fix(converter): stream audio concat and harden error/encoding handling

1 files changed, 27 insertions, 19 deletions

diff --git a/converter/converter.py b/converter/converter.py
index eea84ed..65efb68 100644
--- a/converter/converter.py
+++ b/converter/converter.py
@@ -6,7 +6,7 @@ import time
 import traceback
 from datetime import datetime
 from pathlib import Path
-from typing import Optional
+from typing import Dict, Optional
 
 from . import audio, chunking, config, extractors
 from .tts import QwenTTSClient
@@ -60,17 +60,17 @@ class AudiobookConverter:
         """Validate configuration settings."""
         if self.voice_mode == "voice_clone":
             if not self.voice_clone_ref_audio:
-                print("[ERROR] Configuration Error!")
-                print("Voice Clone mode requires a reference audio file.")
-                print("Use --voice-sample <path> to specify the reference audio.")
-                sys.exit(1)
+                raise ValueError(
+                    "Voice Clone mode requires a reference audio file. "
+                    "Use --voice-sample <path> to specify it."
+                )
 
             if not Path(self.voice_clone_ref_audio).exists():
-                print("[ERROR] Configuration Error!")
-                print(f"Reference audio file not found: {self.voice_clone_ref_audio}")
-                sys.exit(1)
+                raise ValueError(
+                    f"Reference audio file not found: {self.voice_clone_ref_audio}"
+                )
 
-    def convert_book(self, file_path: Path) -> bool:
+    def convert_book(self, file_path: Path, output_name: Optional[str] = None) -> bool:
         """Convert a single book to an audiobook."""
         logger.info("Converting: %s", file_path.name)
         start_time = time.time()
@@ -136,7 +136,6 @@ class AudiobookConverter:
 
             if successful_chunks == 0:
                 logger.error("No chunks were successfully processed")
-                audio.cleanup_chunks()  # Cleanup even on failure
                 return False
 
             if successful_chunks < total_chunks:
@@ -144,7 +143,7 @@ class AudiobookConverter:
                                successful_chunks, total_chunks)
 
             # Combine chunks (only the successful ones)
-            output_path = config.AUDIOBOOKS_FOLDER / f"{file_path.stem}.{config.AUDIO_FORMAT}"
+            output_path = config.AUDIOBOOKS_FOLDER / f"{output_name or file_path.stem}.{config.AUDIO_FORMAT}"
             success = audio.combine_chunks(total_chunks, output_path, results, speed=self.speed)
 
             if success:
@@ -156,19 +155,18 @@ class AudiobookConverter:
             else:
                 logger.error("Failed to combine chunks into final audiobook")
 
-            # Always cleanup, even on failure
-            audio.cleanup_chunks()
             return success
 
         except Exception as exc:
             logger.error("Conversion failed: %s", exc)
             logger.error(traceback.format_exc())
-            # Cleanup on exception
-            audio.cleanup_chunks()
             return False
+        finally:
+            # Always cleanup, even on failure or interrupt
+            audio.cleanup_chunks()
 
-    def run(self) -> None:
-        """Main conversion process."""
+    def run(self) -> bool:
+        """Main conversion process. Returns True if all books converted."""
         api_url = config.VOICE_CLONE_API_URL if self.voice_mode == "voice_clone" else config.QWEN_API_URL
 
         print("=" * 70)
@@ -209,15 +207,23 @@ class AudiobookConverter:
                 encoding="utf-8",
             )
             print(f"[INFO] Created sample file: {sample_file}")
-            return
+            return True
 
         print(f"[INFO] Found {len(book_files)} books to convert")
 
+        # Avoid output collisions when two books share a stem (e.g. dune.txt + dune.epub).
+        stem_counts: Dict[str, int] = {}
+        for book_file in book_files:
+            stem_counts[book_file.stem] = stem_counts.get(book_file.stem, 0) + 1
+
         # Convert each book
         results = {}
         for book_file in book_files:
+            output_name = book_file.stem
+            if stem_counts[book_file.stem] > 1:
+                output_name = f"{book_file.stem}_{book_file.suffix.lstrip('.')}"
             try:
-                success = self.convert_book(book_file)
+                success = self.convert_book(book_file, output_name=output_name)
                 results[book_file.name] = success
             except KeyboardInterrupt:
                 print("\n[WARNING] Conversion interrupted by user")
@@ -242,3 +248,5 @@ class AudiobookConverter:
 
         if successful > 0:
             print(f"\n[INFO] Audiobooks saved to: {config.AUDIOBOOKS_FOLDER}/")
+
+        return total > 0 and successful == total