v1.0.8

CHANGELOG.md

@@ -1,3 +1,18 @@
+# v1.0.8
+- Added support for AMD GPUs in Linux (Special thanks to @hg000125 for his contribution in #23)
+- Added voice preview caching system that stores generated previews in the cache folder, mentioned by @jborza in #22
+- Added extra metadata support for chaptered M4B files, ensuring better compatibility with audiobook players.
+- Added new option: `Separate chapters audio format`, allowing to choose between `wav`, `mp4`, `flac` and `opus` formats for chaptered audio files.
+- Added a download tracker that displays informative messages while downloading Kokoro models or voices from HuggingFace.
+- Skipping PyTorch CUDA installation if GPU is not NVIDIA in WINDOWS_INSTALL.bat script, preventing unnecessary installation of PyTorch.
+- Removed `abogen_` prefix that was adding to converted books in temp directory.
+- Fixed voice preview player keeps playing silently at the background after preview ends.
+- Fixed not writing separate chapters audio when output is OPUS.
+- Improved input box background color handling, fixed display issues in Linux.
+- Updated profile and voice mixer icons, better visibility and aesthetics in voice mixer.
+- Better sleep state handling for Linux.
+- Improvements in documentation and code.
+
 # v1.0.7
 - Improve chaptered audio generation by outputting directly as `m4b` instead of converting from `wav`.
 - Ignore chapter markers and single newlines when calculating text length, improving the accuracy of the text length calculation.

README.md

@@ -23,6 +23,9 @@ Go to [espeak-ng latest release](https://github.com/espeak-ng/espeak-ng/releases
 # For NVIDIA GPUs:
 pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu128
 
+# For AMD GPUs:
+# Not supported yet, because ROCm is not available on Windows. Use Linux if you have AMD GPU.
+
 # Install abogen
 pip install abogen
 ```
@@ -54,6 +57,13 @@ sudo dnf install espeak-ng
 
 # Install abogen
 pip install abogen
+
+# For NVIDIA GPUs:
+# Already supported, no need to install CUDA separately.
+
+# For AMD GPUs:
+pip uninstall torch # After installing abogen, we need to uninstall the existing torch package
+pip install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/rocm6.4
 ```
 > If you get `WARNING: The script abogen-cli is installed in '/home/username/.local/bin' which is not on PATH.` error, run the following command to add it to your PATH:
 >```bash
@@ -62,6 +72,8 @@ pip install abogen
 
 > If you get "No matching distribution found" error, try installing it on supported Python (3.10 to 3.12). You can use [pyenv](https://github.com/pyenv/pyenv) to manage multiple Python versions easily in Linux. Watch this [video](https://www.youtube.com/watch?v=MVyb-nI4KyI) by NetworkChuck for a quick guide.
 
+> Special thanks to [@hg000125](https://github.com/hg000125) for his contribution in [#23](https://github.com/denizsafak/abogen/issues/23). AMD GPU support is possible thanks to his work.
+
 ## `How to run?`
 If you installed using pip, you can simply run the following command to start Abogen:
 
@@ -99,10 +111,11 @@ Here’s Abogen in action: in this demo, it processes ∼3,000 characters of tex
 - **Options**:
     - **Replace single newlines with spaces**: Replaces single newlines with spaces in the text. This is useful for texts that have imaginary line breaks.
     - **Configure max words per subtitle**: Configures the maximum number of words per subtitle entry.
+    - **Separate chapters audio format**: Configures the audio format for separate chapters (e.g., `WAV`, `FLAC`, `MP3`, `OPUS`).
     - **Create desktop shortcut**: Creates a shortcut on your desktop for easy access.
     - **Open config.json directory**: Opens the directory where the configuration file is stored.
     - **Open temp directory**: Opens the temporary directory where converted text files are stored.
-    - **Clear all temporary files**: Deletes all temporary files created during the conversion process.
+    - **Clear temporary files**: Deletes temporary files created during the conversion or preview.
     - **Check for updates at startup**: Automatically checks for updates when the program starts.
 - **After conversion**: `Open file`, `Go to folder`, `New conversion`, or `Go back`.
 
@@ -205,6 +218,7 @@ Abogen is a standalone project, but it is inspired by and shares some similariti
 - [autiobooks](https://github.com/plusuncold/autiobooks): Automatically convert epubs to audiobooks
 - [pdf-narrator](https://github.com/mateogon/pdf-narrator): Convert your PDFs and EPUBs into audiobooks effortlessly.
 - [epub_to_audiobook](https://github.com/p0n1/epub_to_audiobook): EPUB to audiobook converter, optimized for Audiobookshelf
+- [ebook2audiobook](https://github.com/DrewThomasson/ebook2audiobook): Convert ebooks to audiobooks with chapters and metadata using dynamic AI models and voice cloning
 
 ## `Roadmap`
 - [ ] Add OCR scan feature for PDF files using docling.
@@ -239,7 +253,7 @@ Feel free to explore the code and make any changes you like.
 - Thanks to [@wojiushixiaobai](https://github.com/wojiushixiaobai) for [Embedded Python](https://github.com/wojiushixiaobai/Python-Embed-Win64) packages. These modified packages include pip pre-installed, enabling Abogen to function as a standalone application without requiring users to separately install Python in Windows.
 - Thanks to creators of [EbookLib](https://github.com/aerkalov/ebooklib), a Python library for reading and writing ePub files, which is used for extracting text from ePub files.
 - Special thanks to the [PyQt](https://www.riverbankcomputing.com/software/pyqt/) team for providing the cross-platform GUI toolkit that powers Abogen's interface.
-- Icons: [US](https://icons8.com/icon/aRiu1GGi6Aoe/usa), [Great Britain](https://icons8.com/icon/t3NE3BsOAQwq/great-britain), [Spain](https://icons8.com/icon/ly7tzANRt33n/spain), [France](https://icons8.com/icon/3muzEmi4dpD5/france), [India](https://icons8.com/icon/esGVrxg9VCJ1/india), [Italy](https://icons8.com/icon/PW8KZnP7qXzO/italy), [Japan](https://icons8.com/icon/McQbrq9qaQye/japan), [Brazil](https://icons8.com/icon/zHmH8HpOmM90/brazil), [China](https://icons8.com/icon/Ej50Oe3crXwF/china), [Female](https://icons8.com/icon/uI49hxbpxTkp/female), [Male](https://icons8.com/icon/12351/male) and [Voice Id](https://icons8.com/icon/GskSeVoroQ7u/voice-id) icons by [Icons8](https://icons8.com/).
+- Icons: [US](https://icons8.com/icon/aRiu1GGi6Aoe/usa), [Great Britain](https://icons8.com/icon/t3NE3BsOAQwq/great-britain), [Spain](https://icons8.com/icon/ly7tzANRt33n/spain), [France](https://icons8.com/icon/3muzEmi4dpD5/france), [India](https://icons8.com/icon/esGVrxg9VCJ1/india), [Italy](https://icons8.com/icon/PW8KZnP7qXzO/italy), [Japan](https://icons8.com/icon/McQbrq9qaQye/japan), [Brazil](https://icons8.com/icon/zHmH8HpOmM90/brazil), [China](https://icons8.com/icon/Ej50Oe3crXwF/china), [Female](https://icons8.com/icon/uI49hxbpxTkp/female), [Male](https://icons8.com/icon/12351/male), [Adjust](https://icons8.com/icon/21698/adjust) and [Voice Id](https://icons8.com/icon/GskSeVoroQ7u/voice-id) icons by [Icons8](https://icons8.com/).
 
 ## `License`
 This project is available under the MIT License - see the [LICENSE](https://github.com/denizsafak/abogen/blob/main/LICENSE) file for details.

WINDOWS_INSTALL.bat

@@ -230,13 +230,22 @@ if errorlevel 1 (
 
 :: Install setup requirements
 echo Installing setup requirements...
-%PYTHON_CONSOLE_PATH% -m pip install --upgrade setuptools wheel sphinx hatchling --no-warn-script-location
+%PYTHON_CONSOLE_PATH% -m pip install --upgrade setuptools setuptools-scm wheel sphinx hatchling --no-warn-script-location
 if errorlevel 1 (
     echo Failed to install setup requirements.
     pause
     exit /b
 )
 
+:: Install gpustat
+echo Installing gpustat...
+%PYTHON_CONSOLE_PATH% -m pip install gpustat --no-warn-script-location
+if errorlevel 1 (
+    echo Failed to install gpustat.
+    pause
+    exit /b
+)
+
 :: Install project and dependencies from pyproject.toml
 echo Checking and installing project dependencies...
 if exist %PYPROJECT_FILE% (
@@ -263,11 +272,15 @@ if "%MISAKI_LANG%" NEQ "en" (
     )
 )
 
+:: Check for NVIDIA GPU via is_nvidia()
+for /f %%i in ('%PYTHON_CONSOLE_PATH% -c "from abogen.is_nvidia import check; print(check())"') do set IS_NVIDIA=%%i
+
 :: Check if torch is installed with CUDA support
 echo Checking CUDA availability...
-for /f %%i in ('%PYTHON_CONSOLE_PATH% -c "from torch.cuda import is_available; print(is_available())"') do set cuda_available=%%i
+if /I "%IS_NVIDIA%"=="true" (
+    for /f %%i in ('%PYTHON_CONSOLE_PATH% -c "from torch.cuda import is_available; print(is_available())"') do set cuda_available=%%i
 
-if "%cuda_available%"=="False" (
+    if "%cuda_available%"=="False" (
         echo Installing PyTorch with CUDA %CUDA_VERSION% support...
         %PYTHON_CONSOLE_PATH% -m pip install torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/cu%CUDA_VERSION% --no-warn-script-location
         echo.
@@ -276,8 +289,11 @@ if "%cuda_available%"=="False" (
             pause
             exit /b
         )
+    ) else (
+        echo CUDA is available on NVIDIA GPU.
+    )
 ) else (
-    echo CUDA is available.
+    echo GPU is not NVIDIA. Skipping PyTorch CUDA installation.
 )
 
 :: Ask user if they want to create a desktop shortcut

abogen/VERSION

@@ -1 +1 @@
-1.0.7
+1.0.8

abogen/book_handler.py

@@ -1494,8 +1494,13 @@ class HandlerDialog(QDialog):
             authors_text = ", ".join(self.book_metadata["authors"])
             html_content += f"<p style='text-align: center; font-style: italic;'>By {authors_text}</p>"
 
+        if self.book_metadata["publisher"] or self.book_metadata.get("publication_year"):
+            pub_info = []
             if self.book_metadata["publisher"]:
-            html_content += f"<p style='text-align: center;'>Published by {self.book_metadata['publisher']}</p>"
+                pub_info.append(f"Published by {self.book_metadata['publisher']}")
+            if self.book_metadata.get("publication_year"):
+                pub_info.append(f"Year: {self.book_metadata['publication_year']}")
+            html_content += f"<p style='text-align: center;'>{' | '.join(pub_info)}</p>"
 
         html_content += "<hr/>"
 
@@ -1517,24 +1522,51 @@ class HandlerDialog(QDialog):
             "description": None,
             "cover_image": None,
             "publisher": None,
+            "publication_year": None,
         }
 
         if self.file_type == "epub":
+            try:
                 title_items = self.book.get_metadata("DC", "title")
-            if title_items:
+                if title_items and len(title_items) > 0:
                     metadata["title"] = title_items[0][0]
+            except Exception as e:
+                logging.warning(f"Error extracting title metadata: {e}")
 
+            try:
                 author_items = self.book.get_metadata("DC", "creator")
                 if author_items:
-                metadata["authors"] = [author[0] for author in author_items]
+                    metadata["authors"] = [author[0] for author in author_items if len(author) > 0]
+            except Exception as e:
+                logging.warning(f"Error extracting author metadata: {e}")
 
+            try:
                 desc_items = self.book.get_metadata("DC", "description")
-            if desc_items:
+                if desc_items and len(desc_items) > 0:
                     metadata["description"] = desc_items[0][0]
+            except Exception as e:
+                logging.warning(f"Error extracting description metadata: {e}")
 
+            try:
                 publisher_items = self.book.get_metadata("DC", "publisher")
-            if publisher_items:
+                if publisher_items and len(publisher_items) > 0:
                     metadata["publisher"] = publisher_items[0][0]
+            except Exception as e:
+                logging.warning(f"Error extracting publisher metadata: {e}")
+                
+            # Try to extract publication year
+            try:
+                date_items = self.book.get_metadata("DC", "date")
+                if date_items and len(date_items) > 0:
+                    date_str = date_items[0][0]
+                    # Try to extract just the year from the date string
+                    year_match = re.search(r'\b(19|20)\d{2}\b', date_str)
+                    if year_match:
+                        metadata["publication_year"] = year_match.group(0)
+                    else:
+                        metadata["publication_year"] = date_str
+            except Exception as e:
+                logging.warning(f"Error extracting publication date metadata: {e}")
 
             for item in self.book.get_items_of_type(ebooklib.ITEM_COVER):
                 metadata["cover_image"] = item.get_content()
@@ -1565,6 +1597,18 @@ class HandlerDialog(QDialog):
 
                 metadata["publisher"] = pdf_info.get("creator", None)
                 
+                # Try to extract publication date from PDF metadata
+                if "creationDate" in pdf_info:
+                    date_str = pdf_info["creationDate"]
+                    year_match = re.search(r'D:(\d{4})', date_str)
+                    if year_match:
+                        metadata["publication_year"] = year_match.group(1)
+                elif "modDate" in pdf_info:
+                    date_str = pdf_info["modDate"]
+                    year_match = re.search(r'D:(\d{4})', date_str)
+                    if year_match:
+                        metadata["publication_year"] = year_match.group(1)
+
             if len(self.pdf_doc) > 0:
                 try:
                     pix = self.pdf_doc[0].get_pixmap(matrix=fitz.Matrix(2, 2))
@@ -1580,10 +1624,43 @@ class HandlerDialog(QDialog):
         else:
             return self._get_pdf_selected_text()
 
+    def _format_metadata_tags(self):
+        """Format metadata tags for insertion at the beginning of the text"""
+        import datetime
+        
+        metadata = self.book_metadata
+        filename = os.path.splitext(os.path.basename(self.book_path))[0]
+        current_year = str(datetime.datetime.now().year)
+        
+        # Get values with fallbacks
+        title = metadata.get("title") or filename
+        authors = metadata.get("authors") or ["Unknown"]
+        authors_text = ", ".join(authors)
+        album_artist = authors_text or "Unknown"
+        year = metadata.get("publication_year") or current_year  # Use publication year if available
+        
+        # Count chapters/pages
+        total_chapters = len(self.checked_chapters)
+        chapter_text = f"{total_chapters} {'Chapters' if self.file_type == 'epub' else 'Pages'}"
+        
+        # Format metadata tags
+        metadata_tags = [
+            f"<<METADATA_TITLE:{title}>>",
+            f"<<METADATA_ARTIST:{authors_text}>>",
+            f"<<METADATA_ALBUM:{title} ({chapter_text})>>",
+            f"<<METADATA_YEAR:{year}>>",
+            f"<<METADATA_ALBUM_ARTIST:{album_artist}>>"
+        ]
+        
+        return "\n".join(metadata_tags)
+
     def _get_epub_selected_text(self):
         all_checked_identifiers = set()
         chapter_texts = []
 
+        # Add metadata tags at the beginning
+        metadata_tags = self._format_metadata_tags()
+        
         item_order_counter = 0
         ordered_checked_items = []
 
@@ -1608,7 +1685,7 @@ class HandlerDialog(QDialog):
                 marker = f"<<CHAPTER_MARKER:{title}>>"
                 chapter_texts.append(marker + "\n" + text)
 
-        full_text = "\n\n".join(chapter_texts)
+        full_text = metadata_tags + "\n\n" + "\n\n".join(chapter_texts)
         return full_text, all_checked_identifiers
 
     def _get_pdf_selected_text(self):
@@ -1617,6 +1694,9 @@ class HandlerDialog(QDialog):
         section_titles = []
         all_content = []
 
+        # Add metadata tags at the beginning
+        metadata_tags = self._format_metadata_tags()
+
         pdf_has_no_bookmarks = (
             hasattr(self, "has_pdf_bookmarks") and not self.has_pdf_bookmarks
         )
@@ -1641,7 +1721,7 @@ class HandlerDialog(QDialog):
                     if text:
                         all_content.append(text)
                         included_text_ids.add(page_id)
-            return "\n\n".join(all_content), all_checked_identifiers
+            return metadata_tags + "\n\n" + "\n\n".join(all_content), all_checked_identifiers
 
         iterator = QTreeWidgetItemIterator(self.treeWidget)
         while iterator.value():
@@ -1699,7 +1779,7 @@ class HandlerDialog(QDialog):
                         included_text_ids.add(identifier)
             iterator += 1
 
-        return "\n\n".join([t[1] for t in section_titles]), all_checked_identifiers
+        return metadata_tags + "\n\n" + "\n\n".join([t[1] for t in section_titles]), all_checked_identifiers
 
     def on_save_chapters_changed(self, state):
         self.save_chapters_separately = bool(state)

abogen/conversion.py

@@ -4,6 +4,7 @@ import tempfile
 import time
 import chardet
 import charset_normalizer
+import hashlib  # For generating unique cache filenames
 from platformdirs import user_desktop_dir
 from PyQt5.QtCore import QThread, pyqtSignal, Qt
 from PyQt5.QtWidgets import QCheckBox, QVBoxLayout, QDialog, QLabel, QDialogButtonBox
@@ -11,6 +12,7 @@ import soundfile as sf
 from utils import clean_text, create_process
 from constants import PROGRAM_NAME, LANGUAGE_DESCRIPTIONS, SAMPLE_VOICE_TEXTS
 from voice_formulas import get_new_voice
+import hf_tracker
 import static_ffmpeg
 import threading  # for efficient waiting
 import subprocess
@@ -160,6 +162,7 @@ class ConversionThread(QThread):
             f"\nVoice: {self.voice}\nLanguage: {self.lang_code}\nSpeed: {self.speed}\nGPU: {self.use_gpu}\nFile: {self.file_name}\nSubtitle mode: {self.subtitle_mode}\nOutput format: {self.output_format}\nSave option: {self.save_option}\n"
         )
         try:
+            hf_tracker.set_log_callback(lambda msg: self.log_updated.emit(msg))
             # Show configuration
             self.log_updated.emit("Configuration:")
             # Use display_path for logs if available, otherwise use the actual file name
@@ -196,6 +199,9 @@ class ConversionThread(QThread):
                     self.log_updated.emit(
                         f"- Merge chapters at the end: {'Yes' if merge_at_end else 'No'}"
                     )
+                    # Display the separate chapters format if it's set
+                    separate_format = getattr(self, 'separate_chapters_format', 'wav')
+                    self.log_updated.emit(f"- Separate chapters format: {separate_format}")
 
             if self.save_option == "Choose output folder":
                 self.log_updated.emit(
@@ -221,6 +227,10 @@ class ConversionThread(QThread):
             # Clean up text using utility function
             text = clean_text(text)
             
+            # Remove metadata markers from the text to be processed
+            metadata_pattern = r"<<METADATA_[^:]+:[^>]*>>"
+            text = re.sub(metadata_pattern, "", text)
+
             # --- Chapter splitting logic ---
             chapter_pattern = r"<<CHAPTER_MARKER:(.*?)>>"
             chapter_splits = list(re.finditer(chapter_pattern, text))
@@ -500,18 +510,20 @@ class ConversionThread(QThread):
 
                     # Concatenate chapter audio and save
                     chapter_audio = self.np.concatenate(chapter_audio_segments)
-                    # Determine chapter extension (.wav for m4b output)
+                    
-                    chapter_ext = 'wav' if self.output_format == 'm4b' else self.output_format
+                    # Use separate_chapters_format
+                    separate_format = getattr(self, 'separate_chapters_format', 'wav')
+                    
                     chapter_out_path = os.path.join(
-                        chapters_out_dir, f"{chapter_filename}.{chapter_ext}"
+                        chapters_out_dir, f"{chapter_filename}.{separate_format}"
                     )
-                    if self.output_format == "opus":
+                    if separate_format == "opus":
                         static_ffmpeg.add_paths()
                         proc = create_process(
                             [
                                 "ffmpeg",
                                 "-y",
-                                "-thread_queue_size", "1024", # Increased thread_queue_size for chapter opus
+                                "-thread_queue_size", "32768", # Increased thread_queue_size for chapter opus
                                 "-f",
                                 "f32le",
                                 "-ar",
@@ -537,7 +549,7 @@ class ConversionThread(QThread):
                             chapter_out_path,
                             chapter_audio,
                             24000,
-                            format='wav' if self.output_format == 'm4b' else self.output_format,
+                            format=separate_format,
                         )
 
                     # Generate .srt subtitle file for chapter if not Disabled
@@ -597,7 +609,7 @@ class ConversionThread(QThread):
                     opus_out_path = f"{base_filepath_no_ext}.opus"
                     ffmpeg_cmd_opus = [
                         "ffmpeg", "-y",
-                        "-thread_queue_size", "1024", # Increased thread_queue_size
+                        "-thread_queue_size", "32768", # Increased thread_queue_size
                         "-f", "f32le", "-ar", "24000", "-ac", "1", "-i", "pipe:0",
                         "-c:a", "libopus", "-b:a", "24000", # Original bitrate
                         opus_out_path,
@@ -701,13 +713,20 @@ class ConversionThread(QThread):
             static_ffmpeg.add_paths()
             ffmpeg_cmd = [
                 "ffmpeg", "-y",
-                "-thread_queue_size", "1024", # Increased thread_queue_size
+                "-thread_queue_size", "32768", # Increased thread_queue_size
                 "-f", "f32le", "-ar", "24000", "-ac", "1", "-i", "pipe:0",
                 "-i", chapters_info_path,
                 "-map", "0:a",
                 "-map_metadata", "1", 
                 "-map_chapters", "1",
-                "-c:a", "aac", "-b:a", "128k", # Explicitly AAC with a common bitrate
+                # Add metadata tags
+                "-metadata", "composer=Narrator",
+                "-metadata", "genre=Audiobook",
+                # Extract metadata from text markers if available
+                *self._extract_and_add_metadata_tags_to_ffmpeg_cmd(),
+                "-c:a", "aac",
+                "-q:a", "2", # Quality-based VBR for better quality control (lower = better, range 0.1-2)
+                "-movflags", "+faststart+use_metadata_tags",  # Added for better compatibility
                 output_m4b_path,
             ]
 
@@ -742,6 +761,69 @@ class ConversionThread(QThread):
                 except Exception as e_clean:
                     self.log_updated.emit((f"Warning: Could not delete temporary chapter file {chapters_info_path}: {e_clean}", "orange"))
 
+    def _extract_and_add_metadata_tags_to_ffmpeg_cmd(self):
+        """Extract metadata tags from text content and add them to ffmpeg command"""
+        metadata_options = []
+        
+        # Get the input text (either direct or from file)
+        text = ""
+        if self.is_direct_text:
+            text = self.file_name
+        else:
+            try:
+                encoding = detect_encoding(self.file_name)
+                with open(self.file_name, "r", encoding=encoding, errors="replace") as file:
+                    text = file.read()
+            except Exception as e:
+                self.log_updated.emit(f"Warning: Could not read file for metadata extraction: {e}")
+                return []
+                
+        # Extract metadata tags using regex
+        title_match = re.search(r"<<METADATA_TITLE:([^>]*)>>", text)
+        artist_match = re.search(r"<<METADATA_ARTIST:([^>]*)>>", text) 
+        album_match = re.search(r"<<METADATA_ALBUM:([^>]*)>>", text)
+        year_match = re.search(r"<<METADATA_YEAR:([^>]*)>>", text)
+        album_artist_match = re.search(r"<<METADATA_ALBUM_ARTIST:([^>]*)>>", text)
+        
+        # Use display path or filename as fallback for title
+        filename = os.path.splitext(os.path.basename(self.display_path if self.display_path else self.file_name))[0]
+        
+        # Add title metadata
+        if title_match:
+            metadata_options.extend(["-metadata", f"title=Dinginliğin Gücü"])
+        else:
+            metadata_options.extend(["-metadata", f"title={filename}"])
+            
+        # Add artist metadata
+        if artist_match:
+            metadata_options.extend(["-metadata", f"artist={artist_match.group(1)}"])
+        else:
+            metadata_options.extend(["-metadata", f"artist=Unknown"])
+            
+        # Add album metadata
+        if album_match:
+            metadata_options.extend(["-metadata", f"album={album_match.group(1)}"])
+        else:
+            metadata_options.extend(["-metadata", f"album={filename}"])
+            
+        # Add year metadata
+        if year_match:
+            metadata_options.extend(["-metadata", f"date={year_match.group(1)}"])
+        else:
+            # Use current year if year is not specified
+            import datetime
+            current_year = datetime.datetime.now().year
+            metadata_options.extend(["-metadata", f"date={current_year}"])
+
+        # Add album artist metadata
+        if album_artist_match:
+            metadata_options.extend(["-metadata", f"album_artist={album_artist_match.group(1)}"])
+        else:
+            metadata_options.extend(["-metadata", f"album_artist=Unknown"])
+            
+        # Add these to ffmpeg command
+        return metadata_options
+
     def _srt_time(self, t):
         """Helper function to format time for SRT files"""
         h = int(t // 3600)
@@ -865,10 +947,31 @@ class VoicePreviewThread(QThread):
         self.speed = speed
         self.use_gpu = use_gpu
         
+        # Cache location for preview audio
+        self.cache_dir = os.path.join(tempfile.gettempdir(), PROGRAM_NAME, "preview_cache")
+        os.makedirs(self.cache_dir, exist_ok=True)
+        
+        # Calculate cache path
+        self.cache_path = self._get_cache_path()
+
+    def _get_cache_path(self):
+        """Generate a unique filename for the voice with its parameters"""
+        # For a voice formula, use a hash of the formula
+        if "*" in self.voice:
+            voice_id = f"voice_formula_{hashlib.md5(self.voice.encode()).hexdigest()[:8]}"
+        else:
+            voice_id = self.voice
+            
+        # Create a unique filename based on voice_id, language, and speed
+        filename = f"{voice_id}_{self.lang_code}_{self.speed:.2f}.wav"
+        return os.path.join(self.cache_dir, filename)
+
     def run(self):
         print(
             f"\nVoice: {self.voice}\nLanguage: {self.lang_code}\nSpeed: {self.speed}\nGPU: {self.use_gpu}\n"
         )
+        
+        # Generate the preview and save to cache
         try:
             device = "cuda" if self.use_gpu else "cpu"
             tts = self.kpipeline_class(
@@ -887,15 +990,9 @@ class VoicePreviewThread(QThread):
                 audio_segments.append(result.audio)
             if audio_segments:
                 audio = self.np_module.concatenate(audio_segments)
-                # Create temp wav file in a folder in the system temp directory
+                # Save directly to the cache path
-                temp_dir = os.path.join(tempfile.gettempdir(), PROGRAM_NAME)
+                sf.write(self.cache_path, audio, 24000)
-                os.makedirs(temp_dir, exist_ok=True)
+                self.temp_wav = self.cache_path
-                fd, temp_path = tempfile.mkstemp(
-                    prefix="abogen_", suffix=".wav", dir=temp_dir
-                )
-                os.close(fd)
-                sf.write(temp_path, audio, 24000)
-                self.temp_wav = temp_path
             self.finished.emit()
         except Exception as e:
             self.error.emit(f"Voice preview error: {str(e)}")
@@ -908,6 +1005,7 @@ class PlayAudioThread(QThread):
     def __init__(self, wav_path, parent=None):
         super().__init__(parent)
         self.wav_path = wav_path
+        self.is_canceled = False
 
     def run(self):
         try:
@@ -917,10 +1015,37 @@ class PlayAudioThread(QThread):
             pygame.mixer.init()
             pygame.mixer.music.load(self.wav_path)
             pygame.mixer.music.play()
-            # Wait until playback is finished
+            # Wait until playback is finished or canceled
-            while pygame.mixer.music.get_busy():
+            while pygame.mixer.music.get_busy() and not self.is_canceled:
-                _time.sleep(0.1)
+                _time.sleep(0.2)
+                
+            # Make sure to clean up regardless of how we exited the loop
+            try:
+                pygame.mixer.music.stop()
                 pygame.mixer.music.unload()
+                pygame.mixer.quit()  # Quit the mixer
+            except Exception:
+                # Ignore any errors during cleanup
+                pass
+                
             self.finished.emit()
         except Exception as e:
+            # Handle initialization errors separately to give better error messages
+            if "mixer not initialized" in str(e):
+                self.error.emit("Audio playback error: The audio system was not properly initialized")
+            else:
                 self.error.emit(f"Audio playback error: {str(e)}")
+            
+    def stop(self):
+        """Safely stop playback"""
+        self.is_canceled = True
+        # Try to stop pygame if it's running, but catch all exceptions
+        try:
+            import pygame
+            if pygame.mixer.get_init():
+                if pygame.mixer.music.get_busy():
+                    pygame.mixer.music.stop()
+                pygame.mixer.music.unload()
+        except Exception:
+            # Ignore all errors when stopping since mixer might not be initialized
+            pass

abogen/gui.py

@@ -4,6 +4,8 @@ import tempfile
 import platform
 import base64
 import re
+import hf_tracker
+import hashlib  # Added for cache path generation
 from PyQt5.QtWidgets import (
     QApplication,
     QWidget,
@@ -23,6 +25,7 @@ from PyQt5.QtWidgets import (
     QCheckBox,
     QMenu,
     QAction,
+    QActionGroup,
 )
 from PyQt5.QtCore import (
     Qt,
@@ -78,6 +81,13 @@ if platform.system() == "Windows":
     import ctypes
 
 
+class ShowWarningSignalEmitter(QObject):  # New class to handle signal emission
+    show_warning_signal = pyqtSignal(str, str)
+
+    def emit(self, title, message):
+        self.show_warning_signal.emit(title, message)
+
+
 class ThreadSafeLogSignal(QObject):
     log_signal = pyqtSignal(object)
 
@@ -94,9 +104,9 @@ class IconProvider(QFileIconProvider):
 
 
 class InputBox(QLabel):
-    LABEL_CSS = """QLabel { border:2px dashed #aaa; border-radius:5px; padding:20px; background:#f5f9fc; min-height:100px; } QLabel:hover { background:#e5f1fa; border-color:#6ab0de; }"""
+    LABEL_CSS = """QLabel { border:2px dashed #aaa; border-radius:5px; padding:20px; background:rgba(0, 102, 255, 0.05); min-height:100px; } QLabel:hover { background:rgba(0, 102, 255, 0.1); border-color:#6ab0de; }"""
-    LABEL_CSS_ACTIVE = """QLabel { border:2px dashed #42ad4a; border-radius:5px; padding:20px; background:#e6f7e6;min-height:100px; } QLabel:hover { background:#e5f1fa; border-color:#6ab0de; }"""
+    LABEL_CSS_ACTIVE = """QLabel { border:2px dashed #42ad4a; border-radius:5px; padding:20px; background:rgba(66, 173, 73, 0.1);min-height:100px; } QLabel:hover { background:rgba(66, 173, 73, 0.15); border-color:#42ad4a; }"""
-    LABEL_CSS_ERROR = """QLabel { border:2px dashed #e74c3c; border-radius:5px; padding:20px; background:#ffeaea; min-height:100px; color:#c0392b; } QLabel:hover { background:#ffeaea; border-color:#e74c3c; }"""
+    LABEL_CSS_ERROR = """QLabel { border:2px dashed #e74c3c; border-radius:5px; padding:20px; background:rgba(232, 78, 60, 0.10); min-height:100px; color:#c0392b; } QLabel:hover { background:rgba(232, 78, 60, 0.15); border-color:#e74c3c; }"""
 
     def __init__(self, parent=None):
         super().__init__(parent)
@@ -519,6 +529,7 @@ class abogen(QWidget):
             "max_subtitle_words", 50
         )  # Default max words per subtitle
         self.selected_format = self.config.get("selected_format", "wav")
+        self.separate_chapters_format = self.config.get("separate_chapters_format", "wav")  # Format for individual chapter files
         self.use_gpu = self.config.get(
             "use_gpu", True  # Load GPU setting with default True
         )
@@ -530,6 +541,11 @@ class abogen(QWidget):
         self.log_signal = ThreadSafeLogSignal()
         self.log_signal.log_signal.connect(self._update_log_main_thread)
 
+        # Create warning signal emitter
+        self.warning_signal_emitter = ShowWarningSignalEmitter()
+        self.warning_signal_emitter.show_warning_signal.connect(self.show_model_download_warning)
+        hf_tracker.set_show_warning_signal_emitter(self.warning_signal_emitter)
+
         # Set application icon
         icon_path = get_resource_path("abogen.assets", "icon.ico")
         if icon_path:
@@ -581,6 +597,9 @@ class abogen(QWidget):
         if self.check_updates:
             QTimer.singleShot(1000, self.check_for_updates_startup)
 
+        # Set hf_tracker callbacks
+        hf_tracker.set_log_callback(self.update_log)
+
     def initUI(self):
         self.setWindowTitle(f"{PROGRAM_NAME} v{VERSION}")
         screen = QApplication.primaryScreen().geometry()
@@ -677,6 +696,7 @@ class abogen(QWidget):
         self.btn_preview.clicked.connect(self.preview_voice)
         voice_row.addWidget(self.btn_preview)
         self.preview_playing = False
+        self.play_audio_thread = None  # Keep track of audio playing thread
         voice_layout.addLayout(voice_row)
         controls_layout.addLayout(voice_layout)
         # Subtitle mode
@@ -938,10 +958,12 @@ class abogen(QWidget):
                             title = re.sub(r'<[^>]+>', '', str(dialog.book_metadata.get('title', 'Unknown')))
                             publisher = re.sub(r'<[^>]+>', '', str(dialog.book_metadata.get('publisher', 'Unknown')))
                             authors = [re.sub(r'<[^>]+>', '', str(author)) for author in dialog.book_metadata.get('authors', ['Unknown'])]
+                            publication_year = re.sub(r'<[^>]+>', '', str(dialog.book_metadata.get('publication_year', 'Unknown')))
                             
                             f.write(f"Title: {title}\n")
                             f.write(f"Authors: {', '.join(authors)}\n")
                             f.write(f"Publisher: {publisher}\n")
+                            f.write(f"Publication Year: {publication_year}\n")
                             if dialog.book_metadata.get('description'):
                                 description = re.sub(r'<[^>]+>', '', str(dialog.book_metadata.get('description')))
                                 f.write(f"\nDescription:\n{description}\n")
@@ -956,7 +978,7 @@ class abogen(QWidget):
                 os.makedirs(temp_dir, exist_ok=True)
                 
             fd, tmp = tempfile.mkstemp(
-                prefix=f"abogen_{base_name}_", suffix=".txt", dir=temp_dir
+                prefix=f"{base_name}_", suffix=".txt", dir=temp_dir
             )
             os.close(fd)
             with open(tmp, "w", encoding="utf-8") as f:
@@ -1305,6 +1327,8 @@ class abogen(QWidget):
             self.conversion_thread.replace_single_newlines = (
                 self.replace_single_newlines
             )
+            # Pass separate_chapters_format setting
+            self.conversion_thread.separate_chapters_format = self.separate_chapters_format
             # Pass chapter count for EPUB or PDF files
             if self.selected_file_type in ["epub", "pdf"] and hasattr(
                 self, "selected_chapters"
@@ -1518,20 +1542,100 @@ class abogen(QWidget):
                 "Open File Error", f"Could not open file:\n{e}"
             )
 
+    def _get_preview_cache_path(self):
+        """Generate the expected cache path for the current voice settings."""
+        speed = self.speed_slider.value() / 100.0
+        voice_to_cache = ""
+        lang_to_cache = ""
+
+        if self.mixed_voice_state:
+            components = [f"{name}*{weight}" for name, weight in self.mixed_voice_state]
+            voice_formula = " + ".join(filter(None, components))
+            voice_to_cache = voice_formula
+            if self.selected_profile_name:
+                from voice_profiles import load_profiles
+                entry = load_profiles().get(self.selected_profile_name, {})
+                lang_to_cache = entry.get("language")
+            else:
+                lang_to_cache = self.selected_lang
+            if not lang_to_cache and self.mixed_voice_state:
+                lang_to_cache = (
+                    self.mixed_voice_state[0][0][0]
+                    if self.mixed_voice_state and self.mixed_voice_state[0][0]
+                    else None
+                )
+        elif self.selected_voice:
+            lang_to_cache = self.selected_voice[0]
+            voice_to_cache = self.selected_voice
+        else:  # No voice or profile selected
+            return None
+
+        if not lang_to_cache or not voice_to_cache:  # Not enough info
+            return None
+
+        cache_dir = os.path.join(tempfile.gettempdir(), PROGRAM_NAME, "preview_cache")
+        
+        if "*" in voice_to_cache:  # Voice formula
+            voice_id = f"voice_formula_{hashlib.md5(voice_to_cache.encode()).hexdigest()[:8]}"
+        else:  # Single voice
+            voice_id = voice_to_cache
+            
+        filename = f"{voice_id}_{lang_to_cache}_{speed:.2f}.wav"
+        return os.path.join(cache_dir, filename)
+
     def preview_voice(self):
         if self.preview_playing:
             try:
-                import pygame
+                if self.play_audio_thread and self.play_audio_thread.isRunning():
-
+                    # Call the stop method on PlayAudioThread to safely handle stopping
-                pygame.mixer.music.stop()
+                    self.play_audio_thread.stop()
-            except Exception:
+                    self.play_audio_thread.wait(500)  # Wait a bit
-                pass
+            except Exception as e:
+                print(f"Error stopping preview audio: {e}")
             self._preview_cleanup()
             return
 
         if hasattr(self, "preview_thread") and self.preview_thread.isRunning():
             return
 
+        # Check for cache first
+        cached_path = self._get_preview_cache_path()
+        if cached_path and os.path.exists(cached_path):
+            print(f"Cache hit for {cached_path}")
+            self.btn_preview.setEnabled(False)  # Disable button briefly
+            self.voice_combo.setEnabled(False)
+            self.btn_voice_formula_mixer.setEnabled(False)
+            self.btn_start.setEnabled(False)
+            
+            # Directly play from cache
+            self.preview_playing = True
+            self.btn_preview.setIcon(self.stop_icon)
+            self.btn_preview.setToolTip("Stop preview")
+            self.btn_preview.setEnabled(True)
+
+            def cleanup_cached_play():
+                self._preview_cleanup()
+
+            try:
+                # Ensure pygame mixer is initialized for the audio thread
+                import pygame
+                if not pygame.mixer.get_init():
+                    pygame.mixer.init()
+
+                self.play_audio_thread = PlayAudioThread(cached_path)
+                self.play_audio_thread.finished.connect(cleanup_cached_play)
+                self.play_audio_thread.error.connect(
+                    lambda msg: (self._show_preview_error_box(msg), cleanup_cached_play())
+                )
+                self.play_audio_thread.start()
+            except Exception as e:
+                self._show_error_message_box(
+                    "Preview Error", f"Could not play cached preview audio:\n{e}"
+                )
+                cleanup_cached_play()
+            return
+
+        # If no cache hit, proceed to load pipeline and generate
         self.btn_preview.setEnabled(False)
         self.btn_preview.setToolTip("Loading...")
         self.voice_combo.setEnabled(False)
@@ -1609,42 +1713,60 @@ class abogen(QWidget):
         self.preview_thread.error.connect(self._preview_error)
         self.preview_thread.start()
 
-    def _play_preview_audio(self):
+    def _play_preview_audio(self, from_cache=True):  # from_cache default is now False
+        # If preview_thread is the source, get temp_wav from it
+        if hasattr(self, 'preview_thread') and not from_cache:
             temp_wav = self.preview_thread.temp_wav
-        if not temp_wav:
+        elif from_cache:  # This case is now handled before calling _play_preview_audio
-            self.loading_movie.stop()
+            cached_path = self._get_preview_cache_path()
+            if cached_path and os.path.exists(cached_path):
+                temp_wav = cached_path
+            else:  # Should not happen if cache check was done
+                self._show_error_message_box("Preview Error", "Cache file expected but not found.")
+                self._preview_cleanup()
+                return
+        else:  # Should have temp_wav from preview_thread or handled by cache check
+            self._show_error_message_box("Preview Error", "Preview audio path not found.")
+            self._preview_cleanup()
+            return
 
+        if not temp_wav:
+            if hasattr(self, 'loading_movie'): self.loading_movie.stop()
             self._show_error_message_box(
                 "Preview Error", "Preview error: No audio generated."
             )
-            self.btn_preview.setIcon(self.play_icon)
+            self._preview_cleanup()
-            self.btn_preview.setToolTip("Preview selected voice")
-            self.btn_preview.setEnabled(True)
-            self.voice_combo.setEnabled(True)
-            self.btn_voice_formula_mixer.setEnabled(True)  # Re-enable mixer button
-            self.btn_start.setEnabled(True)
             return
+            
         # stop loading animation, switch to stop icon
-        self.loading_movie.stop()
+        if hasattr(self, 'loading_movie'): self.loading_movie.stop()
         self.preview_playing = True
         self.btn_preview.setIcon(self.stop_icon)
         self.btn_preview.setToolTip("Stop preview")
         self.btn_preview.setEnabled(True)
 
         def cleanup():
+            # Only remove if not from cache AND it's a temp file from VoicePreviewThread
+            if not from_cache and hasattr(self, 'preview_thread') and hasattr(self.preview_thread, 'temp_wav') and self.preview_thread.temp_wav == temp_wav:
                 try:
+                    if os.path.exists(temp_wav):  # Ensure it exists before trying to remove
                         os.remove(temp_wav)
                 except Exception:
                     pass
             self._preview_cleanup()
 
         try:
-            self.audio_thread = PlayAudioThread(temp_wav)
+            # Ensure pygame mixer is initialized for the audio thread
-            self.audio_thread.finished.connect(cleanup)
+            import pygame
-            self.audio_thread.error.connect(
+            if not pygame.mixer.get_init():
+                pygame.mixer.init()
+
+            self.play_audio_thread = PlayAudioThread(temp_wav)
+            self.play_audio_thread.finished.connect(cleanup)
+            self.play_audio_thread.error.connect(
                 lambda msg: (self._show_preview_error_box(msg), cleanup())
             )
-            self.audio_thread.start()
+            self.play_audio_thread.start()
         except Exception as e:
             self._show_error_message_box(
                 "Preview Error", f"Could not play preview audio:\n{e}"
@@ -1667,9 +1789,9 @@ class abogen(QWidget):
 
     def _preview_cleanup(self):
         self.preview_playing = False
-        self.loading_movie.stop()
+        if hasattr(self, 'loading_movie'): self.loading_movie.stop()
         try:
-            self.loading_movie.frameChanged.disconnect()
+            if hasattr(self, 'loading_movie'): self.loading_movie.frameChanged.disconnect()
         except Exception:
             pass  # Ignore error if not connected
         self.btn_preview.setIcon(self.play_icon)
@@ -1803,7 +1925,24 @@ class abogen(QWidget):
         max_words_action.triggered.connect(self.set_max_subtitle_words)
         menu.addAction(max_words_action)
         
-        # Add seperator
+        # Add separate chapters format option
+        separate_chapters_format_menu = QMenu("Separate chapters audio format", self)
+        separate_chapters_format_menu.setToolTip("Choose the format for individual chapter files")
+        
+        format_group = QActionGroup(self)
+        format_group.setExclusive(True)
+        
+        for format_option in ["wav", "mp3", "flac", "opus"]:
+            format_action = QAction(format_option, self)
+            format_action.setCheckable(True)
+            format_action.setChecked(self.separate_chapters_format == format_option)
+            format_action.triggered.connect(lambda checked, fmt=format_option: self.set_separate_chapters_format(fmt))
+            format_group.addAction(format_action)
+            separate_chapters_format_menu.addAction(format_action)
+            
+        menu.addMenu(separate_chapters_format_menu)
+
+        # Add separator
         menu.addSeparator()
 
         # Add shortcut to desktop (Windows or Linux)
@@ -1825,7 +1964,7 @@ class abogen(QWidget):
         menu.addAction(open_temp_action)
 
         # Add clear temporary files option
-        clear_temp_action = QAction("Clear all temporary files", self)
+        clear_temp_action = QAction("Clear temporary files", self)
         clear_temp_action.triggered.connect(self.clear_temp_files)
         menu.addAction(clear_temp_action)
 
@@ -2203,7 +2342,7 @@ Categories=AudioVideo;Audio;Utility;
             pass
 
     def clear_temp_files(self):
-        """Clear all temporary files created by the program."""
+        """Clear temporary files created by the program."""
         import glob
 
         try:
@@ -2217,25 +2356,49 @@ Categories=AudioVideo;Audio;Utility;
             # Count the files
             file_count = len(temp_files)
 
-            if file_count == 0:
+            # Check for preview cache files
+            preview_cache_dir = os.path.join(temp_dir, "preview_cache")
+            preview_files = []
+            if os.path.exists(preview_cache_dir):
+                preview_pattern = os.path.join(preview_cache_dir, "*.wav")
+                preview_files = glob.glob(preview_pattern)
+            
+            preview_count = len(preview_files)
+            
+            if file_count == 0 and preview_count == 0:
                 QMessageBox.information(
                     self, "No Temporary Files", "No temporary files were found."
                 )
                 return
 
-            # Confirm with the user
+            # Create a custom message box with checkbox
-            confirm = QMessageBox.question(
+            msg_box = QMessageBox(self)
-                self,
+            msg_box.setIcon(QMessageBox.Question)
-                "Clear Temporary Files",
+            msg_box.setWindowTitle("Clear Temporary Files")
-                f"Found {file_count} temporary file{'s' if file_count != 1 else ''} in the {PROGRAM_NAME} temp folder.\n"
-                "Do you want to delete them?",
-                QMessageBox.Yes | QMessageBox.No,
-            )
             
-            if confirm != QMessageBox.Yes:
+            msg_text = f"Found {file_count} temporary file{'s' if file_count != 1 else ''} in the {PROGRAM_NAME} temp folder."
+            if preview_count > 0:
+                msg_text += f"\nAlso found {preview_count} preview cache file{'s' if preview_count != 1 else ''}."
+                
+            msg_box.setText(msg_text + "\nDo you want to delete them?")
+            
+            # Add checkbox for preview cache
+            preview_cache_checkbox = QCheckBox("Clean preview cache", msg_box)
+            preview_cache_checkbox.setChecked(False)
+            # Only enable checkbox if preview files exist
+            preview_cache_checkbox.setEnabled(preview_count > 0)
+            
+            # Add the checkbox to the layout
+            msg_box.setCheckBox(preview_cache_checkbox)
+            
+            # Add buttons
+            msg_box.setStandardButtons(QMessageBox.Yes | QMessageBox.No)
+            msg_box.setDefaultButton(QMessageBox.Yes)
+            
+            if msg_box.exec_() != QMessageBox.Yes:
                 return
 
-            # Delete the files
+            # Delete the text files
             deleted_count = 0
             for file_path in temp_files:
                 try:
@@ -2244,11 +2407,24 @@ Categories=AudioVideo;Audio;Utility;
                 except Exception as e:
                     print(f"Error deleting {file_path}: {e}")
             
+            # Delete preview cache files if checkbox is checked
+            deleted_preview_count = 0
+            if preview_cache_checkbox.isChecked() and preview_count > 0:
+                for file_path in preview_files:
+                    try:
+                        os.remove(file_path)
+                        deleted_preview_count += 1
+                    except Exception as e:
+                        print(f"Error deleting preview cache {file_path}: {e}")
+
+            # Build result message
+            result_msg = f"Successfully deleted {deleted_count} temporary file{'s' if deleted_count != 1 else ''}."
+            if preview_cache_checkbox.isChecked() and deleted_preview_count > 0:
+                result_msg += f"\nAlso deleted {deleted_preview_count} preview cache file{'s' if deleted_preview_count != 1 else ''}."
+                
             # Show results
             QMessageBox.information(
-                self,
+                self, "Temporary Files Cleared", result_msg
-                "Temporary Files Cleared",
-                f"Successfully deleted {deleted_count} temporary file{'s' if deleted_count != 1 else ''}.",
             )
 
             # If currently selected file is in the temp directory, clear the UI
@@ -2293,3 +2469,12 @@ Categories=AudioVideo;Audio;Utility;
                 f"Maximum words per subtitle set to {value}.",
             )
 
+    def set_separate_chapters_format(self, fmt):
+        """Set the format for separate chapters audio files."""
+        self.separate_chapters_format = fmt
+        self.config["separate_chapters_format"] = fmt
+        save_config(self.config)
+
+    def show_model_download_warning(self, title, message):
+        QMessageBox.information(self, title, message)
+

abogen/hf_tracker.py

@@ -0,0 +1,36 @@
+log_callback = None
+show_warning_signal_emitter = None  # Renamed for clarity
+
+def set_log_callback(cb):
+    global log_callback
+    log_callback = cb
+
+def set_show_warning_signal_emitter(emitter):  # Renamed for clarity
+    global show_warning_signal_emitter
+    show_warning_signal_emitter = emitter
+
+from huggingface_hub import hf_hub_download
+
+def tracked_hf_hub_download(*args, **kwargs):
+    try:
+        local_kwargs = dict(kwargs)
+        local_kwargs["local_files_only"] = True
+        hf_hub_download(*args, **local_kwargs)
+    except Exception:
+        repo_id = kwargs.get("repo_id", "<unknown repo>")
+        filename = kwargs.get("filename", "<unknown file>")
+        if filename.endswith(".pth"):
+            msg = f"\nDownloading model '{filename}' from Hugging Face ({repo_id}). This may take a while. Please wait..."
+            if show_warning_signal_emitter:  # Check if the emitter is set
+                show_warning_signal_emitter.emit("Downloading Model", f"Downloading model '{filename}' from Hugging Face repository '{repo_id}'. This may take a while, please wait.")
+        else:
+            msg = f"\nDownloading '{filename}' from Hugging Face ({repo_id}). Please wait..."
+        if log_callback:
+            print(msg, flush=True)
+            log_callback(msg)
+        else:
+            print(msg, flush=True)
+    return hf_hub_download(*args, **kwargs)
+
+import huggingface_hub
+huggingface_hub.hf_hub_download = tracked_hf_hub_download

abogen/is_nvidia.py

@@ -0,0 +1,16 @@
+import gpustat
+
+def check():
+    try:
+        stats = gpustat.new_query()
+    except Exception:
+        return False
+
+    for gpu in stats.gpus:
+        print(gpu.name)
+        if 'nvidia' in gpu.name.lower():
+            return True
+    return False
+
+if __name__ == "__main__":
+    print(check())

abogen/main.py

@@ -10,6 +10,10 @@ from gui import abogen
 from utils import get_resource_path
 from constants import PROGRAM_NAME, VERSION
 
+# Set environment variables for AMD ROCm
+os.environ["MIOPEN_FIND_MODE"] = "FAST"
+os.environ["MIOPEN_CONV_PRECISE_ROCM_TUNING"] = "0"
+
 # Ensure sys.stdout and sys.stderr are valid in GUI mode
 if sys.stdout is None:
     sys.stdout = open(os.devnull, "w")
@@ -48,6 +52,13 @@ def main():
     if icon_path:
         app.setWindowIcon(QIcon(icon_path))
 
+    # Set the .desktop name on Linux
+    if platform.system() == "Linux":
+        try:
+            app.setDesktopFileName("abogen.desktop")
+        except AttributeError:
+            pass
+
     ex = abogen()
     ex.show()
     sys.exit(app.exec_())

abogen/utils.py

@@ -105,7 +105,7 @@ def clean_text(text, *args, **kwargs):
 
 default_encoding = sys.getfilesystemencoding()
 
-def create_process(cmd, stdin=None, text=True):
+def create_process(cmd, stdin=None, text=True, capture_output=False):
     import logging
     logger = logging.getLogger(__name__)
     
@@ -154,8 +154,8 @@ def create_process(cmd, stdin=None, text=True):
     
     proc = subprocess.Popen(cmd, **kwargs)
     
-    # Stream output to console in real-time
+    # Stream output to console in real-time if not capturing
-    if proc.stdout:
+    if proc.stdout and not capture_output:
         def _stream_output(stream):
             if text:
                 # For text mode, read character by character for real-time output
@@ -204,11 +204,15 @@ def save_config(config):
 
 def calculate_text_length(text):
     # Ignore chapter markers
-    text_no_markers = re.sub(r"<<CHAPTER_MARKER:.*?>>", "", text)
+    text = re.sub(r"<<CHAPTER_MARKER:.*?>>", "", text)
+    # Ignore metadata patterns
+    text = re.sub(r"<<METADATA_[^:]+:[^>]*>>", "", text)
     # Ignore newlines
-    cleaned_text = text_no_markers.replace("\n", "")
+    text = text.replace("\n", "")
+    # Ignore leading/trailing spaces
+    text = text.strip()
     # Calculate character count
-    char_count = len(cleaned_text)
+    char_count = len(text)
     return char_count
 
 
@@ -227,29 +231,26 @@ def prevent_sleep_start():
     system = platform.system()
     if system == "Windows":
         import ctypes
-
         ctypes.windll.kernel32.SetThreadExecutionState(
             0x80000000 | 0x00000001 | 0x00000040
-        )  # ES_CONTINUOUS | ES_SYSTEM_REQUIRED | ES_AWAYMODE_REQUIRED
-    elif system == "Darwin":
-        _sleep_procs["Darwin"] = create_process("caffeinate")
-    elif system == "Linux":
-        try:
-            _sleep_procs["Linux"] = create_process(
-                "systemd-inhibit --what=sleep --why=TextToAudiobook conversion sleep 999999"
         )
-        except Exception:
+    elif system == "Darwin":
-            try:
+        _sleep_procs["Darwin"] = create_process(["caffeinate"])
-                create_process("xdg-screensaver reset")
+    elif system == "Linux":
-            except Exception:
+        # use a sleep that never exits so inhibition stays active
-                pass
+        _sleep_procs["Linux"] = create_process([
+            "systemd-inhibit",
+            "--what=handle-lid-switch:sleep",
+            "--mode=block",
+            "sleep",
+            "infinity",
+        ])
 
 
 def prevent_sleep_end():
     system = platform.system()
     if system == "Windows":
         import ctypes
-
         ctypes.windll.kernel32.SetThreadExecutionState(0x80000000)  # ES_CONTINUOUS
     elif system in ("Darwin", "Linux") and _sleep_procs[system]:
         try:

abogen/voice_formula_gui.py

@@ -264,14 +264,12 @@ class HoverLabel(QLabel):
         self.voice_name = voice_name
         self.setMouseTracking(True)
         self.setStyleSheet(
-            "background-color: #e0e0e0; border-radius: 4px; padding: 3px 6px 3px 6px; margin: 2px;"
+            "background-color: rgba(140, 140, 140, 0.15); border-radius: 4px; padding: 3px 6px 3px 6px; margin: 2px;"
         )
 
         # Create delete button
         self.delete_button = QPushButton("×", self)
         self.delete_button.setFixedSize(16, 16)
-        self.delete_button = QPushButton("×", self)
-        self.delete_button.setFixedSize(16, 16)
         self.delete_button.setStyleSheet(
             """
             QPushButton {
@@ -310,6 +308,12 @@ class HoverLabel(QLabel):
 class VoiceFormulaDialog(QDialog):
     def __init__(self, parent=None, initial_state=None, selected_profile=None):
         super().__init__(parent)
+        # Store original profile/mix state for restoration on cancel
+        self._original_profile_name = None
+        self._original_mixed_voice_state = None
+        if parent is not None:
+            self._original_profile_name = getattr(parent, "selected_profile_name", None)
+            self._original_mixed_voice_state = getattr(parent, "mixed_voice_state", None)
         profiles = load_profiles()
         self._virtual_new_profile = False
         if not profiles:
@@ -734,7 +738,7 @@ class VoiceFormulaDialog(QDialog):
                 percentage = weight / total * 100
                 # Make the voice name bold and include percentage
                 voice_label = HoverLabel(
-                    f'<b><span style="color:#1976d2">{name}: {percentage:.1f}%</span></b>',
+                    f'<b><span style="color:#007dff">{name}: {percentage:.1f}%</span></b>',
                     name,
                 )
                 voice_label.setSizePolicy(QSizePolicy.Preferred, QSizePolicy.Preferred)
@@ -916,6 +920,13 @@ class VoiceFormulaDialog(QDialog):
         super().accept()
 
     def reject(self):
+        # Restore parent's profile/mix state on cancel
+        parent = self.parent()
+        if parent is not None:
+            if hasattr(self, "_original_profile_name"):
+                parent.selected_profile_name = self._original_profile_name
+            if hasattr(self, "_original_mixed_voice_state"):
+                parent.mixed_voice_state = self._original_mixed_voice_state
         # Prompt to save if unsaved changes, then check for zero-weight error after save
         if self._has_unsaved_changes():
             if not self._prompt_save_changes():
@@ -925,6 +936,13 @@ class VoiceFormulaDialog(QDialog):
         super().reject()
 
     def closeEvent(self, event):
+        # Restore parent's profile/mix state on close
+        parent = self.parent()
+        if parent is not None:
+            if hasattr(self, "_original_profile_name"):
+                parent.selected_profile_name = self._original_profile_name
+            if hasattr(self, "_original_mixed_voice_state"):
+                parent.mixed_voice_state = self._original_mixed_voice_state
         # Prompt to save if unsaved changes, then check for zero-weight error after save
         if self._has_unsaved_changes():
             if not self._prompt_save_changes():
@@ -1326,10 +1344,16 @@ class VoiceFormulaDialog(QDialog):
             item = self.profile_list.item(i)
             name = item.text().lstrip("*")
             if self._virtual_new_profile and name == "New profile":
-                item.setBackground(QColor("#fff59d"))  # yellow
+                color = QColor("#ffff00")  # yellow
+                color.setAlpha(64)  # Set transparency
+                item.setBackground(color)
             elif item.text().startswith("*"):
-                item.setBackground(QColor("#fff59d"))  # yellow
+                color = QColor("#ffff00")  # yellow
+                color.setAlpha(64)  # Set transparency
+                item.setBackground(color)
             else:
+                item.setBackground(self.profile_list.palette().base().color()) # Use default list item background
+
                 weights = profiles.get(name, {}).get("voices", [])
                 # Defensive: only sum if weights is a list of (voice, weight) pairs
                 total = 0
@@ -1342,9 +1366,9 @@ class VoiceFormulaDialog(QDialog):
                         ):
                             total += entry[1]
                 if total == 0:
-                    item.setBackground(QColor("#ffcdd2"))  # light red
+                    color = QColor("#ff0000")  # red
-                else:
+                    color.setAlpha(64)  # Set transparency
-                    item.setBackground(QColor("white"))
+                    item.setBackground(color)
         self.update_profile_save_buttons()
 
     def preview_current_mix(self):