Files
new-api/relay/channel/xunfei/relay-xunfei.go
T
Calcium-Ion 86ac0f7745 refactor: extract protocol conversion layer into standalone relaykit module (#6369)
* test(relayconvert): add golden snapshot matrix and relaykit boundary guard

Phase 0 of the relaykit extraction plan: pin byte-level output of every
registered (from,to) request/response/stream conversion route, and
forbid kit-bound packages from growing host-only imports.

* wip(relayconvert): drop gin.Context from converter signatures; add convmeta draft

Phase 1 in progress: relayconvert now takes context.Context; host media
resolver adapts gin.Context back at the service boundary.

* refactor(relayconvert): decouple converters from RelayInfo, gin, and settings

Phase 1 of the relaykit extraction plan:
- converters now depend on convmeta.Meta (implemented by RelayInfo) instead
  of *relaycommon.RelayInfo; ClaudeConvertInfo and the format guesser move
  to convmeta with aliases left behind
- host settings reach converters via a convmeta.Options snapshot built in
  RelayInfo.ConvOptions; no more model_setting/reasoning global reads inside
  the conversion layer
- effort-suffix helpers move to service/relayconvert/reasoning (old package
  forwards); chat-to-responses upgrade policy moves to service (host routing
  logic, not conversion)
- golden conversion matrix unchanged

* test(relayconvert): tighten boundary — kit packages now free of gin/setting imports

* refactor(dto): drop gin and logger dependencies

Phase 2 (part 1): dto.Request.IsStream now takes *http.Request instead of
*gin.Context (Gemini's impl reads query/path off the std request); dto's
three logger calls become common.SysError. Boundary test allowlist is now
empty — kit-bound packages import no gin/setting/logger/model.

* refactor(kit): extract dependency-free kitutil; dto/types/relayconvert stop importing common

Phase 2 of the relaykit extraction plan:
- new service/relayconvert/kitutil holds the pure helpers the kit needs
  (JSON wrappers, pointer/string/uuid/timestamp utils, MaskSensitiveInfo,
  pluggable LogInfo/LogError hooks, Debug flag)
- dto, types, and all relayconvert packages now use kitutil; their only
  remaining internal deps are dto/types/constant
- common keeps every original symbol (MaskSensitiveInfo delegates to
  kitutil) so host code is untouched; main.go routes kit logging into
  common.SysLog/SysError and mirrors DebugEnabled
- golden conversion matrix unchanged

* refactor(kit): move EndpointType/FinishReason to types; OpenRouter dialect via Options

Kit packages (dto/types/relayconvert/reasonmap) no longer import constant:
- EndpointType and finish-reason values live in types; constant re-exports
- the OpenRouter special-case in claude->openai request conversion reads
  Options.OpenRouterDialect, set by the host from the channel type;
  InitChannelMeta invalidates the cached snapshot on channel switch

* refactor: extract relaykit submodule (dto/types/relayconvert/reasonmap)

Phase 3 of the relaykit extraction plan:
- new go module github.com/QuantumNous/new-api/relaykit containing dto
  (minus task family), types, relayconvert (with convmeta/kitutil/reasoning),
  and reasonmap; host consumes it via require + replace, go.work for dev
- task-family dto (task/suno/midjourney/video) stays in the host dto
  package; dual-consumer host files alias it as taskdto
- relaykit builds and tests standalone (GOWORK=off): no host imports,
  no gin, no DB, no settings
- golden conversion matrix unchanged

* build(docker): copy relaykit/go.mod before go mod download

The local-replace submodule's go.mod must exist inside the build context
for the main module graph to resolve.

* fix: address relaykit extraction regressions

* fix: address relaykit review regressions

* docs: document Meta nil receiver contract

* fix(relaykit): fail OpenAI→Claude conversion without max_tokens; reject negative default_max_tokens

The Claude Messages API requires max_tokens (omitting it is a 400
"Field required"), but with a nil Options.Claude.DefaultMaxTokens hook
the converters silently emitted a request the upstream is guaranteed to
reject. Both OpenAI Chat and Responses → Claude conversions now return
sharedclaude.ErrMissingMaxTokens when no path (client value, default
hook, thinking-adapter floor) supplied one. Unreachable in the host,
which always configures the hook.

Host side, claude.default_max_tokens now rejects negative values at the
option API before persisting — they would wrap into huge unsigned values
during conversion. Zero stays allowed: the current API treats
max_tokens: 0 as cache pre-warming.

* fix: make Gemini safety settings read path race-free
2026-07-27 15:56:21 +08:00

293 lines
8.7 KiB
Go

package xunfei
import (
"crypto/hmac"
"crypto/sha256"
"encoding/base64"
"encoding/json"
"fmt"
"io"
"net/url"
"strings"
"time"
"github.com/QuantumNous/new-api/common"
"github.com/QuantumNous/new-api/constant"
"github.com/QuantumNous/new-api/relay/helper"
"github.com/QuantumNous/new-api/relaykit/dto"
"github.com/QuantumNous/new-api/relaykit/types"
"github.com/samber/lo"
"github.com/gin-gonic/gin"
"github.com/gorilla/websocket"
)
// https://console.xfyun.cn/services/cbm
// https://www.xfyun.cn/doc/spark/Web.html
func requestOpenAI2Xunfei(request dto.GeneralOpenAIRequest, xunfeiAppId string, domain string) *XunfeiChatRequest {
messages := make([]XunfeiMessage, 0, len(request.Messages))
shouldCovertSystemMessage := !strings.HasSuffix(request.Model, "3.5")
for _, message := range request.Messages {
if message.Role == "system" && shouldCovertSystemMessage {
messages = append(messages, XunfeiMessage{
Role: "user",
Content: message.StringContent(),
})
messages = append(messages, XunfeiMessage{
Role: "assistant",
Content: "Okay",
})
} else {
messages = append(messages, XunfeiMessage{
Role: message.Role,
Content: message.StringContent(),
})
}
}
xunfeiRequest := XunfeiChatRequest{}
xunfeiRequest.Header.AppId = xunfeiAppId
xunfeiRequest.Parameter.Chat.Domain = domain
xunfeiRequest.Parameter.Chat.Temperature = request.Temperature
xunfeiRequest.Parameter.Chat.TopK = lo.FromPtrOr(request.N, 0)
xunfeiRequest.Parameter.Chat.MaxTokens = request.GetMaxTokens()
xunfeiRequest.Payload.Message.Text = messages
return &xunfeiRequest
}
func responseXunfei2OpenAI(response *XunfeiChatResponse) *dto.OpenAITextResponse {
if len(response.Payload.Choices.Text) == 0 {
response.Payload.Choices.Text = []XunfeiChatResponseTextItem{
{
Content: "",
},
}
}
choice := dto.OpenAITextResponseChoice{
Index: 0,
Message: dto.Message{
Role: "assistant",
Content: response.Payload.Choices.Text[0].Content,
},
FinishReason: constant.FinishReasonStop,
}
fullTextResponse := dto.OpenAITextResponse{
Object: "chat.completion",
Created: common.GetTimestamp(),
Choices: []dto.OpenAITextResponseChoice{choice},
Usage: response.Payload.Usage.Text,
}
return &fullTextResponse
}
func streamResponseXunfei2OpenAI(xunfeiResponse *XunfeiChatResponse) *dto.ChatCompletionsStreamResponse {
if len(xunfeiResponse.Payload.Choices.Text) == 0 {
xunfeiResponse.Payload.Choices.Text = []XunfeiChatResponseTextItem{
{
Content: "",
},
}
}
var choice dto.ChatCompletionsStreamResponseChoice
choice.Delta.SetContentString(xunfeiResponse.Payload.Choices.Text[0].Content)
if xunfeiResponse.Payload.Choices.Status == 2 {
choice.FinishReason = &constant.FinishReasonStop
}
response := dto.ChatCompletionsStreamResponse{
Object: "chat.completion.chunk",
Created: common.GetTimestamp(),
Model: "SparkDesk",
Choices: []dto.ChatCompletionsStreamResponseChoice{choice},
}
return &response
}
func buildXunfeiAuthUrl(hostUrl string, apiKey, apiSecret string) string {
HmacWithShaToBase64 := func(algorithm, data, key string) string {
mac := hmac.New(sha256.New, []byte(key))
mac.Write([]byte(data))
encodeData := mac.Sum(nil)
return base64.StdEncoding.EncodeToString(encodeData)
}
ul, err := url.Parse(hostUrl)
if err != nil {
fmt.Println(err)
}
date := time.Now().UTC().Format(time.RFC1123)
signString := []string{"host: " + ul.Host, "date: " + date, "GET " + ul.Path + " HTTP/1.1"}
sign := strings.Join(signString, "\n")
sha := HmacWithShaToBase64("hmac-sha256", sign, apiSecret)
authUrl := fmt.Sprintf("hmac username=\"%s\", algorithm=\"%s\", headers=\"%s\", signature=\"%s\"", apiKey,
"hmac-sha256", "host date request-line", sha)
authorization := base64.StdEncoding.EncodeToString([]byte(authUrl))
v := url.Values{}
v.Add("host", ul.Host)
v.Add("date", date)
v.Add("authorization", authorization)
callUrl := hostUrl + "?" + v.Encode()
return callUrl
}
func xunfeiStreamHandler(c *gin.Context, textRequest dto.GeneralOpenAIRequest, appId string, apiSecret string, apiKey string) (*dto.Usage, *types.NewAPIError) {
domain, authUrl := getXunfeiAuthUrl(c, apiKey, apiSecret, textRequest.Model)
dataChan, stopChan, err := xunfeiMakeRequest(textRequest, domain, authUrl, appId)
if err != nil {
return nil, types.NewError(err, types.ErrorCodeDoRequestFailed)
}
helper.SetEventStreamHeaders(c)
var usage dto.Usage
c.Stream(func(w io.Writer) bool {
select {
case xunfeiResponse := <-dataChan:
usage.PromptTokens += xunfeiResponse.Payload.Usage.Text.PromptTokens
usage.CompletionTokens += xunfeiResponse.Payload.Usage.Text.CompletionTokens
usage.TotalTokens += xunfeiResponse.Payload.Usage.Text.TotalTokens
response := streamResponseXunfei2OpenAI(&xunfeiResponse)
jsonResponse, err := json.Marshal(response)
if err != nil {
common.SysLog("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
}
})
return &usage, nil
}
func xunfeiHandler(c *gin.Context, textRequest dto.GeneralOpenAIRequest, appId string, apiSecret string, apiKey string) (*dto.Usage, *types.NewAPIError) {
domain, authUrl := getXunfeiAuthUrl(c, apiKey, apiSecret, textRequest.Model)
dataChan, stopChan, err := xunfeiMakeRequest(textRequest, domain, authUrl, appId)
if err != nil {
return nil, types.NewError(err, types.ErrorCodeDoRequestFailed)
}
var usage dto.Usage
var content string
var xunfeiResponse XunfeiChatResponse
stop := false
for !stop {
select {
case xunfeiResponse = <-dataChan:
if len(xunfeiResponse.Payload.Choices.Text) == 0 {
continue
}
content += xunfeiResponse.Payload.Choices.Text[0].Content
usage.PromptTokens += xunfeiResponse.Payload.Usage.Text.PromptTokens
usage.CompletionTokens += xunfeiResponse.Payload.Usage.Text.CompletionTokens
usage.TotalTokens += xunfeiResponse.Payload.Usage.Text.TotalTokens
case stop = <-stopChan:
}
}
if len(xunfeiResponse.Payload.Choices.Text) == 0 {
xunfeiResponse.Payload.Choices.Text = []XunfeiChatResponseTextItem{
{
Content: "",
},
}
}
xunfeiResponse.Payload.Choices.Text[0].Content = content
response := responseXunfei2OpenAI(&xunfeiResponse)
jsonResponse, err := json.Marshal(response)
if err != nil {
return nil, types.NewError(err, types.ErrorCodeBadResponseBody)
}
c.Writer.Header().Set("Content-Type", "application/json")
_, _ = c.Writer.Write(jsonResponse)
return &usage, nil
}
func xunfeiMakeRequest(textRequest dto.GeneralOpenAIRequest, domain, authUrl, appId string) (chan XunfeiChatResponse, chan bool, error) {
d := websocket.Dialer{
HandshakeTimeout: 5 * time.Second,
}
conn, resp, err := d.Dial(authUrl, nil)
if err != nil || resp.StatusCode != 101 {
return nil, nil, err
}
data := requestOpenAI2Xunfei(textRequest, appId, domain)
err = conn.WriteJSON(data)
if err != nil {
return nil, nil, err
}
dataChan := make(chan XunfeiChatResponse)
stopChan := make(chan bool)
go func() {
defer func() {
conn.Close()
}()
for {
_, msg, err := conn.ReadMessage()
if err != nil {
common.SysLog("error reading stream response: " + err.Error())
break
}
var response XunfeiChatResponse
err = json.Unmarshal(msg, &response)
if err != nil {
common.SysLog("error unmarshalling stream response: " + err.Error())
break
}
dataChan <- response
if response.Payload.Choices.Status == 2 {
if err != nil {
common.SysLog("error closing websocket connection: " + err.Error())
}
break
}
}
stopChan <- true
}()
return dataChan, stopChan, nil
}
func apiVersion2domain(apiVersion string) string {
switch apiVersion {
case "v1.1":
return "lite"
case "v2.1":
return "generalv2"
case "v3.1":
return "generalv3"
case "v3.5":
return "generalv3.5"
case "v4.0":
return "4.0Ultra"
}
return "general" + apiVersion
}
func getXunfeiAuthUrl(c *gin.Context, apiKey string, apiSecret string, modelName string) (string, string) {
apiVersion := getAPIVersion(c, modelName)
domain := apiVersion2domain(apiVersion)
authUrl := buildXunfeiAuthUrl(fmt.Sprintf("wss://spark-api.xf-yun.com/%s/chat", apiVersion), apiKey, apiSecret)
return domain, authUrl
}
func getAPIVersion(c *gin.Context, modelName string) string {
query := c.Request.URL.Query()
apiVersion := query.Get("api-version")
if apiVersion != "" {
return apiVersion
}
parts := strings.Split(modelName, "-")
if len(parts) == 2 {
apiVersion = parts[1]
return apiVersion
}
apiVersion = c.GetString("api_version")
if apiVersion != "" {
return apiVersion
}
apiVersion = "v1.1"
common.SysLog("api_version not found, using default: " + apiVersion)
return apiVersion
}