refactor: overhaul config.py, remove cli default options

1 files changed, 30 insertions, 20 deletions

diff --git a/app/converter/converter.py b/app/converter/converter.py
index 5235991..a9ea7eb 100644
--- a/app/converter/converter.py
+++ b/app/converter/converter.py
@@ -31,7 +31,7 @@ from .clients import (
     FasterTTSClient,
     QwenTTSClient,
     normalize_language,
-    speaker_display_name,
+    speaker_display_name_for,
 )
 
 logger = logging.getLogger(__name__)
@@ -207,7 +207,7 @@ class AudiobookConverter:
     def __init__(self, voice_mode: str = VOICE_MODE_CUSTOM, voice_clone_ref_audio: Optional[str] = None,
                  voice_clone_ref_text: Optional[str] = None, skip_transcription: bool = False,
                  speed: float = 1.0, single_file: bool = False, output_format: str = config.AUDIO_FORMAT,
-                 language: Optional[str] = None, backend: str = config.BACKEND,
+                 language: Optional[str] = None, backend: str = None,
                  voice: Optional[str] = None, debug: bool = False,
                  model_id: Optional[str] = None,
                  instructions: Optional[str] = None,
@@ -219,6 +219,8 @@ class AudiobookConverter:
             raise ValueError(f"Speed must be a positive number, got {speed}")
         if output_format not in AUDIO_FORMATS:
             raise ValueError(f"Unsupported output format: {output_format}")
+        if backend is None:
+            raise ValueError("backend is required (pass --backend)")
         if backend not in BACKENDS:
             raise ValueError(
                 f"Unknown backend: {backend!r} (expected one of {BACKENDS})"
@@ -257,10 +259,10 @@ class AudiobookConverter:
         elif backend == BACKEND_AUDIOCPP:
             # --voice picks the voice: a built-in speaker name on the
             # CustomVoice entry, or a server-side preset (cloning)
-            # elsewhere. model_id overrides AUDIOCPP_MODEL_ID for
-            # multi-model servers; instructions describe or style the
-            # voice, request_options pass per-model controls through to
-            # the server.
+            # elsewhere. model_id picks the server entry per run
+            # (auto-selected on single-entry servers); instructions
+            # describe or style the voice, request_options pass
+            # per-model controls through to the server.
             self.tts = AudioCppTTSClient(chunks_dir=CHUNKS_FOLDER,
                                          voice=voice, language=self.language,
                                          model_id=model_id,
@@ -270,6 +272,7 @@ class AudiobookConverter:
         else:
             # Qwen: the voice mode picks the request shape (built-in
             # speaker, clone from a reference .wav, or a designed voice);
+            # --voice names the built-in speaker in speaker mode and
             # instructions describe the voice in design mode.
             self.tts = QwenTTSClient(
                 chunks_dir=CHUNKS_FOLDER,
@@ -281,6 +284,7 @@ class AudiobookConverter:
                 instructions=self.instructions,
                 api_url=api_url,
                 quiet=quiet,
+                voice=voice,
             )
         self._progress = progress
         self.tts.cancel = cancel
@@ -314,6 +318,13 @@ class AudiobookConverter:
                 "Voice Design mode requires a voice description. "
                 "Use --instructions \"...\" to describe the voice to synthesize with."
             )
+        if self.backend == BACKEND_QWEN and self.voice_mode == VOICE_MODE_CUSTOM \
+                and not (self.voice or "").strip():
+            raise ValueError(
+                "CustomVoice mode requires a speaker. Use --voice SPEAKER "
+                "(e.g. Vivian) to pick one, or --clone / --instructions "
+                "for the other voice modes."
+            )
         if self.voice_mode == VOICE_MODE_CLONE and self.backend == BACKEND_QWEN:
             if not self.voice_clone_ref_audio:
                 raise ValueError(
@@ -348,19 +359,18 @@ class AudiobookConverter:
 
         Custom voice mode uses the built-in speaker's display name; voice
         clone mode uses the reference audio file's stem; the faster and
-        audiocpp backends use the server-side voice name (or, for audiocpp's
-        speaker mode, the selected built-in CustomVoice speaker, falling back
-        to the configured one). An instruction without a voice (voice design,
-        or instruction-defined voices on families without built-in speakers)
-        uses "designed". Spaces become underscores (e.g. "Uncle Fu" ->
-        "Uncle_Fu").
+        audiocpp backends use the server-side voice name (for audiocpp's
+        speaker mode, the selected built-in CustomVoice speaker). An
+        instruction without a voice (voice design, or instruction-defined
+        voices on families without built-in speakers) uses "designed".
+        Spaces become underscores (e.g. "Uncle Fu" -> "Uncle_Fu").
 
         Pure (no I/O, no server) so the pre-flight overwrite check can
         compute the exact output names a run would produce before spending
         time connecting to a TTS server.
         """
         if backend == BACKEND_FASTER:
-            narrator = voice or config.FASTER_VOICE
+            narrator = voice or "default"
         elif backend == BACKEND_AUDIOCPP:
             if voice:
                 narrator = voice
@@ -368,7 +378,10 @@ class AudiobookConverter:
                 # The voice comes from the instruction, not a speaker name.
                 narrator = "designed"
             else:
-                narrator = speaker_display_name()
+                # Unreachable in a valid run (the audiocpp client refuses a
+                # speaker-capable entry without --voice); keep a stable tag
+                # for the pre-flight of runs that will fail at connect time.
+                narrator = "narrator"
         elif voice_mode == VOICE_MODE_DESIGN:
             # Qwen's VoiceDesign model: the voice is described by an
             # instruction and has no speaker name.
@@ -376,7 +389,7 @@ class AudiobookConverter:
         elif voice_mode == VOICE_MODE_CLONE:
             narrator = Path(voice_clone_ref_audio).stem
         else:
-            narrator = speaker_display_name()
+            narrator = speaker_display_name_for(voice or "")
         return AudiobookConverter._sanitize_filename(
             narrator, fallback="narrator").replace(" ", "_")
 
@@ -736,7 +749,7 @@ class AudiobookConverter:
         if self.backend == BACKEND_FASTER:
             self._say(f"Faster TTS endpoint: {config.FASTER_API_URL}")
             self._say("Backend: faster (voice cloning, reference configured on server)")
-            self._say(f"Voice: {self.voice or config.FASTER_VOICE}")
+            self._say(f"Voice: {self.voice}")
         elif self.backend == BACKEND_AUDIOCPP:
             self._say(f"audio.cpp endpoint: {config.AUDIOCPP_API_URL}")
             self._say(f"Model id: {self.tts.model_id}")
@@ -750,9 +763,6 @@ class AudiobookConverter:
             elif self.instructions:
                 self._say("Backend: audio.cpp (voice from --instructions description)")
                 self._say(f"Instruction: {self.instructions}")
-            else:
-                self._say("Backend: audio.cpp (custom voice, built-in speaker)")
-                self._say(f"Speaker: {config.SPEAKER}")
             if self.request_options:
                 self._say(f"Request options: {self.request_options}")
             self._say(f"Language: {self.language}")
@@ -764,7 +774,7 @@ class AudiobookConverter:
             self._say(f"Voice mode: {self.voice_mode}")
             self._say(f"Model size: {MODEL_SIZE} (always)")
             if self.voice_mode == VOICE_MODE_CUSTOM:
-                self._say(f"Speaker: {config.SPEAKER}")
+                self._say(f"Speaker: {self.voice}")
                 self._say(f"Language: {self.language}")
             elif self.voice_mode == VOICE_MODE_CLONE:
                 self._say(f"Reference audio: {Path(self.voice_clone_ref_audio).name}")