailConsumer()
+	}
+}
+
+func Emit(channelId int, success bool) {
+	if !config.EnableMetric {
+		return
+	}
+	go func() {
+		if success {
+			metricSuccessChan <- channelId
+		} else {
+			metricFailChan <- channelId
+		}
+	}()
+}

relay/channel/ai360/constants.go

@@ -0,0 +1,8 @@
+package ai360
+
+var ModelList = []string{
+	"360GPT_S2_V9",
+	"embedding-bert-512-v1",
+	"embedding_s1_v1",
+	"semantic_similarity_s1_v1",
+}

relay/channel/aiproxy/adaptor.go

@@ -1,22 +1,60 @@
 package aiproxy
 
 import (
+	"errors"
+	"fmt"
 	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/relay/channel"
+	"github.com/songquanpeng/one-api/relay/model"
+	"github.com/songquanpeng/one-api/relay/util"
+	"io"
 	"net/http"
-	"one-api/relay/channel/openai"
 )
 
 type Adaptor struct {
 }
 
-func (a *Adaptor) Auth(c *gin.Context) error {
+func (a *Adaptor) Init(meta *util.RelayMeta) {
+
+}
+
+func (a *Adaptor) GetRequestURL(meta *util.RelayMeta) (string, error) {
+	return fmt.Sprintf("%s/api/library/ask", meta.BaseURL), nil
+}
+
+func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *util.RelayMeta) error {
+	channel.SetupCommonRequestHeader(c, req, meta)
+	req.Header.Set("Authorization", "Bearer "+meta.APIKey)
 	return nil
 }
 
-func (a *Adaptor) ConvertRequest(request *openai.GeneralOpenAIRequest) (any, error) {
-	return nil, 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")
+	}
+	aiProxyLibraryRequest := ConvertRequest(*request)
+	aiProxyLibraryRequest.LibraryId = c.GetString(common.ConfigKeyLibraryID)
+	return aiProxyLibraryRequest, nil
 }
 
-func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response) (*openai.ErrorWithStatusCode, *openai.Usage, error) {
-	return nil, nil, nil
+func (a *Adaptor) DoRequest(c *gin.Context, meta *util.RelayMeta, requestBody io.Reader) (*http.Response, error) {
+	return channel.DoRequestHelper(a, c, meta, requestBody)
+}
+
+func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *util.RelayMeta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
+	if meta.IsStream {
+		err, usage = StreamHandler(c, resp)
+	} else {
+		err, usage = Handler(c, resp)
+	}
+	return
+}
+
+func (a *Adaptor) GetModelList() []string {
+	return ModelList
+}
+
+func (a *Adaptor) GetChannelName() string {
+	return "aiproxy"
 }

relay/channel/aiproxy/constants.go

@@ -0,0 +1,9 @@
+package aiproxy
+
+import "github.com/songquanpeng/one-api/relay/channel/openai"
+
+var ModelList = []string{""}
+
+func init() {
+	ModelList = openai.ModelList
+}

relay/channel/aiproxy/main.go

@@ -5,20 +5,21 @@ import (
 	"encoding/json"
 	"fmt"
 	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/helper"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/relay/channel/openai"
+	"github.com/songquanpeng/one-api/relay/constant"
+	"github.com/songquanpeng/one-api/relay/model"
 	"io"
 	"net/http"
-	"one-api/common"
-	"one-api/common/helper"
-	"one-api/common/logger"
-	"one-api/relay/channel/openai"
-	"one-api/relay/constant"
 	"strconv"
 	"strings"
 )
 
 // https://docs.aiproxy.io/dev/library#使用已经定制好的知识库进行对话问答
 
-func ConvertRequest(request openai.GeneralOpenAIRequest) *LibraryRequest {
+func ConvertRequest(request model.GeneralOpenAIRequest) *LibraryRequest {
 	query := ""
 	if len(request.Messages) != 0 {
 		query = request.Messages[len(request.Messages)-1].StringContent()
@@ -45,14 +46,14 @@ func responseAIProxyLibrary2OpenAI(response *LibraryResponse) *openai.TextRespon
 	content := response.Answer + aiProxyDocuments2Markdown(response.Documents)
 	choice := openai.TextResponseChoice{
 		Index: 0,
-		Message: openai.Message{
+		Message: model.Message{
 			Role:    "assistant",
 			Content: content,
 		},
 		FinishReason: "stop",
 	}
 	fullTextResponse := openai.TextResponse{
-		Id:      helper.GetUUID(),
+		Id:      fmt.Sprintf("chatcmpl-%s", helper.GetUUID()),
 		Object:  "chat.completion",
 		Created: helper.GetTimestamp(),
 		Choices: []openai.TextResponseChoice{choice},
@@ -65,7 +66,7 @@ func documentsAIProxyLibrary(documents []LibraryDocument) *openai.ChatCompletion
 	choice.Delta.Content = aiProxyDocuments2Markdown(documents)
 	choice.FinishReason = &constant.StopFinishReason
 	return &openai.ChatCompletionsStreamResponse{
-		Id:      helper.GetUUID(),
+		Id:      fmt.Sprintf("chatcmpl-%s", helper.GetUUID()),
 		Object:  "chat.completion.chunk",
 		Created: helper.GetTimestamp(),
 		Model:   "",
@@ -77,7 +78,7 @@ func streamResponseAIProxyLibrary2OpenAI(response *LibraryStreamResponse) *opena
 	var choice openai.ChatCompletionsStreamResponseChoice
 	choice.Delta.Content = response.Content
 	return &openai.ChatCompletionsStreamResponse{
-		Id:      helper.GetUUID(),
+		Id:      fmt.Sprintf("chatcmpl-%s", helper.GetUUID()),
 		Object:  "chat.completion.chunk",
 		Created: helper.GetTimestamp(),
 		Model:   response.Model,
@@ -85,8 +86,8 @@ func streamResponseAIProxyLibrary2OpenAI(response *LibraryStreamResponse) *opena
 	}
 }
 
-func StreamHandler(c *gin.Context, resp *http.Response) (*openai.ErrorWithStatusCode, *openai.Usage) {
-	var usage openai.Usage
+func StreamHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, *model.Usage) {
+	var usage model.Usage
 	scanner := bufio.NewScanner(resp.Body)
 	scanner.Split(func(data []byte, atEOF bool) (advance int, token []byte, err error) {
 		if atEOF && len(data) == 0 {
@@ -157,7 +158,7 @@ func StreamHandler(c *gin.Context, resp *http.Response) (*openai.ErrorWithStatus
 	return nil, &usage
 }
 
-func Handler(c *gin.Context, resp *http.Response) (*openai.ErrorWithStatusCode, *openai.Usage) {
+func Handler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, *model.Usage) {
 	var AIProxyLibraryResponse LibraryResponse
 	responseBody, err := io.ReadAll(resp.Body)
 	if err != nil {
@@ -172,8 +173,8 @@ func Handler(c *gin.Context, resp *http.Response) (*openai.ErrorWithStatusCode,
 		return openai.ErrorWrapper(err, "unmarshal_response_body_failed", http.StatusInternalServerError), nil
 	}
 	if AIProxyLibraryResponse.ErrCode != 0 {
-		return &openai.ErrorWithStatusCode{
-			Error: openai.Error{
+		return &model.ErrorWithStatusCode{
+			Error: model.Error{
 				Message: AIProxyLibraryResponse.Message,
 				Type:    strconv.Itoa(AIProxyLibraryResponse.ErrCode),
 				Code:    AIProxyLibraryResponse.ErrCode,
@@ -189,5 +190,8 @@ func Handler(c *gin.Context, resp *http.Response) (*openai.ErrorWithStatusCode,
 	c.Writer.Header().Set("Content-Type", "application/json")
 	c.Writer.WriteHeader(resp.StatusCode)
 	_, err = c.Writer.Write(jsonResponse)
+	if err != nil {
+		return openai.ErrorWrapper(err, "write_response_body_failed", http.StatusInternalServerError), nil
+	}
 	return nil, &fullTextResponse.Usage
 }

relay/channel/ali/adaptor.go

@@ -1,22 +1,86 @@
 package ali
 
 import (
+	"errors"
+	"fmt"
 	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/relay/channel"
+	"github.com/songquanpeng/one-api/relay/constant"
+	"github.com/songquanpeng/one-api/relay/model"
+	"github.com/songquanpeng/one-api/relay/util"
+	"io"
 	"net/http"
-	"one-api/relay/channel/openai"
 )
 
+// https://help.aliyun.com/zh/dashscope/developer-reference/api-details
+
 type Adaptor struct {
 }
 
-func (a *Adaptor) Auth(c *gin.Context) error {
+func (a *Adaptor) Init(meta *util.RelayMeta) {
+
+}
+
+func (a *Adaptor) GetRequestURL(meta *util.RelayMeta) (string, error) {
+	fullRequestURL := fmt.Sprintf("%s/api/v1/services/aigc/text-generation/generation", meta.BaseURL)
+	if meta.Mode == constant.RelayModeEmbeddings {
+		fullRequestURL = fmt.Sprintf("%s/api/v1/services/embeddings/text-embedding/text-embedding", meta.BaseURL)
+	}
+	return fullRequestURL, nil
+}
+
+func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *util.RelayMeta) error {
+	channel.SetupCommonRequestHeader(c, req, meta)
+	if meta.IsStream {
+		req.Header.Set("Accept", "text/event-stream")
+	}
+	req.Header.Set("Authorization", "Bearer "+meta.APIKey)
+	if meta.IsStream {
+		req.Header.Set("X-DashScope-SSE", "enable")
+	}
+	if c.GetString(common.ConfigKeyPlugin) != "" {
+		req.Header.Set("X-DashScope-Plugin", c.GetString(common.ConfigKeyPlugin))
+	}
 	return nil
 }
 
-func (a *Adaptor) ConvertRequest(request *openai.GeneralOpenAIRequest) (any, error) {
-	return nil, 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")
+	}
+	switch relayMode {
+	case constant.RelayModeEmbeddings:
+		baiduEmbeddingRequest := ConvertEmbeddingRequest(*request)
+		return baiduEmbeddingRequest, nil
+	default:
+		baiduRequest := ConvertRequest(*request)
+		return baiduRequest, nil
+	}
 }
 
-func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response) (*openai.ErrorWithStatusCode, *openai.Usage, error) {
-	return nil, nil, nil
+func (a *Adaptor) DoRequest(c *gin.Context, meta *util.RelayMeta, requestBody io.Reader) (*http.Response, error) {
+	return channel.DoRequestHelper(a, c, meta, requestBody)
+}
+
+func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *util.RelayMeta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
+	if meta.IsStream {
+		err, usage = StreamHandler(c, resp)
+	} else {
+		switch meta.Mode {
+		case constant.RelayModeEmbeddings:
+			err, usage = EmbeddingHandler(c, resp)
+		default:
+			err, usage = Handler(c, resp)
+		}
+	}
+	return
+}
+
+func (a *Adaptor) GetModelList() []string {
+	return ModelList
+}
+
+func (a *Adaptor) GetChannelName() string {
+	return "ali"
 }

relay/channel/ali/constants.go

@@ -0,0 +1,6 @@
+package ali
+
+var ModelList = []string{
+	"qwen-turbo", "qwen-plus", "qwen-max", "qwen-max-longcontext",
+	"text-embedding-v1",
+}

relay/channel/ali/main.go

@@ -4,12 +4,13 @@ import (
 	"bufio"
 	"encoding/json"
 	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/helper"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/relay/channel/openai"
+	"github.com/songquanpeng/one-api/relay/model"
 	"io"
 	"net/http"
-	"one-api/common"
-	"one-api/common/helper"
-	"one-api/common/logger"
-	"one-api/relay/channel/openai"
 	"strings"
 )
 
@@ -17,7 +18,7 @@ import (
 
 const EnableSearchModelSuffix = "-internet"
 
-func ConvertRequest(request openai.GeneralOpenAIRequest) *ChatRequest {
+func ConvertRequest(request model.GeneralOpenAIRequest) *ChatRequest {
 	messages := make([]Message, 0, len(request.Messages))
 	for i := 0; i < len(request.Messages); i++ {
 		message := request.Messages[i]
@@ -32,6 +33,9 @@ func ConvertRequest(request openai.GeneralOpenAIRequest) *ChatRequest {
 		enableSearch = true
 		aliModel = strings.TrimSuffix(aliModel, EnableSearchModelSuffix)
 	}
+	if request.TopP >= 1 {
+		request.TopP = 0.9999
+	}
 	return &ChatRequest{
 		Model: aliModel,
 		Input: Input{
@@ -40,11 +44,15 @@ func ConvertRequest(request openai.GeneralOpenAIRequest) *ChatRequest {
 		Parameters: Parameters{
 			EnableSearch:      enableSearch,
 			IncrementalOutput: request.Stream,
+			Seed:              uint64(request.Seed),
+			MaxTokens:         request.MaxTokens,
+			Temperature:       request.Temperature,
+			TopP:              request.TopP,
 		},
 	}
 }
 
-func ConvertEmbeddingRequest(request openai.GeneralOpenAIRequest) *EmbeddingRequest {
+func ConvertEmbeddingRequest(request model.GeneralOpenAIRequest) *EmbeddingRequest {
 	return &EmbeddingRequest{
 		Model: "text-embedding-v1",
 		Input: struct {
@@ -55,7 +63,7 @@ func ConvertEmbeddingRequest(request openai.GeneralOpenAIRequest) *EmbeddingRequ
 	}
 }
 
-func EmbeddingHandler(c *gin.Context, resp *http.Response) (*openai.ErrorWithStatusCode, *openai.Usage) {
+func EmbeddingHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, *model.Usage) {
 	var aliResponse EmbeddingResponse
 	err := json.NewDecoder(resp.Body).Decode(&aliResponse)
 	if err != nil {
@@ -68,8 +76,8 @@ func EmbeddingHandler(c *gin.Context, resp *http.Response) (*openai.ErrorWithSta
 	}
 
 	if aliResponse.Code != "" {
-		return &openai.ErrorWithStatusCode{
-			Error: openai.Error{
+		return &model.ErrorWithStatusCode{
+			Error: model.Error{
 				Message: aliResponse.Message,
 				Type:    aliResponse.Code,
 				Param:   aliResponse.RequestId,
@@ -95,7 +103,7 @@ func embeddingResponseAli2OpenAI(response *EmbeddingResponse) *openai.EmbeddingR
 		Object: "list",
 		Data:   make([]openai.EmbeddingResponseItem, 0, len(response.Output.Embeddings)),
 		Model:  "text-embedding-v1",
-		Usage:  openai.Usage{TotalTokens: response.Usage.TotalTokens},
+		Usage:  model.Usage{TotalTokens: response.Usage.TotalTokens},
 	}
 
 	for _, item := range response.Output.Embeddings {
@@ -111,7 +119,7 @@ func embeddingResponseAli2OpenAI(response *EmbeddingResponse) *openai.EmbeddingR
 func responseAli2OpenAI(response *ChatResponse) *openai.TextResponse {
 	choice := openai.TextResponseChoice{
 		Index: 0,
-		Message: openai.Message{
+		Message: model.Message{
 			Role:    "assistant",
 			Content: response.Output.Text,
 		},
@@ -122,7 +130,7 @@ func responseAli2OpenAI(response *ChatResponse) *openai.TextResponse {
 		Object:  "chat.completion",
 		Created: helper.GetTimestamp(),
 		Choices: []openai.TextResponseChoice{choice},
-		Usage: openai.Usage{
+		Usage: model.Usage{
 			PromptTokens:     response.Usage.InputTokens,
 			CompletionTokens: response.Usage.OutputTokens,
 			TotalTokens:      response.Usage.InputTokens + response.Usage.OutputTokens,
@@ -148,8 +156,8 @@ func streamResponseAli2OpenAI(aliResponse *ChatResponse) *openai.ChatCompletions
 	return &response
 }
 
-func StreamHandler(c *gin.Context, resp *http.Response) (*openai.ErrorWithStatusCode, *openai.Usage) {
-	var usage openai.Usage
+func StreamHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, *model.Usage) {
+	var usage model.Usage
 	scanner := bufio.NewScanner(resp.Body)
 	scanner.Split(func(data []byte, atEOF bool) (advance int, token []byte, err error) {
 		if atEOF && len(data) == 0 {
@@ -217,7 +225,7 @@ func StreamHandler(c *gin.Context, resp *http.Response) (*openai.ErrorWithStatus
 	return nil, &usage
 }
 
-func Handler(c *gin.Context, resp *http.Response) (*openai.ErrorWithStatusCode, *openai.Usage) {
+func Handler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, *model.Usage) {
 	var aliResponse ChatResponse
 	responseBody, err := io.ReadAll(resp.Body)
 	if err != nil {
@@ -232,8 +240,8 @@ func Handler(c *gin.Context, resp *http.Response) (*openai.ErrorWithStatusCode,
 		return openai.ErrorWrapper(err, "unmarshal_response_body_failed", http.StatusInternalServerError), nil
 	}
 	if aliResponse.Code != "" {
-		return &openai.ErrorWithStatusCode{
-			Error: openai.Error{
+		return &model.ErrorWithStatusCode{
+			Error: model.Error{
 				Message: aliResponse.Message,
 				Type:    aliResponse.Code,
 				Param:   aliResponse.RequestId,

relay/channel/ali/model.go

@@ -16,6 +16,8 @@ type Parameters struct {
 	Seed              uint64  `json:"seed,omitempty"`
 	EnableSearch      bool    `json:"enable_search,omitempty"`
 	IncrementalOutput bool    `json:"incremental_output,omitempty"`
+	MaxTokens         int     `json:"max_tokens,omitempty"`
+	Temperature       float64 `json:"temperature,omitempty"`
 }
 
 type ChatRequest struct {

relay/channel/anthropic/adaptor.go

@@ -1,22 +1,63 @@
 package anthropic
 
 import (
+	"errors"
+	"fmt"
 	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/relay/channel"
+	"github.com/songquanpeng/one-api/relay/model"
+	"github.com/songquanpeng/one-api/relay/util"
+	"io"
 	"net/http"
-	"one-api/relay/channel/openai"
 )
 
 type Adaptor struct {
 }
 
-func (a *Adaptor) Auth(c *gin.Context) error {
+func (a *Adaptor) Init(meta *util.RelayMeta) {
+
+}
+
+func (a *Adaptor) GetRequestURL(meta *util.RelayMeta) (string, error) {
+	return fmt.Sprintf("%s/v1/messages", meta.BaseURL), nil
+}
+
+func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *util.RelayMeta) error {
+	channel.SetupCommonRequestHeader(c, req, meta)
+	req.Header.Set("x-api-key", meta.APIKey)
+	anthropicVersion := c.Request.Header.Get("anthropic-version")
+	if anthropicVersion == "" {
+		anthropicVersion = "2023-06-01"
+	}
+	req.Header.Set("anthropic-version", anthropicVersion)
+	req.Header.Set("anthropic-beta", "messages-2023-12-15")
 	return nil
 }
 
-func (a *Adaptor) ConvertRequest(request *openai.GeneralOpenAIRequest) (any, error) {
-	return nil, 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")
+	}
+	return ConvertRequest(*request), nil
 }
 
-func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response) (*openai.ErrorWithStatusCode, *openai.Usage, error) {
-	return nil, nil, nil
+func (a *Adaptor) DoRequest(c *gin.Context, meta *util.RelayMeta, requestBody io.Reader) (*http.Response, error) {
+	return channel.DoRequestHelper(a, c, meta, requestBody)
+}
+
+func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *util.RelayMeta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
+	if meta.IsStream {
+		err, usage = StreamHandler(c, resp)
+	} else {
+		err, usage = Handler(c, resp, meta.PromptTokens, meta.ActualModelName)
+	}
+	return
+}
+
+func (a *Adaptor) GetModelList() []string {
+	return ModelList
+}
+
+func (a *Adaptor) GetChannelName() string {
+	return "anthropic"
 }

relay/channel/anthropic/constants.go

@@ -0,0 +1,8 @@
+package anthropic
+
+var ModelList = []string{
+	"claude-instant-1.2", "claude-2.0", "claude-2.1",
+	"claude-3-haiku-20240307",
+	"claude-3-sonnet-20240229",
+	"claude-3-opus-20240229",
+}

relay/channel/anthropic/main.go

@@ -5,82 +5,146 @@ import (
 	"encoding/json"
 	"fmt"
 	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/helper"
+	"github.com/songquanpeng/one-api/common/image"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/relay/channel/openai"
+	"github.com/songquanpeng/one-api/relay/model"
 	"io"
 	"net/http"
-	"one-api/common"
-	"one-api/common/helper"
-	"one-api/common/logger"
-	"one-api/relay/channel/openai"
 	"strings"
 )
 
-func stopReasonClaude2OpenAI(reason string) string {
-	switch reason {
+func stopReasonClaude2OpenAI(reason *string) string {
+	if reason == nil {
+		return ""
+	}
+	switch *reason {
+	case "end_turn":
+		return "stop"
 	case "stop_sequence":
 		return "stop"
 	case "max_tokens":
 		return "length"
 	default:
-		return reason
+		return *reason
 	}
 }
 
-func ConvertRequest(textRequest openai.GeneralOpenAIRequest) *Request {
+func ConvertRequest(textRequest model.GeneralOpenAIRequest) *Request {
 	claudeRequest := Request{
-		Model:             textRequest.Model,
-		Prompt:            "",
-		MaxTokensToSample: textRequest.MaxTokens,
-		StopSequences:     nil,
-		Temperature:       textRequest.Temperature,
-		TopP:              textRequest.TopP,
-		Stream:            textRequest.Stream,
+		Model:       textRequest.Model,
+		MaxTokens:   textRequest.MaxTokens,
+		Temperature: textRequest.Temperature,
+		TopP:        textRequest.TopP,
+		Stream:      textRequest.Stream,
 	}
-	if claudeRequest.MaxTokensToSample == 0 {
-		claudeRequest.MaxTokensToSample = 1000000
+	if claudeRequest.MaxTokens == 0 {
+		claudeRequest.MaxTokens = 4096
+	}
+	// legacy model name mapping
+	if claudeRequest.Model == "claude-instant-1" {
+		claudeRequest.Model = "claude-instant-1.1"
+	} else if claudeRequest.Model == "claude-2" {
+		claudeRequest.Model = "claude-2.1"
 	}
-	prompt := ""
 	for _, message := range textRequest.Messages {
-		if message.Role == "user" {
-			prompt += fmt.Sprintf("\n\nHuman: %s", message.Content)
-		} else if message.Role == "assistant" {
-			prompt += fmt.Sprintf("\n\nAssistant: %s", message.Content)
-		} else if message.Role == "system" {
-			if prompt == "" {
-				prompt = message.StringContent()
-			}
+		if message.Role == "system" && claudeRequest.System == "" {
+			claudeRequest.System = message.StringContent()
+			continue
 		}
+		claudeMessage := Message{
+			Role: message.Role,
+		}
+		var content Content
+		if message.IsStringContent() {
+			content.Type = "text"
+			content.Text = message.StringContent()
+			claudeMessage.Content = append(claudeMessage.Content, content)
+			claudeRequest.Messages = append(claudeRequest.Messages, claudeMessage)
+			continue
+		}
+		var contents []Content
+		openaiContent := message.ParseContent()
+		for _, part := range openaiContent {
+			var content Content
+			if part.Type == model.ContentTypeText {
+				content.Type = "text"
+				content.Text = part.Text
+			} else if part.Type == model.ContentTypeImageURL {
+				content.Type = "image"
+				content.Source = &ImageSource{
+					Type: "base64",
+				}
+				mimeType, data, _ := image.GetImageFromUrl(part.ImageURL.Url)
+				content.Source.MediaType = mimeType
+				content.Source.Data = data
+			}
+			contents = append(contents, content)
+		}
+		claudeMessage.Content = contents
+		claudeRequest.Messages = append(claudeRequest.Messages, claudeMessage)
 	}
-	prompt += "\n\nAssistant:"
-	claudeRequest.Prompt = prompt
 	return &claudeRequest
 }
 
-func streamResponseClaude2OpenAI(claudeResponse *Response) *openai.ChatCompletionsStreamResponse {
+// https://docs.anthropic.com/claude/reference/messages-streaming
+func streamResponseClaude2OpenAI(claudeResponse *StreamResponse) (*openai.ChatCompletionsStreamResponse, *Response) {
+	var response *Response
+	var responseText string
+	var stopReason string
+	switch claudeResponse.Type {
+	case "message_start":
+		return nil, claudeResponse.Message
+	case "content_block_start":
+		if claudeResponse.ContentBlock != nil {
+			responseText = claudeResponse.ContentBlock.Text
+		}
+	case "content_block_delta":
+		if claudeResponse.Delta != nil {
+			responseText = claudeResponse.Delta.Text
+		}
+	case "message_delta":
+		if claudeResponse.Usage != nil {
+			response = &Response{
+				Usage: *claudeResponse.Usage,
+			}
+		}
+		if claudeResponse.Delta != nil && claudeResponse.Delta.StopReason != nil {
+			stopReason = *claudeResponse.Delta.StopReason
+		}
+	}
 	var choice openai.ChatCompletionsStreamResponseChoice
-	choice.Delta.Content = claudeResponse.Completion
-	finishReason := stopReasonClaude2OpenAI(claudeResponse.StopReason)
+	choice.Delta.Content = responseText
+	choice.Delta.Role = "assistant"
+	finishReason := stopReasonClaude2OpenAI(&stopReason)
 	if finishReason != "null" {
 		choice.FinishReason = &finishReason
 	}
-	var response openai.ChatCompletionsStreamResponse
-	response.Object = "chat.completion.chunk"
-	response.Model = claudeResponse.Model
-	response.Choices = []openai.ChatCompletionsStreamResponseChoice{choice}
-	return &response
+	var openaiResponse openai.ChatCompletionsStreamResponse
+	openaiResponse.Object = "chat.completion.chunk"
+	openaiResponse.Choices = []openai.ChatCompletionsStreamResponseChoice{choice}
+	return &openaiResponse, response
 }
 
 func responseClaude2OpenAI(claudeResponse *Response) *openai.TextResponse {
+	var responseText string
+	if len(claudeResponse.Content) > 0 {
+		responseText = claudeResponse.Content[0].Text
+	}
 	choice := openai.TextResponseChoice{
 		Index: 0,
-		Message: openai.Message{
+		Message: model.Message{
 			Role:    "assistant",
-			Content: strings.TrimPrefix(claudeResponse.Completion, " "),
+			Content: responseText,
 			Name:    nil,
 		},
 		FinishReason: stopReasonClaude2OpenAI(claudeResponse.StopReason),
 	}
 	fullTextResponse := openai.TextResponse{
-		Id:      fmt.Sprintf("chatcmpl-%s", helper.GetUUID()),
+		Id:      fmt.Sprintf("chatcmpl-%s", claudeResponse.Id),
+		Model:   claudeResponse.Model,
 		Object:  "chat.completion",
 		Created: helper.GetTimestamp(),
 		Choices: []openai.TextResponseChoice{choice},
@@ -88,17 +152,15 @@ func responseClaude2OpenAI(claudeResponse *Response) *openai.TextResponse {
 	return &fullTextResponse
 }
 
-func StreamHandler(c *gin.Context, resp *http.Response) (*openai.ErrorWithStatusCode, string) {
-	responseText := ""
-	responseId := fmt.Sprintf("chatcmpl-%s", helper.GetUUID())
+func StreamHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, *model.Usage) {
 	createdTime := helper.GetTimestamp()
 	scanner := bufio.NewScanner(resp.Body)
 	scanner.Split(func(data []byte, atEOF bool) (advance int, token []byte, err error) {
 		if atEOF && len(data) == 0 {
 			return 0, nil, nil
 		}
-		if i := strings.Index(string(data), "\r\n\r\n"); i >= 0 {
-			return i + 4, data[0:i], nil
+		if i := strings.Index(string(data), "\n"); i >= 0 {
+			return i + 1, data[0:i], nil
 		}
 		if atEOF {
 			return len(data), data, nil
@@ -110,29 +172,45 @@ func StreamHandler(c *gin.Context, resp *http.Response) (*openai.ErrorWithStatus
 	go func() {
 		for scanner.Scan() {
 			data := scanner.Text()
-			if !strings.HasPrefix(data, "event: completion") {
+			if len(data) < 6 {
 				continue
 			}
-			data = strings.TrimPrefix(data, "event: completion\r\ndata: ")
+			if !strings.HasPrefix(data, "data: ") {
+				continue
+			}
+			data = strings.TrimPrefix(data, "data: ")
 			dataChan <- data
 		}
 		stopChan <- true
 	}()
 	common.SetEventStreamHeaders(c)
+	var usage model.Usage
+	var modelName string
+	var id string
 	c.Stream(func(w io.Writer) bool {
 		select {
 		case data := <-dataChan:
 			// some implementations may add \r at the end of data
 			data = strings.TrimSuffix(data, "\r")
-			var claudeResponse Response
+			var claudeResponse StreamResponse
 			err := json.Unmarshal([]byte(data), &claudeResponse)
 			if err != nil {
 				logger.SysError("error unmarshalling stream response: " + err.Error())
 				return true
 			}
-			responseText += claudeResponse.Completion
-			response := streamResponseClaude2OpenAI(&claudeResponse)
-			response.Id = responseId
+			response, meta := streamResponseClaude2OpenAI(&claudeResponse)
+			if meta != nil {
+				usage.PromptTokens += meta.Usage.InputTokens
+				usage.CompletionTokens += meta.Usage.OutputTokens
+				modelName = meta.Model
+				id = fmt.Sprintf("chatcmpl-%s", meta.Id)
+				return true
+			}
+			if response == nil {
+				return true
+			}
+			response.Id = id
+			response.Model = modelName
 			response.Created = createdTime
 			jsonStr, err := json.Marshal(response)
 			if err != nil {
@@ -146,14 +224,11 @@ func StreamHandler(c *gin.Context, resp *http.Response) (*openai.ErrorWithStatus
 			return false
 		}
 	})
-	err := resp.Body.Close()
-	if err != nil {
-		return openai.ErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), ""
-	}
-	return nil, responseText
+	_ = resp.Body.Close()
+	return nil, &usage
 }
 
-func Handler(c *gin.Context, resp *http.Response, promptTokens int, model string) (*openai.ErrorWithStatusCode, *openai.Usage) {
+func Handler(c *gin.Context, resp *http.Response, promptTokens int, modelName string) (*model.ErrorWithStatusCode, *model.Usage) {
 	responseBody, err := io.ReadAll(resp.Body)
 	if err != nil {
 		return openai.ErrorWrapper(err, "read_response_body_failed", http.StatusInternalServerError), nil
@@ -168,8 +243,8 @@ func Handler(c *gin.Context, resp *http.Response, promptTokens int, model string
 		return openai.ErrorWrapper(err, "unmarshal_response_body_failed", http.StatusInternalServerError), nil
 	}
 	if claudeResponse.Error.Type != "" {
-		return &openai.ErrorWithStatusCode{
-			Error: openai.Error{
+		return &model.ErrorWithStatusCode{
+			Error: model.Error{
 				Message: claudeResponse.Error.Message,
 				Type:    claudeResponse.Error.Type,
 				Param:   "",
@@ -179,12 +254,11 @@ func Handler(c *gin.Context, resp *http.Response, promptTokens int, model string
 		}, nil
 	}
 	fullTextResponse := responseClaude2OpenAI(&claudeResponse)
-	fullTextResponse.Model = model
-	completionTokens := openai.CountTokenText(claudeResponse.Completion, model)
-	usage := openai.Usage{
-		PromptTokens:     promptTokens,
-		CompletionTokens: completionTokens,
-		TotalTokens:      promptTokens + completionTokens,
+	fullTextResponse.Model = modelName
+	usage := model.Usage{
+		PromptTokens:     claudeResponse.Usage.InputTokens,
+		CompletionTokens: claudeResponse.Usage.OutputTokens,
+		TotalTokens:      claudeResponse.Usage.InputTokens + claudeResponse.Usage.OutputTokens,
 	}
 	fullTextResponse.Usage = usage
 	jsonResponse, err := json.Marshal(fullTextResponse)

relay/channel/anthropic/model.go

@@ -1,19 +1,44 @@
 package anthropic
 
+// https://docs.anthropic.com/claude/reference/messages_post
+
 type Metadata struct {
 	UserId string `json:"user_id"`
 }
 
+type ImageSource struct {
+	Type      string `json:"type"`
+	MediaType string `json:"media_type"`
+	Data      string `json:"data"`
+}
+
+type Content struct {
+	Type   string       `json:"type"`
+	Text   string       `json:"text,omitempty"`
+	Source *ImageSource `json:"source,omitempty"`
+}
+
+type Message struct {
+	Role    string    `json:"role"`
+	Content []Content `json:"content"`
+}
+
 type Request struct {
-	Model             string   `json:"model"`
-	Prompt            string   `json:"prompt"`
-	MaxTokensToSample int      `json:"max_tokens_to_sample"`
-	StopSequences     []string `json:"stop_sequences,omitempty"`
-	Temperature       float64  `json:"temperature,omitempty"`
-	TopP              float64  `json:"top_p,omitempty"`
-	TopK              int      `json:"top_k,omitempty"`
+	Model         string    `json:"model"`
+	Messages      []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"`
 	//Metadata    `json:"metadata,omitempty"`
-	Stream bool `json:"stream,omitempty"`
+}
+
+type Usage struct {
+	InputTokens  int `json:"input_tokens"`
+	OutputTokens int `json:"output_tokens"`
 }
 
 type Error struct {
@@ -22,8 +47,29 @@ type Error struct {
 }
 
 type Response struct {
-	Completion string `json:"completion"`
-	StopReason string `json:"stop_reason"`
-	Model      string `json:"model"`
-	Error      Error  `json:"error"`
+	Id           string    `json:"id"`
+	Type         string    `json:"type"`
+	Role         string    `json:"role"`
+	Content      []Content `json:"content"`
+	Model        string    `json:"model"`
+	StopReason   *string   `json:"stop_reason"`
+	StopSequence *string   `json:"stop_sequence"`
+	Usage        Usage     `json:"usage"`
+	Error        Error     `json:"error"`
+}
+
+type Delta struct {
+	Type         string  `json:"type"`
+	Text         string  `json:"text"`
+	StopReason   *string `json:"stop_reason"`
+	StopSequence *string `json:"stop_sequence"`
+}
+
+type StreamResponse struct {
+	Type         string    `json:"type"`
+	Message      *Response `json:"message"`
+	Index        int       `json:"index"`
+	ContentBlock *Content  `json:"content_block"`
+	Delta        *Delta    `json:"delta"`
+	Usage        *Usage    `json:"usage"`
 }

relay/channel/baichuan/constants.go

@@ -0,0 +1,7 @@
+package baichuan
+
+var ModelList = []string{
+	"Baichuan2-Turbo",
+	"Baichuan2-Turbo-192k",
+	"Baichuan-Text-Embedding",
+}

relay/channel/baidu/adaptor.go

@@ -1,22 +1,118 @@
 package baidu
 
 import (
-	"github.com/gin-gonic/gin"
+	"errors"
+	"fmt"
+	"io"
 	"net/http"
-	"one-api/relay/channel/openai"
+	"strings"
+
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/relay/channel"
+	"github.com/songquanpeng/one-api/relay/constant"
+	"github.com/songquanpeng/one-api/relay/model"
+	"github.com/songquanpeng/one-api/relay/util"
 )
 
 type Adaptor struct {
 }
 
-func (a *Adaptor) Auth(c *gin.Context) error {
+func (a *Adaptor) Init(meta *util.RelayMeta) {
+
+}
+
+func (a *Adaptor) GetRequestURL(meta *util.RelayMeta) (string, error) {
+	// https://cloud.baidu.com/doc/WENXINWORKSHOP/s/clntwmv7t
+	suffix := "chat/"
+	if strings.HasPrefix(meta.ActualModelName, "Embedding") {
+		suffix = "embeddings/"
+	}
+	if strings.HasPrefix(meta.ActualModelName, "bge-large") {
+		suffix = "embeddings/"
+	}
+	if strings.HasPrefix(meta.ActualModelName, "tao-8k") {
+		suffix = "embeddings/"
+	}
+	switch meta.ActualModelName {
+	case "ERNIE-4.0":
+		suffix += "completions_pro"
+	case "ERNIE-Bot-4":
+		suffix += "completions_pro"
+	case "ERNIE-3.5-8K":
+		suffix += "completions"
+	case "ERNIE-Bot-8K":
+		suffix += "ernie_bot_8k"
+	case "ERNIE-Bot":
+		suffix += "completions"
+	case "ERNIE-Speed":
+		suffix += "ernie_speed"
+	case "ERNIE-Bot-turbo":
+		suffix += "eb-instant"
+	case "BLOOMZ-7B":
+		suffix += "bloomz_7b1"
+	case "Embedding-V1":
+		suffix += "embedding-v1"
+	case "bge-large-zh":
+		suffix += "bge_large_zh"
+	case "bge-large-en":
+		suffix += "bge_large_en"
+	case "tao-8k":
+		suffix += "tao_8k"
+	default:
+		suffix += meta.ActualModelName
+	}
+	fullRequestURL := fmt.Sprintf("%s/rpc/2.0/ai_custom/v1/wenxinworkshop/%s", meta.BaseURL, suffix)
+	var accessToken string
+	var err error
+	if accessToken, err = GetAccessToken(meta.APIKey); err != nil {
+		return "", err
+	}
+	fullRequestURL += "?access_token=" + accessToken
+	return fullRequestURL, nil
+}
+
+func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *util.RelayMeta) error {
+	channel.SetupCommonRequestHeader(c, req, meta)
+	req.Header.Set("Authorization", "Bearer "+meta.APIKey)
 	return nil
 }
 
-func (a *Adaptor) ConvertRequest(request *openai.GeneralOpenAIRequest) (any, error) {
-	return nil, 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")
+	}
+	switch relayMode {
+	case constant.RelayModeEmbeddings:
+		baiduEmbeddingRequest := ConvertEmbeddingRequest(*request)
+		return baiduEmbeddingRequest, nil
+	default:
+		baiduRequest := ConvertRequest(*request)
+		return baiduRequest, nil
+	}
 }
 
-func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response) (*openai.ErrorWithStatusCode, *openai.Usage, error) {
-	return nil, nil, nil
+func (a *Adaptor) DoRequest(c *gin.Context, meta *util.RelayMeta, requestBody io.Reader) (*http.Response, error) {
+	return channel.DoRequestHelper(a, c, meta, requestBody)
+}
+
+func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *util.RelayMeta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
+	if meta.IsStream {
+		err, usage = StreamHandler(c, resp)
+	} else {
+		switch meta.Mode {
+		case constant.RelayModeEmbeddings:
+			err, usage = EmbeddingHandler(c, resp)
+		default:
+			err, usage = Handler(c, resp)
+		}
+	}
+	return
+}
+
+func (a *Adaptor) GetModelList() []string {
+	return ModelList
+}
+
+func (a *Adaptor) GetChannelName() string {
+	return "baidu"
 }

relay/channel/baidu/constants.go

@@ -0,0 +1,13 @@
+package baidu
+
+var ModelList = []string{
+	"ERNIE-Bot-4",
+	"ERNIE-Bot-8K",
+	"ERNIE-Bot",
+	"ERNIE-Speed",
+	"ERNIE-Bot-turbo",
+	"Embedding-V1",
+	"bge-large-zh",
+	"bge-large-en",
+	"tao-8k",
+}

relay/channel/baidu/main.go

@@ -6,13 +6,14 @@ import (
 	"errors"
 	"fmt"
 	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/relay/channel/openai"
+	"github.com/songquanpeng/one-api/relay/constant"
+	"github.com/songquanpeng/one-api/relay/model"
+	"github.com/songquanpeng/one-api/relay/util"
 	"io"
 	"net/http"
-	"one-api/common"
-	"one-api/common/logger"
-	"one-api/relay/channel/openai"
-	"one-api/relay/constant"
-	"one-api/relay/util"
 	"strings"
 	"sync"
 	"time"
@@ -31,9 +32,16 @@ type Message struct {
 }
 
 type ChatRequest struct {
-	Messages []Message `json:"messages"`
-	Stream   bool      `json:"stream"`
-	UserId   string    `json:"user_id,omitempty"`
+	Messages        []Message `json:"messages"`
+	Temperature     float64   `json:"temperature,omitempty"`
+	TopP            float64   `json:"top_p,omitempty"`
+	PenaltyScore    float64   `json:"penalty_score,omitempty"`
+	Stream          bool      `json:"stream,omitempty"`
+	System          string    `json:"system,omitempty"`
+	DisableSearch   bool      `json:"disable_search,omitempty"`
+	EnableCitation  bool      `json:"enable_citation,omitempty"`
+	MaxOutputTokens int       `json:"max_output_tokens,omitempty"`
+	UserId          string    `json:"user_id,omitempty"`
 }
 
 type Error struct {
@@ -43,35 +51,35 @@ type Error struct {
 
 var baiduTokenStore sync.Map
 
-func ConvertRequest(request openai.GeneralOpenAIRequest) *ChatRequest {
-	messages := make([]Message, 0, len(request.Messages))
+func ConvertRequest(request model.GeneralOpenAIRequest) *ChatRequest {
+	baiduRequest := ChatRequest{
+		Messages:        make([]Message, 0, len(request.Messages)),
+		Temperature:     request.Temperature,
+		TopP:            request.TopP,
+		PenaltyScore:    request.FrequencyPenalty,
+		Stream:          request.Stream,
+		DisableSearch:   false,
+		EnableCitation:  false,
+		MaxOutputTokens: request.MaxTokens,
+		UserId:          request.User,
+	}
 	for _, message := range request.Messages {
 		if message.Role == "system" {
-			messages = append(messages, Message{
-				Role:    "user",
-				Content: message.StringContent(),
-			})
-			messages = append(messages, Message{
-				Role:    "assistant",
-				Content: "Okay",
-			})
+			baiduRequest.System = message.StringContent()
 		} else {
-			messages = append(messages, Message{
+			baiduRequest.Messages = append(baiduRequest.Messages, Message{
 				Role:    message.Role,
 				Content: message.StringContent(),
 			})
 		}
 	}
-	return &ChatRequest{
-		Messages: messages,
-		Stream:   request.Stream,
-	}
+	return &baiduRequest
 }
 
 func responseBaidu2OpenAI(response *ChatResponse) *openai.TextResponse {
 	choice := openai.TextResponseChoice{
 		Index: 0,
-		Message: openai.Message{
+		Message: model.Message{
 			Role:    "assistant",
 			Content: response.Result,
 		},
@@ -103,7 +111,7 @@ func streamResponseBaidu2OpenAI(baiduResponse *ChatStreamResponse) *openai.ChatC
 	return &response
 }
 
-func ConvertEmbeddingRequest(request openai.GeneralOpenAIRequest) *EmbeddingRequest {
+func ConvertEmbeddingRequest(request model.GeneralOpenAIRequest) *EmbeddingRequest {
 	return &EmbeddingRequest{
 		Input: request.ParseInput(),
 	}
@@ -126,8 +134,8 @@ func embeddingResponseBaidu2OpenAI(response *EmbeddingResponse) *openai.Embeddin
 	return &openAIEmbeddingResponse
 }
 
-func StreamHandler(c *gin.Context, resp *http.Response) (*openai.ErrorWithStatusCode, *openai.Usage) {
-	var usage openai.Usage
+func StreamHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, *model.Usage) {
+	var usage model.Usage
 	scanner := bufio.NewScanner(resp.Body)
 	scanner.Split(func(data []byte, atEOF bool) (advance int, token []byte, err error) {
 		if atEOF && len(data) == 0 {
@@ -189,7 +197,7 @@ func StreamHandler(c *gin.Context, resp *http.Response) (*openai.ErrorWithStatus
 	return nil, &usage
 }
 
-func Handler(c *gin.Context, resp *http.Response) (*openai.ErrorWithStatusCode, *openai.Usage) {
+func Handler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, *model.Usage) {
 	var baiduResponse ChatResponse
 	responseBody, err := io.ReadAll(resp.Body)
 	if err != nil {
@@ -204,8 +212,8 @@ func Handler(c *gin.Context, resp *http.Response) (*openai.ErrorWithStatusCode,
 		return openai.ErrorWrapper(err, "unmarshal_response_body_failed", http.StatusInternalServerError), nil
 	}
 	if baiduResponse.ErrorMsg != "" {
-		return &openai.ErrorWithStatusCode{
-			Error: openai.Error{
+		return &model.ErrorWithStatusCode{
+			Error: model.Error{
 				Message: baiduResponse.ErrorMsg,
 				Type:    "baidu_error",
 				Param:   "",
@@ -226,7 +234,7 @@ func Handler(c *gin.Context, resp *http.Response) (*openai.ErrorWithStatusCode,
 	return nil, &fullTextResponse.Usage
 }
 
-func EmbeddingHandler(c *gin.Context, resp *http.Response) (*openai.ErrorWithStatusCode, *openai.Usage) {
+func EmbeddingHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, *model.Usage) {
 	var baiduResponse EmbeddingResponse
 	responseBody, err := io.ReadAll(resp.Body)
 	if err != nil {
@@ -241,8 +249,8 @@ func EmbeddingHandler(c *gin.Context, resp *http.Response) (*openai.ErrorWithSta
 		return openai.ErrorWrapper(err, "unmarshal_response_body_failed", http.StatusInternalServerError), nil
 	}
 	if baiduResponse.ErrorMsg != "" {
-		return &openai.ErrorWithStatusCode{
-			Error: openai.Error{
+		return &model.ErrorWithStatusCode{
+			Error: model.Error{
 				Message: baiduResponse.ErrorMsg,
 				Type:    "baidu_error",
 				Param:   "",

relay/channel/baidu/model.go

@@ -1,18 +1,18 @@
 package baidu
 
 import (
-	"one-api/relay/channel/openai"
+	"github.com/songquanpeng/one-api/relay/model"
 	"time"
 )
 
 type ChatResponse struct {
-	Id               string       `json:"id"`
-	Object           string       `json:"object"`
-	Created          int64        `json:"created"`
-	Result           string       `json:"result"`
-	IsTruncated      bool         `json:"is_truncated"`
-	NeedClearHistory bool         `json:"need_clear_history"`
-	Usage            openai.Usage `json:"usage"`
+	Id               string      `json:"id"`
+	Object           string      `json:"object"`
+	Created          int64       `json:"created"`
+	Result           string      `json:"result"`
+	IsTruncated      bool        `json:"is_truncated"`
+	NeedClearHistory bool        `json:"need_clear_history"`
+	Usage            model.Usage `json:"usage"`
 	Error
 }
 
@@ -37,7 +37,7 @@ type EmbeddingResponse struct {
 	Object  string          `json:"object"`
 	Created int64           `json:"created"`
 	Data    []EmbeddingData `json:"data"`
-	Usage   openai.Usage    `json:"usage"`
+	Usage   model.Usage     `json:"usage"`
 	Error
 }
 

relay/channel/common.go

@@ -0,0 +1,51 @@
+package channel
+
+import (
+	"errors"
+	"fmt"
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/relay/util"
+	"io"
+	"net/http"
+)
+
+func SetupCommonRequestHeader(c *gin.Context, req *http.Request, meta *util.RelayMeta) {
+	req.Header.Set("Content-Type", c.Request.Header.Get("Content-Type"))
+	req.Header.Set("Accept", c.Request.Header.Get("Accept"))
+	if meta.IsStream && c.Request.Header.Get("Accept") == "" {
+		req.Header.Set("Accept", "text/event-stream")
+	}
+}
+
+func DoRequestHelper(a Adaptor, c *gin.Context, meta *util.RelayMeta, requestBody io.Reader) (*http.Response, error) {
+	fullRequestURL, err := a.GetRequestURL(meta)
+	if err != nil {
+		return nil, fmt.Errorf("get request url failed: %w", err)
+	}
+	req, err := http.NewRequest(c.Request.Method, fullRequestURL, requestBody)
+	if err != nil {
+		return nil, fmt.Errorf("new request failed: %w", err)
+	}
+	err = a.SetupRequestHeader(c, req, meta)
+	if err != nil {
+		return nil, fmt.Errorf("setup request header failed: %w", err)
+	}
+	resp, err := DoRequest(c, req)
+	if err != nil {
+		return nil, fmt.Errorf("do request failed: %w", err)
+	}
+	return resp, nil
+}
+
+func DoRequest(c *gin.Context, req *http.Request) (*http.Response, error) {
+	resp, err := util.HTTPClient.Do(req)
+	if err != nil {
+		return nil, err
+	}
+	if resp == nil {
+		return nil, errors.New("resp is nil")
+	}
+	_ = req.Body.Close()
+	_ = c.Request.Body.Close()
+	return resp, nil
+}

relay/channel/gemini/adaptor.go

@@ -0,0 +1,66 @@
+package gemini
+
+import (
+	"errors"
+	"fmt"
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common/helper"
+	channelhelper "github.com/songquanpeng/one-api/relay/channel"
+	"github.com/songquanpeng/one-api/relay/channel/openai"
+	"github.com/songquanpeng/one-api/relay/model"
+	"github.com/songquanpeng/one-api/relay/util"
+	"io"
+	"net/http"
+)
+
+type Adaptor struct {
+}
+
+func (a *Adaptor) Init(meta *util.RelayMeta) {
+
+}
+
+func (a *Adaptor) GetRequestURL(meta *util.RelayMeta) (string, error) {
+	version := helper.AssignOrDefault(meta.APIVersion, "v1")
+	action := "generateContent"
+	if meta.IsStream {
+		action = "streamGenerateContent"
+	}
+	return fmt.Sprintf("%s/%s/models/%s:%s", meta.BaseURL, version, meta.ActualModelName, action), nil
+}
+
+func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *util.RelayMeta) error {
+	channelhelper.SetupCommonRequestHeader(c, req, meta)
+	req.Header.Set("x-goog-api-key", meta.APIKey)
+	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")
+	}
+	return ConvertRequest(*request), nil
+}
+
+func (a *Adaptor) DoRequest(c *gin.Context, meta *util.RelayMeta, requestBody io.Reader) (*http.Response, error) {
+	return channelhelper.DoRequestHelper(a, c, meta, requestBody)
+}
+
+func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *util.RelayMeta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
+	if meta.IsStream {
+		var responseText string
+		err, responseText = StreamHandler(c, resp)
+		usage = openai.ResponseText2Usage(responseText, meta.ActualModelName, meta.PromptTokens)
+	} else {
+		err, usage = Handler(c, resp, meta.PromptTokens, meta.ActualModelName)
+	}
+	return
+}
+
+func (a *Adaptor) GetModelList() []string {
+	return ModelList
+}
+
+func (a *Adaptor) GetChannelName() string {
+	return "google gemini"
+}

relay/channel/gemini/constants.go

@@ -0,0 +1,6 @@
+package gemini
+
+var ModelList = []string{
+	"gemini-pro", "gemini-1.0-pro-001",
+	"gemini-pro-vision", "gemini-1.0-pro-vision-001",
+}

relay/channel/gemini/main.go

@@ -1,18 +1,19 @@
-package google
+package gemini
 
 import (
 	"bufio"
 	"encoding/json"
 	"fmt"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/helper"
+	"github.com/songquanpeng/one-api/common/image"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/relay/channel/openai"
+	"github.com/songquanpeng/one-api/relay/constant"
+	"github.com/songquanpeng/one-api/relay/model"
 	"io"
 	"net/http"
-	"one-api/common"
-	"one-api/common/config"
-	"one-api/common/helper"
-	"one-api/common/image"
-	"one-api/common/logger"
-	"one-api/relay/channel/openai"
-	"one-api/relay/constant"
 	"strings"
 
 	"github.com/gin-gonic/gin"
@@ -21,14 +22,14 @@ import (
 // https://ai.google.dev/docs/gemini_api_overview?hl=zh-cn
 
 const (
-	GeminiVisionMaxImageNum = 16
+	VisionMaxImageNum = 16
 )
 
 // Setting safety to the lowest possible values since Gemini is already powerless enough
-func ConvertGeminiRequest(textRequest openai.GeneralOpenAIRequest) *GeminiChatRequest {
-	geminiRequest := GeminiChatRequest{
-		Contents: make([]GeminiChatContent, 0, len(textRequest.Messages)),
-		SafetySettings: []GeminiChatSafetySettings{
+func ConvertRequest(textRequest model.GeneralOpenAIRequest) *ChatRequest {
+	geminiRequest := ChatRequest{
+		Contents: make([]ChatContent, 0, len(textRequest.Messages)),
+		SafetySettings: []ChatSafetySettings{
 			{
 				Category:  "HARM_CATEGORY_HARASSMENT",
 				Threshold: config.GeminiSafetySetting,
@@ -46,14 +47,14 @@ func ConvertGeminiRequest(textRequest openai.GeneralOpenAIRequest) *GeminiChatRe
 				Threshold: config.GeminiSafetySetting,
 			},
 		},
-		GenerationConfig: GeminiChatGenerationConfig{
+		GenerationConfig: ChatGenerationConfig{
 			Temperature:     textRequest.Temperature,
 			TopP:            textRequest.TopP,
 			MaxOutputTokens: textRequest.MaxTokens,
 		},
 	}
 	if textRequest.Functions != nil {
-		geminiRequest.Tools = []GeminiChatTools{
+		geminiRequest.Tools = []ChatTools{
 			{
 				FunctionDeclarations: textRequest.Functions,
 			},
@@ -61,30 +62,30 @@ func ConvertGeminiRequest(textRequest openai.GeneralOpenAIRequest) *GeminiChatRe
 	}
 	shouldAddDummyModelMessage := false
 	for _, message := range textRequest.Messages {
-		content := GeminiChatContent{
+		content := ChatContent{
 			Role: message.Role,
-			Parts: []GeminiPart{
+			Parts: []Part{
 				{
 					Text: message.StringContent(),
 				},
 			},
 		}
 		openaiContent := message.ParseContent()
-		var parts []GeminiPart
+		var parts []Part
 		imageNum := 0
 		for _, part := range openaiContent {
-			if part.Type == openai.ContentTypeText {
-				parts = append(parts, GeminiPart{
+			if part.Type == model.ContentTypeText {
+				parts = append(parts, Part{
 					Text: part.Text,
 				})
-			} else if part.Type == openai.ContentTypeImageURL {
+			} else if part.Type == model.ContentTypeImageURL {
 				imageNum += 1
-				if imageNum > GeminiVisionMaxImageNum {
+				if imageNum > VisionMaxImageNum {
 					continue
 				}
 				mimeType, data, _ := image.GetImageFromUrl(part.ImageURL.Url)
-				parts = append(parts, GeminiPart{
-					InlineData: &GeminiInlineData{
+				parts = append(parts, Part{
+					InlineData: &InlineData{
 						MimeType: mimeType,
 						Data:     data,
 					},
@@ -106,9 +107,9 @@ func ConvertGeminiRequest(textRequest openai.GeneralOpenAIRequest) *GeminiChatRe
 
 		// If a system message is the last message, we need to add a dummy model message to make gemini happy
 		if shouldAddDummyModelMessage {
-			geminiRequest.Contents = append(geminiRequest.Contents, GeminiChatContent{
+			geminiRequest.Contents = append(geminiRequest.Contents, ChatContent{
 				Role: "model",
-				Parts: []GeminiPart{
+				Parts: []Part{
 					{
 						Text: "Okay",
 					},
@@ -121,12 +122,12 @@ func ConvertGeminiRequest(textRequest openai.GeneralOpenAIRequest) *GeminiChatRe
 	return &geminiRequest
 }
 
-type GeminiChatResponse struct {
-	Candidates     []GeminiChatCandidate    `json:"candidates"`
-	PromptFeedback GeminiChatPromptFeedback `json:"promptFeedback"`
+type ChatResponse struct {
+	Candidates     []ChatCandidate    `json:"candidates"`
+	PromptFeedback ChatPromptFeedback `json:"promptFeedback"`
 }
 
-func (g *GeminiChatResponse) GetResponseText() string {
+func (g *ChatResponse) GetResponseText() string {
 	if g == nil {
 		return ""
 	}
@@ -136,23 +137,23 @@ func (g *GeminiChatResponse) GetResponseText() string {
 	return ""
 }
 
-type GeminiChatCandidate struct {
-	Content       GeminiChatContent        `json:"content"`
-	FinishReason  string                   `json:"finishReason"`
-	Index         int64                    `json:"index"`
-	SafetyRatings []GeminiChatSafetyRating `json:"safetyRatings"`
+type ChatCandidate struct {
+	Content       ChatContent        `json:"content"`
+	FinishReason  string             `json:"finishReason"`
+	Index         int64              `json:"index"`
+	SafetyRatings []ChatSafetyRating `json:"safetyRatings"`
 }
 
-type GeminiChatSafetyRating struct {
+type ChatSafetyRating struct {
 	Category    string `json:"category"`
 	Probability string `json:"probability"`
 }
 
-type GeminiChatPromptFeedback struct {
-	SafetyRatings []GeminiChatSafetyRating `json:"safetyRatings"`
+type ChatPromptFeedback struct {
+	SafetyRatings []ChatSafetyRating `json:"safetyRatings"`
 }
 
-func responseGeminiChat2OpenAI(response *GeminiChatResponse) *openai.TextResponse {
+func responseGeminiChat2OpenAI(response *ChatResponse) *openai.TextResponse {
 	fullTextResponse := openai.TextResponse{
 		Id:      fmt.Sprintf("chatcmpl-%s", helper.GetUUID()),
 		Object:  "chat.completion",
@@ -162,7 +163,7 @@ func responseGeminiChat2OpenAI(response *GeminiChatResponse) *openai.TextRespons
 	for i, candidate := range response.Candidates {
 		choice := openai.TextResponseChoice{
 			Index: i,
-			Message: openai.Message{
+			Message: model.Message{
 				Role:    "assistant",
 				Content: "",
 			},
@@ -176,7 +177,7 @@ func responseGeminiChat2OpenAI(response *GeminiChatResponse) *openai.TextRespons
 	return &fullTextResponse
 }
 
-func streamResponseGeminiChat2OpenAI(geminiResponse *GeminiChatResponse) *openai.ChatCompletionsStreamResponse {
+func streamResponseGeminiChat2OpenAI(geminiResponse *ChatResponse) *openai.ChatCompletionsStreamResponse {
 	var choice openai.ChatCompletionsStreamResponseChoice
 	choice.Delta.Content = geminiResponse.GetResponseText()
 	choice.FinishReason = &constant.StopFinishReason
@@ -187,7 +188,7 @@ func streamResponseGeminiChat2OpenAI(geminiResponse *GeminiChatResponse) *openai
 	return &response
 }
 
-func StreamHandler(c *gin.Context, resp *http.Response) (*openai.ErrorWithStatusCode, string) {
+func StreamHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, string) {
 	responseText := ""
 	dataChan := make(chan string)
 	stopChan := make(chan bool)
@@ -257,7 +258,7 @@ func StreamHandler(c *gin.Context, resp *http.Response) (*openai.ErrorWithStatus
 	return nil, responseText
 }
 
-func GeminiHandler(c *gin.Context, resp *http.Response, promptTokens int, model string) (*openai.ErrorWithStatusCode, *openai.Usage) {
+func Handler(c *gin.Context, resp *http.Response, promptTokens int, modelName string) (*model.ErrorWithStatusCode, *model.Usage) {
 	responseBody, err := io.ReadAll(resp.Body)
 	if err != nil {
 		return openai.ErrorWrapper(err, "read_response_body_failed", http.StatusInternalServerError), nil
@@ -266,14 +267,14 @@ func GeminiHandler(c *gin.Context, resp *http.Response, promptTokens int, model
 	if err != nil {
 		return openai.ErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), nil
 	}
-	var geminiResponse GeminiChatResponse
+	var geminiResponse ChatResponse
 	err = json.Unmarshal(responseBody, &geminiResponse)
 	if err != nil {
 		return openai.ErrorWrapper(err, "unmarshal_response_body_failed", http.StatusInternalServerError), nil
 	}
 	if len(geminiResponse.Candidates) == 0 {
-		return &openai.ErrorWithStatusCode{
-			Error: openai.Error{
+		return &model.ErrorWithStatusCode{
+			Error: model.Error{
 				Message: "No candidates returned",
 				Type:    "server_error",
 				Param:   "",
@@ -283,9 +284,9 @@ func GeminiHandler(c *gin.Context, resp *http.Response, promptTokens int, model
 		}, nil
 	}
 	fullTextResponse := responseGeminiChat2OpenAI(&geminiResponse)
-	fullTextResponse.Model = model
-	completionTokens := openai.CountTokenText(geminiResponse.GetResponseText(), model)
-	usage := openai.Usage{
+	fullTextResponse.Model = modelName
+	completionTokens := openai.CountTokenText(geminiResponse.GetResponseText(), modelName)
+	usage := model.Usage{
 		PromptTokens:     promptTokens,
 		CompletionTokens: completionTokens,
 		TotalTokens:      promptTokens + completionTokens,

relay/channel/gemini/model.go

@@ -0,0 +1,41 @@
+package gemini
+
+type ChatRequest struct {
+	Contents         []ChatContent        `json:"contents"`
+	SafetySettings   []ChatSafetySettings `json:"safety_settings,omitempty"`
+	GenerationConfig ChatGenerationConfig `json:"generation_config,omitempty"`
+	Tools            []ChatTools          `json:"tools,omitempty"`
+}
+
+type InlineData struct {
+	MimeType string `json:"mimeType"`
+	Data     string `json:"data"`
+}
+
+type Part struct {
+	Text       string      `json:"text,omitempty"`
+	InlineData *InlineData `json:"inlineData,omitempty"`
+}
+
+type ChatContent struct {
+	Role  string `json:"role,omitempty"`
+	Parts []Part `json:"parts"`
+}
+
+type ChatSafetySettings struct {
+	Category  string `json:"category"`
+	Threshold string `json:"threshold"`
+}
+
+type ChatTools struct {
+	FunctionDeclarations any `json:"functionDeclarations,omitempty"`
+}
+
+type ChatGenerationConfig struct {
+	Temperature     float64  `json:"temperature,omitempty"`
+	TopP            float64  `json:"topP,omitempty"`
+	TopK            float64  `json:"topK,omitempty"`
+	MaxOutputTokens int      `json:"maxOutputTokens,omitempty"`
+	CandidateCount  int      `json:"candidateCount,omitempty"`
+	StopSequences   []string `json:"stopSequences,omitempty"`
+}

relay/channel/google/adaptor.go

@@ -1,22 +0,0 @@
-package google
-
-import (
-	"github.com/gin-gonic/gin"
-	"net/http"
-	"one-api/relay/channel/openai"
-)
-
-type Adaptor struct {
-}
-
-func (a *Adaptor) Auth(c *gin.Context) error {
-	return nil
-}
-
-func (a *Adaptor) ConvertRequest(request *openai.GeneralOpenAIRequest) (any, error) {
-	return nil, nil
-}
-
-func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response) (*openai.ErrorWithStatusCode, *openai.Usage, error) {
-	return nil, nil, nil
-}

relay/channel/google/model.go

@@ -1,80 +0,0 @@
-package google
-
-import (
-	"one-api/relay/channel/openai"
-)
-
-type GeminiChatRequest struct {
-	Contents         []GeminiChatContent        `json:"contents"`
-	SafetySettings   []GeminiChatSafetySettings `json:"safety_settings,omitempty"`
-	GenerationConfig GeminiChatGenerationConfig `json:"generation_config,omitempty"`
-	Tools            []GeminiChatTools          `json:"tools,omitempty"`
-}
-
-type GeminiInlineData struct {
-	MimeType string `json:"mimeType"`
-	Data     string `json:"data"`
-}
-
-type GeminiPart struct {
-	Text       string            `json:"text,omitempty"`
-	InlineData *GeminiInlineData `json:"inlineData,omitempty"`
-}
-
-type GeminiChatContent struct {
-	Role  string       `json:"role,omitempty"`
-	Parts []GeminiPart `json:"parts"`
-}
-
-type GeminiChatSafetySettings struct {
-	Category  string `json:"category"`
-	Threshold string `json:"threshold"`
-}
-
-type GeminiChatTools struct {
-	FunctionDeclarations any `json:"functionDeclarations,omitempty"`
-}
-
-type GeminiChatGenerationConfig struct {
-	Temperature     float64  `json:"temperature,omitempty"`
-	TopP            float64  `json:"topP,omitempty"`
-	TopK            float64  `json:"topK,omitempty"`
-	MaxOutputTokens int      `json:"maxOutputTokens,omitempty"`
-	CandidateCount  int      `json:"candidateCount,omitempty"`
-	StopSequences   []string `json:"stopSequences,omitempty"`
-}
-
-type PaLMChatMessage struct {
-	Author  string `json:"author"`
-	Content string `json:"content"`
-}
-
-type PaLMFilter struct {
-	Reason  string `json:"reason"`
-	Message string `json:"message"`
-}
-
-type PaLMPrompt struct {
-	Messages []PaLMChatMessage `json:"messages"`
-}
-
-type PaLMChatRequest struct {
-	Prompt         PaLMPrompt `json:"prompt"`
-	Temperature    float64    `json:"temperature,omitempty"`
-	CandidateCount int        `json:"candidateCount,omitempty"`
-	TopP           float64    `json:"topP,omitempty"`
-	TopK           int        `json:"topK,omitempty"`
-}
-
-type PaLMError struct {
-	Code    int    `json:"code"`
-	Message string `json:"message"`
-	Status  string `json:"status"`
-}
-
-type PaLMChatResponse struct {
-	Candidates []PaLMChatMessage `json:"candidates"`
-	Messages   []openai.Message  `json:"messages"`
-	Filters    []PaLMFilter      `json:"filters"`
-	Error      PaLMError         `json:"error"`
-}

relay/channel/groq/constants.go

@@ -0,0 +1,10 @@
+package groq
+
+// https://console.groq.com/docs/models
+
+var ModelList = []string{
+	"gemma-7b-it",
+	"llama2-7b-2048",
+	"llama2-70b-4096",
+	"mixtral-8x7b-32768",
+}

relay/channel/interface.go

@@ -2,14 +2,19 @@ package channel
 
 import (
 	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/relay/model"
+	"github.com/songquanpeng/one-api/relay/util"
+	"io"
 	"net/http"
-	"one-api/relay/channel/openai"
 )
 
 type Adaptor interface {
-	GetRequestURL() string
-	Auth(c *gin.Context) error
-	ConvertRequest(request *openai.GeneralOpenAIRequest) (any, error)
-	DoRequest(request *openai.GeneralOpenAIRequest) error
-	DoResponse(c *gin.Context, resp *http.Response) (*openai.ErrorWithStatusCode, *openai.Usage, error)
+	Init(meta *util.RelayMeta)
+	GetRequestURL(meta *util.RelayMeta) (string, error)
+	SetupRequestHeader(c *gin.Context, req *http.Request, meta *util.RelayMeta) error
+	ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error)
+	DoRequest(c *gin.Context, meta *util.RelayMeta, requestBody io.Reader) (*http.Response, error)
+	DoResponse(c *gin.Context, resp *http.Response, meta *util.RelayMeta) (usage *model.Usage, err *model.ErrorWithStatusCode)
+	GetModelList() []string
+	GetChannelName() string
 }

relay/channel/lingyiwanwu/constants.go

@@ -0,0 +1,9 @@
+package lingyiwanwu
+
+// https://platform.lingyiwanwu.com/docs
+
+var ModelList = []string{
+	"yi-34b-chat-0205",
+	"yi-34b-chat-200k",
+	"yi-vl-plus",
+}

relay/channel/minimax/constants.go

@@ -0,0 +1,7 @@
+package minimax
+
+var ModelList = []string{
+	"abab5.5s-chat",
+	"abab5.5-chat",
+	"abab6-chat",
+}

relay/channel/minimax/main.go

@@ -0,0 +1,14 @@
+package minimax
+
+import (
+	"fmt"
+	"github.com/songquanpeng/one-api/relay/constant"
+	"github.com/songquanpeng/one-api/relay/util"
+)
+
+func GetRequestURL(meta *util.RelayMeta) (string, error) {
+	if meta.Mode == constant.RelayModeChatCompletions {
+		return fmt.Sprintf("%s/v1/text/chatcompletion_v2", meta.BaseURL), nil
+	}
+	return "", fmt.Errorf("unsupported relay mode %d for minimax", meta.Mode)
+}

relay/channel/mistral/constants.go

@@ -0,0 +1,10 @@
+package mistral
+
+var ModelList = []string{
+	"open-mistral-7b",
+	"open-mixtral-8x7b",
+	"mistral-small-latest",
+	"mistral-medium-latest",
+	"mistral-large-latest",
+	"mistral-embed",
+}

relay/channel/moonshot/constants.go

@@ -0,0 +1,7 @@
+package moonshot
+
+var ModelList = []string{
+	"moonshot-v1-8k",
+	"moonshot-v1-32k",
+	"moonshot-v1-128k",
+}

relay/channel/ollama/adaptor.go

@@ -0,0 +1,65 @@
+package ollama
+
+import (
+	"errors"
+	"fmt"
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/relay/channel"
+	"github.com/songquanpeng/one-api/relay/constant"
+	"github.com/songquanpeng/one-api/relay/model"
+	"github.com/songquanpeng/one-api/relay/util"
+	"io"
+	"net/http"
+)
+
+type Adaptor struct {
+}
+
+func (a *Adaptor) Init(meta *util.RelayMeta) {
+
+}
+
+func (a *Adaptor) GetRequestURL(meta *util.RelayMeta) (string, error) {
+	// https://github.com/ollama/ollama/blob/main/docs/api.md
+	fullRequestURL := fmt.Sprintf("%s/api/chat", meta.BaseURL)
+	return fullRequestURL, nil
+}
+
+func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *util.RelayMeta) error {
+	channel.SetupCommonRequestHeader(c, req, meta)
+	req.Header.Set("Authorization", "Bearer "+meta.APIKey)
+	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")
+	}
+	switch relayMode {
+	case constant.RelayModeEmbeddings:
+		return nil, errors.New("not supported")
+	default:
+		return ConvertRequest(*request), nil
+	}
+}
+
+func (a *Adaptor) DoRequest(c *gin.Context, meta *util.RelayMeta, requestBody io.Reader) (*http.Response, error) {
+	return channel.DoRequestHelper(a, c, meta, requestBody)
+}
+
+func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *util.RelayMeta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
+	if meta.IsStream {
+		err, usage = StreamHandler(c, resp)
+	} else {
+		err, usage = Handler(c, resp)
+	}
+	return
+}
+
+func (a *Adaptor) GetModelList() []string {
+	return ModelList
+}
+
+func (a *Adaptor) GetChannelName() string {
+	return "ollama"
+}

relay/channel/ollama/constants.go

@@ -0,0 +1,5 @@
+package ollama
+
+var ModelList = []string{
+	"qwen:0.5b-chat",
+}

relay/channel/ollama/main.go

@@ -0,0 +1,178 @@
+package ollama
+
+import (
+	"bufio"
+	"context"
+	"encoding/json"
+	"fmt"
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/helper"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/relay/channel/openai"
+	"github.com/songquanpeng/one-api/relay/constant"
+	"github.com/songquanpeng/one-api/relay/model"
+	"io"
+	"net/http"
+	"strings"
+)
+
+func ConvertRequest(request model.GeneralOpenAIRequest) *ChatRequest {
+	ollamaRequest := ChatRequest{
+		Model: request.Model,
+		Options: &Options{
+			Seed:             int(request.Seed),
+			Temperature:      request.Temperature,
+			TopP:             request.TopP,
+			FrequencyPenalty: request.FrequencyPenalty,
+			PresencePenalty:  request.PresencePenalty,
+		},
+		Stream: request.Stream,
+	}
+	for _, message := range request.Messages {
+		ollamaRequest.Messages = append(ollamaRequest.Messages, Message{
+			Role:    message.Role,
+			Content: message.StringContent(),
+		})
+	}
+	return &ollamaRequest
+}
+
+func responseOllama2OpenAI(response *ChatResponse) *openai.TextResponse {
+	choice := openai.TextResponseChoice{
+		Index: 0,
+		Message: model.Message{
+			Role:    response.Message.Role,
+			Content: response.Message.Content,
+		},
+	}
+	if response.Done {
+		choice.FinishReason = "stop"
+	}
+	fullTextResponse := openai.TextResponse{
+		Id:      fmt.Sprintf("chatcmpl-%s", helper.GetUUID()),
+		Object:  "chat.completion",
+		Created: helper.GetTimestamp(),
+		Choices: []openai.TextResponseChoice{choice},
+		Usage: model.Usage{
+			PromptTokens:     response.PromptEvalCount,
+			CompletionTokens: response.EvalCount,
+			TotalTokens:      response.PromptEvalCount + response.EvalCount,
+		},
+	}
+	return &fullTextResponse
+}
+
+func streamResponseOllama2OpenAI(ollamaResponse *ChatResponse) *openai.ChatCompletionsStreamResponse {
+	var choice openai.ChatCompletionsStreamResponseChoice
+	choice.Delta.Role = ollamaResponse.Message.Role
+	choice.Delta.Content = ollamaResponse.Message.Content
+	if ollamaResponse.Done {
+		choice.FinishReason = &constant.StopFinishReason
+	}
+	response := openai.ChatCompletionsStreamResponse{
+		Id:      fmt.Sprintf("chatcmpl-%s", helper.GetUUID()),
+		Object:  "chat.completion.chunk",
+		Created: helper.GetTimestamp(),
+		Model:   ollamaResponse.Model,
+		Choices: []openai.ChatCompletionsStreamResponseChoice{choice},
+	}
+	return &response
+}
+
+func StreamHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, *model.Usage) {
+	var usage model.Usage
+	scanner := bufio.NewScanner(resp.Body)
+	scanner.Split(func(data []byte, atEOF bool) (advance int, token []byte, err error) {
+		if atEOF && len(data) == 0 {
+			return 0, nil, nil
+		}
+		if i := strings.Index(string(data), "}\n"); i >= 0 {
+			return i + 2, data[0:i], nil
+		}
+		if atEOF {
+			return len(data), data, nil
+		}
+		return 0, nil, nil
+	})
+	dataChan := make(chan string)
+	stopChan := make(chan bool)
+	go func() {
+		for scanner.Scan() {
+			data := strings.TrimPrefix(scanner.Text(), "}")
+			dataChan <- data + "}"
+		}
+		stopChan <- true
+	}()
+	common.SetEventStreamHeaders(c)
+	c.Stream(func(w io.Writer) bool {
+		select {
+		case data := <-dataChan:
+			var ollamaResponse ChatResponse
+			err := json.Unmarshal([]byte(data), &ollamaResponse)
+			if err != nil {
+				logger.SysError("error unmarshalling stream response: " + err.Error())
+				return true
+			}
+			if ollamaResponse.EvalCount != 0 {
+				usage.PromptTokens = ollamaResponse.PromptEvalCount
+				usage.CompletionTokens = ollamaResponse.EvalCount
+				usage.TotalTokens = ollamaResponse.PromptEvalCount + ollamaResponse.EvalCount
+			}
+			response := streamResponseOllama2OpenAI(&ollamaResponse)
+			jsonResponse, 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(jsonResponse)})
+			return true
+		case <-stopChan:
+			c.Render(-1, common.CustomEvent{Data: "data: [DONE]"})
+			return false
+		}
+	})
+	err := resp.Body.Close()
+	if err != nil {
+		return openai.ErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), nil
+	}
+	return nil, &usage
+}
+
+func Handler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, *model.Usage) {
+	ctx := context.TODO()
+	var ollamaResponse ChatResponse
+	responseBody, err := io.ReadAll(resp.Body)
+	if err != nil {
+		return openai.ErrorWrapper(err, "read_response_body_failed", http.StatusInternalServerError), nil
+	}
+	logger.Debugf(ctx, "ollama response: %s", string(responseBody))
+	err = resp.Body.Close()
+	if err != nil {
+		return openai.ErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), nil
+	}
+	err = json.Unmarshal(responseBody, &ollamaResponse)
+	if err != nil {
+		return openai.ErrorWrapper(err, "unmarshal_response_body_failed", http.StatusInternalServerError), nil
+	}
+	if ollamaResponse.Error != "" {
+		return &model.ErrorWithStatusCode{
+			Error: model.Error{
+				Message: ollamaResponse.Error,
+				Type:    "ollama_error",
+				Param:   "",
+				Code:    "ollama_error",
+			},
+			StatusCode: resp.StatusCode,
+		}, nil
+	}
+	fullTextResponse := responseOllama2OpenAI(&ollamaResponse)
+	jsonResponse, err := json.Marshal(fullTextResponse)
+	if err != nil {
+		return openai.ErrorWrapper(err, "marshal_response_body_failed", http.StatusInternalServerError), nil
+	}
+	c.Writer.Header().Set("Content-Type", "application/json")
+	c.Writer.WriteHeader(resp.StatusCode)
+	_, err = c.Writer.Write(jsonResponse)
+	return nil, &fullTextResponse.Usage
+}