feat(doubao): Support Doubao Speech Large Model 2.0 (Seed-TTS) #41
There are no files selected for viewing
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change | ||||
|---|---|---|---|---|---|---|
| @@ -0,0 +1,210 @@ | ||||||
| package doubao | ||||||
|
|
||||||
| import ( | ||||||
| "bytes" | ||||||
| "encoding/base64" | ||||||
| "encoding/json" | ||||||
| "log/slog" | ||||||
| "net/http" | ||||||
| "strings" | ||||||
|
|
||||||
| "github.com/google/uuid" | ||||||
| "github.com/labstack/echo/v4" | ||||||
| "github.com/moeru-ai/unspeech/pkg/apierrors" | ||||||
| "github.com/moeru-ai/unspeech/pkg/backend/types" | ||||||
| "github.com/moeru-ai/unspeech/pkg/utils" | ||||||
| "github.com/samber/lo" | ||||||
| "github.com/samber/mo" | ||||||
| ) | ||||||
|
|
||||||
| // DoubaoSpeechRequest represents the request structure for Doubao TTS API v3 | ||||||
| // Reference: https://www.volcengine.com/docs/6561/1329505 | ||||||
| // Note: Although the protocol is v3, the endpoint URL remains /api/v1/tts | ||||||
| type DoubaoSpeechRequest struct { | ||||||
| App *DoubaoSpeechRequestApp `json:"app,omitempty"` | ||||||
| // User configuration | ||||||
| User *DoubaoSpeechRequestUser `json:"user,omitempty"` | ||||||
| // Audio synthesis parameters | ||||||
| Audio *DoubaoSpeechRequestAudio `json:"audio,omitempty"` | ||||||
| // Request metadata | ||||||
| Request *DoubaoSpeechRequestRequest `json:"request,omitempty"` | ||||||
| } | ||||||
|
|
||||||
| // DoubaoSpeechRequestApp contains application credentials | ||||||
| type DoubaoSpeechRequestApp struct { | ||||||
| AppID string `json:"appid,omitempty"` | ||||||
| Token string `json:"token,omitempty"` | ||||||
| Cluster string `json:"cluster,omitempty"` | ||||||
| } | ||||||
|
|
||||||
| // DoubaoSpeechRequestUser contains user information | ||||||
| type DoubaoSpeechRequestUser struct { | ||||||
| UserID string `json:"uid,omitempty"` | ||||||
| } | ||||||
|
|
||||||
| // DoubaoSpeechRequestAudio contains audio synthesis parameters | ||||||
| type DoubaoSpeechRequestAudio struct { | ||||||
| VoiceType string `json:"voice_type,omitempty"` | ||||||
| Encoding *string `json:"encoding,omitempty"` | ||||||
| SpeedRatio *float64 `json:"speed_ratio,omitempty"` | ||||||
| Rate *int `json:"rate,omitempty"` | ||||||
| BitRate *int `json:"bit_rate,omitempty"` | ||||||
| ExplicitLanguage *string `json:"explicit_language,omitempty"` | ||||||
| ContextLanguage *string `json:"context_language,omitempty"` | ||||||
| LoudnessRatio *float64 `json:"loudness_ratio,omitempty"` | ||||||
| Pitch *float64 `json:"pitch,omitempty"` | ||||||
| Emotion *string `json:"emotion,omitempty"` | ||||||
| ResourceID *string `json:"resource_id,omitempty"` | ||||||
| } | ||||||
|
|
||||||
| // DoubaoSpeechRequestRequest contains request metadata | ||||||
| type DoubaoSpeechRequestRequest struct { | ||||||
| RequestID string `json:"reqid,omitempty"` | ||||||
| Text string `json:"text"` | ||||||
| TextType *string `json:"text_type,omitempty"` | ||||||
| SilenceDuration *float64 `json:"silence_duration,omitempty"` | ||||||
| WithTimestamp *string `json:"with_timestamp,omitempty"` | ||||||
| Operation *string `json:"operation,omitempty"` | ||||||
| ExtraParam *string `json:"extra_param,omitempty"` | ||||||
| DisableMarkdownFilter *bool `json:"disable_markdown_filter,omitempty"` | ||||||
| EnableLatexTone *bool `json:"enable_latex_tn,omitempty"` | ||||||
| CacheConfig map[string]any `json:"cache_config,omitempty"` | ||||||
| UseCache *bool `json:"use_cache,omitempty"` | ||||||
| } | ||||||
|
|
||||||
| // HandleSpeech handles the speech synthesis request for Doubao TTS | ||||||
| func HandleSpeech(c echo.Context, options mo.Option[types.SpeechRequestOptions]) mo.Result[any] { | ||||||
| opts := options.MustGet() | ||||||
|
|
||||||
| // Extract token from Authorization header | ||||||
| token := strings.TrimPrefix(c.Request().Header.Get("Authorization"), "Bearer ") | ||||||
|
|
||||||
| // Doubao TTS uses doubao_tts cluster by default | ||||||
| cluster := utils.GetByJSONPath[string](opts.ExtraBody, "{ .app.cluster }") | ||||||
| if cluster == "" { | ||||||
| cluster = "doubao_tts" | ||||||
| } | ||||||
|
|
||||||
| // Generate or get user ID | ||||||
| userID := utils.GetByJSONPath[string](opts.ExtraBody, "{ .user.uid }") | ||||||
| if userID == "" { | ||||||
| userID = uuid.New().String() | ||||||
| } | ||||||
|
|
||||||
| // Generate or get request ID | ||||||
| requestID := utils.GetByJSONPath[string](opts.ExtraBody, "{ .request.reqid }") | ||||||
| if requestID == "" { | ||||||
| requestID = uuid.New().String() | ||||||
| } | ||||||
|
|
||||||
| // Get operation type, default to "query" | ||||||
| operation := utils.GetByJSONPath[*string](opts.ExtraBody, "{ .request.operation }") | ||||||
| if operation == nil || *operation == "" { | ||||||
| operation = lo.ToPtr("query") | ||||||
| } | ||||||
|
|
||||||
| // Get speed ratio, default to 1.0 | ||||||
| speedRatio := utils.GetByJSONPath[*float64](opts.ExtraBody, "{ .audio.speed_ratio }") | ||||||
| if speedRatio == nil || *speedRatio == 0 { | ||||||
| speedRatio = lo.ToPtr(1.0) | ||||||
| } | ||||||
|
|
||||||
| // Build the request payload | ||||||
| reqBody := &DoubaoSpeechRequest{ | ||||||
| App: &DoubaoSpeechRequestApp{ | ||||||
| AppID: utils.GetByJSONPath[string](opts.ExtraBody, "{ .app.appid }"), | ||||||
| Token: token, | ||||||
| Cluster: cluster, | ||||||
| }, | ||||||
| User: &DoubaoSpeechRequestUser{ | ||||||
| UserID: userID, | ||||||
| }, | ||||||
| Audio: &DoubaoSpeechRequestAudio{ | ||||||
| VoiceType: opts.Voice, | ||||||
| Encoding: lo.Ternary(opts.ResponseFormat != "", lo.ToPtr(opts.ResponseFormat), lo.ToPtr("mp3")), | ||||||
| SpeedRatio: speedRatio, | ||||||
| Rate: utils.GetByJSONPath[*int](opts.ExtraBody, "{ .audio.rate }"), | ||||||
| BitRate: utils.GetByJSONPath[*int](opts.ExtraBody, "{ .audio.bit_rate }"), | ||||||
| ExplicitLanguage: utils.GetByJSONPath[*string](opts.ExtraBody, "{ .audio.explicit_language }"), | ||||||
| ContextLanguage: utils.GetByJSONPath[*string](opts.ExtraBody, "{ .audio.context_language }"), | ||||||
| LoudnessRatio: utils.GetByJSONPath[*float64](opts.ExtraBody, "{ .audio.loudness_ratio }"), | ||||||
| Pitch: utils.GetByJSONPath[*float64](opts.ExtraBody, "{ .audio.pitch }"), | ||||||
| Emotion: utils.GetByJSONPath[*string](opts.ExtraBody, "{ .audio.emotion }"), | ||||||
| ResourceID: utils.GetByJSONPath[*string](opts.ExtraBody, "{ .audio.resource_id }"), | ||||||
| }, | ||||||
| Request: &DoubaoSpeechRequestRequest{ | ||||||
| RequestID: requestID, | ||||||
| Text: opts.Input, | ||||||
| TextType: utils.GetByJSONPath[*string](opts.ExtraBody, "{ .request.text_type }"), | ||||||
| SilenceDuration: utils.GetByJSONPath[*float64](opts.ExtraBody, "{ .request.silence_duration }"), | ||||||
| WithTimestamp: utils.GetByJSONPath[*string](opts.ExtraBody, "{ .request.with_timestamp }"), | ||||||
| Operation: operation, | ||||||
| ExtraParam: utils.GetByJSONPath[*string](opts.ExtraBody, "{ .request.extra_param }"), | ||||||
| DisableMarkdownFilter: utils.GetByJSONPath[*bool](opts.ExtraBody, "{ .request.disable_markdown_filter }"), | ||||||
| EnableLatexTone: utils.GetByJSONPath[*bool](opts.ExtraBody, "{ .request.enable_latex_tn }"), | ||||||
| CacheConfig: utils.GetByJSONPath[map[string]any](opts.ExtraBody, "{ .request.cache_config }"), | ||||||
| }, | ||||||
| } | ||||||
|
|
||||||
| jsonBytes, err := json.Marshal(reqBody) | ||||||
| if err != nil { | ||||||
| return mo.Err[any](apierrors.NewErrInternal().WithDetail(err.Error()).WithCaller()) | ||||||
| } | ||||||
|
|
||||||
| req, err := http.NewRequestWithContext(c.Request().Context(), http.MethodPost, "https://openspeech.bytedance.com/api/v1/tts", bytes.NewBuffer(jsonBytes)) | ||||||
| if err != nil { | ||||||
| return mo.Err[any](apierrors.NewErrInternal().WithDetail(err.Error()).WithCaller()) | ||||||
| } | ||||||
|
|
||||||
| // Set authorization header - Doubao uses Bearer;token format | ||||||
| req.Header.Set("Authorization", "Bearer;"+token) | ||||||
|
There was a problem hiding this comment. The Authorization header is constructed using
Suggested change
|
||||||
| req.Header.Set("Content-Type", "application/json") | ||||||
|
|
||||||
| resp, err := http.DefaultClient.Do(req) | ||||||
| if err != nil { | ||||||
| return mo.Err[any](apierrors.NewErrInternal().WithDetail(err.Error()).WithCaller()) | ||||||
| } | ||||||
|
|
||||||
| defer func() { _ = resp.Body.Close() }() | ||||||
|
|
||||||
| // Handle error responses | ||||||
| if resp.StatusCode >= 400 && resp.StatusCode < 600 { | ||||||
| switch { | ||||||
| case strings.HasPrefix(resp.Header.Get("Content-Type"), "application/json"): | ||||||
| return mo.Err[any](apierrors. | ||||||
| NewUpstreamError(resp.StatusCode). | ||||||
| WithDetail(utils.NewJSONResponseError(resp.StatusCode, resp.Body).OrEmpty().Error())) | ||||||
| case strings.HasPrefix(resp.Header.Get("Content-Type"), "text/"): | ||||||
| return mo.Err[any](apierrors. | ||||||
| NewUpstreamError(resp.StatusCode). | ||||||
| WithDetail(utils.NewTextResponseError(resp.StatusCode, resp.Body).OrEmpty().Error())) | ||||||
| default: | ||||||
| slog.Warn("unknown upstream error with unknown Content-Type", | ||||||
| slog.Int("status", resp.StatusCode), | ||||||
| slog.String("content_type", resp.Header.Get("Content-Type")), | ||||||
| slog.String("content_length", resp.Header.Get("Content-Length")), | ||||||
| ) | ||||||
| } | ||||||
| } | ||||||
|
|
||||||
| var resBody map[string]any | ||||||
|
|
||||||
| err = json.NewDecoder(resp.Body).Decode(&resBody) | ||||||
| if err != nil { | ||||||
| return mo.Err[any](apierrors.NewErrInternal().WithDetail(err.Error()).WithError(err).WithCaller()) | ||||||
| } | ||||||
|
|
||||||
| // Extract base64 encoded audio data from response | ||||||
| audioBase64String := utils.GetByJSONPath[string](resBody, "{ .data }") | ||||||
| if audioBase64String == "" { | ||||||
| return mo.Err[any](apierrors.NewErrInternal().WithDetail("upstream returned empty audio base64 string").WithCaller()) | ||||||
| } | ||||||
|
|
||||||
| // Decode base64 audio to binary | ||||||
| audioBytes, err := base64.StdEncoding.DecodeString(audioBase64String) | ||||||
| if err != nil { | ||||||
| return mo.Err[any](apierrors.NewErrInternal().WithDetail(err.Error()).WithError(err).WithCaller()) | ||||||
| } | ||||||
|
|
||||||
| return mo.Ok[any](c.Blob(http.StatusOK, "audio/mp3", audioBytes)) | ||||||
|
There was a problem hiding this comment. The
Suggested change
|
||||||
| } | ||||||
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,85 @@ | ||
| package doubao | ||
|
|
||
| import ( | ||
| _ "embed" | ||
| "encoding/json" | ||
|
|
||
| "github.com/labstack/echo/v4" | ||
| "github.com/moeru-ai/unspeech/pkg/apierrors" | ||
| "github.com/moeru-ai/unspeech/pkg/backend/types" | ||
| "github.com/samber/mo" | ||
| ) | ||
|
|
||
| var ( | ||
| //go:embed voices.json | ||
| voicesJSON string | ||
| ) | ||
|
|
||
| // VoicesResponseItem represents a voice in the voices.json file | ||
| type VoicesResponseItem struct { | ||
| Name string `json:"name"` | ||
| PreviewAudioURL string `json:"preview_audio_url"` | ||
| Model string `json:"model"` | ||
| Voice string `json:"voice"` | ||
| Scenarios []string `json:"scenarios"` | ||
| Language string `json:"language"` | ||
| Bitrate string `json:"bitrate"` | ||
| Format string `json:"format"` | ||
|
Comment on lines
+26
to
+27
There was a problem hiding this comment. The type VoicesResponseItem struct {
Name string `json:"name"`
PreviewAudioURL string `json:"preview_audio_url"`
Model string `json:"model"`
Voice string `json:"voice"`
Scenarios []string `json:"scenarios"`
Language string `json:"language"`
} |
||
| } | ||
|
|
||
| // HandleVoices handles the voices list request for Doubao TTS | ||
| func HandleVoices(c echo.Context, options mo.Option[types.VoicesRequestOptions]) mo.Result[any] { | ||
| var voicesResponse []VoicesResponseItem | ||
|
|
||
| err := json.Unmarshal([]byte(voicesJSON), &voicesResponse) | ||
| if err != nil { | ||
| return mo.Err[any](apierrors.NewErrInternal().WithDetail(err.Error()).WithCaller()) | ||
| } | ||
|
|
||
| voices := make([]types.Voice, 0, len(voicesResponse)) | ||
|
|
||
| for _, voice := range voicesResponse { | ||
| voices = append(voices, types.Voice{ | ||
| ID: voice.Voice, | ||
| Name: voice.Name, | ||
| Description: voice.Name, | ||
| Labels: map[string]any{ | ||
| "tailoredScenarios": voice.Scenarios, | ||
| }, | ||
| Tags: make([]string, 0), | ||
| Formats: []types.VoiceFormat{ | ||
| // https://www.volcengine.com/docs/6561/1257584 | ||
| {Name: "MP3", Extension: ".mp3", MimeType: "audio/mp3", SampleRate: 8000, Bitrate: 16, FormatCode: "mp3"}, | ||
| {Name: "MP3", Extension: ".mp3", MimeType: "audio/mp3", SampleRate: 16000, Bitrate: 16, FormatCode: "mp3"}, | ||
| {Name: "MP3", Extension: ".mp3", MimeType: "audio/mp3", SampleRate: 22050, Bitrate: 16, FormatCode: "mp3"}, | ||
| {Name: "MP3", Extension: ".mp3", MimeType: "audio/mp3", SampleRate: 24000, Bitrate: 16, FormatCode: "mp3"}, | ||
| {Name: "MP3", Extension: ".mp3", MimeType: "audio/mp3", SampleRate: 44100, Bitrate: 16, FormatCode: "mp3"}, | ||
| {Name: "MP3", Extension: ".mp3", MimeType: "audio/mp3", SampleRate: 48000, Bitrate: 16, FormatCode: "mp3"}, | ||
| {Name: "PCM", Extension: ".pcm", MimeType: "audio/pcm", SampleRate: 8000, Bitrate: 16, FormatCode: "pcm"}, | ||
| {Name: "PCM", Extension: ".pcm", MimeType: "audio/pcm", SampleRate: 16000, Bitrate: 16, FormatCode: "pcm"}, | ||
| {Name: "PCM", Extension: ".pcm", MimeType: "audio/pcm", SampleRate: 22050, Bitrate: 16, FormatCode: "pcm"}, | ||
| {Name: "PCM", Extension: ".pcm", MimeType: "audio/pcm", SampleRate: 24000, Bitrate: 16, FormatCode: "pcm"}, | ||
| {Name: "PCM", Extension: ".pcm", MimeType: "audio/pcm", SampleRate: 44100, Bitrate: 16, FormatCode: "pcm"}, | ||
| {Name: "PCM", Extension: ".pcm", MimeType: "audio/pcm", SampleRate: 48000, Bitrate: 16, FormatCode: "pcm"}, | ||
| {Name: "WAV", Extension: ".wav", MimeType: "audio/wav", SampleRate: 8000, Bitrate: 16, FormatCode: "wav"}, | ||
| {Name: "WAV", Extension: ".wav", MimeType: "audio/wav", SampleRate: 16000, Bitrate: 16, FormatCode: "wav"}, | ||
| {Name: "WAV", Extension: ".wav", MimeType: "audio/wav", SampleRate: 22050, Bitrate: 16, FormatCode: "wav"}, | ||
| {Name: "WAV", Extension: ".wav", MimeType: "audio/wav", SampleRate: 24000, Bitrate: 16, FormatCode: "wav"}, | ||
| {Name: "WAV", Extension: ".wav", MimeType: "audio/wav", SampleRate: 44100, Bitrate: 16, FormatCode: "wav"}, | ||
| {Name: "WAV", Extension: ".wav", MimeType: "audio/wav", SampleRate: 48000, Bitrate: 16, FormatCode: "wav"}, | ||
| }, | ||
| CompatibleModels: []string{voice.Model}, | ||
| PreviewAudioURL: voice.PreviewAudioURL, | ||
| Languages: []types.VoiceLanguage{ | ||
| { | ||
| Title: voice.Language, | ||
| Code: voice.Language, | ||
| }, | ||
| }, | ||
| }) | ||
| } | ||
|
|
||
| return mo.Ok[any](types.ListVoicesResponse{ | ||
| Voices: voices, | ||
| }) | ||
| } | ||
Oops, something went wrong.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The PR description states that
resource_idis required for Seed-TTS 2.0 and should default toseed-tts-2.0if not specified. Currently, the code only passesResourceIDif it's present inopts.ExtraBody. If it's a mandatory field, it should be explicitly defaulted when missing to prevent potential API errors.