deblan-mirror/buzz
1
0
Fork 0
mirror of https://github.com/chidiwilliams/buzz.git synced 2024-06-29 05:00:21 +02:00
Code Issues Packages Projects Releases Wiki Activity

Disable Whisper.cpp when DLL is not loaded (#275)

Browse source
This commit is contained in:
Chidi Williams 2022-12-30 23:11:20 +00:00 committed by GitHub
parent 6e89684ac8
commit cfcb4f6c28
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
6 changed files with 53 additions and 29 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
.github/workflows/ci.yml vendored
View file

@ -49,7 +49,7 @@ jobs:
path: |
~/Library/Caches/Buzz
~/.cache/whisper
~/AppData/Local/Buzz/Cache
~/AppData/Local/Buzz/Buzz/Cache
key: whisper-models-${{ runner.os }}
- uses: FedericoCarboni/setup-ffmpeg@v1

3
assets/circle-plus-icon.svg
View file

@ -1,5 +1,4 @@
<svg xmlns="http://www.w3.org/2000/svg"
viewBox="0 0 512 512"><!--! Font Awesome Pro 6.2.1 by @fontawesome - https://fontawesome.com License - https://fontawesome.com/license (Commercial License) Copyright 2022 Fonticons, Inc. -->
<path fill="#888"
d="M256 512c141.4 0 256-114.6 256-256S397.4 0 256 0S0 114.6 0 256S114.6 512 256 512zM232 344V280H168c-13.3 0-24-10.7-24-24s10.7-24 24-24h64V168c0-13.3 10.7-24 24-24s24 10.7 24 24v64h64c13.3 0 24 10.7 24 24s-10.7 24-24 24H280v64c0 13.3-10.7 24-24 24s-24-10.7-24-24z"/>
<path d="M256 512c141.4 0 256-114.6 256-256S397.4 0 256 0S0 114.6 0 256S114.6 512 256 512zM232 344V280H168c-13.3 0-24-10.7-24-24s10.7-24 24-24h64V168c0-13.3 10.7-24 24-24s24 10.7 24 24v64h64c13.3 0 24 10.7 24 24s-10.7 24-24 24H280v64c0 13.3-10.7 24-24 24s-24-10.7-24-24z"/>
</svg>
Side by side Swipe Overlay

Before

Width:  |  Height:  |  Size: 542 B

After

Width:  |  Height:  |  Size: 520 B

3
assets/record-icon.svg
View file

@ -1,5 +1,4 @@
<svg xmlns="http://www.w3.org/2000/svg"
viewBox="0 0 384 512"><!--! Font Awesome Pro 6.2.1 by @fontawesome - https://fontawesome.com License - https://fontawesome.com/license (Commercial License) Copyright 2022 Fonticons, Inc. -->
<path fill="#888"
d="M192 0C139 0 96 43 96 96V256c0 53 43 96 96 96s96-43 96-96V96c0-53-43-96-96-96zM64 216c0-13.3-10.7-24-24-24s-24 10.7-24 24v40c0 89.1 66.2 162.7 152 174.4V464H120c-13.3 0-24 10.7-24 24s10.7 24 24 24h72 72c13.3 0 24-10.7 24-24s-10.7-24-24-24H216V430.4c85.8-11.7 152-85.3 152-174.4V216c0-13.3-10.7-24-24-24s-24 10.7-24 24v40c0 70.7-57.3 128-128 128s-128-57.3-128-128V216z"/>
<path d="M192 0C139 0 96 43 96 96V256c0 53 43 96 96 96s96-43 96-96V96c0-53-43-96-96-96zM64 216c0-13.3-10.7-24-24-24s-24 10.7-24 24v40c0 89.1 66.2 162.7 152 174.4V464H120c-13.3 0-24 10.7-24 24s10.7 24 24 24h72 72c13.3 0 24-10.7 24-24s-10.7-24-24-24H216V430.4c85.8-11.7 152-85.3 152-174.4V216c0-13.3-10.7-24-24-24s-24 10.7-24 24v40c0 70.7-57.3 128-128 128s-128-57.3-128-128V216z"/>
</svg>
Side by side Swipe Overlay

Before

Width:  |  Height:  |  Size: 648 B

After

Width:  |  Height:  |  Size: 626 B

2
assets/trash-icon.svg
View file

@ -1,4 +1,4 @@
<svg xmlns="http://www.w3.org/2000/svg"
viewBox="0 0 448 512"><!--! Font Awesome Pro 6.2.1 by @fontawesome - https://fontawesome.com License - https://fontawesome.com/license (Commercial License) Copyright 2022 Fonticons, Inc. -->
<path fill="#888" d="M135.2 17.7L128 32H32C14.3 32 0 46.3 0 64S14.3 96 32 96H416c17.7 0 32-14.3 32-32s-14.3-32-32-32H320l-7.2-14.3C307.4 6.8 296.3 0 284.2 0H163.8c-12.1 0-23.2 6.8-28.6 17.7zM416 128H32L53.2 467c1.6 25.3 22.6 45 47.9 45H346.9c25.3 0 46.3-19.7 47.9-45L416 128z"/>
<path d="M135.2 17.7L128 32H32C14.3 32 0 46.3 0 64S14.3 96 32 96H416c17.7 0 32-14.3 32-32s-14.3-32-32-32H320l-7.2-14.3C307.4 6.8 296.3 0 284.2 0H163.8c-12.1 0-23.2 6.8-28.6 17.7zM416 128H32L53.2 467c1.6 25.3 22.6 45 47.9 45H346.9c25.3 0 46.3-19.7 47.9-45L416 128z"/>
</svg>
Side by side Swipe Overlay

Before

Width:  |  Height:  |  Size: 525 B

After

Width:  |  Height:  |  Size: 513 B

3
assets/up-down-and-down-left-from-center-icon.svg
View file

@ -1,5 +1,4 @@
<svg xmlns="http://www.w3.org/2000/svg"
viewBox="0 0 512 512"><!--! Font Awesome Pro 6.2.1 by @fontawesome - https://fontawesome.com License - https://fontawesome.com/license (Commercial License) Copyright 2022 Fonticons, Inc. -->
<path fill="#888"
d="M344 0H488c13.3 0 24 10.7 24 24V168c0 9.7-5.8 18.5-14.8 22.2s-19.3 1.7-26.2-5.2l-39-39-87 87c-9.4 9.4-24.6 9.4-33.9 0l-32-32c-9.4-9.4-9.4-24.6 0-33.9l87-87L327 41c-6.9-6.9-8.9-17.2-5.2-26.2S334.3 0 344 0zM184 496H40c-13.3 0-24-10.7-24-24V328c0-9.7 5.8-18.5 14.8-22.2s19.3-1.7 26.2 5.2l39 39 87-87c9.4-9.4 24.6-9.4 33.9 0l32 32c9.4 9.4 9.4 24.6 0 33.9l-87 87 39 39c6.9 6.9 8.9 17.2 5.2 26.2s-12.5 14.8-22.2 14.8z"/>
<path d="M344 0H488c13.3 0 24 10.7 24 24V168c0 9.7-5.8 18.5-14.8 22.2s-19.3 1.7-26.2-5.2l-39-39-87 87c-9.4 9.4-24.6 9.4-33.9 0l-32-32c-9.4-9.4-9.4-24.6 0-33.9l87-87L327 41c-6.9-6.9-8.9-17.2-5.2-26.2S334.3 0 344 0zM184 496H40c-13.3 0-24-10.7-24-24V328c0-9.7 5.8-18.5 14.8-22.2s19.3-1.7 26.2 5.2l39 39 87-87c9.4-9.4 24.6-9.4 33.9 0l32 32c9.4 9.4 9.4 24.6 0 33.9l-87 87 39 39c6.9 6.9 8.9 17.2 5.2 26.2s-12.5 14.8-22.2 14.8z"/>
</svg>
Side by side Swipe Overlay

Before

Width:  |  Height:  |  Size: 692 B

After

Width:  |  Height:  |  Size: 670 B

69
buzz/gui.py
View file

@ -16,14 +16,14 @@ from PyQt6.QtCore import (QDateTime, QObject, Qt, QThread,
QTimer, QUrl, pyqtSignal, QModelIndex, QSize, QPoint,
QUrlQuery, QMetaObject, QEvent)
from PyQt6.QtGui import (QAction, QCloseEvent, QDesktopServices, QIcon,
QKeySequence, QPixmap, QTextCursor, QValidator, QKeyEvent)
QKeySequence, QPixmap, QTextCursor, QValidator, QKeyEvent, QPainter, QColor)
from PyQt6.QtNetwork import QNetworkAccessManager, QNetworkReply, QNetworkRequest
from PyQt6.QtWidgets import (QApplication, QCheckBox, QComboBox, QDialog,
QDialogButtonBox, QFileDialog, QLabel, QLineEdit,
QMainWindow, QMessageBox, QPlainTextEdit,
QProgressDialog, QPushButton, QVBoxLayout, QHBoxLayout, QMenu,
QWidget, QGroupBox, QToolBar, QTableWidget, QMenuBar, QFormLayout, QTableWidgetItem,
QHeaderView, QAbstractItemView, QListWidget, QListWidgetItem)
QHeaderView, QAbstractItemView, QListWidget, QListWidgetItem, QToolButton)
from requests import get
from whisper import tokenizer
@ -34,7 +34,7 @@ from .transcriber import (SUPPORTED_OUTPUT_FORMATS, FileTranscriptionOptions, Ou
Task,
WhisperCppFileTranscriber, WhisperFileTranscriber,
get_default_output_file_path, segments_to_text, write_output, TranscriptionOptions,
FileTranscriberQueueWorker, FileTranscriptionTask, RecordingTranscriber)
FileTranscriberQueueWorker, FileTranscriptionTask, RecordingTranscriber, LOADED_WHISPER_DLL)
APP_NAME = 'Buzz'
@ -64,26 +64,28 @@ class AudioDevicesComboBox(QComboBox):
device_changed = pyqtSignal(int)
audio_devices: List[Tuple[int, str]]
def __init__(self, parent: Optional[QWidget] = None, *args) -> None:
super().__init__(parent, *args)
def __init__(self, parent: Optional[QWidget] = None) -> None:
super().__init__(parent)
self.audio_devices = self.get_audio_devices()
self.addItems(map(lambda device: device[1], self.audio_devices))
self.addItems([device[1] for device in self.audio_devices])
self.currentIndexChanged.connect(self.on_index_changed)
if self.get_default_device_id() != -1:
default_device_index = next(i for i, device in enumerate(
self.audio_devices) if device[0] == self.get_default_device_id())
self.setCurrentIndex(default_device_index)
default_device_id = self.get_default_device_id()
if default_device_id != -1:
for i, device in enumerate(self.audio_devices):
if device[0] == default_device_id:
self.setCurrentIndex(i)
def get_audio_devices(self) -> List[Tuple[int, str]]:
try:
devices: sounddevice.DeviceList = sounddevice.query_devices()
input_devices = filter(
lambda device: device.get('max_input_channels') > 0, devices)
return list(map(lambda device: (device.get('index'), device.get('name')), input_devices))
return [(device.get('index'), device.get('name'))
for device in devices if device.get('max_input_channels') > 0]
except UnicodeDecodeError:
QMessageBox.critical(
self, '',
'An error occured while loading your audio devices. Please check the application logs for more information.')
'An error occurred while loading your audio devices. Please check the application logs for more '
'information.')
return []
def on_index_changed(self, index: int):
@ -441,8 +443,9 @@ class RecordingTranscriberWidget(QWidget):
self.setWindowTitle('Live Recording')
self.transcription_options = TranscriptionOptions(model=TranscriptionModel(model_type=ModelType.WHISPER_CPP,
whisper_model_size=WhisperModelSize.TINY))
self.transcription_options = TranscriptionOptions(
model=TranscriptionModel(model_type=ModelType.WHISPER_CPP if LOADED_WHISPER_DLL else ModelType.WHISPER,
whisper_model_size=WhisperModelSize.TINY))
self.audio_devices_combo_box = AudioDevicesComboBox(self)
self.audio_devices_combo_box.device_changed.connect(
@ -742,19 +745,19 @@ class MainWindowToolbar(QToolBar):
def __init__(self, parent: Optional[QWidget]):
super().__init__(parent)
record_action = QAction(QIcon(RECORD_ICON_PATH), 'Record', self)
record_action = QAction(self.load_icon(RECORD_ICON_PATH), 'Record', self)
record_action.triggered.connect(self.on_record_action_triggered)
new_transcription_action = QAction(
QIcon(ADD_ICON_PATH), 'New Transcription', self)
self.load_icon(ADD_ICON_PATH), 'New Transcription', self)
self.new_transcription_action_triggered = new_transcription_action.triggered
self.open_transcript_action = QAction(QIcon(EXPAND_ICON_PATH),
self.open_transcript_action = QAction(self.load_icon(EXPAND_ICON_PATH),
'Open Transcript', self)
self.open_transcript_action_triggered = self.open_transcript_action.triggered
self.open_transcript_action.setDisabled(True)
self.clear_history_action = QAction(QIcon(TRASH_ICON_PATH), 'Clear History', self)
self.clear_history_action = QAction(self.load_icon(TRASH_ICON_PATH), 'Clear History', self)
self.clear_history_action_triggered = self.clear_history_action.triggered
self.clear_history_action.setDisabled(True)
@ -767,11 +770,30 @@ class MainWindowToolbar(QToolBar):
self.setIconSize(QSize(16, 16))
self.setContentsMargins(0, 2, 0, 2)
for action in self.actions():
widget = self.widgetForAction(action)
if isinstance(widget, QToolButton):
widget.setToolButtonStyle(Qt.ToolButtonStyle.ToolButtonIconOnly)
# Fix spacing issue on Mac
if platform.system() == 'Darwin':
self.widgetForAction(self.actions()[0]).setStyleSheet(
'QToolButton { margin-left: 9px; margin-right: 1px; }')
def load_icon(self, file_path: str):
is_dark_background = self.palette().window().color().black() > 127
return self.load_icon_with_color(file_path, '#888' if is_dark_background else '#444')
@staticmethod
def load_icon_with_color(file_path: str, color: str):
"""Adapted from https://stackoverflow.com/questions/15123544/change-the-color-of-an-svg-in-qt"""
pixmap = QPixmap(file_path)
painter = QPainter(pixmap)
painter.setCompositionMode(QPainter.CompositionMode.CompositionMode_SourceIn)
painter.fillRect(pixmap.rect(), QColor(color))
painter.end()
return QIcon(pixmap)
def on_record_action_triggered(self):
recording_transcriber_window = RecordingTranscriberWidget(
self, flags=Qt.WindowType.Window)
@ -1082,7 +1104,12 @@ class TranscriptionOptionsGroupBox(QGroupBox):
self.hugging_face_search_line_edit.model_selected.connect(self.on_hugging_face_model_changed)
self.model_type_combo_box = QComboBox(self)
self.model_type_combo_box.addItems([model_type.value for model_type in ModelType])
for model_type in ModelType:
# Hide Whisper.cpp option is whisper.dll did not load correctly.
# See: https://github.com/chidiwilliams/buzz/issues/274, https://github.com/chidiwilliams/buzz/issues/197
if model_type == ModelType.WHISPER_CPP and LOADED_WHISPER_DLL is False:
continue
self.model_type_combo_box.addItem(model_type.value)
self.model_type_combo_box.setCurrentText(default_transcription_options.model.model_type.value)
self.model_type_combo_box.currentTextChanged.connect(self.on_model_type_changed)