Files
new-api/relay/channel/claude/relay-claude.go

412 lines
13 KiB
Go
Raw Normal View History

2024-02-29 01:08:18 +08:00
package claude
2023-07-22 16:18:03 +08:00
2023-07-22 17:12:13 +08:00
import (
2023-07-22 17:36:40 +08:00
"bufio"
"encoding/json"
2023-07-22 17:12:13 +08:00
"fmt"
2023-07-22 17:36:40 +08:00
"github.com/gin-gonic/gin"
"io"
"net/http"
2023-07-22 17:12:13 +08:00
"one-api/common"
"one-api/constant"
2024-02-29 01:08:18 +08:00
"one-api/dto"
relaycommon "one-api/relay/common"
2024-02-29 01:08:18 +08:00
"one-api/service"
2023-07-22 17:12:13 +08:00
"strings"
"time"
2023-07-22 17:12:13 +08:00
)
2023-07-22 16:18:03 +08:00
func stopReasonClaude2OpenAI(reason string) string {
switch reason {
case "stop_sequence":
return "stop"
2024-03-08 19:43:33 +08:00
case "end_turn":
return "stop"
2023-07-22 16:18:03 +08:00
case "max_tokens":
2024-04-24 18:49:56 +08:00
return "max_tokens"
2023-07-22 16:18:03 +08:00
default:
return reason
}
}
2023-07-22 17:12:13 +08:00
2024-04-23 11:44:40 +08:00
func RequestOpenAI2ClaudeComplete(textRequest dto.GeneralOpenAIRequest) *ClaudeRequest {
2023-07-22 17:12:13 +08:00
claudeRequest := ClaudeRequest{
2024-04-23 12:19:23 +08:00
Model: textRequest.Model,
Prompt: "",
StopSequences: nil,
Temperature: textRequest.Temperature,
TopP: textRequest.TopP,
TopK: textRequest.TopK,
Stream: textRequest.Stream,
2023-07-22 17:12:13 +08:00
}
2024-04-28 20:45:34 +08:00
if claudeRequest.MaxTokensToSample == 0 {
claudeRequest.MaxTokensToSample = 4096
2023-07-22 17:12:13 +08:00
}
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()
}
2023-07-22 17:12:13 +08:00
}
}
prompt += "\n\nAssistant:"
2023-07-22 17:12:13 +08:00
claudeRequest.Prompt = prompt
return &claudeRequest
}
2024-04-23 11:44:40 +08:00
func RequestOpenAI2ClaudeMessage(textRequest dto.GeneralOpenAIRequest) (*ClaudeRequest, error) {
2024-03-08 18:25:57 +08:00
claudeRequest := ClaudeRequest{
Model: textRequest.Model,
MaxTokens: textRequest.MaxTokens,
StopSequences: nil,
Temperature: textRequest.Temperature,
TopP: textRequest.TopP,
2024-04-01 14:33:58 +08:00
TopK: textRequest.TopK,
2024-03-08 18:25:57 +08:00
Stream: textRequest.Stream,
}
if claudeRequest.MaxTokens == 0 {
claudeRequest.MaxTokens = 4096
}
formatMessages := make([]dto.Message, 0)
var lastMessage *dto.Message
2024-04-25 20:37:50 +08:00
for i, message := range textRequest.Messages {
2024-04-25 16:04:53 +08:00
//if message.Role == "system" {
// if i != 0 {
// message.Role = "user"
// }
//}
if message.Role == "" {
2024-04-25 22:57:11 +08:00
textRequest.Messages[i].Role = "user"
}
fmtMessage := dto.Message{
Role: message.Role,
Content: message.Content,
}
if lastMessage != nil && lastMessage.Role == message.Role {
if lastMessage.IsStringContent() && message.IsStringContent() {
content, _ := json.Marshal(strings.Trim(fmt.Sprintf("%s %s", lastMessage.StringContent(), message.StringContent()), "\""))
fmtMessage.Content = content
// delete last message
formatMessages = formatMessages[:len(formatMessages)-1]
}
}
if fmtMessage.Content == nil {
content, _ := json.Marshal("...")
fmtMessage.Content = content
}
formatMessages = append(formatMessages, fmtMessage)
2024-04-25 20:37:50 +08:00
lastMessage = &textRequest.Messages[i]
}
2024-03-08 18:25:57 +08:00
claudeMessages := make([]ClaudeMessage, 0)
for _, message := range formatMessages {
2024-03-08 18:25:57 +08:00
if message.Role == "system" {
2024-04-25 20:37:50 +08:00
if message.IsStringContent() {
claudeRequest.System = message.StringContent()
} else {
contents := message.ParseContent()
content := ""
for _, ctx := range contents {
if ctx.Type == "text" {
content += ctx.Text
}
}
claudeRequest.System = content
}
2024-03-08 18:25:57 +08:00
} else {
claudeMessage := ClaudeMessage{
Role: message.Role,
}
if message.IsStringContent() {
claudeMessage.Content = message.StringContent()
} else {
claudeMediaMessages := make([]ClaudeMediaMessage, 0)
for _, mediaMessage := range message.ParseContent() {
claudeMediaMessage := ClaudeMediaMessage{
Type: mediaMessage.Type,
}
if mediaMessage.Type == "text" {
claudeMediaMessage.Text = mediaMessage.Text
} else {
imageUrl := mediaMessage.ImageUrl.(dto.MessageImageUrl)
claudeMediaMessage.Type = "image"
claudeMediaMessage.Source = &ClaudeMessageSource{
Type: "base64",
}
// 判断是否是url
if strings.HasPrefix(imageUrl.Url, "http") {
// 是url获取图片的类型和base64编码的数据
mimeType, data, _ := service.GetImageFromUrl(imageUrl.Url)
2024-03-08 18:25:57 +08:00
claudeMediaMessage.Source.MediaType = mimeType
claudeMediaMessage.Source.Data = data
} else {
_, format, base64String, err := service.DecodeBase64ImageData(imageUrl.Url)
2024-03-08 18:25:57 +08:00
if err != nil {
return nil, err
}
claudeMediaMessage.Source.MediaType = "image/" + format
claudeMediaMessage.Source.Data = base64String
}
}
claudeMediaMessages = append(claudeMediaMessages, claudeMediaMessage)
}
claudeMessage.Content = claudeMediaMessages
}
claudeMessages = append(claudeMessages, claudeMessage)
}
}
claudeRequest.Prompt = ""
claudeRequest.Messages = claudeMessages
return &claudeRequest, nil
}
2024-04-23 11:44:40 +08:00
func StreamResponseClaude2OpenAI(reqMode int, claudeResponse *ClaudeResponse) (*dto.ChatCompletionsStreamResponse, *ClaudeUsage) {
2024-02-29 01:08:18 +08:00
var response dto.ChatCompletionsStreamResponse
2024-03-08 19:43:33 +08:00
var claudeUsage *ClaudeUsage
2023-07-22 17:12:13 +08:00
response.Object = "chat.completion.chunk"
response.Model = claudeResponse.Model
2024-03-08 19:43:33 +08:00
response.Choices = make([]dto.ChatCompletionsStreamResponseChoice, 0)
var choice dto.ChatCompletionsStreamResponseChoice
if reqMode == RequestModeCompletion {
2024-04-25 23:57:39 +08:00
choice.Delta.SetContentString(claudeResponse.Completion)
2024-03-08 19:43:33 +08:00
finishReason := stopReasonClaude2OpenAI(claudeResponse.StopReason)
if finishReason != "null" {
choice.FinishReason = &finishReason
}
} else {
if claudeResponse.Type == "message_start" {
response.Id = claudeResponse.Message.Id
response.Model = claudeResponse.Message.Model
claudeUsage = &claudeResponse.Message.Usage
2024-04-25 23:57:39 +08:00
choice.Delta.SetContentString("")
choice.Delta.Role = "assistant"
2024-04-26 02:56:35 +08:00
} else if claudeResponse.Type == "content_block_start" {
return nil, nil
2024-03-08 19:43:33 +08:00
} else if claudeResponse.Type == "content_block_delta" {
choice.Index = claudeResponse.Index
2024-04-25 23:57:39 +08:00
choice.Delta.SetContentString(claudeResponse.Delta.Text)
2024-03-08 19:43:33 +08:00
} else if claudeResponse.Type == "message_delta" {
finishReason := stopReasonClaude2OpenAI(*claudeResponse.Delta.StopReason)
if finishReason != "null" {
choice.FinishReason = &finishReason
}
claudeUsage = &claudeResponse.Usage
2024-04-23 11:44:40 +08:00
} else if claudeResponse.Type == "message_stop" {
return nil, nil
2024-04-25 23:57:39 +08:00
} else {
return nil, nil
2024-03-08 19:43:33 +08:00
}
}
2024-03-20 21:28:45 +08:00
if claudeUsage == nil {
claudeUsage = &ClaudeUsage{}
}
2024-03-08 19:43:33 +08:00
response.Choices = append(response.Choices, choice)
2024-04-25 23:57:39 +08:00
2024-03-08 19:43:33 +08:00
return &response, claudeUsage
2023-07-22 17:12:13 +08:00
}
2024-04-23 11:44:40 +08:00
func ResponseClaude2OpenAI(reqMode int, claudeResponse *ClaudeResponse) *dto.OpenAITextResponse {
2024-03-08 18:25:57 +08:00
choices := make([]dto.OpenAITextResponseChoice, 0)
2024-02-29 01:08:18 +08:00
fullTextResponse := dto.OpenAITextResponse{
2023-07-22 17:12:13 +08:00
Id: fmt.Sprintf("chatcmpl-%s", common.GetUUID()),
Object: "chat.completion",
Created: common.GetTimestamp(),
}
2024-03-08 18:25:57 +08:00
if reqMode == RequestModeCompletion {
content, _ := json.Marshal(strings.TrimPrefix(claudeResponse.Completion, " "))
choice := dto.OpenAITextResponseChoice{
Index: 0,
Message: dto.Message{
Role: "assistant",
Content: content,
Name: nil,
},
FinishReason: stopReasonClaude2OpenAI(claudeResponse.StopReason),
}
choices = append(choices, choice)
} else {
fullTextResponse.Id = claudeResponse.Id
for i, message := range claudeResponse.Content {
content, _ := json.Marshal(message.Text)
choice := dto.OpenAITextResponseChoice{
Index: i,
Message: dto.Message{
Role: "assistant",
Content: content,
},
FinishReason: stopReasonClaude2OpenAI(claudeResponse.StopReason),
}
choices = append(choices, choice)
}
}
fullTextResponse.Choices = choices
2023-07-22 17:12:13 +08:00
return &fullTextResponse
}
2023-07-22 17:36:40 +08:00
func claudeStreamHandler(c *gin.Context, resp *http.Response, info *relaycommon.RelayInfo, requestMode int) (*dto.OpenAIErrorWithStatusCode, *dto.Usage) {
2023-07-22 17:36:40 +08:00
responseId := fmt.Sprintf("chatcmpl-%s", common.GetUUID())
2024-03-20 17:07:42 +08:00
var usage *dto.Usage
2024-03-20 21:32:33 +08:00
usage = &dto.Usage{}
2024-03-08 19:43:33 +08:00
responseText := ""
2023-07-22 17:36:40 +08:00
createdTime := common.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
}
2024-03-08 19:43:33 +08:00
if i := strings.Index(string(data), "\n"); i >= 0 {
return i + 1, data[0:i], nil
2023-07-22 17:36:40 +08:00
}
if atEOF {
return len(data), data, nil
}
return 0, nil, nil
})
dataChan := make(chan string, 5)
stopChan := make(chan bool, 2)
2023-07-22 17:36:40 +08:00
go func() {
for scanner.Scan() {
data := scanner.Text()
2024-03-08 19:43:33 +08:00
if !strings.HasPrefix(data, "data: ") {
2023-07-22 17:36:40 +08:00
continue
}
2024-03-08 19:43:33 +08:00
data = strings.TrimPrefix(data, "data: ")
if !common.SafeSendStringTimeout(dataChan, data, constant.StreamingTimeout) {
// send data timeout, stop the stream
common.LogError(c, "send data timeout, stop the stream")
break
}
2023-07-22 17:36:40 +08:00
}
stopChan <- true
}()
isFirst := true
2024-02-29 01:08:18 +08:00
service.SetEventStreamHeaders(c)
2023-07-22 17:36:40 +08:00
c.Stream(func(w io.Writer) bool {
select {
case data := <-dataChan:
if isFirst {
isFirst = false
info.FirstResponseTime = time.Now()
}
2023-07-22 17:36:40 +08:00
// some implementations may add \r at the end of data
data = strings.TrimSuffix(data, "\r")
var claudeResponse ClaudeResponse
err := json.Unmarshal([]byte(data), &claudeResponse)
if err != nil {
common.SysError("error unmarshalling stream response: " + err.Error())
return true
}
2024-03-08 19:43:33 +08:00
2024-04-23 11:44:40 +08:00
response, claudeUsage := StreamResponseClaude2OpenAI(requestMode, &claudeResponse)
if response == nil {
return true
}
2024-03-08 19:43:33 +08:00
if requestMode == RequestModeCompletion {
responseText += claudeResponse.Completion
responseId = response.Id
} else {
if claudeResponse.Type == "message_start" {
// message_start, 获取usage
responseId = claudeResponse.Message.Id
info.UpstreamModelName = claudeResponse.Message.Model
2024-03-08 19:43:33 +08:00
usage.PromptTokens = claudeUsage.InputTokens
} else if claudeResponse.Type == "content_block_delta" {
responseText += claudeResponse.Delta.Text
} else if claudeResponse.Type == "message_delta" {
usage.CompletionTokens = claudeUsage.OutputTokens
usage.TotalTokens = claudeUsage.InputTokens + claudeUsage.OutputTokens
} else {
return true
}
}
//response.Id = responseId
2023-07-22 17:36:40 +08:00
response.Id = responseId
response.Created = createdTime
response.Model = info.UpstreamModelName
2024-03-08 19:43:33 +08:00
2023-07-22 17:36:40 +08:00
jsonStr, err := json.Marshal(response)
if err != nil {
common.SysError("error marshalling stream response: " + err.Error())
return true
}
c.Render(-1, common.CustomEvent{Data: "data: " + string(jsonStr)})
return true
case <-stopChan:
c.Render(-1, common.CustomEvent{Data: "data: [DONE]"})
return false
}
})
err := resp.Body.Close()
if err != nil {
2024-03-08 19:43:33 +08:00
return service.OpenAIErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), nil
}
if requestMode == RequestModeCompletion {
usage, _ = service.ResponseText2Usage(responseText, info.UpstreamModelName, info.PromptTokens)
} else {
2024-06-15 01:30:51 +08:00
if usage.PromptTokens == 0 {
usage.PromptTokens = info.PromptTokens
2024-06-15 01:30:51 +08:00
}
if usage.CompletionTokens == 0 {
usage, _ = service.ResponseText2Usage(responseText, info.UpstreamModelName, usage.PromptTokens)
}
2023-07-22 17:36:40 +08:00
}
2024-03-20 17:07:42 +08:00
return nil, usage
2023-07-22 17:36:40 +08:00
}
2024-03-08 18:25:57 +08:00
func claudeHandler(requestMode int, c *gin.Context, resp *http.Response, promptTokens int, model string) (*dto.OpenAIErrorWithStatusCode, *dto.Usage) {
2023-07-22 17:36:40 +08:00
responseBody, err := io.ReadAll(resp.Body)
if err != nil {
2024-02-29 01:08:18 +08:00
return service.OpenAIErrorWrapper(err, "read_response_body_failed", http.StatusInternalServerError), nil
2023-07-22 17:36:40 +08:00
}
err = resp.Body.Close()
if err != nil {
2024-02-29 01:08:18 +08:00
return service.OpenAIErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), nil
2023-07-22 17:36:40 +08:00
}
var claudeResponse ClaudeResponse
err = json.Unmarshal(responseBody, &claudeResponse)
if err != nil {
2024-02-29 01:08:18 +08:00
return service.OpenAIErrorWrapper(err, "unmarshal_response_body_failed", http.StatusInternalServerError), nil
2023-07-22 17:36:40 +08:00
}
if claudeResponse.Error.Type != "" {
2024-02-29 01:08:18 +08:00
return &dto.OpenAIErrorWithStatusCode{
Error: dto.OpenAIError{
2023-07-22 17:36:40 +08:00
Message: claudeResponse.Error.Message,
Type: claudeResponse.Error.Type,
Param: "",
Code: claudeResponse.Error.Type,
},
StatusCode: resp.StatusCode,
}, nil
}
2024-04-23 11:44:40 +08:00
fullTextResponse := ResponseClaude2OpenAI(requestMode, &claudeResponse)
2024-05-23 23:59:55 +08:00
completionTokens, err := service.CountTokenText(claudeResponse.Completion, model)
2024-03-20 17:07:42 +08:00
if err != nil {
return service.OpenAIErrorWrapper(err, "count_token_text_failed", http.StatusInternalServerError), nil
}
2024-03-08 18:25:57 +08:00
usage := dto.Usage{}
if requestMode == RequestModeCompletion {
usage.PromptTokens = promptTokens
usage.CompletionTokens = completionTokens
usage.TotalTokens = promptTokens + completionTokens
} else {
usage.PromptTokens = claudeResponse.Usage.InputTokens
usage.CompletionTokens = claudeResponse.Usage.OutputTokens
usage.TotalTokens = claudeResponse.Usage.InputTokens + claudeResponse.Usage.OutputTokens
2023-07-22 17:36:40 +08:00
}
fullTextResponse.Usage = usage
jsonResponse, err := json.Marshal(fullTextResponse)
if err != nil {
2024-02-29 01:08:18 +08:00
return service.OpenAIErrorWrapper(err, "marshal_response_body_failed", http.StatusInternalServerError), nil
2023-07-22 17:36:40 +08:00
}
c.Writer.Header().Set("Content-Type", "application/json")
c.Writer.WriteHeader(resp.StatusCode)
_, err = c.Writer.Write(jsonResponse)
return nil, &usage
}