Fix bilingual initial greeting playback

main.py

@@ -97,25 +97,32 @@ def _single_initial_greeting_enabled() -> bool:
     return raw in {"1", "true", "yes", "on"}
 
 
+def _env_text_or_default(name: str, default: str) -> str:
+    value = os.getenv(name)
+    if value is None or not str(value).strip():
+        return default
+    return str(value).strip()
+
+
 def _initial_greeting_segments() -> tuple[tuple[str, str | None], ...]:
     raw = str(os.getenv("REALTIME_VOICE_INITIAL_GREETING_TEXT", "")).strip()
     if raw and _single_initial_greeting_enabled():
         language = str(os.getenv("REALTIME_VOICE_INITIAL_GREETING_LANGUAGE_CODE", "")).strip() or None
         return ((raw, language),)
 
-    ru_text = (
-        str(os.getenv("REALTIME_VOICE_INITIAL_GREETING_RU_TEXT") or "").strip()
-        or (
+    ru_text = _env_text_or_default(
+        "REALTIME_VOICE_INITIAL_GREETING_RU_TEXT",
+        (
             "Здравствуйте! Меня зовут Айнур, я эй-ай-ассистент компании ДиджиОпс. "
             "Как я могу к вам обращаться и на каком языке вам удобнее продолжить: на русском или на казахском?"
-        )
+        ),
     )
-    kk_text = (
-        str(os.getenv("REALTIME_VOICE_INITIAL_GREETING_KK_TEXT") or "").strip()
-        or (
+    kk_text = _env_text_or_default(
+        "REALTIME_VOICE_INITIAL_GREETING_KK_TEXT",
+        (
             "Сәлеметсіз бе! Менің атым Айнұр, мен ДиджиОпс компаниясының эй-ай ассистентімін. "
             "Сізге қалай жүгінсем болады және қай тілде сөйлескен ыңғайлы: орысша ма, қазақша ма?"
-        )
+        ),
     )
     segments: list[tuple[str, str | None]] = []
     if ru_text:
@@ -213,12 +220,32 @@ class RealtimeVoiceService:
                 len(greeting_text),
                 greeting_text[:120],
             )
-            clip = await self._filler_audio.synthesize_segments(self._tts, greeting_segments)
-            if not clip:
-                LOGGER.warning("initial greeting audio preload produced empty clip")
+            total_bytes = 0
+            for index, (segment_text, language_code) in enumerate(greeting_segments, start=1):
+                segment_clip = await self._filler_audio.synthesize_segments(
+                    self._tts,
+                    ((segment_text, language_code),),
+                )
+                if not segment_clip:
+                    LOGGER.warning(
+                        "initial greeting segment preload produced empty clip: index=%s language=%s",
+                        index,
+                        language_code,
+                    )
+                    continue
+                clip_key = f"initial_greeting_{index}_{language_code or 'default'}"
+                self._filler_audio.add_clip(clip_key, segment_clip)
+                total_bytes += len(segment_clip)
+                LOGGER.info(
+                    "initial greeting segment preload done: key=%s language=%s bytes=%s",
+                    clip_key,
+                    language_code,
+                    len(segment_clip),
+                )
+            if total_bytes <= 0:
+                LOGGER.warning("initial greeting audio preload produced no segment clips")
                 return
-            self._filler_audio.add_clip("initial_greeting", clip)
-            LOGGER.info("initial greeting audio preload done: bytes=%s", len(clip))
+            LOGGER.info("initial greeting audio preload done: bytes=%s", total_bytes)
         except Exception:
             LOGGER.exception("failed to preload initial greeting audio")