feat: opt minimax tts req struct
This commit is contained in:
@@ -34,17 +34,20 @@ func (a *Adaptor) ConvertAudioRequest(c *gin.Context, info *relaycommon.RelayInf
|
|||||||
return nil, errors.New("unsupported audio relay mode")
|
return nil, errors.New("unsupported audio relay mode")
|
||||||
}
|
}
|
||||||
|
|
||||||
voiceID := mapVoiceType(request.Voice)
|
voiceID := request.Voice
|
||||||
speed := request.Speed
|
speed := request.Speed
|
||||||
outputFormat := mapOutputFormat(request.ResponseFormat)
|
outputFormat := request.ResponseFormat
|
||||||
|
|
||||||
c.Set("response_format", outputFormat)
|
|
||||||
|
|
||||||
minimaxRequest := MiniMaxTTSRequest{
|
minimaxRequest := MiniMaxTTSRequest{
|
||||||
Model: getTTSModel(info.OriginModelName),
|
Model: info.OriginModelName,
|
||||||
Text: request.Input,
|
Text: request.Input,
|
||||||
VoiceID: voiceID,
|
VoiceSetting: VoiceSetting{
|
||||||
Speed: speed,
|
VoiceID: voiceID,
|
||||||
|
Speed: speed,
|
||||||
|
},
|
||||||
|
AudioSetting: &AudioSetting{
|
||||||
|
Format: outputFormat,
|
||||||
|
},
|
||||||
OutputFormat: outputFormat,
|
OutputFormat: outputFormat,
|
||||||
}
|
}
|
||||||
|
|
||||||
@@ -59,6 +62,11 @@ func (a *Adaptor) ConvertAudioRequest(c *gin.Context, info *relaycommon.RelayInf
|
|||||||
if err != nil {
|
if err != nil {
|
||||||
return nil, fmt.Errorf("error marshalling minimax request: %w", err)
|
return nil, fmt.Errorf("error marshalling minimax request: %w", err)
|
||||||
}
|
}
|
||||||
|
if outputFormat != "hex" {
|
||||||
|
outputFormat = "url"
|
||||||
|
}
|
||||||
|
|
||||||
|
c.Set("response_format", outputFormat)
|
||||||
|
|
||||||
// Debug: log the request structure
|
// Debug: log the request structure
|
||||||
fmt.Printf("MiniMax TTS Request: %s\n", string(jsonData))
|
fmt.Printf("MiniMax TTS Request: %s\n", string(jsonData))
|
||||||
@@ -79,12 +87,6 @@ func (a *Adaptor) GetRequestURL(info *relaycommon.RelayInfo) (string, error) {
|
|||||||
|
|
||||||
func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Header, info *relaycommon.RelayInfo) error {
|
func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Header, info *relaycommon.RelayInfo) error {
|
||||||
channel.SetupApiRequestHeader(info, c, req)
|
channel.SetupApiRequestHeader(info, c, req)
|
||||||
|
|
||||||
if info.RelayMode == constant.RelayModeAudioSpeech {
|
|
||||||
req.Set("Content-Type", "application/json")
|
|
||||||
return nil
|
|
||||||
}
|
|
||||||
|
|
||||||
req.Set("Authorization", "Bearer "+info.ApiKey)
|
req.Set("Authorization", "Bearer "+info.ApiKey)
|
||||||
return nil
|
return nil
|
||||||
}
|
}
|
||||||
|
|||||||
@@ -1,11 +1,13 @@
|
|||||||
package minimax
|
package minimax
|
||||||
|
|
||||||
import (
|
import (
|
||||||
"encoding/base64"
|
"encoding/hex"
|
||||||
"encoding/json"
|
"encoding/json"
|
||||||
"errors"
|
"errors"
|
||||||
|
"fmt"
|
||||||
"io"
|
"io"
|
||||||
"net/http"
|
"net/http"
|
||||||
|
"strings"
|
||||||
|
|
||||||
"github.com/QuantumNous/new-api/dto"
|
"github.com/QuantumNous/new-api/dto"
|
||||||
relaycommon "github.com/QuantumNous/new-api/relay/common"
|
relaycommon "github.com/QuantumNous/new-api/relay/common"
|
||||||
@@ -14,96 +16,78 @@ import (
|
|||||||
)
|
)
|
||||||
|
|
||||||
type MiniMaxTTSRequest struct {
|
type MiniMaxTTSRequest struct {
|
||||||
Model string `json:"model"`
|
Model string `json:"model"`
|
||||||
Text string `json:"text"`
|
Text string `json:"text"`
|
||||||
VoiceID string `json:"voice_id"`
|
Stream bool `json:"stream,omitempty"`
|
||||||
Speed float64 `json:"speed,omitempty"`
|
StreamOptions *StreamOptions `json:"stream_options,omitempty"`
|
||||||
Vol float64 `json:"vol,omitempty"`
|
VoiceSetting VoiceSetting `json:"voice_setting"`
|
||||||
Pitch int `json:"pitch,omitempty"`
|
PronunciationDict *PronunciationDict `json:"pronunciation_dict,omitempty"`
|
||||||
AudioSampleRate int `json:"audio_sample_rate,omitempty"`
|
AudioSetting *AudioSetting `json:"audio_setting,omitempty"`
|
||||||
OutputFormat string `json:"output_format,omitempty"`
|
TimbreWeights []TimbreWeight `json:"timbre_weights,omitempty"`
|
||||||
|
LanguageBoost string `json:"language_boost,omitempty"`
|
||||||
|
VoiceModify *VoiceModify `json:"voice_modify,omitempty"`
|
||||||
|
SubtitleEnable bool `json:"subtitle_enable,omitempty"`
|
||||||
|
OutputFormat string `json:"output_format,omitempty"`
|
||||||
|
AigcWatermark bool `json:"aigc_watermark,omitempty"`
|
||||||
|
}
|
||||||
|
|
||||||
|
type StreamOptions struct {
|
||||||
|
ExcludeAggregatedAudio bool `json:"exclude_aggregated_audio,omitempty"`
|
||||||
|
}
|
||||||
|
|
||||||
|
type VoiceSetting struct {
|
||||||
|
VoiceID string `json:"voice_id"`
|
||||||
|
Speed float64 `json:"speed,omitempty"`
|
||||||
|
Vol float64 `json:"vol,omitempty"`
|
||||||
|
Pitch int `json:"pitch,omitempty"`
|
||||||
|
Emotion string `json:"emotion,omitempty"`
|
||||||
|
TextNormalization bool `json:"text_normalization,omitempty"`
|
||||||
|
LatexRead bool `json:"latex_read,omitempty"`
|
||||||
|
}
|
||||||
|
|
||||||
|
type PronunciationDict struct {
|
||||||
|
Tone []string `json:"tone,omitempty"`
|
||||||
|
}
|
||||||
|
|
||||||
|
type AudioSetting struct {
|
||||||
|
SampleRate int `json:"sample_rate,omitempty"`
|
||||||
|
Bitrate int `json:"bitrate,omitempty"`
|
||||||
|
Format string `json:"format,omitempty"`
|
||||||
|
Channel int `json:"channel,omitempty"`
|
||||||
|
ForceCbr bool `json:"force_cbr,omitempty"`
|
||||||
|
}
|
||||||
|
|
||||||
|
type TimbreWeight struct {
|
||||||
|
VoiceID string `json:"voice_id"`
|
||||||
|
Weight int `json:"weight"`
|
||||||
|
}
|
||||||
|
|
||||||
|
type VoiceModify struct {
|
||||||
|
Pitch int `json:"pitch,omitempty"`
|
||||||
|
Intensity int `json:"intensity,omitempty"`
|
||||||
|
Timbre int `json:"timbre,omitempty"`
|
||||||
|
SoundEffects string `json:"sound_effects,omitempty"`
|
||||||
}
|
}
|
||||||
|
|
||||||
type MiniMaxTTSResponse struct {
|
type MiniMaxTTSResponse struct {
|
||||||
Created int `json:"created"`
|
Data MiniMaxTTSData `json:"data"`
|
||||||
Data []MiniMaxTTSData `json:"data"`
|
ExtraInfo MiniMaxExtraInfo `json:"extra_info"`
|
||||||
ID string `json:"id"`
|
TraceID string `json:"trace_id"`
|
||||||
Model string `json:"model"`
|
BaseResp MiniMaxBaseResp `json:"base_resp"`
|
||||||
Object string `json:"object"`
|
|
||||||
Usage MiniMaxTTSUsage `json:"usage"`
|
|
||||||
}
|
}
|
||||||
|
|
||||||
type MiniMaxTTSData struct {
|
type MiniMaxTTSData struct {
|
||||||
Index int `json:"index"`
|
Audio string `json:"audio"`
|
||||||
Audio string `json:"audio"`
|
Status int `json:"status"`
|
||||||
Text string `json:"text"`
|
|
||||||
FinishReason string `json:"finish_reason"`
|
|
||||||
}
|
}
|
||||||
|
|
||||||
type MiniMaxTTSUsage struct {
|
type MiniMaxExtraInfo struct {
|
||||||
TotalTokens int `json:"total_tokens"`
|
UsageCharacters int64 `json:"usage_characters"`
|
||||||
}
|
}
|
||||||
|
|
||||||
type MiniMaxTTSErrorResponse struct {
|
type MiniMaxBaseResp struct {
|
||||||
Error MiniMaxTTSError `json:"error"`
|
StatusCode int64 `json:"status_code"`
|
||||||
}
|
StatusMsg string `json:"status_msg"`
|
||||||
|
|
||||||
type MiniMaxTTSError struct {
|
|
||||||
Code string `json:"code"`
|
|
||||||
Message string `json:"message"`
|
|
||||||
Type string `json:"type"`
|
|
||||||
}
|
|
||||||
|
|
||||||
// OpenAI voice to MiniMax voice_id mapping
|
|
||||||
var openAIToMiniMaxVoiceMap = map[string]string{
|
|
||||||
"alloy": "male-qn-qingse",
|
|
||||||
"echo": "male-qn-jingying",
|
|
||||||
"fable": "female-shaonv",
|
|
||||||
"onyx": "male-qn-badao",
|
|
||||||
"nova": "female-shaonv-jingpin",
|
|
||||||
"shimmer": "female-yujie",
|
|
||||||
// Add some standard MiniMax voice IDs
|
|
||||||
"voice-1": "male-qn-qingse",
|
|
||||||
"voice-2": "female-shaonv",
|
|
||||||
}
|
|
||||||
|
|
||||||
// OpenAI response format to MiniMax output format mapping
|
|
||||||
var responseFormatToOutputFormatMap = map[string]string{
|
|
||||||
"mp3": "mp3",
|
|
||||||
"opus": "mp3",
|
|
||||||
"aac": "aac",
|
|
||||||
"flac": "flac",
|
|
||||||
"wav": "wav",
|
|
||||||
"pcm": "pcm",
|
|
||||||
}
|
|
||||||
|
|
||||||
// TTS model mapping - MiniMax uses speech-01 or speech-01-turbo
|
|
||||||
var modelToTTSModelMap = map[string]string{
|
|
||||||
"speech-01": "speech-01",
|
|
||||||
"speech-01-turbo": "speech-01-turbo",
|
|
||||||
"tts-1": "speech-01-turbo",
|
|
||||||
"tts-1-hd": "speech-01",
|
|
||||||
}
|
|
||||||
|
|
||||||
func mapVoiceType(openAIVoice string) string {
|
|
||||||
if voice, ok := openAIToMiniMaxVoiceMap[openAIVoice]; ok {
|
|
||||||
return voice
|
|
||||||
}
|
|
||||||
return "female-shaonv" // default voice
|
|
||||||
}
|
|
||||||
|
|
||||||
func mapOutputFormat(responseFormat string) string {
|
|
||||||
if format, ok := responseFormatToOutputFormatMap[responseFormat]; ok {
|
|
||||||
return format
|
|
||||||
}
|
|
||||||
return "mp3" // default format
|
|
||||||
}
|
|
||||||
|
|
||||||
func getTTSModel(modelName string) string {
|
|
||||||
if ttsModel, ok := modelToTTSModelMap[modelName]; ok {
|
|
||||||
return ttsModel
|
|
||||||
}
|
|
||||||
return "speech-01-turbo" // default model
|
|
||||||
}
|
}
|
||||||
|
|
||||||
func getContentTypeByFormat(format string) string {
|
func getContentTypeByFormat(format string) string {
|
||||||
@@ -124,66 +108,64 @@ func handleTTSResponse(c *gin.Context, resp *http.Response, info *relaycommon.Re
|
|||||||
body, readErr := io.ReadAll(resp.Body)
|
body, readErr := io.ReadAll(resp.Body)
|
||||||
if readErr != nil {
|
if readErr != nil {
|
||||||
return nil, types.NewErrorWithStatusCode(
|
return nil, types.NewErrorWithStatusCode(
|
||||||
errors.New("failed to read minimax response"),
|
fmt.Errorf("failed to read minimax response: %w", readErr),
|
||||||
types.ErrorCodeReadResponseBodyFailed,
|
types.ErrorCodeReadResponseBodyFailed,
|
||||||
http.StatusInternalServerError,
|
http.StatusInternalServerError,
|
||||||
)
|
)
|
||||||
}
|
}
|
||||||
defer resp.Body.Close()
|
defer resp.Body.Close()
|
||||||
|
|
||||||
// First try to parse as error response
|
// Parse response
|
||||||
var errorResp MiniMaxTTSErrorResponse
|
|
||||||
if unmarshalErr := json.Unmarshal(body, &errorResp); unmarshalErr == nil && errorResp.Error.Code != "" {
|
|
||||||
return nil, types.NewErrorWithStatusCode(
|
|
||||||
errors.New(errorResp.Error.Message),
|
|
||||||
types.ErrorCodeBadResponse,
|
|
||||||
http.StatusBadRequest,
|
|
||||||
)
|
|
||||||
}
|
|
||||||
|
|
||||||
// Parse as successful response
|
|
||||||
var minimaxResp MiniMaxTTSResponse
|
var minimaxResp MiniMaxTTSResponse
|
||||||
if unmarshalErr := json.Unmarshal(body, &minimaxResp); unmarshalErr != nil {
|
if unmarshalErr := json.Unmarshal(body, &minimaxResp); unmarshalErr != nil {
|
||||||
return nil, types.NewErrorWithStatusCode(
|
return nil, types.NewErrorWithStatusCode(
|
||||||
errors.New("failed to parse minimax response"),
|
fmt.Errorf("failed to unmarshal minimax TTS response: %w", unmarshalErr),
|
||||||
types.ErrorCodeBadResponseBody,
|
types.ErrorCodeBadResponseBody,
|
||||||
http.StatusInternalServerError,
|
http.StatusInternalServerError,
|
||||||
)
|
)
|
||||||
}
|
}
|
||||||
|
|
||||||
|
// Check base_resp status code
|
||||||
|
if minimaxResp.BaseResp.StatusCode != 0 {
|
||||||
|
return nil, types.NewErrorWithStatusCode(
|
||||||
|
fmt.Errorf("minimax TTS error: %d - %s", minimaxResp.BaseResp.StatusCode, minimaxResp.BaseResp.StatusMsg),
|
||||||
|
types.ErrorCodeBadResponse,
|
||||||
|
http.StatusBadRequest,
|
||||||
|
)
|
||||||
|
}
|
||||||
|
|
||||||
// Check if we have audio data
|
// Check if we have audio data
|
||||||
if len(minimaxResp.Data) == 0 || minimaxResp.Data[0].Audio == "" {
|
if minimaxResp.Data.Audio == "" {
|
||||||
return nil, types.NewErrorWithStatusCode(
|
return nil, types.NewErrorWithStatusCode(
|
||||||
errors.New("no audio data in response"),
|
fmt.Errorf("no audio data in minimax TTS response"),
|
||||||
types.ErrorCodeBadResponse,
|
types.ErrorCodeBadResponse,
|
||||||
http.StatusBadRequest,
|
http.StatusBadRequest,
|
||||||
)
|
)
|
||||||
}
|
}
|
||||||
|
|
||||||
// Decode base64 audio data
|
if strings.HasPrefix(minimaxResp.Data.Audio, "http") {
|
||||||
audioData, decodeErr := base64.StdEncoding.DecodeString(minimaxResp.Data[0].Audio)
|
c.Redirect(http.StatusFound, minimaxResp.Data.Audio)
|
||||||
if decodeErr != nil {
|
} else {
|
||||||
return nil, types.NewErrorWithStatusCode(
|
// Handle hex-encoded audio data
|
||||||
errors.New("failed to decode audio data"),
|
audioData, decodeErr := hex.DecodeString(minimaxResp.Data.Audio)
|
||||||
types.ErrorCodeBadResponseBody,
|
if decodeErr != nil {
|
||||||
http.StatusInternalServerError,
|
return nil, types.NewErrorWithStatusCode(
|
||||||
)
|
fmt.Errorf("failed to decode hex audio data: %w", decodeErr),
|
||||||
}
|
types.ErrorCodeBadResponse,
|
||||||
|
http.StatusInternalServerError,
|
||||||
|
)
|
||||||
|
}
|
||||||
|
|
||||||
// Get output format from context or default to mp3
|
// Determine content type - default to mp3
|
||||||
outputFormat := c.GetString("response_format")
|
contentType := "audio/mpeg"
|
||||||
if outputFormat == "" {
|
|
||||||
outputFormat = "mp3"
|
|
||||||
}
|
|
||||||
|
|
||||||
contentType := getContentTypeByFormat(outputFormat)
|
c.Data(http.StatusOK, contentType, audioData)
|
||||||
c.Header("Content-Type", contentType)
|
}
|
||||||
c.Data(http.StatusOK, contentType, audioData)
|
|
||||||
|
|
||||||
usage = &dto.Usage{
|
usage = &dto.Usage{
|
||||||
PromptTokens: info.PromptTokens,
|
PromptTokens: info.PromptTokens,
|
||||||
CompletionTokens: 0,
|
CompletionTokens: 0,
|
||||||
TotalTokens: minimaxResp.Usage.TotalTokens,
|
TotalTokens: int(minimaxResp.ExtraInfo.UsageCharacters),
|
||||||
}
|
}
|
||||||
|
|
||||||
return usage, nil
|
return usage, nil
|
||||||
|
|||||||
Reference in New Issue
Block a user