Add support for Azure TTS and document usage

.gitignore

@@ -1,2 +1,3 @@
 account.json
 .env
+webtts

README.md

@@ -1,3 +1,58 @@
 # Luzifer / webtts
 
-This project is a simple wrapper around the [Google Cloud Text-To-Speech](https://cloud.google.com/text-to-speech) API to output OGG Vorbis Audio to be used with OBS overlays.
+This project is a simple wrapper around the [Google Cloud Text-To-Speech](https://cloud.google.com/text-to-speech) and [Azure Text-To-Speech](https://azure.microsoft.com/en-us/services/cognitive-services/speech-service/) API to output OGG Vorbis Audio to be used with OBS overlays.
+
+## Usage
+
+### Google Cloud Text-To-Speech
+
+- Create a project in the [Google Cloud Console](https://console.cloud.google.com/).
+- Enable the [Text-to-Speech API](https://cloud.google.com/text-to-speech/docs/apis).
+- Create credentials (Service Account Key) and download it as JSON.
+- Set the environment variable `GOOGLE_APPLICATION_CREDENTIALS` to the path of the downloaded JSON file.
+
+### Azure Text-To-Speech
+
+- Create a Text-To-Speech resource in the [Azure Portal](https://portal.azure.com/).
+- Navigate to the resource and find the "Keys" section. Copy one of the keys.
+- Search for a voice in the [Speech Studio](https://speech.microsoft.com/portal)
+- Set the environment variable `AZURE_SPEECH_RESOURCE_KEY` to the copied key and `AZURE_SPEECH_REGION` to the region where your resource is located.
+
+### Request
+
+```
+GET /tts.ogg
+  ?provider=google|azure
+  &lang=en-US
+  &text=The%20text%20to%20convert%20to%20speech
+  &valid-to=<RFC3339-Timestamp>
+  &voice=<name-of-the-voice>
+  &signature=<HMAC-SHA256-Signature>
+```
+
+- The `signature` is an HMAC-SHA256 signature of the request parameters, using the secret key
+  - It contains all parameters except for the signature itself
+  - The parameters are sorted by name
+  - The HMAC is generated over the parameters concatinated with `\n`: `param1=value1\nparam2=value2\n...`
+  - The signature is a lower-case hex encoding of the HMAC
+- The `valid-to` timestamp is a RFC3339 timestamp indicating when the request is valid
+
+So for example these would be valid URLs for the key `topsecret`:
+
+```
+http://localhost:3000/tts.ogg
+  ?lang=en-EN
+  &provider=google
+  &text=Hello%20there%2C%20general%20Kenobi%21
+  &valid-to=2025-01-31T01%3A22%3A17.405263Z
+  &voice=de-DE-Standard-G
+  &signature=afb82dc41b444f9573d585094cf4a22a517853b307d98031c9a324f294db026e
+
+http://localhost:3000/tts.ogg
+  ?lang=en-EN
+  &provider=azure
+  &text=Hello%20there%2C%20general%20Kenobi%21
+  &valid-to=2025-01-31T01%3A23%3A57.905524Z
+  &voice=en-US-AvaMultilingualNeural
+  &signature=ad4b15b78acd7d59a9d659b6b7a67dce2eee070a49634f9be66be3727eb1f5fc
+```

go.mod

@@ -9,6 +9,7 @@ require (
 	github.com/Luzifer/go_helpers/v2 v2.25.0
 	github.com/Luzifer/rconfig/v2 v2.5.2
 	github.com/sirupsen/logrus v1.9.3
+	github.com/stretchr/testify v1.10.0
 )
 
 require (
@@ -17,6 +18,7 @@ require (
 	cloud.google.com/go/auth/oauth2adapt v0.2.7 // indirect
 	cloud.google.com/go/compute/metadata v0.6.0 // indirect
 	cloud.google.com/go/longrunning v0.6.4 // indirect
+	github.com/davecgh/go-spew v1.1.1 // indirect
 	github.com/felixge/httpsnoop v1.0.4 // indirect
 	github.com/go-logr/logr v1.4.2 // indirect
 	github.com/go-logr/stdr v1.2.2 // indirect
@@ -24,6 +26,7 @@ require (
 	github.com/googleapis/enterprise-certificate-proxy v0.3.4 // indirect
 	github.com/googleapis/gax-go/v2 v2.14.1 // indirect
 	github.com/pkg/errors v0.9.1 // indirect
+	github.com/pmezard/go-difflib v1.0.0 // indirect
 	github.com/spf13/pflag v1.0.6 // indirect
 	go.opentelemetry.io/auto/sdk v1.1.0 // indirect
 	go.opentelemetry.io/contrib/instrumentation/google.golang.org/grpc/otelgrpc v0.59.0 // indirect

main.go

@@ -15,6 +15,7 @@ import (
 	httpHelper "github.com/Luzifer/go_helpers/v2/http"
 	"github.com/Luzifer/rconfig/v2"
 	"github.com/Luzifer/webtts/pkg/synth"
+	"github.com/Luzifer/webtts/pkg/synth/azure"
 	"github.com/Luzifer/webtts/pkg/synth/google"
 )
 
@@ -22,7 +23,7 @@ var (
 	cfg = struct {
 		Listen         string `flag:"listen" default:":3000" description:"Port/IP to listen on"`
 		LogLevel       string `flag:"log-level" default:"info" description:"Log level (debug, info, warn, error, fatal)"`
-		SignatureKey   string `flag:"signature-key" default:"" description:"Key to sign requests with" validate:"nonzero"`
+		SignatureKey   string `flag:"signature-key" default:"" description:"Key to sign requests with"`
 		VersionAndExit bool   `flag:"version" default:"false" description:"Prints current version and exits"`
 	}{}
 
@@ -41,6 +42,10 @@ func initApp() (err error) {
 	}
 	logrus.SetLevel(l)
 
+	if cfg.SignatureKey == "" {
+		logrus.Warn("no signature key is set, all requests are valid, do not do this in production!")
+	}
+
 	return nil
 }
 
@@ -101,6 +106,13 @@ func handleTTS(w http.ResponseWriter, r *http.Request) {
 
 	var p synth.Provider
 	switch provider {
+	case "azure":
+		if p, err = azure.New(); err != nil {
+			logrus.WithError(err).Error("creating azure provider")
+			http.Error(w, "creating provider", http.StatusInternalServerError)
+			return
+		}
+
 	case "google", "gcp":
 		if p, err = google.New(); err != nil {
 			logrus.WithError(err).Error("creating google provider")
@@ -128,6 +140,11 @@ func handleTTS(w http.ResponseWriter, r *http.Request) {
 }
 
 func checkSignature(signature string, r *http.Request) error {
+	if cfg.SignatureKey == "" {
+		// No key given, should only happen in development
+		return nil
+	}
+
 	keys := []string{}
 	for k := range r.URL.Query() {
 		if k == "signature" {

pkg/synth/azure/azure.go

@@ -0,0 +1,106 @@
+// Package azure provides a text-to-speech synthesis provider for Azure.
+package azure
+
+import (
+	"bytes"
+	"context"
+	"encoding/xml"
+	"fmt"
+	"io"
+	"net/http"
+	"os"
+
+	"github.com/Luzifer/webtts/pkg/synth"
+	"github.com/sirupsen/logrus"
+)
+
+type (
+	// Provider represents the Azure text-to-speech synthesis provider.
+	Provider struct{}
+
+	ssmlRequest struct {
+		XMLName xml.Name `xml:"speak"`
+		Text    string   `xml:",chardata"`
+		Version string   `xml:"version,attr"`
+		Lang    string   `xml:"xml:lang,attr"`
+		Voice   struct {
+			Text string `xml:",chardata"`
+			Name string `xml:"name,attr"`
+		} `xml:"voice"`
+	}
+)
+
+var _ synth.Provider = (*Provider)(nil)
+
+// New creates a new instance of the Azure text-to-speech synthesis provider.
+func New() (*Provider, error) {
+	speechKey := os.Getenv("AZURE_SPEECH_RESOURCE_KEY")
+	speechRegion := os.Getenv("AZURE_SPEECH_REGION")
+
+	if speechKey == "" || speechRegion == "" {
+		return nil, fmt.Errorf("missing environment variables: AZURE_SPEECH_RESOURCE_KEY and AZURE_SPEECH_REGION")
+	}
+
+	return &Provider{}, nil
+}
+
+// GenerateAudio generates audio from the given text using Azure's Text-to-Speech service.
+func (p Provider) GenerateAudio(ctx context.Context, voice, language, text string) ([]byte, error) {
+	speechKey := os.Getenv("AZURE_SPEECH_RESOURCE_KEY")
+	speechRegion := os.Getenv("AZURE_SPEECH_REGION")
+
+	body, err := p.requestSSML(voice, language, text)
+	if err != nil {
+		return nil, fmt.Errorf("generating SSML: %w", err)
+	}
+
+	req, err := http.NewRequestWithContext(ctx, http.MethodPost, p.apiURL(speechRegion), body)
+	if err != nil {
+		return nil, fmt.Errorf("creating request: %w", err)
+	}
+
+	req.Header.Set("Content-Type", "application/ssml+xml")
+	req.Header.Set("Ocp-Apim-Subscription-Key", speechKey)
+	req.Header.Set("User-Agent", "webtts/0.x (https://github.com/Luzifer/webtts)")
+	req.Header.Set("X-Microsoft-OutputFormat", "ogg-48khz-16bit-mono-opus")
+
+	resp, err := http.DefaultClient.Do(req)
+	if err != nil {
+		return nil, fmt.Errorf("requesting audio: %w", err)
+	}
+	defer func() {
+		if err := resp.Body.Close(); err != nil {
+			logrus.WithError(err).Error("closing response body")
+		}
+	}()
+
+	if resp.StatusCode != http.StatusOK {
+		return nil, fmt.Errorf("unexpected status code: %d", resp.StatusCode)
+	}
+
+	audioData, err := io.ReadAll(resp.Body)
+	if err != nil {
+		return nil, fmt.Errorf("reading audio data:%w", err)
+	}
+
+	return audioData, nil
+}
+
+func (Provider) apiURL(region string) string {
+	return fmt.Sprintf("https://%s.tts.speech.microsoft.com/cognitiveservices/v1", region)
+}
+
+func (Provider) requestSSML(voice, language, text string) (io.Reader, error) {
+	var req ssmlRequest
+	req.Lang = language
+	req.Version = "1.0"
+	req.Voice.Name = voice
+	req.Voice.Text = text
+
+	data, err := xml.Marshal(req)
+	if err != nil {
+		return nil, fmt.Errorf("marshalling ssml request: %w", err)
+	}
+
+	return bytes.NewReader(data), nil
+}

pkg/synth/azure/azure_test.go

@@ -0,0 +1,18 @@
+package azure
+
+import (
+	"io"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+func TestSSMLGeneration(t *testing.T) {
+	r, err := Provider{}.requestSSML("en-US-ChristopherNeural", "en-US", "excited to try text to speech!")
+	require.NoError(t, err)
+
+	data, err := io.ReadAll(r)
+	require.NoError(t, err)
+	assert.Equal(t, string(data), `<speak version="1.0" xml:lang="en-US"><voice name="en-US-ChristopherNeural">excited to try text to speech!</voice></speak>`)
+}