Nik Afiq
5238298b55
- Implemented core model components in `modules.py` including various convolutional layers and normalization techniques. - Added transformation functions in `transforms.py` for piecewise rational quadratic transformations. - Created utility functions in `utils.py` for checkpoint management, logging, and hyperparameter handling. - Introduced monotonic alignment functionality with Cython optimization in `monotonic_align`. - Developed a minimal inference server in `server.py` to handle synthesis requests. - Updated requirements to include necessary dependencies for Cython and scipy.
65 lines
2.0 KiB
Go
65 lines
2.0 KiB
Go
package grpc
|
|
|
|
import (
|
|
"context"
|
|
"errors"
|
|
|
|
"google.golang.org/grpc/codes"
|
|
"google.golang.org/grpc/status"
|
|
|
|
ttsv1 "gitea.nik4nao.com/nik/home-services/gen/tts/v1"
|
|
"gitea.nik4nao.com/nik/home-services/tts-gateway/internal/app"
|
|
"gitea.nik4nao.com/nik/home-services/tts-gateway/internal/core/domain"
|
|
"gitea.nik4nao.com/nik/home-services/tts-gateway/internal/core/ports/driving"
|
|
)
|
|
|
|
type TTSGRPC struct {
|
|
ttsv1.UnimplementedTTSServiceServer
|
|
svc driving.TTSService
|
|
}
|
|
|
|
// NewTTSGRPC constructs the gRPC adapter for TTSService.
|
|
func NewTTSGRPC(svc driving.TTSService) *TTSGRPC {
|
|
return &TTSGRPC{svc: svc}
|
|
}
|
|
|
|
// Synthesize translates a protobuf request into a domain call, applying
|
|
// uma-tts-api's sampling defaults for any knob the client omits.
|
|
func (h *TTSGRPC) Synthesize(ctx context.Context, req *ttsv1.SynthesizeRequest) (*ttsv1.SynthesizeResponse, error) {
|
|
params := domain.SynthesisParams{
|
|
NoiseScale: domain.DefaultNoiseScale,
|
|
NoiseScaleW: domain.DefaultNoiseScaleW,
|
|
LengthScale: domain.DefaultLengthScale,
|
|
}
|
|
if req.NoiseScale != nil {
|
|
params.NoiseScale = *req.NoiseScale
|
|
}
|
|
if req.NoiseScaleW != nil {
|
|
params.NoiseScaleW = *req.NoiseScaleW
|
|
}
|
|
if req.LengthScale != nil {
|
|
params.LengthScale = *req.LengthScale
|
|
}
|
|
|
|
clip, err := h.svc.Synthesize(ctx, req.SpeakerName, req.Text, params)
|
|
if err != nil {
|
|
return nil, grpcError(err)
|
|
}
|
|
return &ttsv1.SynthesizeResponse{Audio: clip.Data, MimeType: clip.MimeType}, nil
|
|
}
|
|
|
|
// ListSpeakers returns the (optionally filtered) speaker roster.
|
|
func (h *TTSGRPC) ListSpeakers(ctx context.Context, req *ttsv1.ListSpeakersRequest) (*ttsv1.ListSpeakersResponse, error) {
|
|
names := h.svc.ListSpeakers(ctx, req.Search)
|
|
return &ttsv1.ListSpeakersResponse{SpeakerNames: names}, nil
|
|
}
|
|
|
|
// grpcError maps domain errors to appropriate gRPC status codes, mirroring
|
|
// uma-tts-api's /synthesize 400 response for an unknown speaker name.
|
|
func grpcError(err error) error {
|
|
if errors.Is(err, app.ErrSpeakerNotFound) {
|
|
return status.Errorf(codes.InvalidArgument, "%v", err)
|
|
}
|
|
return status.Errorf(codes.Internal, "%v", err)
|
|
}
|