1 Commits
Author SHA1 Message Date
Deniz Şafak 7a8df9b34e v1.1.1 2025-07-11 11:51:41 +03:00
14 changed files with 881 additions and 406 deletions
+4
View File
@@ -1,3 +1,7 @@
# v1.1.1
- Fixed adding wrong file in queue for EPUB and PDF files, ensuring the correct file is added to the queue.
- Reformatted the code using Black.
# v1.1.0
- Added queue system for processing multiple items, allowing users to add multiple files and process them in a queue, mentioned by @jborza in #30 (Special thanks to @jborza for implementing this feature in PR #35)
- Added a feature that allows selecting multiple items in book handler (in right click menu) by @jborza in #31, that fixes #28
+1 -1
View File
@@ -1 +1 @@
1.1.0
1.1.1
+64 -20
View File
@@ -997,10 +997,21 @@ class HandlerDialog(QDialog):
bookmark_item.setData(0, Qt.UserRole, page_id)
# only allow checking if this chapter has content
if self.content_lengths.get(page_id, 0) > 0:
bookmark_item.setFlags(bookmark_item.flags() | Qt.ItemIsUserCheckable)
bookmark_item.setCheckState(0, Qt.Checked if page_id in self.checked_chapters else Qt.Unchecked)
bookmark_item.setFlags(
bookmark_item.flags() | Qt.ItemIsUserCheckable
)
bookmark_item.setCheckState(
0,
(
Qt.Checked
if page_id in self.checked_chapters
else Qt.Unchecked
),
)
else:
bookmark_item.setFlags(bookmark_item.flags() & ~Qt.ItemIsUserCheckable)
bookmark_item.setFlags(
bookmark_item.flags() & ~Qt.ItemIsUserCheckable
)
# map for uncategorized pages
self.bookmark_items_map[page_num] = bookmark_item
@@ -1027,10 +1038,21 @@ class HandlerDialog(QDialog):
page_item.setData(0, Qt.UserRole, page_id)
# only allow checking if this sub-page has content
if self.content_lengths.get(page_id, 0) > 0:
page_item.setFlags(page_item.flags() | Qt.ItemIsUserCheckable)
page_item.setCheckState(0, Qt.Checked if page_id in self.checked_chapters else Qt.Unchecked)
page_item.setFlags(
page_item.flags() | Qt.ItemIsUserCheckable
)
page_item.setCheckState(
0,
(
Qt.Checked
if page_id in self.checked_chapters
else Qt.Unchecked
),
)
else:
page_item.setFlags(page_item.flags() & ~Qt.ItemIsUserCheckable)
page_item.setFlags(
page_item.flags() & ~Qt.ItemIsUserCheckable
)
added_pages.add(sub_page_num)
@@ -1038,11 +1060,15 @@ class HandlerDialog(QDialog):
covered_pages = set(added_pages)
# attach any pages without direct bookmarks under nearest preceding chapter
uncategorized_pages = [i for i in range(len(self.pdf_doc)) if i not in covered_pages]
uncategorized_pages = [
i for i in range(len(self.pdf_doc)) if i not in covered_pages
]
for page_num in uncategorized_pages:
# find nearest previous bookmark
prev_nums = [n for n in sorted(self.bookmark_items_map) if n < page_num]
parent_item = self.bookmark_items_map[prev_nums[-1]] if prev_nums else self.treeWidget
parent_item = (
self.bookmark_items_map[prev_nums[-1]] if prev_nums else self.treeWidget
)
page_id = f"page_{page_num+1}"
title = f"Page {page_num+1}"
text = self.content_texts.get(page_id, "").strip()
@@ -1055,7 +1081,9 @@ class HandlerDialog(QDialog):
# only allow checking if uncategorized page has content
if self.content_lengths.get(page_id, 0) > 0:
page_item.setFlags(page_item.flags() | Qt.ItemIsUserCheckable)
page_item.setCheckState(0, Qt.Checked if page_id in self.checked_chapters else Qt.Unchecked)
page_item.setCheckState(
0, Qt.Checked if page_id in self.checked_chapters else Qt.Unchecked
)
else:
page_item.setFlags(page_item.flags() & ~Qt.ItemIsUserCheckable)
@@ -1081,7 +1109,9 @@ class HandlerDialog(QDialog):
# only allow checking if standalone page has content
if self.content_lengths.get(page_id, 0) > 0:
page_item.setFlags(page_item.flags() | Qt.ItemIsUserCheckable)
page_item.setCheckState(0, Qt.Checked if page_id in self.checked_chapters else Qt.Unchecked)
page_item.setCheckState(
0, Qt.Checked if page_id in self.checked_chapters else Qt.Unchecked
)
else:
page_item.setFlags(page_item.flags() & ~Qt.ItemIsUserCheckable)
@@ -1501,7 +1531,9 @@ 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"):
if self.book_metadata["publisher"] or self.book_metadata.get(
"publication_year"
):
pub_info = []
if self.book_metadata["publisher"]:
pub_info.append(f"Published by {self.book_metadata['publisher']}")
@@ -1543,7 +1575,9 @@ class HandlerDialog(QDialog):
try:
author_items = self.book.get_metadata("DC", "creator")
if author_items:
metadata["authors"] = [author[0] for author in author_items if len(author) > 0]
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}")
@@ -1567,7 +1601,7 @@ class HandlerDialog(QDialog):
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)
year_match = re.search(r"\b(19|20)\d{2}\b", date_str)
if year_match:
metadata["publication_year"] = year_match.group(0)
else:
@@ -1607,12 +1641,12 @@ class HandlerDialog(QDialog):
# 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)
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)
year_match = re.search(r"D:(\d{4})", date_str)
if year_match:
metadata["publication_year"] = year_match.group(1)
@@ -1644,11 +1678,15 @@ class HandlerDialog(QDialog):
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
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'}"
chapter_text = (
f"{total_chapters} {'Chapters' if self.file_type == 'epub' else 'Pages'}"
)
# Format metadata tags
metadata_tags = [
@@ -1658,7 +1696,7 @@ class HandlerDialog(QDialog):
f"<<METADATA_YEAR:{year}>>",
f"<<METADATA_ALBUM_ARTIST:{album_artist}>>",
f"<<METADATA_COMPOSER:Narrator>>",
f"<<METADATA_GENRE:Audiobook>>"
f"<<METADATA_GENRE:Audiobook>>",
]
return "\n".join(metadata_tags)
@@ -1730,7 +1768,10 @@ class HandlerDialog(QDialog):
if text:
all_content.append(text)
included_text_ids.add(page_id)
return metadata_tags + "\n\n" + "\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():
@@ -1788,7 +1829,10 @@ class HandlerDialog(QDialog):
included_text_ids.add(identifier)
iterator += 1
return metadata_tags + "\n\n" + "\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)
+219 -79
View File
@@ -10,13 +10,20 @@ from PyQt5.QtCore import QThread, pyqtSignal, Qt, QTimer
from PyQt5.QtWidgets import QCheckBox, QVBoxLayout, QDialog, QLabel, QDialogButtonBox
import soundfile as sf
from abogen.utils import clean_text, create_process
from abogen.constants import PROGRAM_NAME, LANGUAGE_DESCRIPTIONS, SAMPLE_VOICE_TEXTS, COLORS, CHAPTER_OPTIONS_COUNTDOWN
from abogen.constants import (
PROGRAM_NAME,
LANGUAGE_DESCRIPTIONS,
SAMPLE_VOICE_TEXTS,
COLORS,
CHAPTER_OPTIONS_COUNTDOWN,
)
from abogen.voice_formulas import get_new_voice
import abogen.hf_tracker as hf_tracker
import static_ffmpeg
import threading # for efficient waiting
import subprocess
def get_sample_voice_text(lang_code):
return SAMPLE_VOICE_TEXTS.get(lang_code, SAMPLE_VOICE_TEXTS["a"])
@@ -73,7 +80,9 @@ class ChapterOptionsDialog(QDialog):
# Countdown label
self.countdown_seconds = CHAPTER_OPTIONS_COUNTDOWN
self.countdown_label = QLabel(f"Auto-accepting in {self.countdown_seconds} seconds...")
self.countdown_label = QLabel(
f"Auto-accepting in {self.countdown_seconds} seconds..."
)
self.countdown_label.setStyleSheet(f"color: {COLORS['GREEN']};")
layout.addWidget(self.countdown_label)
@@ -96,7 +105,9 @@ class ChapterOptionsDialog(QDialog):
def _on_timer_tick(self):
self.countdown_seconds -= 1
if self.countdown_seconds > 0:
self.countdown_label.setText(f"Auto-accepting in {self.countdown_seconds} seconds...")
self.countdown_label.setText(
f"Auto-accepting in {self.countdown_seconds} seconds..."
)
else:
self._timer.stop()
self._button_box.accepted.emit() # Simulate OK click
@@ -152,7 +163,7 @@ class ConversionThread(QThread):
start_time,
total_char_count,
use_gpu=True,
from_queue=False
from_queue=False,
): # Add use_gpu parameter
super().__init__()
self._chapter_options_event = threading.Event()
@@ -182,7 +193,9 @@ class ConversionThread(QThread):
self.use_gpu = use_gpu # Store the GPU setting
self.max_subtitle_words = 50 # Default value, will be overridden from GUI
def _stream_audio_in_chunks(self, segments, process_func, progress_prefix="Processing"):
def _stream_audio_in_chunks(
self, segments, process_func, progress_prefix="Processing"
):
"""
Process audio segments in memory-efficient chunks
@@ -204,16 +217,18 @@ class ConversionThread(QThread):
for i, segment in enumerate(segments):
try:
# Handle both NumPy arrays and PyTorch tensors
if hasattr(segment, 'astype'):
if hasattr(segment, "astype"):
segment_bytes = segment.astype("float32").tobytes()
else:
segment_bytes = segment.cpu().numpy().astype("float32").tobytes()
is_last = (i == len(segments) - 1)
is_last = i == len(segments) - 1
# Update progress periodically - skip if there's only one segment
if (i % 20 == 0 or is_last) and len(segments) > 1:
progress_percent = int((samples_processed / total_samples) * 100)
self.log_updated.emit(f"{progress_prefix} segment {i+1}/{len(segments)} ({progress_percent}% complete)")
self.log_updated.emit(
f"{progress_prefix} segment {i+1}/{len(segments)} ({progress_percent}% complete)"
)
# Process this segment
process_func(segment_bytes, is_last)
@@ -229,7 +244,14 @@ class ConversionThread(QThread):
return samples_processed
def _process_audio_segments(self, audio_segments, output_path, output_format, use_ffmpeg=False, ffmpeg_args=None):
def _process_audio_segments(
self,
audio_segments,
output_path,
output_format,
use_ffmpeg=False,
ffmpeg_args=None,
):
"""
Process audio segments to a target format with memory-efficient handling
@@ -254,7 +276,7 @@ class ConversionThread(QThread):
# Append remaining segments
for i, segment in enumerate(audio_segments[1:], 1):
with sf.SoundFile(output_path, mode='r+') as f:
with sf.SoundFile(output_path, mode="r+") as f:
f.seek(0, sf.SEEK_END)
f.write(segment)
return True, output_path
@@ -268,12 +290,18 @@ class ConversionThread(QThread):
# Basic FFmpeg command
cmd = [
"ffmpeg", "-y",
"-thread_queue_size", "32768",
"-f", "f32le",
"-ar", "24000",
"-ac", "1",
"-i", "pipe:0"
"ffmpeg",
"-y",
"-thread_queue_size",
"32768",
"-f",
"f32le",
"-ar",
"24000",
"-ac",
"1",
"-i",
"pipe:0",
]
# Add custom FFmpeg arguments if provided
@@ -302,19 +330,29 @@ class ConversionThread(QThread):
def process_chunk(chunk_bytes, is_last):
proc.stdin.write(chunk_bytes)
self._stream_audio_in_chunks(audio_segments, process_chunk,
progress_prefix=f"Processing {output_format.upper()}")
self._stream_audio_in_chunks(
audio_segments,
process_chunk,
progress_prefix=f"Processing {output_format.upper()}",
)
# Close stdin and wait for process to complete
proc.stdin.close()
if proc.wait() != 0:
self.log_updated.emit((f"{output_format.upper()} conversion failed.", "red"))
self.log_updated.emit(
(f"{output_format.upper()} conversion failed.", "red")
)
return False, None
return True, output_path
except Exception as e:
self.log_updated.emit((f"Error during {output_format.upper()} conversion: {str(e)}", "red"))
self.log_updated.emit(
(
f"Error during {output_format.upper()} conversion: {str(e)}",
"red",
)
)
proc.stdin.close()
try:
proc.terminate()
@@ -325,12 +363,16 @@ class ConversionThread(QThread):
# For formats supported by soundfile (mp3, flac)
else:
try:
with sf.SoundFile(output_path, 'w', samplerate=24000, channels=1, format=output_format) as f:
with sf.SoundFile(
output_path, "w", samplerate=24000, channels=1, format=output_format
) as f:
for i, segment in enumerate(audio_segments):
f.write(segment)
return True, output_path
except Exception as e:
self.log_updated.emit((f"Error processing {output_format.upper()} file: {str(e)}", "red"))
self.log_updated.emit(
(f"Error processing {output_format.upper()} file: {str(e)}", "red")
)
return False, None
def run(self):
@@ -346,7 +388,9 @@ class ConversionThread(QThread):
if getattr(self, "from_queue", False):
display_file = self.file_name
else:
display_file = self.display_path if self.display_path else self.file_name
display_file = (
self.display_path if self.display_path else self.file_name
)
self.log_updated.emit(f"- Input File: {display_file}")
@@ -363,7 +407,9 @@ class ConversionThread(QThread):
self.log_updated.emit(f"- Speed: {self.speed}")
self.log_updated.emit(f"- Subtitle mode: {self.subtitle_mode}")
self.log_updated.emit(f"- Output format: {self.output_format}")
self.log_updated.emit(f"- Subtitle format: {getattr(self, 'subtitle_format', 'srt')}")
self.log_updated.emit(
f"- Subtitle format: {getattr(self, 'subtitle_format', 'srt')}"
)
self.log_updated.emit(f"- Save option: {self.save_option}")
if self.replace_single_newlines:
self.log_updated.emit(f"- Replace single newlines: Yes")
@@ -382,8 +428,10 @@ class ConversionThread(QThread):
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}")
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(
@@ -478,7 +526,6 @@ class ConversionThread(QThread):
chapters_out_dir = None
suffix = ""
# Use file_name for logs if from_queue, otherwise use display_path if available
if getattr(self, "from_queue", False):
base_path = self.file_name
@@ -701,11 +748,11 @@ class ConversionThread(QThread):
# Find last word boundary before limit
pos = sanitized[:MAX_LEN].rfind("_")
# Use word boundary if found, otherwise use hard limit
sanitized = sanitized[:pos if pos > 0 else MAX_LEN].rstrip("_")
sanitized = sanitized[: pos if pos > 0 else MAX_LEN].rstrip("_")
chapter_filename = f"{chapter_idx:02d}_{sanitized}"
# Use separate_chapters_format
separate_format = getattr(self, 'separate_chapters_format', 'wav')
separate_format = getattr(self, "separate_chapters_format", "wav")
chapter_out_path = os.path.join(
chapters_out_dir, f"{chapter_filename}.{separate_format}"
@@ -713,30 +760,49 @@ class ConversionThread(QThread):
# Process audio segments using the unified function
success, chapter_out_path = self._process_audio_segments(
chapter_audio_segments, chapter_out_path, separate_format, use_ffmpeg=(separate_format in ["opus", "m4b"])
chapter_audio_segments,
chapter_out_path,
separate_format,
use_ffmpeg=(separate_format in ["opus", "m4b"]),
)
if not success:
self.log_updated.emit((f"Failed to write {separate_format.upper()} file.", "red"))
self.log_updated.emit(
(f"Failed to write {separate_format.upper()} file.", "red")
)
continue
# Generate subtitle file for chapter if not Disabled
if self.subtitle_mode != "Disabled" and chapter_subtitle_entries:
subtitle_format = getattr(self, 'subtitle_format', 'srt')
file_extension = 'ass' if 'ass' in subtitle_format else 'srt'
subtitle_format = getattr(self, "subtitle_format", "srt")
file_extension = "ass" if "ass" in subtitle_format else "srt"
chapter_subtitle_path = os.path.join(
chapters_out_dir, f"{chapter_filename}.{file_extension}"
)
if 'ass' in subtitle_format:
if "ass" in subtitle_format:
# Generate ASS subtitle
is_centered = subtitle_format in ("ass_centered_wide", "ass_centered_narrow")
is_narrow = subtitle_format in ("ass_narrow", "ass_centered_narrow")
self._write_ass_subtitle(chapter_subtitle_path, chapter_subtitle_entries, is_centered, is_narrow)
is_centered = subtitle_format in (
"ass_centered_wide",
"ass_centered_narrow",
)
is_narrow = subtitle_format in (
"ass_narrow",
"ass_centered_narrow",
)
self._write_ass_subtitle(
chapter_subtitle_path,
chapter_subtitle_entries,
is_centered,
is_narrow,
)
else:
# Generate SRT subtitle (default)
with open(
chapter_subtitle_path, "w", encoding="utf-8", errors="replace"
chapter_subtitle_path,
"w",
encoding="utf-8",
errors="replace",
) as srt_file:
for i, (start, end, text) in enumerate(
chapter_subtitle_entries, 1
@@ -788,9 +854,10 @@ class ConversionThread(QThread):
else:
# Process audio segments using the unified function
success, final_out_path = self._process_audio_segments(
audio_segments, f"{base_filepath_no_ext}.{intended_output_format}",
audio_segments,
f"{base_filepath_no_ext}.{intended_output_format}",
intended_output_format,
use_ffmpeg=(intended_output_format in ["opus", "m4b"])
use_ffmpeg=(intended_output_format in ["opus", "m4b"]),
)
if not success:
@@ -798,25 +865,41 @@ class ConversionThread(QThread):
if not final_out_path:
self.log_updated.emit(("Audio generation failed.", "red"))
self.conversion_finished.emit(("Audio generation failed.", "red"), None)
self.conversion_finished.emit(
("Audio generation failed.", "red"), None
)
return
# Subtitle and final message logic
if final_out_path:
if self.subtitle_mode != "Disabled":
subtitle_format = getattr(self, 'subtitle_format', 'srt')
file_extension = 'ass' if 'ass' in subtitle_format else 'srt'
subtitle_path = os.path.splitext(final_out_path)[0] + f".{file_extension}"
subtitle_format = getattr(self, "subtitle_format", "srt")
file_extension = "ass" if "ass" in subtitle_format else "srt"
subtitle_path = (
os.path.splitext(final_out_path)[0] + f".{file_extension}"
)
if 'ass' in subtitle_format:
if "ass" in subtitle_format:
# Generate ASS subtitle
is_centered = subtitle_format in ("ass_centered_wide", "ass_centered_narrow")
is_narrow = subtitle_format in ("ass_narrow", "ass_centered_narrow")
self._write_ass_subtitle(subtitle_path, subtitle_entries, is_centered, is_narrow)
is_centered = subtitle_format in (
"ass_centered_wide",
"ass_centered_narrow",
)
is_narrow = subtitle_format in (
"ass_narrow",
"ass_centered_narrow",
)
self._write_ass_subtitle(
subtitle_path, subtitle_entries, is_centered, is_narrow
)
else:
# Generate SRT subtitle (default)
with open(subtitle_path, "w", encoding="utf-8", errors="replace") as srt_file:
for i, (start, end, text) in enumerate(subtitle_entries, 1):
with open(
subtitle_path, "w", encoding="utf-8", errors="replace"
) as srt_file:
for i, (start, end, text) in enumerate(
subtitle_entries, 1
):
srt_file.write(
f"{i}\n{self._srt_time(start)} --> {self._srt_time(end)}\n{text}\n\n"
)
@@ -829,11 +912,16 @@ class ConversionThread(QThread):
)
else:
self.conversion_finished.emit(
(f"\nAudiobook saved to: {final_out_path}", "green"), final_out_path
(f"\nAudiobook saved to: {final_out_path}", "green"),
final_out_path,
)
else:
self.log_updated.emit(("Audio generation failed (final_out_path was not set).", "red"))
self.conversion_finished.emit(("Audio generation failed.", "red"), None)
self.log_updated.emit(
("Audio generation failed (final_out_path was not set).", "red")
)
self.conversion_finished.emit(
("Audio generation failed.", "red"), None
)
elif audio_segments and not merge_chapters:
self.conversion_finished.emit(
(
@@ -856,7 +944,9 @@ class ConversionThread(QThread):
self.waiting_for_user_input = False
self._chapter_options_event.set()
def _generate_m4b_with_chapters(self, audio_segments, chapters_time, base_filepath_no_ext):
def _generate_m4b_with_chapters(
self, audio_segments, chapters_time, base_filepath_no_ext
):
"""Generate M4B file with chapters from audio segments"""
final_wav_path = f"{base_filepath_no_ext}.wav"
output_m4b_path = f"{base_filepath_no_ext}.m4b"
@@ -877,7 +967,9 @@ class ConversionThread(QThread):
if success:
return wav_path
else:
self.log_updated.emit((f"\nFailed to save single/no chapter audio as WAV", "red"))
self.log_updated.emit(
(f"\nFailed to save single/no chapter audio as WAV", "red")
)
return None
try:
@@ -885,7 +977,7 @@ class ConversionThread(QThread):
with open(chapters_info_path, "w", encoding="utf-8") as f:
f.write(";FFMETADATA1\n")
for chapter in chapters_time:
chapter_title = chapter['chapter'].replace('=', '\\=')
chapter_title = chapter["chapter"].replace("=", "\\=")
f.write(f"[CHAPTER]\n")
f.write(f"TIMEBASE=1/1000\n")
f.write(f"START={int(chapter['start']*1000)}\n")
@@ -898,19 +990,30 @@ class ConversionThread(QThread):
# Use pipe-based approach for audio input with special args for M4B chapters
ffmpeg_args = [
"-i", chapters_info_path,
"-map", "0:a",
"-map_metadata", "1",
"-map_chapters", "1",
"-i",
chapters_info_path,
"-map",
"0:a",
"-map_metadata",
"1",
"-map_chapters",
"1",
*metadata_options,
"-c:a", "aac",
"-q:a", "2", # Quality-based VBR for better quality control
"-movflags", "+faststart+use_metadata_tags", # Added for better compatibility
"-c:a",
"aac",
"-q:a",
"2", # Quality-based VBR for better quality control
"-movflags",
"+faststart+use_metadata_tags", # Added for better compatibility
]
# Use the established unified method with M4B-specific args
success, out_path = self._process_audio_segments(
audio_segments, output_m4b_path, "m4b", use_ffmpeg=True, ffmpeg_args=ffmpeg_args
audio_segments,
output_m4b_path,
"m4b",
use_ffmpeg=True,
ffmpeg_args=ffmpeg_args,
)
# Clean up the temporary chapter metadata file
@@ -918,13 +1021,17 @@ class ConversionThread(QThread):
try:
os.remove(chapters_info_path)
except Exception as e:
self.log_updated.emit((f"Warning: Could not delete chapters file: {e}", "orange"))
self.log_updated.emit(
(f"Warning: Could not delete chapters file: {e}", "orange")
)
if success:
return out_path
# If M4B generation failed, fallback to WAV
self.log_updated.emit((f"M4B conversion failed. Falling back to WAV.\n", "red"))
self.log_updated.emit(
(f"M4B conversion failed. Falling back to WAV.\n", "red")
)
success, wav_path = self._process_audio_segments(
audio_segments, final_wav_path, "wav", use_ffmpeg=False
)
@@ -936,7 +1043,12 @@ class ConversionThread(QThread):
except Exception as e:
# General error during M4B generation - create WAV file directly as final output
self.log_updated.emit((f"Error during M4B generation: {str(e)}.\n\nFalling back to WAV.\n", "red"))
self.log_updated.emit(
(
f"Error during M4B generation: {str(e)}.\n\nFalling back to WAV.\n",
"red",
)
)
# Use the established unified method to create a WAV file as fallback
success, wav_path = self._process_audio_segments(
@@ -953,7 +1065,9 @@ class ConversionThread(QThread):
if success:
return wav_path
else:
self.log_updated.emit((f"Critical error: Failed to save WAV fallback", "red"))
self.log_updated.emit(
(f"Critical error: Failed to save WAV fallback", "red")
)
return None
def _extract_and_add_metadata_tags_to_ffmpeg_cmd(self):
@@ -967,10 +1081,14 @@ class ConversionThread(QThread):
else:
try:
encoding = detect_encoding(self.file_name)
with open(self.file_name, "r", encoding=encoding, errors="replace") as file:
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}")
self.log_updated.emit(
f"Warning: Could not read file for metadata extraction: {e}"
)
return []
# Extract metadata tags using regex
@@ -988,7 +1106,11 @@ class ConversionThread(QThread):
if getattr(self, "from_queue", False):
filename = os.path.splitext(os.path.basename(self.file_name))[0]
else:
filename = os.path.splitext(os.path.basename(self.display_path if self.display_path else self.file_name))[0]
filename = os.path.splitext(
os.path.basename(
self.display_path if self.display_path else self.file_name
)
)[0]
if title_match:
metadata_options.extend(["-metadata", f"title={title_match.group(1)}"])
@@ -1013,18 +1135,23 @@ class ConversionThread(QThread):
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)}"])
metadata_options.extend(
["-metadata", f"album_artist={album_artist_match.group(1)}"]
)
else:
metadata_options.extend(["-metadata", f"album_artist=Unknown"])
# Add composer metadata
if composer_match:
metadata_options.extend(["-metadata", f"composer={composer_match.group(1)}"])
metadata_options.extend(
["-metadata", f"composer={composer_match.group(1)}"]
)
else:
metadata_options.extend(["-metadata", f"composer=Narrator"])
@@ -1139,7 +1266,9 @@ class ConversionThread(QThread):
(group[0]["start"], group[-1]["end"], text.strip())
)
def _write_ass_subtitle(self, file_path, subtitle_entries, is_centered=False, is_narrow=False):
def _write_ass_subtitle(
self, file_path, subtitle_entries, is_centered=False, is_narrow=False
):
with open(file_path, "w", encoding="utf-8", errors="replace") as f:
# Minimal ASS header
f.write("[Script Info]\n")
@@ -1148,7 +1277,9 @@ class ConversionThread(QThread):
# Only events section, use override tags for positioning
f.write("[Events]\n")
f.write("Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV, Effect, Text\n")
f.write(
"Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV, Effect, Text\n"
)
# Set margin based on is_narrow parameter
margin = "90" if is_narrow else ""
@@ -1161,7 +1292,9 @@ class ConversionThread(QThread):
for i, (start, end, text) in enumerate(subtitle_entries, 1):
start_time = self._ass_time(start)
end_time = self._ass_time(end)
f.write(f"Dialogue: 0,{start_time},{end_time},Default,,{margin},{margin},0,,{alignment_tag}{text}\n")
f.write(
f"Dialogue: 0,{start_time},{end_time},Default,,{margin},{margin},0,,{alignment_tag}{text}\n"
)
def cancel(self):
self.cancel_requested = True
@@ -1198,7 +1331,9 @@ class VoicePreviewThread(QThread):
self.use_gpu = use_gpu
# Cache location for preview audio
self.cache_dir = os.path.join(tempfile.gettempdir(), PROGRAM_NAME, "preview_cache")
self.cache_dir = os.path.join(
tempfile.gettempdir(), PROGRAM_NAME, "preview_cache"
)
os.makedirs(self.cache_dir, exist_ok=True)
# Calculate cache path
@@ -1208,7 +1343,9 @@ class VoicePreviewThread(QThread):
"""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]}"
voice_id = (
f"voice_formula_{hashlib.md5(self.voice.encode()).hexdigest()[:8]}"
)
else:
voice_id = self.voice
@@ -1282,7 +1419,9 @@ class PlayAudioThread(QThread):
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")
self.error.emit(
"Audio playback error: The audio system was not properly initialized"
)
else:
self.error.emit(f"Audio playback error: {str(e)}")
@@ -1292,6 +1431,7 @@ class PlayAudioThread(QThread):
# 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()
+302 -115
View File
@@ -74,7 +74,7 @@ from abogen.constants import (
LANGUAGE_DESCRIPTIONS,
VOICES_INTERNAL,
SUPPORTED_LANGUAGES_FOR_SUBTITLE_GENERATION,
COLORS
COLORS,
)
import threading
from abogen.voice_formula_gui import VoiceFormulaDialog
@@ -85,6 +85,7 @@ if platform.system() == "Windows":
import ctypes
import winreg
class DarkTitleBarEventFilter(QObject):
def __init__(self, is_windows, get_dark_mode_func, set_title_bar_dark_mode_func):
super().__init__()
@@ -100,6 +101,7 @@ class DarkTitleBarEventFilter(QObject):
self.set_title_bar_dark_mode(obj, True)
return False
class ShowWarningSignalEmitter(QObject): # New class to handle signal emission
show_warning_signal = pyqtSignal(str, str)
@@ -128,10 +130,14 @@ class InputBox(QLabel):
STYLE_DEFAULT_HOVER = f"background:{COLORS['BLUE_BG_HOVER']}; border-color:{COLORS['BLUE_BORDER_HOVER']};"
STYLE_ACTIVE = f"border:2px dashed {COLORS['GREEN']}; border-radius:5px; padding:20px; background:{COLORS['GREEN_BG']}; min-height:100px;"
STYLE_ACTIVE_HOVER = f"background:{COLORS['GREEN_BG_HOVER']}; border-color:{COLORS['GREEN_BORDER']};"
STYLE_ACTIVE_HOVER = (
f"background:{COLORS['GREEN_BG_HOVER']}; border-color:{COLORS['GREEN_BORDER']};"
)
STYLE_ERROR = f"border:2px dashed {COLORS['RED']}; border-radius:5px; padding:20px; background:{COLORS['RED_BG']}; min-height:100px; color:{COLORS['RED']};"
STYLE_ERROR_HOVER = f"background:{COLORS['RED_BG_HOVER']}; border-color:{COLORS['RED']};"
STYLE_ERROR_HOVER = (
f"background:{COLORS['RED_BG_HOVER']}; border-color:{COLORS['RED']};"
)
def __init__(self, parent=None):
super().__init__(parent)
@@ -140,7 +146,9 @@ class InputBox(QLabel):
self.setText(
"Drag and drop your file here or click to browse.\n(.txt, .epub, .pdf)"
)
self.setStyleSheet(f"QLabel {{ {self.STYLE_DEFAULT} }} QLabel:hover {{ {self.STYLE_DEFAULT_HOVER} }}")
self.setStyleSheet(
f"QLabel {{ {self.STYLE_DEFAULT} }} QLabel:hover {{ {self.STYLE_DEFAULT_HOVER} }}"
)
self.setSizePolicy(QSizePolicy.Expanding, QSizePolicy.Expanding)
self.setCursor(Qt.PointingHandCursor)
@@ -172,7 +180,9 @@ class InputBox(QLabel):
# Add Go to folder button
self.go_to_folder_btn = QPushButton("Go to folder", self)
self.go_to_folder_btn.setStyleSheet("QPushButton { padding: 6px 10px; }")
self.go_to_folder_btn.setToolTip("Open the folder that contains the converted file")
self.go_to_folder_btn.setToolTip(
"Open the folder that contains the converted file"
)
self.go_to_folder_btn.clicked.connect(self.on_go_to_folder_clicked)
self.go_to_folder_btn.hide()
@@ -215,7 +225,6 @@ class InputBox(QLabel):
return float(match.group(1))
raise ValueError(f"Invalid size format: {size_str}")
# Format numbers with commas
def format_num(n):
try:
@@ -265,7 +274,9 @@ class InputBox(QLabel):
)
# Set fixed width to force wrapping
self.setWordWrap(True)
self.setStyleSheet(f"QLabel {{ {self.STYLE_ACTIVE} }} QLabel:hover {{ {self.STYLE_ACTIVE_HOVER} }}")
self.setStyleSheet(
f"QLabel {{ {self.STYLE_ACTIVE} }} QLabel:hover {{ {self.STYLE_ACTIVE_HOVER} }}"
)
self.clear_btn.show()
is_document = self.window().selected_file_type in ["epub", "pdf"]
self.chapters_btn.setVisible(is_document)
@@ -278,7 +289,6 @@ class InputBox(QLabel):
else: # PDF - always use Pages
self.chapters_btn.setText(f"Pages ({chapter_count})")
# Hide textbox and show edit only for .txt files
self.textbox_btn.hide()
# Show edit button for txt files directly
@@ -296,21 +306,23 @@ class InputBox(QLabel):
self.go_to_folder_btn.show()
# Enable add to queue button only when file is accepted (input box is green)
self.resizeEvent(None)
if hasattr(self.window(), 'btn_add_to_queue'):
if hasattr(self.window(), "btn_add_to_queue"):
self.window().btn_add_to_queue.setEnabled(True)
self.chapters_btn.adjustSize()
# Reset the input_box_cleared_by_queue flag after setting file info
if hasattr(self.window(), 'input_box_cleared_by_queue'):
if hasattr(self.window(), "input_box_cleared_by_queue"):
self.window().input_box_cleared_by_queue = False
def set_error(self, message):
self.setText(message)
self.setStyleSheet(f"QLabel {{ {self.STYLE_ERROR} }} QLabel:hover {{ {self.STYLE_ERROR_HOVER} }}")
self.setStyleSheet(
f"QLabel {{ {self.STYLE_ERROR} }} QLabel:hover {{ {self.STYLE_ERROR_HOVER} }}"
)
# Show textbox button in error state as well
self.textbox_btn.show()
# Disable add to queue button on error
if hasattr(self.window(), 'btn_add_to_queue'):
if hasattr(self.window(), "btn_add_to_queue"):
self.window().btn_add_to_queue.setEnabled(False)
def clear_input(self):
@@ -321,7 +333,9 @@ class InputBox(QLabel):
self.setText(
"Drag and drop your file here or click to browse.\n(.txt, .epub, .pdf)"
)
self.setStyleSheet(f"QLabel {{ {self.STYLE_DEFAULT} }} QLabel:hover {{ {self.STYLE_DEFAULT_HOVER} }}")
self.setStyleSheet(
f"QLabel {{ {self.STYLE_DEFAULT} }} QLabel:hover {{ {self.STYLE_DEFAULT_HOVER} }}"
)
self.clear_btn.hide()
self.chapters_btn.hide()
self.chapters_btn.setText("Chapters") # Reset text
@@ -330,10 +344,10 @@ class InputBox(QLabel):
self.edit_btn.hide()
self.go_to_folder_btn.hide()
# Disable add to queue button when input is cleared
if hasattr(self.window(), 'btn_add_to_queue'):
if hasattr(self.window(), "btn_add_to_queue"):
self.window().btn_add_to_queue.setEnabled(False)
# Reset the input_box_cleared_by_queue flag after setting file info
if hasattr(self.window(), 'input_box_cleared_by_queue'):
if hasattr(self.window(), "input_box_cleared_by_queue"):
self.window().input_box_cleared_by_queue = True
def _human_readable_size(self, size, decimal_places=2):
@@ -352,26 +366,42 @@ class InputBox(QLabel):
urls = event.mimeData().urls()
if urls:
ext = urls[0].toLocalFile().lower()
if ext.endswith(".txt") or ext.endswith(".epub") or ext.endswith(".pdf"):
if (
ext.endswith(".txt")
or ext.endswith(".epub")
or ext.endswith(".pdf")
):
event.acceptProposedAction()
# Set hover style based on current state
if self.styleSheet().find(self.STYLE_ACTIVE) != -1:
self.setStyleSheet(f"QLabel {{ {self.STYLE_ACTIVE} }} QLabel:hover {{ {self.STYLE_ACTIVE_HOVER} }} {self.STYLE_ACTIVE_HOVER}")
self.setStyleSheet(
f"QLabel {{ {self.STYLE_ACTIVE} }} QLabel:hover {{ {self.STYLE_ACTIVE_HOVER} }} {self.STYLE_ACTIVE_HOVER}"
)
elif self.styleSheet().find(self.STYLE_ERROR) != -1:
self.setStyleSheet(f"QLabel {{ {self.STYLE_ERROR} }} QLabel:hover {{ {self.STYLE_ERROR_HOVER} }} {self.STYLE_ERROR_HOVER}")
self.setStyleSheet(
f"QLabel {{ {self.STYLE_ERROR} }} QLabel:hover {{ {self.STYLE_ERROR_HOVER} }} {self.STYLE_ERROR_HOVER}"
)
else:
self.setStyleSheet(f"QLabel {{ {self.STYLE_DEFAULT} }} QLabel:hover {{ {self.STYLE_DEFAULT_HOVER} }} {self.STYLE_DEFAULT_HOVER}")
self.setStyleSheet(
f"QLabel {{ {self.STYLE_DEFAULT} }} QLabel:hover {{ {self.STYLE_DEFAULT_HOVER} }} {self.STYLE_DEFAULT_HOVER}"
)
return
event.ignore()
def dragLeaveEvent(self, event):
# Restore the style based on current state
if self.styleSheet().find(self.STYLE_ACTIVE) != -1:
self.setStyleSheet(f"QLabel {{ {self.STYLE_ACTIVE} }} QLabel:hover {{ {self.STYLE_ACTIVE_HOVER} }}")
self.setStyleSheet(
f"QLabel {{ {self.STYLE_ACTIVE} }} QLabel:hover {{ {self.STYLE_ACTIVE_HOVER} }}"
)
elif self.styleSheet().find(self.STYLE_ERROR) != -1:
self.setStyleSheet(f"QLabel {{ {self.STYLE_ERROR} }} QLabel:hover {{ {self.STYLE_ERROR_HOVER} }}")
self.setStyleSheet(
f"QLabel {{ {self.STYLE_ERROR} }} QLabel:hover {{ {self.STYLE_ERROR_HOVER} }}"
)
else:
self.setStyleSheet(f"QLabel {{ {self.STYLE_DEFAULT} }} QLabel:hover {{ {self.STYLE_DEFAULT_HOVER} }}")
self.setStyleSheet(
f"QLabel {{ {self.STYLE_DEFAULT} }} QLabel:hover {{ {self.STYLE_DEFAULT_HOVER} }}"
)
event.accept()
def dropEvent(self, event):
@@ -433,7 +463,11 @@ class InputBox(QLabel):
# win.selected_file holds the path to the text that is converted.
file_to_check = win.selected_file
if file_to_check and os.path.exists(file_to_check) and os.path.isfile(file_to_check):
if (
file_to_check
and os.path.exists(file_to_check)
and os.path.isfile(file_to_check)
):
folder_path = os.path.dirname(file_to_check)
QDesktopServices.openUrl(QUrl.fromLocalFile(folder_path))
else:
@@ -582,6 +616,7 @@ class TextboxDialog(QDialog):
self.update_char_count()
self.text_edit.setFocus()
def migrate_subtitle_format(config):
"""Convert old subtitle_format values to new internal keys."""
old_to_new = {
@@ -596,6 +631,7 @@ def migrate_subtitle_format(config):
config["subtitle_format"] = old_to_new[val]
save_config(config)
class abogen(QWidget):
def __init__(self):
super().__init__()
@@ -631,7 +667,9 @@ 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.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
)
@@ -645,7 +683,9 @@ class abogen(QWidget):
# Create warning signal emitter
self.warning_signal_emitter = ShowWarningSignalEmitter()
self.warning_signal_emitter.show_warning_signal.connect(self.show_model_download_warning)
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
@@ -793,6 +833,7 @@ class abogen(QWidget):
self.btn_voice_formula_mixer.setStyleSheet("QPushButton { padding: 6px 12px; }")
self.btn_voice_formula_mixer.clicked.connect(self.show_voice_formula_dialog)
voice_layout.addWidget(self.btn_voice_formula_mixer)
# Play/Stop icons
def make_icon(color, shape):
pix = QPixmap(20, 20)
@@ -900,13 +941,15 @@ class abogen(QWidget):
self.subtitle_format_combo.setStyleSheet(
"QComboBox { min-height: 20px; padding: 6px 12px; }"
)
self.subtitle_format_combo.setSizePolicy(QSizePolicy.Expanding, QSizePolicy.Fixed)
self.subtitle_format_combo.setSizePolicy(
QSizePolicy.Expanding, QSizePolicy.Fixed
)
subtitle_formats = [
("srt", "SRT (standard)"),
("ass_wide", "ASS (wide)"),
("ass_narrow", "ASS (narrow)"),
("ass_centered_wide", "ASS (centered wide)"),
("ass_centered_narrow", "ASS (centered narrow)")
("ass_centered_narrow", "ASS (centered narrow)"),
]
for value, text in subtitle_formats:
self.subtitle_format_combo.addItem(text, value)
@@ -927,13 +970,19 @@ class abogen(QWidget):
replace_newlines_layout.addWidget(replace_newlines_label)
self.replace_newlines_combo = QComboBox(self)
self.replace_newlines_combo.addItems(["Disabled", "Enabled"])
self.replace_newlines_combo.setToolTip("Replace single newlines in the input text with spaces before processing.")
self.replace_newlines_combo.setToolTip(
"Replace single newlines in the input text with spaces before processing."
)
self.replace_newlines_combo.setStyleSheet(
"QComboBox { min-height: 20px; padding: 6px 12px; }"
)
self.replace_newlines_combo.setSizePolicy(QSizePolicy.Expanding, QSizePolicy.Fixed)
self.replace_newlines_combo.setSizePolicy(
QSizePolicy.Expanding, QSizePolicy.Fixed
)
# Set initial value based on config
self.replace_newlines_combo.setCurrentIndex(1 if self.replace_single_newlines else 0)
self.replace_newlines_combo.setCurrentIndex(
1 if self.replace_single_newlines else 0
)
self.replace_newlines_combo.currentIndexChanged.connect(
lambda idx: self.toggle_replace_single_newlines(idx == 1)
)
@@ -969,9 +1018,7 @@ class abogen(QWidget):
selected_folder_label = QLabel("Selected folder:", self.save_path_row_widget)
save_path_row.addWidget(selected_folder_label)
self.save_path_label = QLabel("", self.save_path_row_widget)
self.save_path_label.setStyleSheet(
f"QLabel {{ color: {COLORS['GREEN']}; }}"
)
self.save_path_label.setStyleSheet(f"QLabel {{ color: {COLORS['GREEN']}; }}")
self.save_path_label.setSizePolicy(QSizePolicy.Expanding, QSizePolicy.Preferred)
save_path_row.addWidget(self.save_path_label)
self.save_path_row_widget.hide() # Hide the whole row by default
@@ -1119,6 +1166,7 @@ class abogen(QWidget):
dialog.setWindowModality(Qt.NonModal)
dialog.setModal(False)
dialog.show() # We'll handle the dialog result asynchronously
def on_dialog_finished(result):
if result != QDialog.Accepted:
return False
@@ -1161,22 +1209,49 @@ class abogen(QWidget):
# Save metadata if available
meta_dir = os.path.join(project_dir, "metadata")
os.makedirs(meta_dir, exist_ok=True) # Save book metadata if available
os.makedirs(
meta_dir, exist_ok=True
) # Save book metadata if available
if hasattr(dialog, "book_metadata"):
meta_path = os.path.join(meta_dir, "book_info.txt")
with open(meta_path, "w", encoding="utf-8") as f:
# Clean HTML tags from metadata
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')))
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')))
if dialog.book_metadata.get("description"):
description = re.sub(
r"<[^>]+>",
"",
str(dialog.book_metadata.get("description")),
)
f.write(f"\nDescription:\n{description}\n")
# Save cover image if available
@@ -1404,7 +1479,7 @@ class abogen(QWidget):
self._update_log_main_thread(message)
def _update_log_main_thread(self, message):
if not hasattr(self, '_log_lines'):
if not hasattr(self, "_log_lines"):
self._log_lines = []
# Always produce a single HTML line (with color if tuple)
@@ -1423,7 +1498,7 @@ class abogen(QWidget):
# Trim buffer if needed
if len(self._log_lines) > self.log_window_max_lines:
self._log_lines = self._log_lines[-self.log_window_max_lines:]
self._log_lines = self._log_lines[-self.log_window_max_lines :]
self.log_text.clear()
for line in self._log_lines:
self.log_text.insertHtml(line)
@@ -1438,7 +1513,11 @@ class abogen(QWidget):
def _get_queue_progress_format(self, value=None):
"""Return the progress bar format string for queue mode."""
if hasattr(self, 'queued_items') and self.queued_items and hasattr(self, 'current_queue_index'):
if (
hasattr(self, "queued_items")
and self.queued_items
and hasattr(self, "current_queue_index")
):
N = self.current_queue_index + 1
M = len(self.queued_items)
percent = value if value is not None else self.progress_bar.value()
@@ -1453,7 +1532,11 @@ class abogen(QWidget):
value = 99
self.progress_bar.setValue(value)
# Show queue progress if in queue mode
if hasattr(self, 'queued_items') and self.queued_items and hasattr(self, 'current_queue_index'):
if (
hasattr(self, "queued_items")
and self.queued_items
and hasattr(self, "current_queue_index")
):
N = self.current_queue_index + 1
M = len(self.queued_items)
self.progress_bar.setFormat(f"{value}% ({N}/{M})")
@@ -1499,9 +1582,9 @@ class abogen(QWidget):
pass
self.btn_start.clicked.connect(self.start_conversion)
def enqueue(self, item : QueuedItem):
def enqueue(self, item: QueuedItem):
self.queued_items.append(item)
#self.update_log((f"Enqueued: {item.file_name}", True))
# self.update_log((f"Enqueued: {item.file_name}", True))
# enable start queue button, manage queue button
self.enable_disable_queue_buttons()
@@ -1509,8 +1592,15 @@ class abogen(QWidget):
return self.queued_items
def add_to_queue(self):
# Use the file currently displayed in the input box
file_to_queue = self.displayed_file_path if self.displayed_file_path else self.selected_file
# For epub/pdf, always use the converted txt file (selected_file)
if self.selected_file_type in ["epub", "pdf"]:
file_to_queue = self.selected_file
else:
file_to_queue = (
self.displayed_file_path
if self.displayed_file_path
else self.selected_file
)
if not file_to_queue:
self.input_box.set_error("Please add a file.")
@@ -1529,23 +1619,26 @@ class abogen(QWidget):
subtitle_mode=actual_subtitle_mode,
output_format=self.selected_format,
total_char_count=self.char_count,
replace_single_newlines=self.replace_single_newlines
replace_single_newlines=self.replace_single_newlines,
)
# Prevent adding duplicate items to the queue
for queued_item in self.queued_items:
if (
queued_item.file_name == item_queue.file_name and
queued_item.lang_code == item_queue.lang_code and
queued_item.speed == item_queue.speed and
queued_item.voice == item_queue.voice and
queued_item.save_option == item_queue.save_option and
queued_item.output_folder == item_queue.output_folder and
queued_item.subtitle_mode == item_queue.subtitle_mode and
queued_item.output_format == item_queue.output_format and
getattr(queued_item, "replace_single_newlines", False) == item_queue.replace_single_newlines
queued_item.file_name == item_queue.file_name
and queued_item.lang_code == item_queue.lang_code
and queued_item.speed == item_queue.speed
and queued_item.voice == item_queue.voice
and queued_item.save_option == item_queue.save_option
and queued_item.output_folder == item_queue.output_folder
and queued_item.subtitle_mode == item_queue.subtitle_mode
and queued_item.output_format == item_queue.output_format
and getattr(queued_item, "replace_single_newlines", False)
== item_queue.replace_single_newlines
):
QMessageBox.warning(self, "Duplicate Item", "This item is already in the queue.")
QMessageBox.warning(
self, "Duplicate Item", "This item is already in the queue."
)
return
self.enqueue(item_queue)
@@ -1562,7 +1655,7 @@ class abogen(QWidget):
"Confirm Clear Queue",
f"Are you sure you want to clear {len(self.queued_items)} items from the queue?",
QMessageBox.Yes | QMessageBox.No,
QMessageBox.No
QMessageBox.No,
)
if reply != QMessageBox.Yes:
return
@@ -1598,7 +1691,9 @@ class abogen(QWidget):
self.subtitle_mode = queued_item.subtitle_mode
self.selected_format = queued_item.output_format
self.char_count = queued_item.total_char_count
self.replace_single_newlines = getattr(queued_item, "replace_single_newlines", False)
self.replace_single_newlines = getattr(
queued_item, "replace_single_newlines", False
)
self.start_conversion(from_queue=True)
else:
# Queue finished, reset index
@@ -1612,7 +1707,6 @@ class abogen(QWidget):
else:
self.current_queue_index = 0 # Reset for next time
def get_voice_formula(self) -> str:
if self.mixed_voice_state:
formula_components = [
@@ -1625,6 +1719,7 @@ class abogen(QWidget):
def get_selected_lang(self, voice_formula) -> str:
if self.selected_profile_name:
from abogen.voice_profiles import load_profiles
entry = load_profiles().get(self.selected_profile_name, {})
selected_lang = entry.get("language")
else:
@@ -1636,11 +1731,7 @@ class abogen(QWidget):
return selected_lang
def get_actual_subtitle_mode(self) -> str:
return (
"Disabled"
if not self.subtitle_combo.isEnabled()
else self.subtitle_mode
)
return "Disabled" if not self.subtitle_combo.isEnabled() else self.subtitle_mode
def start_conversion(self, from_queue=False):
if not self.selected_file:
@@ -1651,7 +1742,12 @@ class abogen(QWidget):
self.convert_input_box_to_log()
self.progress_bar.setValue(0)
# Show queue progress if in queue mode
if from_queue and hasattr(self, 'queued_items') and self.queued_items and hasattr(self, 'current_queue_index'):
if (
from_queue
and hasattr(self, "queued_items")
and self.queued_items
and hasattr(self, "current_queue_index")
):
N = self.current_queue_index + 1
M = len(self.queued_items)
self.progress_bar.setFormat(f"0% ({N}/{M})")
@@ -1725,9 +1821,13 @@ class abogen(QWidget):
self.replace_single_newlines
)
# Pass separate_chapters_format setting
self.conversion_thread.separate_chapters_format = self.separate_chapters_format
self.conversion_thread.separate_chapters_format = (
self.separate_chapters_format
)
# Pass subtitle format setting
self.conversion_thread.subtitle_format = self.config.get("subtitle_format", "ass_centered_narrow")
self.conversion_thread.subtitle_format = self.config.get(
"subtitle_format", "ass_centered_narrow"
)
# Pass chapter count for EPUB or PDF files
if self.selected_file_type in ["epub", "pdf"] and hasattr(
self, "selected_chapters"
@@ -1781,7 +1881,6 @@ class abogen(QWidget):
f"Processed {len(self.queued_items)} items:<br><br>"
)
for idx, item in enumerate(self.queued_items, 1):
output = getattr(item, "output_path", None)
if not output:
@@ -1834,7 +1933,7 @@ class abogen(QWidget):
else self.selected_file
)
# Only repopulate if not cleared by queue
if not getattr(self, 'input_box_cleared_by_queue', False):
if not getattr(self, "input_box_cleared_by_queue", False):
if display_path and os.path.exists(display_path):
self.input_box.set_file_info(display_path)
else:
@@ -1873,7 +1972,10 @@ class abogen(QWidget):
self.open_file_btn.setVisible(show_open_file_button)
# Only show finish_widget if queue is done
if self.current_queue_index + 1 >= len(self.queued_items) or not self.queued_items:
if (
self.current_queue_index + 1 >= len(self.queued_items)
or not self.queued_items
):
# Queue finished, show finish screen
self.controls_widget.hide()
self.finish_widget.show()
@@ -1938,7 +2040,7 @@ class abogen(QWidget):
)
# Only repopulate if not cleared by queue
if not getattr(self, 'input_box_cleared_by_queue', False):
if not getattr(self, "input_box_cleared_by_queue", False):
if display_path and os.path.exists(display_path):
self.input_box.set_file_info(display_path)
else:
@@ -2031,6 +2133,7 @@ class abogen(QWidget):
voice_to_cache = voice_formula
if self.selected_profile_name:
from abogen.voice_profiles import load_profiles
entry = load_profiles().get(self.selected_profile_name, {})
lang_to_cache = entry.get("language")
else:
@@ -2053,7 +2156,9 @@ class abogen(QWidget):
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]}"
voice_id = (
f"voice_formula_{hashlib.md5(voice_to_cache.encode()).hexdigest()[:8]}"
)
else: # Single voice
voice_id = voice_to_cache
@@ -2096,13 +2201,17 @@ class abogen(QWidget):
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())
lambda msg: (
self._show_preview_error_box(msg),
cleanup_cached_play(),
)
)
self.play_audio_thread.start()
except Exception as e:
@@ -2192,23 +2301,29 @@ class abogen(QWidget):
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:
if hasattr(self, "preview_thread") and not from_cache:
temp_wav = self.preview_thread.temp_wav
elif from_cache: # This case is now handled before calling _play_preview_audio
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, please try again.")
self._show_error_message_box(
"Preview Error",
"Cache file expected but not found, please try again.",
)
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._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()
if hasattr(self, "loading_movie"):
self.loading_movie.stop()
self._show_error_message_box(
"Preview Error", "Preview error: No audio generated."
)
@@ -2216,7 +2331,8 @@ class abogen(QWidget):
return
# stop loading animation, switch to stop icon
if hasattr(self, 'loading_movie'): 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")
@@ -2224,9 +2340,16 @@ class abogen(QWidget):
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:
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
if os.path.exists(
temp_wav
): # Ensure it exists before trying to remove
os.remove(temp_wav)
except Exception:
pass
@@ -2235,6 +2358,7 @@ class abogen(QWidget):
try:
# Ensure pygame mixer is initialized for the audio thread
import pygame
if not pygame.mixer.get_init():
pygame.mixer.init()
@@ -2266,9 +2390,11 @@ class abogen(QWidget):
def _preview_cleanup(self):
self.preview_playing = False
if hasattr(self, 'loading_movie'): self.loading_movie.stop()
if hasattr(self, "loading_movie"):
self.loading_movie.stop()
try:
if hasattr(self, 'loading_movie'): 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)
@@ -2301,10 +2427,12 @@ class abogen(QWidget):
):
if not hasattr(self, "_conversion_lock"):
self._conversion_lock = threading.Lock()
def _cancel():
with self._conversion_lock:
self.conversion_thread.cancel() # <-- Use cancel() method
self.conversion_thread.wait()
threading.Thread(target=_cancel, daemon=True).start()
self.is_converting = False
@@ -2323,7 +2451,7 @@ class abogen(QWidget):
else self.selected_file
)
# Only repopulate if not cleared by queue
if not getattr(self, 'input_box_cleared_by_queue', False):
if not getattr(self, "input_box_cleared_by_queue", False):
if display_path and os.path.exists(display_path):
self.input_box.set_file_info(display_path)
else:
@@ -2409,9 +2537,10 @@ class abogen(QWidget):
def is_windows_dark_mode():
try:
import winreg
with winreg.OpenKey(
winreg.HKEY_CURRENT_USER,
r"Software\Microsoft\Windows\CurrentVersion\Themes\Personalize"
r"Software\Microsoft\Windows\CurrentVersion\Themes\Personalize",
) as key:
value, _ = winreg.QueryValueEx(key, "AppsUseLightTheme")
return value == 0
@@ -2436,14 +2565,23 @@ class abogen(QWidget):
palette.setColor(QPalette.ColorRole.Button, button_bg)
palette.setColor(QPalette.ColorRole.ButtonText, Qt.GlobalColor.white)
# Disabled roles
palette.setColor(QPalette.ColorGroup.Disabled, QPalette.ColorRole.WindowText, disabled_fg)
palette.setColor(QPalette.ColorGroup.Disabled, QPalette.ColorRole.Text, disabled_fg)
palette.setColor(QPalette.ColorGroup.Disabled, QPalette.ColorRole.ButtonText, disabled_fg)
palette.setColor(QPalette.ColorGroup.Disabled, QPalette.ColorRole.Base, dark_bg)
palette.setColor(QPalette.ColorGroup.Disabled, QPalette.ColorRole.Button, dark_bg)
palette.setColor(
QPalette.ColorGroup.Disabled, QPalette.ColorRole.WindowText, disabled_fg
)
palette.setColor(
QPalette.ColorGroup.Disabled, QPalette.ColorRole.Text, disabled_fg
)
palette.setColor(
QPalette.ColorGroup.Disabled, QPalette.ColorRole.ButtonText, disabled_fg
)
palette.setColor(
QPalette.ColorGroup.Disabled, QPalette.ColorRole.Base, dark_bg
)
palette.setColor(
QPalette.ColorGroup.Disabled, QPalette.ColorRole.Button, dark_bg
)
app.setPalette(palette)
def set_light_palette():
palette = QPalette()
disabled_fg = QColor(COLORS["LIGHT_DISABLED"])
@@ -2457,11 +2595,25 @@ class abogen(QWidget):
palette.setColor(QPalette.ColorRole.Button, Qt.GlobalColor.white)
palette.setColor(QPalette.ColorRole.ButtonText, Qt.GlobalColor.black)
# Disabled roles
palette.setColor(QPalette.ColorGroup.Disabled, QPalette.ColorRole.WindowText, disabled_fg)
palette.setColor(QPalette.ColorGroup.Disabled, QPalette.ColorRole.Text, disabled_fg)
palette.setColor(QPalette.ColorGroup.Disabled, QPalette.ColorRole.ButtonText, disabled_fg)
palette.setColor(QPalette.ColorGroup.Disabled, QPalette.ColorRole.Base, Qt.GlobalColor.white)
palette.setColor(QPalette.ColorGroup.Disabled, QPalette.ColorRole.Button, Qt.GlobalColor.white)
palette.setColor(
QPalette.ColorGroup.Disabled, QPalette.ColorRole.WindowText, disabled_fg
)
palette.setColor(
QPalette.ColorGroup.Disabled, QPalette.ColorRole.Text, disabled_fg
)
palette.setColor(
QPalette.ColorGroup.Disabled, QPalette.ColorRole.ButtonText, disabled_fg
)
palette.setColor(
QPalette.ColorGroup.Disabled,
QPalette.ColorRole.Base,
Qt.GlobalColor.white,
)
palette.setColor(
QPalette.ColorGroup.Disabled,
QPalette.ColorRole.Button,
Qt.GlobalColor.white,
)
app.setPalette(palette)
# --- Dark title bar support for Windows ---
@@ -2473,12 +2625,19 @@ class abogen(QWidget):
set_window_attribute = ctypes.windll.dwmapi.DwmSetWindowAttribute
hwnd = int(window.winId())
value = ctypes.c_int(2 if enable else 0)
set_window_attribute(hwnd, DWMWA_USE_IMMERSIVE_DARK_MODE, ctypes.byref(value), ctypes.sizeof(value))
set_window_attribute(
hwnd,
DWMWA_USE_IMMERSIVE_DARK_MODE,
ctypes.byref(value),
ctypes.sizeof(value),
)
except Exception:
pass
# Main logic
dark_mode = theme == "dark" or (theme == "system" and is_windows and is_windows_dark_mode())
dark_mode = theme == "dark" or (
theme == "system" and is_windows and is_windows_dark_mode()
)
if dark_mode:
app.setStyle("Fusion")
set_dark_palette()
@@ -2512,7 +2671,10 @@ class abogen(QWidget):
delattr(app, "_dark_titlebar_event_filter")
def get_dark_mode():
return theme == "dark" or (theme == "system" and is_windows and is_windows_dark_mode())
return theme == "dark" or (
theme == "system" and is_windows and is_windows_dark_mode()
)
app._dark_titlebar_event_filter = DarkTitleBarEventFilter(
is_windows, get_dark_mode, set_title_bar_dark_mode
)
@@ -2554,7 +2716,9 @@ class abogen(QWidget):
# 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")
separate_chapters_format_menu.setToolTip(
"Choose the format for individual chapter files"
)
format_group = QActionGroup(self)
format_group.setExclusive(True)
@@ -2563,7 +2727,11 @@ class abogen(QWidget):
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_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)
@@ -2584,7 +2752,11 @@ class abogen(QWidget):
# Add shortcut to desktop (Windows or Linux)
if platform.system() == "Windows" or platform.system() == "Linux":
# Use extended label on Linux
label = "Create desktop shortcut and install" if platform.system() == "Linux" else "Create desktop shortcut"
label = (
"Create desktop shortcut and install"
if platform.system() == "Linux"
else "Create desktop shortcut"
)
add_shortcut_action = QAction(label, self)
add_shortcut_action.triggered.connect(self.add_shortcut_to_desktop)
menu.addAction(add_shortcut_action)
@@ -2673,7 +2845,9 @@ class abogen(QWidget):
target = os.path.join(python_dir, "Scripts", "abogen.exe")
if not os.path.exists(target):
QMessageBox.critical(
self, "Shortcut Error", f"Could not find abogen.exe at:\n{target}"
self,
"Shortcut Error",
f"Could not find abogen.exe at:\n{target}",
)
return
@@ -2683,13 +2857,19 @@ class abogen(QWidget):
icon = target # Create a more direct PowerShell command
shortcut_ps = shortcut_path.replace("'", "''").replace("\\", "\\\\")
target_ps = target.replace("'", "''").replace("\\", "\\\\")
workdir_ps = os.path.dirname(target).replace("'", "''").replace("\\", "\\\\")
workdir_ps = (
os.path.dirname(target).replace("'", "''").replace("\\", "\\\\")
)
icon_ps = icon.replace("'", "''").replace("\\", "\\\\")
# Create PowerShell script as a single line with no line breaks (more reliable)
ps_cmd = f"$s=New-Object -ComObject WScript.Shell; $lnk=$s.CreateShortcut('{shortcut_ps}'); $lnk.TargetPath='{target_ps}'; $lnk.WorkingDirectory='{workdir_ps}'; $lnk.IconLocation='{icon_ps}'; $lnk.Save()"
# Run PowerShell with the command directly
proc = create_process("powershell -NoProfile -ExecutionPolicy Bypass -Command \"" + ps_cmd + "\"")
proc = create_process(
'powershell -NoProfile -ExecutionPolicy Bypass -Command "'
+ ps_cmd
+ '"'
)
proc.wait()
if proc.returncode == 0:
@@ -2702,17 +2882,20 @@ class abogen(QWidget):
QMessageBox.critical(
self,
"Shortcut Error",
f"PowerShell failed with exit code: {proc.returncode}"
f"PowerShell failed with exit code: {proc.returncode}",
)
elif platform.system() == "Linux":
desktop = user_desktop_dir()
if not desktop or not os.path.isdir(desktop):
QMessageBox.critical(self, "Shortcut Error", "Could not determine desktop directory.")
QMessageBox.critical(
self, "Shortcut Error", "Could not determine desktop directory."
)
return
shortcut_path = os.path.join(desktop, "abogen.desktop")
import shutil
found = shutil.which("abogen")
if found:
target = found
@@ -2728,7 +2911,11 @@ class abogen(QWidget):
if os.path.exists(target_fallback):
target = target_fallback
else:
QMessageBox.critical(self, "Shortcut Error", "Could not find abogen executable in PATH or common installation directories.")
QMessageBox.critical(
self,
"Shortcut Error",
"Could not find abogen executable in PATH or common installation directories.",
)
return
icon_path = get_resource_path("abogen.assets", "icon.png")
@@ -2763,6 +2950,7 @@ Categories=AudioVideo;Audio;Utility;
)
if reply == QMessageBox.Yes:
import shutil
user_app_dir = os.path.expanduser("~/.local/share/applications")
os.makedirs(user_app_dir, exist_ok=True)
user_entry = os.path.join(user_app_dir, "abogen.desktop")
@@ -2782,7 +2970,9 @@ Categories=AudioVideo;Audio;Utility;
)
else:
QMessageBox.information(
self, "Unsupported OS", "Desktop shortcut creation is not supported on this operating system."
self,
"Unsupported OS",
"Desktop shortcut creation is not supported on this operating system.",
)
except Exception as e:
@@ -3057,9 +3247,7 @@ Categories=AudioVideo;Audio;Utility;
result_msg += f"\nAlso deleted {deleted_preview_count} preview cache file{'s' if deleted_preview_count != 1 else ''}."
# Show results
QMessageBox.information(
self, "Temporary Files Cleared", result_msg
)
QMessageBox.information(self, "Temporary Files Cleared", result_msg)
# If currently selected file is in the temp directory, clear the UI
if (
@@ -3139,4 +3327,3 @@ Categories=AudioVideo;Audio;Utility;
def show_model_download_warning(self, title, message):
QMessageBox.information(self, title, message)
+10 -1
View File
@@ -1,16 +1,20 @@
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)
@@ -22,7 +26,10 @@ def tracked_hf_hub_download(*args, **kwargs):
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.")
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:
@@ -32,5 +39,7 @@ def tracked_hf_hub_download(*args, **kwargs):
print(msg, flush=True)
return hf_hub_download(*args, **kwargs)
import huggingface_hub
huggingface_hub.hf_hub_download = tracked_hf_hub_download
+3 -1
View File
@@ -1,18 +1,20 @@
import gpustat
def check():
try:
stats = gpustat.new_query()
except Exception:
return False
nvidia_keywords = ['nvidia', 'rtx', 'gtx', 'quadro', 'tesla', 'titan', 'mx']
nvidia_keywords = ["nvidia", "rtx", "gtx", "quadro", "tesla", "titan", "mx"]
for gpu in stats.gpus:
name = gpu.name.lower()
if any(keyword in name for keyword in nvidia_keywords):
return True
return False
if __name__ == "__main__":
stats = gpustat.new_query()
for gpu in stats.gpus:
+1
View File
@@ -25,6 +25,7 @@ if sys.stderr is None:
if platform.system() == "Darwin" and platform.processor() == "arm":
os.environ["PYTORCH_ENABLE_MPS_FALLBACK"] = "1"
# Custom message handler to filter out specific Qt warnings
def qt_message_handler(mode, context, message):
if "Wayland does not support QWindow::requestActivate()" in message:
+113 -51
View File
@@ -20,6 +20,7 @@ from abogen.constants import COLORS
from copy import deepcopy
from PyQt5.QtGui import QFontMetrics
class ElidedLabel(QLabel):
def __init__(self, text, parent=None):
super().__init__(text, parent)
@@ -40,6 +41,7 @@ class ElidedLabel(QLabel):
def fullText(self):
return self._full_text
class QueueListItemWidget(QWidget):
def __init__(self, file_name, char_count):
super().__init__()
@@ -47,6 +49,7 @@ class QueueListItemWidget(QWidget):
layout.setContentsMargins(12, 0, 6, 0)
layout.setSpacing(0)
import os
name_label = ElidedLabel(os.path.basename(file_name))
char_label = QLabel(f"Chars: {char_count}")
char_label.setStyleSheet(f"color: {COLORS['LIGHT_DISABLED']};")
@@ -56,6 +59,7 @@ class QueueListItemWidget(QWidget):
layout.addWidget(char_label, 0)
self.setLayout(layout)
class DroppableQueueListWidget(QListWidget):
def __init__(self, parent_dialog):
super().__init__()
@@ -73,7 +77,7 @@ class DroppableQueueListWidget(QListWidget):
def dragEnterEvent(self, event):
if event.mimeData().hasUrls():
for url in event.mimeData().urls():
if url.isLocalFile() and url.toLocalFile().lower().endswith('.txt'):
if url.isLocalFile() and url.toLocalFile().lower().endswith(".txt"):
self.drag_overlay.resize(self.size())
self.drag_overlay.setVisible(True)
event.acceptProposedAction()
@@ -84,7 +88,7 @@ class DroppableQueueListWidget(QListWidget):
def dragMoveEvent(self, event):
if event.mimeData().hasUrls():
for url in event.mimeData().urls():
if url.isLocalFile() and url.toLocalFile().lower().endswith('.txt'):
if url.isLocalFile() and url.toLocalFile().lower().endswith(".txt"):
event.acceptProposedAction()
return
event.ignore()
@@ -96,7 +100,11 @@ class DroppableQueueListWidget(QListWidget):
def dropEvent(self, event):
self.drag_overlay.setVisible(False)
if event.mimeData().hasUrls():
file_paths = [url.toLocalFile() for url in event.mimeData().urls() if url.isLocalFile() and url.toLocalFile().lower().endswith('.txt')]
file_paths = [
url.toLocalFile()
for url in event.mimeData().urls()
if url.isLocalFile() and url.toLocalFile().lower().endswith(".txt")
]
if file_paths:
self.parent_dialog.add_files_from_paths(file_paths)
event.acceptProposedAction()
@@ -107,14 +115,17 @@ class DroppableQueueListWidget(QListWidget):
def resizeEvent(self, event):
super().resizeEvent(event)
if hasattr(self, 'drag_overlay'):
if hasattr(self, "drag_overlay"):
self.drag_overlay.resize(self.size())
class QueueManager(QDialog):
def __init__(self, parent, queue: list, title="Queue Manager", size=(600, 700)):
super().__init__()
self.queue = queue
self._original_queue = deepcopy(queue) # Store a deep copy of the original queue
self._original_queue = deepcopy(
queue
) # Store a deep copy of the original queue
self.parent = parent
layout = QVBoxLayout()
layout.setContentsMargins(15, 15, 15, 15) # set main layout margins
@@ -141,10 +152,12 @@ class QueueManager(QDialog):
# Overlay label for empty queue
self.empty_overlay = QLabel(
"Drag and drop your text files here or use the 'Add files' button.",
self.listwidget
self.listwidget,
)
self.empty_overlay.setAlignment(Qt.AlignCenter)
self.empty_overlay.setStyleSheet(f"color: {COLORS['LIGHT_DISABLED']}; background: transparent; padding: 20px;")
self.empty_overlay.setStyleSheet(
f"color: {COLORS['LIGHT_DISABLED']}; background: transparent; padding: 20px;"
)
self.empty_overlay.setWordWrap(True)
self.empty_overlay.setAttribute(Qt.WA_TransparentForMouseEvents, True)
self.empty_overlay.hide()
@@ -208,13 +221,14 @@ class QueueManager(QDialog):
file_name = item.file_name
display_name = file_name
import os
if os.path.sep in file_name:
display_name = os.path.basename(file_name)
# Get icon for the file
icon = icon_provider.icon(QFileInfo(file_name))
list_item = QListWidgetItem()
# Set tooltip with detailed info
output_folder = getattr(item, 'output_folder', '')
output_folder = getattr(item, "output_folder", "")
tooltip = (
f"<b>Path:</b> {file_name}<br>"
f"<b>Language:</b> {getattr(item, 'lang_code', '')}<br>"
@@ -222,7 +236,7 @@ class QueueManager(QDialog):
f"<b>Voice:</b> {getattr(item, 'voice', '')}<br>"
f"<b>Save Option:</b> {getattr(item, 'save_option', '')}<br>"
)
if output_folder not in (None, '', 'None'):
if output_folder not in (None, "", "None"):
tooltip += f"<b>Output Folder:</b> {output_folder}<br>"
tooltip += (
f"<b>Subtitle Mode:</b> {getattr(item, 'subtitle_mode', '')}<br>"
@@ -233,7 +247,7 @@ class QueueManager(QDialog):
list_item.setToolTip(tooltip)
list_item.setIcon(icon)
# Use custom widget for display
char_count = getattr(item, 'total_char_count', 0)
char_count = getattr(item, "total_char_count", 0)
widget = QueueListItemWidget(file_name, char_count)
self.listwidget.addItem(list_item)
self.listwidget.setItemWidget(list_item, widget)
@@ -244,6 +258,7 @@ class QueueManager(QDialog):
if not items:
return
from PyQt5.QtWidgets import QMessageBox
# Remove by index to ensure correct mapping
rows = sorted([self.listwidget.row(item) for item in items], reverse=True)
# Warn user if removing multiple files
@@ -253,7 +268,7 @@ class QueueManager(QDialog):
"Confirm Remove",
f"Are you sure you want to remove {len(rows)} selected items from the queue?",
QMessageBox.Yes | QMessageBox.No,
QMessageBox.No
QMessageBox.No,
)
if reply != QMessageBox.Yes:
return
@@ -265,19 +280,22 @@ class QueueManager(QDialog):
def clear_queue(self):
from PyQt5.QtWidgets import QMessageBox
if len(self.queue) > 1:
reply = QMessageBox.question(
self,
"Confirm Clear Queue",
f"Are you sure you want to clear {len(self.queue)} items from the queue?",
QMessageBox.Yes | QMessageBox.No,
QMessageBox.No
QMessageBox.No,
)
if reply != QMessageBox.Yes:
return
self.queue.clear()
self.listwidget.clear()
self.empty_overlay.resize(self.listwidget.size()) # Ensure overlay is sized correctly
self.empty_overlay.resize(
self.listwidget.size()
) # Ensure overlay is sized correctly
self.empty_overlay.show() # Show the overlay when queue is empty
self.update_button_states()
@@ -290,56 +308,74 @@ class QueueManager(QDialog):
parent = self.parent
if parent is not None:
# lang_code: use parent's get_voice_formula and get_selected_lang
if hasattr(parent, 'get_voice_formula') and hasattr(parent, 'get_selected_lang'):
if hasattr(parent, "get_voice_formula") and hasattr(
parent, "get_selected_lang"
):
voice_formula = parent.get_voice_formula()
attrs['lang_code'] = parent.get_selected_lang(voice_formula)
attrs['voice'] = voice_formula
attrs["lang_code"] = parent.get_selected_lang(voice_formula)
attrs["voice"] = voice_formula
else:
attrs['lang_code'] = getattr(parent, 'selected_lang', '')
attrs['voice'] = getattr(parent, 'selected_voice', '')
attrs["lang_code"] = getattr(parent, "selected_lang", "")
attrs["voice"] = getattr(parent, "selected_voice", "")
# speed
if hasattr(parent, 'speed_slider'):
attrs['speed'] = parent.speed_slider.value() / 100.0
if hasattr(parent, "speed_slider"):
attrs["speed"] = parent.speed_slider.value() / 100.0
else:
attrs['speed'] = getattr(parent, 'speed', 1.0)
attrs["speed"] = getattr(parent, "speed", 1.0)
# save_option
attrs['save_option'] = getattr(parent, 'save_option', '')
attrs["save_option"] = getattr(parent, "save_option", "")
# output_folder
attrs['output_folder'] = getattr(parent, 'selected_output_folder', '')
attrs["output_folder"] = getattr(parent, "selected_output_folder", "")
# subtitle_mode
if hasattr(parent, 'get_actual_subtitle_mode'):
attrs['subtitle_mode'] = parent.get_actual_subtitle_mode()
if hasattr(parent, "get_actual_subtitle_mode"):
attrs["subtitle_mode"] = parent.get_actual_subtitle_mode()
else:
attrs['subtitle_mode'] = getattr(parent, 'subtitle_mode', '')
attrs["subtitle_mode"] = getattr(parent, "subtitle_mode", "")
# output_format
attrs['output_format'] = getattr(parent, 'selected_format', '')
attrs["output_format"] = getattr(parent, "selected_format", "")
# total_char_count
attrs['total_char_count'] = getattr(parent, 'char_count', '')
attrs["total_char_count"] = getattr(parent, "char_count", "")
# replace_single_newlines
attrs['replace_single_newlines'] = getattr(parent, 'replace_single_newlines', False)
attrs["replace_single_newlines"] = getattr(
parent, "replace_single_newlines", False
)
else:
# fallback: empty values
attrs = {k: '' for k in [
'lang_code', 'speed', 'voice', 'save_option',
'output_folder', 'subtitle_mode', 'output_format', 'total_char_count', 'replace_single_newlines']}
attrs = {
k: ""
for k in [
"lang_code",
"speed",
"voice",
"save_option",
"output_folder",
"subtitle_mode",
"output_format",
"total_char_count",
"replace_single_newlines",
]
}
return attrs
def add_files_from_paths(self, file_paths):
from abogen.utils import calculate_text_length
from PyQt5.QtWidgets import QMessageBox
import os
current_attrs = self.get_current_attributes()
duplicates = []
for file_path in file_paths:
class QueueItem:
pass
item = QueueItem()
item.file_name = file_path
for attr, value in current_attrs.items():
setattr(item, attr, value)
# Read file content and calculate total_char_count using calculate_text_length
try:
with open(file_path, 'r', encoding='utf-8', errors='ignore') as f:
with open(file_path, "r", encoding="utf-8", errors="ignore") as f:
file_content = f.read()
item.total_char_count = calculate_text_length(file_content)
except Exception:
@@ -348,16 +384,26 @@ class QueueManager(QDialog):
is_duplicate = False
for queued_item in self.queue:
if (
getattr(queued_item, 'file_name', None) == getattr(item, 'file_name', None) and
getattr(queued_item, 'lang_code', None) == getattr(item, 'lang_code', None) and
getattr(queued_item, 'speed', None) == getattr(item, 'speed', None) and
getattr(queued_item, 'voice', None) == getattr(item, 'voice', None) and
getattr(queued_item, 'save_option', None) == getattr(item, 'save_option', None) and
getattr(queued_item, 'output_folder', None) == getattr(item, 'output_folder', None) and
getattr(queued_item, 'subtitle_mode', None) == getattr(item, 'subtitle_mode', None) and
getattr(queued_item, 'output_format', None) == getattr(item, 'output_format', None) and
getattr(queued_item, 'total_char_count', None) == getattr(item, 'total_char_count', None) and
getattr(queued_item, 'replace_single_newlines', False) == getattr(item, 'replace_single_newlines', False)
getattr(queued_item, "file_name", None)
== getattr(item, "file_name", None)
and getattr(queued_item, "lang_code", None)
== getattr(item, "lang_code", None)
and getattr(queued_item, "speed", None)
== getattr(item, "speed", None)
and getattr(queued_item, "voice", None)
== getattr(item, "voice", None)
and getattr(queued_item, "save_option", None)
== getattr(item, "save_option", None)
and getattr(queued_item, "output_folder", None)
== getattr(item, "output_folder", None)
and getattr(queued_item, "subtitle_mode", None)
== getattr(item, "subtitle_mode", None)
and getattr(queued_item, "output_format", None)
== getattr(item, "output_format", None)
and getattr(queued_item, "total_char_count", None)
== getattr(item, "total_char_count", None)
and getattr(queued_item, "replace_single_newlines", False)
== getattr(item, "replace_single_newlines", False)
):
is_duplicate = True
break
@@ -369,7 +415,7 @@ class QueueManager(QDialog):
QMessageBox.warning(
self,
"Duplicate Item(s)",
f"Skipping {len(duplicates)} file(s) with the same attributes, already in the queue."
f"Skipping {len(duplicates)} file(s) with the same attributes, already in the queue.",
)
self.process_queue()
self.update_button_states()
@@ -377,20 +423,23 @@ class QueueManager(QDialog):
def add_more_files(self):
from PyQt5.QtWidgets import QFileDialog
from abogen.utils import calculate_text_length # import the function
# Only allow .txt files
files, _ = QFileDialog.getOpenFileNames(self, "Select .txt files", "", "Text Files (*.txt)")
files, _ = QFileDialog.getOpenFileNames(
self, "Select .txt files", "", "Text Files (*.txt)"
)
if not files:
return
self.add_files_from_paths(files)
def resizeEvent(self, event):
super().resizeEvent(event)
if hasattr(self, 'empty_overlay'):
if hasattr(self, "empty_overlay"):
self.empty_overlay.resize(self.listwidget.size())
def update_button_states(self):
# Enable Remove if at least one item is selected, else disable
if hasattr(self, 'remove_button'):
if hasattr(self, "remove_button"):
selected_count = len(self.listwidget.selectedItems())
self.remove_button.setEnabled(selected_count > 0)
if selected_count > 1:
@@ -398,7 +447,7 @@ class QueueManager(QDialog):
else:
self.remove_button.setText("Remove selected")
# Disable Clear if queue is empty
if hasattr(self, 'clear_button'):
if hasattr(self, "clear_button"):
self.clear_button.setEnabled(bool(self.queue))
def show_context_menu(self, pos):
@@ -406,6 +455,7 @@ class QueueManager(QDialog):
from PyQt5.QtGui import QDesktopServices
from PyQt5.QtCore import QUrl
import os
global_pos = self.listwidget.viewport().mapToGlobal(pos)
selected_items = self.listwidget.selectedItems()
menu = QMenu(self)
@@ -417,35 +467,45 @@ class QueueManager(QDialog):
# Add Open file action
open_file_action = QAction("Open file", self)
def open_file():
from PyQt5.QtWidgets import QMessageBox
item = selected_items[0]
display_name = item.text()
for q in self.queue:
if os.path.basename(q.file_name) == display_name:
if not os.path.exists(q.file_name):
QMessageBox.warning(self, "File Not Found", f"The file does not exist.")
QMessageBox.warning(
self, "File Not Found", f"The file does not exist."
)
return
QDesktopServices.openUrl(QUrl.fromLocalFile(q.file_name))
break
open_file_action.triggered.connect(open_file)
menu.addAction(open_file_action)
# Add Go to folder action
go_to_folder_action = QAction("Go to folder", self)
def go_to_folder():
from PyQt5.QtWidgets import QMessageBox
item = selected_items[0]
display_name = item.text()
for q in self.queue:
if os.path.basename(q.file_name) == display_name:
if not os.path.exists(q.file_name):
QMessageBox.warning(self, "File Not Found", f"The file does not exist.")
QMessageBox.warning(
self, "File Not Found", f"The file does not exist."
)
return
folder = os.path.dirname(q.file_name)
if os.path.exists(folder):
QDesktopServices.openUrl(QUrl.fromLocalFile(folder))
break
go_to_folder_action.triggered.connect(go_to_folder)
menu.addAction(go_to_folder_action)
@@ -466,6 +526,7 @@ class QueueManager(QDialog):
def reject(self):
# Cancel: restore original queue
from PyQt5.QtWidgets import QMessageBox
# Warn if user changed a lot (e.g., more than 1 items difference)
original_count = len(self._original_queue)
current_count = len(self.queue)
@@ -475,7 +536,7 @@ class QueueManager(QDialog):
"Confirm Cancel",
f"Are you sure you want to cancel and discard all changes?",
QMessageBox.Yes | QMessageBox.No,
QMessageBox.No
QMessageBox.No,
)
if reply != QMessageBox.Yes:
return
@@ -485,6 +546,7 @@ class QueueManager(QDialog):
def keyPressEvent(self, event):
from PyQt5.QtCore import Qt
if event.key() == Qt.Key_Delete:
self.remove_item()
else:
+1
View File
@@ -1,6 +1,7 @@
# represents a queued item - book, chapters, voice, etc.
from dataclasses import dataclass
@dataclass
class QueuedItem:
file_name: str
+19 -5
View File
@@ -105,15 +105,17 @@ def clean_text(text, *args, **kwargs):
default_encoding = sys.getfilesystemencoding()
def create_process(cmd, stdin=None, text=True, capture_output=False):
import logging
logger = logging.getLogger(__name__)
# Configure root logger to output to console if not already configured
root = logging.getLogger()
if not root.handlers:
handler = logging.StreamHandler(sys.stdout)
formatter = logging.Formatter('%(message)s')
formatter = logging.Formatter("%(message)s")
handler.setFormatter(formatter)
root.addHandler(handler)
root.setLevel(logging.INFO)
@@ -156,6 +158,7 @@ def create_process(cmd, stdin=None, text=True, capture_output=False):
# Stream output to console in real-time if not capturing
if proc.stdout and not capture_output:
def _stream_output(stream):
if text:
# For text mode, read character by character for real-time output
@@ -174,7 +177,9 @@ def create_process(cmd, stdin=None, text=True, capture_output=False):
break
try:
# Try to decode binary data for display
sys.stdout.write(chunk.decode(default_encoding, errors='replace'))
sys.stdout.write(
chunk.decode(default_encoding, errors="replace")
)
sys.stdout.flush()
except Exception:
pass
@@ -220,6 +225,7 @@ def get_gpu_acceleration(enabled):
try:
import torch
from torch.cuda import is_available
if not enabled:
return "CUDA GPU available but using CPU.", False
if is_available():
@@ -227,7 +233,11 @@ def get_gpu_acceleration(enabled):
# Gather more diagnostic info if CUDA is not available
try:
cuda_devices = torch.cuda.device_count()
cuda_error = torch.cuda.get_device_name(0) if cuda_devices > 0 else "No devices found"
cuda_error = (
torch.cuda.get_device_name(0)
if cuda_devices > 0
else "No devices found"
)
except Exception as e:
cuda_error = str(e)
return f"CUDA GPU is not available. Using CPU. ({cuda_error})", False
@@ -239,6 +249,7 @@ def prevent_sleep_start():
system = platform.system()
if system == "Windows":
import ctypes
ctypes.windll.kernel32.SetThreadExecutionState(
0x80000000 | 0x00000001 | 0x00000040
)
@@ -246,19 +257,22 @@ def prevent_sleep_start():
_sleep_procs["Darwin"] = create_process(["caffeinate"])
elif system == "Linux":
# use a sleep that never exits so inhibition stays active
_sleep_procs["Linux"] = create_process([
_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:
+15 -5
View File
@@ -230,13 +230,15 @@ class VoiceMixer(QWidget):
appstyle = QApplication.instance().style().objectName().lower()
if appstyle != "windowsvista":
# Set custom groove color for disabled state using COLORS["GREY_BACKGROUND"]
self.slider.setStyleSheet(f"""
self.slider.setStyleSheet(
f"""
QSlider::groove:vertical:disabled {{
background: {COLORS.get("GREY_BACKGROUND")};
width: 4px;
border-radius: 4px;
}}
""")
"""
)
# Connect controls
self.slider.valueChanged.connect(lambda val: self.spin_box.setValue(val / 100))
@@ -330,7 +332,9 @@ class VoiceFormulaDialog(QDialog):
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)
self._original_mixed_voice_state = getattr(
parent, "mixed_voice_state", None
)
profiles = load_profiles()
self._virtual_new_profile = False
if not profiles:
@@ -369,7 +373,9 @@ class VoiceFormulaDialog(QDialog):
self.profile_list = QListWidget()
self.profile_list.setSelectionMode(QListWidget.SingleSelection)
self.profile_list.setSelectionBehavior(QListWidget.SelectRows)
self.profile_list.setStyleSheet("QListWidget::item:selected { background: palette(highlight); color: palette(highlighted-text); }")
self.profile_list.setStyleSheet(
"QListWidget::item:selected { background: palette(highlight); color: palette(highlighted-text); }"
)
icon = QIcon(get_resource_path("abogen.assets", "profile.png"))
if self._virtual_new_profile:
item = QListWidgetItem(icon, "New profile")
@@ -976,6 +982,7 @@ class VoiceFormulaDialog(QDialog):
def _parse_rgba_to_qcolor(self, rgba_str):
from PyQt5.QtCore import Qt
from PyQt5.QtGui import QColor
"""Helper to convert 'rgba(R,G,B,A_float)' string to QColor."""
match = re.match(r"rgba\((\d+),\s*(\d+),\s*(\d+),\s*([\d.]+)\)", rgba_str)
if match:
@@ -1372,6 +1379,7 @@ class VoiceFormulaDialog(QDialog):
def update_profile_list_colors(self):
from PyQt5.QtCore import Qt
profiles = load_profiles()
for i in range(self.profile_list.count()):
item = self.profile_list.item(i)
@@ -1383,7 +1391,9 @@ class VoiceFormulaDialog(QDialog):
color = self._parse_rgba_to_qcolor(COLORS.get("YELLOW_BACKGROUND"))
item.setData(Qt.BackgroundRole, color)
else:
item.setData(Qt.BackgroundRole, self.profile_list.palette().base().color())
item.setData(
Qt.BackgroundRole, self.profile_list.palette().base().color()
)
weights = profiles.get(name, {}).get("voices", [])
total = 0
if isinstance(weights, list):
+1
View File
@@ -1,6 +1,7 @@
import re
from abogen.constants import VOICES_INTERNAL
# Calls parsing and loads the voice to gpu or cpu
def get_new_voice(pipeline, formula, use_gpu):
try: