summaryrefslogtreecommitdiff
path: root/internal/audio
diff options
context:
space:
mode:
authorPaul Buetow <paul@buetow.org>2026-04-02 22:26:43 +0300
committerPaul Buetow <paul@buetow.org>2026-04-02 22:26:43 +0300
commit5a1a6b863c3adaa8ec9d087ba130f7676fc9575b (patch)
tree31494a962d15f46bed22f8ba090afe7511dc54c7 /internal/audio
parent842ce63eda42a5dfa1599bf69b5ea22b332a954d (diff)
task zz: add Voices()/BuildAttribution() to Provider interface [OCP]
Extend audio.Provider with Voices() []string and BuildAttribution() string so all provider-specific behaviour is encapsulated in the implementation rather than scattered as switch-cases across callers. Add package-level VoicesFor(name) and BuildAttributionFor(name, params) for callers (processor, GUI) that need these before constructing a Provider instance. Add AttributionParamsFrom(config, word, ...) so callers can build AttributionParams from the flat Config without a manual provider switch. Implement both new interface methods in OpenAIProvider and GeminiProvider. Update all Provider mock/fake types in tests. Migrate audioVoicesForProvider() and saveAudioAttribution() in both processor.go and gui/generator.go to use the new package-level helpers, replacing the 10+ duplicated switch blocks. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Diffstat (limited to 'internal/audio')
-rw-r--r--internal/audio/attribution.go28
-rw-r--r--internal/audio/gemini_provider.go10
-rw-r--r--internal/audio/openai_provider.go10
-rw-r--r--internal/audio/provider.go36
-rw-r--r--internal/audio/provider_test.go8
5 files changed, 88 insertions, 4 deletions
diff --git a/internal/audio/attribution.go b/internal/audio/attribution.go
index 7c3cda2..57f8c1d 100644
--- a/internal/audio/attribution.go
+++ b/internal/audio/attribution.go
@@ -18,6 +18,34 @@ type AttributionParams struct {
GeneratedAt time.Time
}
+// AttributionParamsFrom builds an AttributionParams from a flat Config and the
+// word being transcribed, reading provider-appropriate sub-config fields to
+// eliminate the need for callers to switch on config.Provider themselves.
+func AttributionParamsFrom(config *Config, word, instruction, processedText string, generatedAt time.Time) AttributionParams {
+ base := AttributionParams{
+ Word: word,
+ Instruction: instruction,
+ ProcessedText: processedText,
+ GeneratedAt: generatedAt,
+ }
+ if config == nil {
+ return base
+ }
+ switch strings.ToLower(strings.TrimSpace(config.Provider)) {
+ case "gemini":
+ g := geminiAudioConfigFrom(config)
+ base.Model = g.TTSModel
+ base.Voice = g.Voice
+ base.Speed = g.Speed
+ default:
+ o := openAIAudioConfigFrom(config)
+ base.Model = o.Model
+ base.Voice = o.Voice
+ base.Speed = o.Speed
+ }
+ return base
+}
+
// AttributionPath returns the sidecar attribution file path for a generated audio file.
func AttributionPath(audioFile string) string {
return strings.TrimSuffix(audioFile, filepath.Ext(audioFile)) + "_attribution.txt"
diff --git a/internal/audio/gemini_provider.go b/internal/audio/gemini_provider.go
index bda01f5..5100c5e 100644
--- a/internal/audio/gemini_provider.go
+++ b/internal/audio/gemini_provider.go
@@ -97,6 +97,16 @@ func (p *GeminiProvider) Name() string {
return "gemini"
}
+// Voices returns the list of Gemini voices supported by the app.
+func (p *GeminiProvider) Voices() []string {
+ return GeminiVoices
+}
+
+// BuildAttribution returns the Gemini attribution text for a generated audio file.
+func (p *GeminiProvider) BuildAttribution(params AttributionParams) string {
+ return BuildGeminiAttribution(params)
+}
+
// IsAvailable checks if the Google API key is configured.
func (p *GeminiProvider) IsAvailable() error {
if p == nil || strings.TrimSpace(p.config.APIKey) == "" {
diff --git a/internal/audio/openai_provider.go b/internal/audio/openai_provider.go
index 9c6d3a4..e6212f8 100644
--- a/internal/audio/openai_provider.go
+++ b/internal/audio/openai_provider.go
@@ -147,6 +147,16 @@ func (p *OpenAIProvider) IsAvailable() error {
return nil
}
+// Voices returns the list of OpenAI voices supported by the app.
+func (p *OpenAIProvider) Voices() []string {
+ return OpenAIVoices
+}
+
+// BuildAttribution returns the OpenAI attribution text for a generated audio file.
+func (p *OpenAIProvider) BuildAttribution(params AttributionParams) string {
+ return BuildOpenAIAttribution(params)
+}
+
// preprocessBulgarianText prepares Bulgarian text for clearer TTS pronunciation
func (p *OpenAIProvider) preprocessBulgarianText(text string) string {
return openAIProcessedText(text)
diff --git a/internal/audio/provider.go b/internal/audio/provider.go
index fac0cd9..60be848 100644
--- a/internal/audio/provider.go
+++ b/internal/audio/provider.go
@@ -3,18 +3,27 @@ package audio
import (
"context"
"fmt"
+ "strings"
)
-// Provider defines the interface for text-to-speech providers
+// Provider defines the interface for text-to-speech providers.
+// All provider-specific behavior (voices, attribution) is encapsulated here
+// so callers never need to switch on the provider name (OCP).
type Provider interface {
- // GenerateAudio generates audio from text and saves it to the specified file
+ // GenerateAudio generates audio from text and saves it to the specified file.
GenerateAudio(ctx context.Context, text string, outputFile string) error
- // Name returns the provider name
+ // Name returns the provider name.
Name() string
- // IsAvailable checks if the provider is properly configured and available
+ // IsAvailable checks if the provider is properly configured and available.
IsAvailable() error
+
+ // Voices returns the list of voice names supported by this provider.
+ Voices() []string
+
+ // BuildAttribution returns the attribution text for a generated audio file.
+ BuildAttribution(params AttributionParams) string
}
// OpenAIAudioConfig holds settings specific to the OpenAI TTS backend.
@@ -58,6 +67,25 @@ type Config struct {
GeminiSpeed float64 // Prompt hint for desired speech speed
}
+// VoicesFor returns the voice list for the named provider. This is a
+// convenience for callers that need voices before constructing a Provider.
+func VoicesFor(providerName string) []string {
+ if strings.ToLower(strings.TrimSpace(providerName)) == "gemini" {
+ return GeminiVoices
+ }
+ return OpenAIVoices
+}
+
+// BuildAttributionFor builds the attribution text for the named provider
+// without requiring a Provider instance. Use Provider.BuildAttribution when
+// you already have an instance.
+func BuildAttributionFor(providerName string, params AttributionParams) string {
+ if strings.ToLower(strings.TrimSpace(providerName)) == "gemini" {
+ return BuildGeminiAttribution(params)
+ }
+ return BuildOpenAIAttribution(params)
+}
+
// openAIAudioConfigFrom extracts the OpenAI-specific sub-config from the flat Config.
// A nil Config produces a zero-value OpenAIAudioConfig.
func openAIAudioConfigFrom(c *Config) OpenAIAudioConfig {
diff --git a/internal/audio/provider_test.go b/internal/audio/provider_test.go
index 5702016..9b8037c 100644
--- a/internal/audio/provider_test.go
+++ b/internal/audio/provider_test.go
@@ -28,6 +28,14 @@ func (m *mockProvider) IsAvailable() error {
return m.availableErr
}
+func (m *mockProvider) Voices() []string {
+ return nil
+}
+
+func (m *mockProvider) BuildAttribution(params AttributionParams) string {
+ return ""
+}
+
func TestDefaultProviderConfig(t *testing.T) {
config := DefaultProviderConfig()