File size: 15,272 Bytes
4674012 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 |
package controller
import (
"bytes"
"fmt"
"io"
"log"
"net/http"
"strings"
"github.com/QuantumNous/new-api/common"
"github.com/QuantumNous/new-api/constant"
"github.com/QuantumNous/new-api/dto"
"github.com/QuantumNous/new-api/logger"
"github.com/QuantumNous/new-api/middleware"
"github.com/QuantumNous/new-api/model"
"github.com/QuantumNous/new-api/relay"
relaycommon "github.com/QuantumNous/new-api/relay/common"
relayconstant "github.com/QuantumNous/new-api/relay/constant"
"github.com/QuantumNous/new-api/relay/helper"
"github.com/QuantumNous/new-api/service"
"github.com/QuantumNous/new-api/setting"
"github.com/QuantumNous/new-api/types"
"github.com/bytedance/gopkg/util/gopool"
"github.com/gin-gonic/gin"
"github.com/gorilla/websocket"
)
func relayHandler(c *gin.Context, info *relaycommon.RelayInfo) *types.NewAPIError {
var err *types.NewAPIError
switch info.RelayMode {
case relayconstant.RelayModeImagesGenerations, relayconstant.RelayModeImagesEdits:
err = relay.ImageHelper(c, info)
case relayconstant.RelayModeAudioSpeech:
fallthrough
case relayconstant.RelayModeAudioTranslation:
fallthrough
case relayconstant.RelayModeAudioTranscription:
err = relay.AudioHelper(c, info)
case relayconstant.RelayModeRerank:
err = relay.RerankHelper(c, info)
case relayconstant.RelayModeEmbeddings:
err = relay.EmbeddingHelper(c, info)
case relayconstant.RelayModeResponses:
err = relay.ResponsesHelper(c, info)
default:
err = relay.TextHelper(c, info)
}
return err
}
func geminiRelayHandler(c *gin.Context, info *relaycommon.RelayInfo) *types.NewAPIError {
var err *types.NewAPIError
if strings.Contains(c.Request.URL.Path, "embed") {
err = relay.GeminiEmbeddingHandler(c, info)
} else {
err = relay.GeminiHelper(c, info)
}
return err
}
func Relay(c *gin.Context, relayFormat types.RelayFormat) {
requestId := c.GetString(common.RequestIdKey)
group := common.GetContextKeyString(c, constant.ContextKeyUsingGroup)
originalModel := common.GetContextKeyString(c, constant.ContextKeyOriginalModel)
var (
newAPIError *types.NewAPIError
ws *websocket.Conn
)
if relayFormat == types.RelayFormatOpenAIRealtime {
var err error
ws, err = upgrader.Upgrade(c.Writer, c.Request, nil)
if err != nil {
helper.WssError(c, ws, types.NewError(err, types.ErrorCodeGetChannelFailed, types.ErrOptionWithSkipRetry()).ToOpenAIError())
return
}
defer ws.Close()
}
defer func() {
if newAPIError != nil {
logger.LogError(c, fmt.Sprintf("relay error: %s", newAPIError.Error()))
newAPIError.SetMessage(common.MessageWithRequestId(newAPIError.Error(), requestId))
switch relayFormat {
case types.RelayFormatOpenAIRealtime:
helper.WssError(c, ws, newAPIError.ToOpenAIError())
case types.RelayFormatClaude:
c.JSON(newAPIError.StatusCode, gin.H{
"type": "error",
"error": newAPIError.ToClaudeError(),
})
default:
c.JSON(newAPIError.StatusCode, gin.H{
"error": newAPIError.ToOpenAIError(),
})
}
}
}()
request, err := helper.GetAndValidateRequest(c, relayFormat)
if err != nil {
newAPIError = types.NewError(err, types.ErrorCodeInvalidRequest)
return
}
relayInfo, err := relaycommon.GenRelayInfo(c, relayFormat, request, ws)
if err != nil {
newAPIError = types.NewError(err, types.ErrorCodeGenRelayInfoFailed)
return
}
meta := request.GetTokenCountMeta()
if setting.ShouldCheckPromptSensitive() {
contains, words := service.CheckSensitiveText(meta.CombineText)
if contains {
logger.LogWarn(c, fmt.Sprintf("user sensitive words detected: %s", strings.Join(words, ", ")))
newAPIError = types.NewError(err, types.ErrorCodeSensitiveWordsDetected)
return
}
}
tokens, err := service.CountRequestToken(c, meta, relayInfo)
if err != nil {
newAPIError = types.NewError(err, types.ErrorCodeCountTokenFailed)
return
}
relayInfo.SetPromptTokens(tokens)
priceData, err := helper.ModelPriceHelper(c, relayInfo, tokens, meta)
if err != nil {
newAPIError = types.NewError(err, types.ErrorCodeModelPriceError)
return
}
// common.SetContextKey(c, constant.ContextKeyTokenCountMeta, meta)
if priceData.FreeModel {
logger.LogInfo(c, fmt.Sprintf("模型 %s 免费,跳过预扣费", relayInfo.OriginModelName))
} else {
newAPIError = service.PreConsumeQuota(c, priceData.QuotaToPreConsume, relayInfo)
if newAPIError != nil {
return
}
}
defer func() {
// Only return quota if downstream failed and quota was actually pre-consumed
if newAPIError != nil && relayInfo.FinalPreConsumedQuota != 0 {
service.ReturnPreConsumedQuota(c, relayInfo)
}
}()
for i := 0; i <= common.RetryTimes; i++ {
channel, err := getChannel(c, group, originalModel, i)
if err != nil {
logger.LogError(c, err.Error())
newAPIError = err
break
}
addUsedChannel(c, channel.Id)
requestBody, _ := common.GetRequestBody(c)
c.Request.Body = io.NopCloser(bytes.NewBuffer(requestBody))
switch relayFormat {
case types.RelayFormatOpenAIRealtime:
newAPIError = relay.WssHelper(c, relayInfo)
case types.RelayFormatClaude:
newAPIError = relay.ClaudeHelper(c, relayInfo)
case types.RelayFormatGemini:
newAPIError = geminiRelayHandler(c, relayInfo)
default:
newAPIError = relayHandler(c, relayInfo)
}
if newAPIError == nil {
return
}
processChannelError(c, *types.NewChannelError(channel.Id, channel.Type, channel.Name, channel.ChannelInfo.IsMultiKey, common.GetContextKeyString(c, constant.ContextKeyChannelKey), channel.GetAutoBan()), newAPIError)
if !shouldRetry(c, newAPIError, common.RetryTimes-i) {
break
}
}
useChannel := c.GetStringSlice("use_channel")
if len(useChannel) > 1 {
retryLogStr := fmt.Sprintf("重试:%s", strings.Trim(strings.Join(strings.Fields(fmt.Sprint(useChannel)), "->"), "[]"))
logger.LogInfo(c, retryLogStr)
}
}
var upgrader = websocket.Upgrader{
Subprotocols: []string{"realtime"}, // WS 握手支持的协议,如果有使用 Sec-WebSocket-Protocol,则必须在此声明对应的 Protocol TODO add other protocol
CheckOrigin: func(r *http.Request) bool {
return true // 允许跨域
},
}
func addUsedChannel(c *gin.Context, channelId int) {
useChannel := c.GetStringSlice("use_channel")
useChannel = append(useChannel, fmt.Sprintf("%d", channelId))
c.Set("use_channel", useChannel)
}
func getChannel(c *gin.Context, group, originalModel string, retryCount int) (*model.Channel, *types.NewAPIError) {
if retryCount == 0 {
autoBan := c.GetBool("auto_ban")
autoBanInt := 1
if !autoBan {
autoBanInt = 0
}
return &model.Channel{
Id: c.GetInt("channel_id"),
Type: c.GetInt("channel_type"),
Name: c.GetString("channel_name"),
AutoBan: &autoBanInt,
}, nil
}
channel, selectGroup, err := service.CacheGetRandomSatisfiedChannel(c, group, originalModel, retryCount)
if err != nil {
return nil, types.NewError(fmt.Errorf("获取分组 %s 下模型 %s 的可用渠道失败(retry): %s", selectGroup, originalModel, err.Error()), types.ErrorCodeGetChannelFailed, types.ErrOptionWithSkipRetry())
}
if channel == nil {
return nil, types.NewError(fmt.Errorf("分组 %s 下模型 %s 的可用渠道不存在(retry)", selectGroup, originalModel), types.ErrorCodeGetChannelFailed, types.ErrOptionWithSkipRetry())
}
newAPIError := middleware.SetupContextForSelectedChannel(c, channel, originalModel)
if newAPIError != nil {
return nil, newAPIError
}
return channel, nil
}
func shouldRetry(c *gin.Context, openaiErr *types.NewAPIError, retryTimes int) bool {
if openaiErr == nil {
return false
}
if types.IsChannelError(openaiErr) {
return true
}
if types.IsSkipRetryError(openaiErr) {
return false
}
if retryTimes <= 0 {
return false
}
if _, ok := c.Get("specific_channel_id"); ok {
return false
}
if openaiErr.StatusCode == http.StatusTooManyRequests {
return true
}
if openaiErr.StatusCode == 307 {
return true
}
if openaiErr.StatusCode/100 == 5 {
// 超时不重试
if openaiErr.StatusCode == 504 || openaiErr.StatusCode == 524 {
return false
}
return true
}
if openaiErr.StatusCode == http.StatusBadRequest {
return false
}
if openaiErr.StatusCode == 408 {
// azure处理超时不重试
return false
}
if openaiErr.StatusCode/100 == 2 {
return false
}
return true
}
func processChannelError(c *gin.Context, channelError types.ChannelError, err *types.NewAPIError) {
logger.LogError(c, fmt.Sprintf("channel error (channel #%d, status code: %d): %s", channelError.ChannelId, err.StatusCode, err.Error()))
// 不要使用context获取渠道信息,异步处理时可能会出现渠道信息不一致的情况
// do not use context to get channel info, there may be inconsistent channel info when processing asynchronously
if service.ShouldDisableChannel(channelError.ChannelId, err) && channelError.AutoBan {
gopool.Go(func() {
service.DisableChannel(channelError, err.Error())
})
}
if constant.ErrorLogEnabled && types.IsRecordErrorLog(err) {
// 保存错误日志到mysql中
userId := c.GetInt("id")
tokenName := c.GetString("token_name")
modelName := c.GetString("original_model")
tokenId := c.GetInt("token_id")
userGroup := c.GetString("group")
channelId := c.GetInt("channel_id")
other := make(map[string]interface{})
if c.Request != nil && c.Request.URL != nil {
other["request_path"] = c.Request.URL.Path
}
other["error_type"] = err.GetErrorType()
other["error_code"] = err.GetErrorCode()
other["status_code"] = err.StatusCode
other["channel_id"] = channelId
other["channel_name"] = c.GetString("channel_name")
other["channel_type"] = c.GetInt("channel_type")
adminInfo := make(map[string]interface{})
adminInfo["use_channel"] = c.GetStringSlice("use_channel")
isMultiKey := common.GetContextKeyBool(c, constant.ContextKeyChannelIsMultiKey)
if isMultiKey {
adminInfo["is_multi_key"] = true
adminInfo["multi_key_index"] = common.GetContextKeyInt(c, constant.ContextKeyChannelMultiKeyIndex)
}
other["admin_info"] = adminInfo
model.RecordErrorLog(c, userId, channelId, modelName, tokenName, err.MaskSensitiveError(), tokenId, 0, false, userGroup, other)
}
}
func RelayMidjourney(c *gin.Context) {
relayInfo, err := relaycommon.GenRelayInfo(c, types.RelayFormatMjProxy, nil, nil)
if err != nil {
c.JSON(http.StatusInternalServerError, gin.H{
"description": fmt.Sprintf("failed to generate relay info: %s", err.Error()),
"type": "upstream_error",
"code": 4,
})
return
}
var mjErr *dto.MidjourneyResponse
switch relayInfo.RelayMode {
case relayconstant.RelayModeMidjourneyNotify:
mjErr = relay.RelayMidjourneyNotify(c)
case relayconstant.RelayModeMidjourneyTaskFetch, relayconstant.RelayModeMidjourneyTaskFetchByCondition:
mjErr = relay.RelayMidjourneyTask(c, relayInfo.RelayMode)
case relayconstant.RelayModeMidjourneyTaskImageSeed:
mjErr = relay.RelayMidjourneyTaskImageSeed(c)
case relayconstant.RelayModeSwapFace:
mjErr = relay.RelaySwapFace(c, relayInfo)
default:
mjErr = relay.RelayMidjourneySubmit(c, relayInfo)
}
//err = relayMidjourneySubmit(c, relayMode)
log.Println(mjErr)
if mjErr != nil {
statusCode := http.StatusBadRequest
if mjErr.Code == 30 {
mjErr.Result = "当前分组负载已饱和,请稍后再试,或升级账户以提升服务质量。"
statusCode = http.StatusTooManyRequests
}
c.JSON(statusCode, gin.H{
"description": fmt.Sprintf("%s %s", mjErr.Description, mjErr.Result),
"type": "upstream_error",
"code": mjErr.Code,
})
channelId := c.GetInt("channel_id")
logger.LogError(c, fmt.Sprintf("relay error (channel #%d, status code %d): %s", channelId, statusCode, fmt.Sprintf("%s %s", mjErr.Description, mjErr.Result)))
}
}
func RelayNotImplemented(c *gin.Context) {
err := dto.OpenAIError{
Message: "API not implemented",
Type: "new_api_error",
Param: "",
Code: "api_not_implemented",
}
c.JSON(http.StatusNotImplemented, gin.H{
"error": err,
})
}
func RelayNotFound(c *gin.Context) {
err := dto.OpenAIError{
Message: fmt.Sprintf("Invalid URL (%s %s)", c.Request.Method, c.Request.URL.Path),
Type: "invalid_request_error",
Param: "",
Code: "",
}
c.JSON(http.StatusNotFound, gin.H{
"error": err,
})
}
func RelayTask(c *gin.Context) {
retryTimes := common.RetryTimes
channelId := c.GetInt("channel_id")
group := c.GetString("group")
originalModel := c.GetString("original_model")
c.Set("use_channel", []string{fmt.Sprintf("%d", channelId)})
relayInfo, err := relaycommon.GenRelayInfo(c, types.RelayFormatTask, nil, nil)
if err != nil {
return
}
taskErr := taskRelayHandler(c, relayInfo)
if taskErr == nil {
retryTimes = 0
}
for i := 0; shouldRetryTaskRelay(c, channelId, taskErr, retryTimes) && i < retryTimes; i++ {
channel, newAPIError := getChannel(c, group, originalModel, i)
if newAPIError != nil {
logger.LogError(c, fmt.Sprintf("CacheGetRandomSatisfiedChannel failed: %s", newAPIError.Error()))
taskErr = service.TaskErrorWrapperLocal(newAPIError.Err, "get_channel_failed", http.StatusInternalServerError)
break
}
channelId = channel.Id
useChannel := c.GetStringSlice("use_channel")
useChannel = append(useChannel, fmt.Sprintf("%d", channelId))
c.Set("use_channel", useChannel)
logger.LogInfo(c, fmt.Sprintf("using channel #%d to retry (remain times %d)", channel.Id, i))
//middleware.SetupContextForSelectedChannel(c, channel, originalModel)
requestBody, _ := common.GetRequestBody(c)
c.Request.Body = io.NopCloser(bytes.NewBuffer(requestBody))
taskErr = taskRelayHandler(c, relayInfo)
}
useChannel := c.GetStringSlice("use_channel")
if len(useChannel) > 1 {
retryLogStr := fmt.Sprintf("重试:%s", strings.Trim(strings.Join(strings.Fields(fmt.Sprint(useChannel)), "->"), "[]"))
logger.LogInfo(c, retryLogStr)
}
if taskErr != nil {
if taskErr.StatusCode == http.StatusTooManyRequests {
taskErr.Message = "当前分组上游负载已饱和,请稍后再试"
}
c.JSON(taskErr.StatusCode, taskErr)
}
}
func taskRelayHandler(c *gin.Context, relayInfo *relaycommon.RelayInfo) *dto.TaskError {
var err *dto.TaskError
switch relayInfo.RelayMode {
case relayconstant.RelayModeSunoFetch, relayconstant.RelayModeSunoFetchByID, relayconstant.RelayModeVideoFetchByID:
err = relay.RelayTaskFetch(c, relayInfo.RelayMode)
default:
err = relay.RelayTaskSubmit(c, relayInfo)
}
return err
}
func shouldRetryTaskRelay(c *gin.Context, channelId int, taskErr *dto.TaskError, retryTimes int) bool {
if taskErr == nil {
return false
}
if retryTimes <= 0 {
return false
}
if _, ok := c.Get("specific_channel_id"); ok {
return false
}
if taskErr.StatusCode == http.StatusTooManyRequests {
return true
}
if taskErr.StatusCode == 307 {
return true
}
if taskErr.StatusCode/100 == 5 {
// 超时不重试
if taskErr.StatusCode == 504 || taskErr.StatusCode == 524 {
return false
}
return true
}
if taskErr.StatusCode == http.StatusBadRequest {
return false
}
if taskErr.StatusCode == 408 {
// azure处理超时不重试
return false
}
if taskErr.LocalError {
return false
}
if taskErr.StatusCode/100 == 2 {
return false
}
return true
}
|