hueso/gajim-plugins
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Support multi models. Moved config stuff into config_dialog

Browse Source
This commit is contained in:
mesonium
2024-06-15 22:11:50 +02:00
committed by hueso
parent 31c2868599
commit f5db7f9809
4 changed files with 117 additions and 136 deletions
Download Patch File Download Diff File Expand all files Collapse all files

108
stt_voice_messages/configs.py
View File

@@ -1,108 +0,0 @@
# This file is part of Gajim.
#
# Gajim is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# Gajim is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with Gajim. If not, see <http://www.gnu.org/licenses/>.
import logging
import typing
from gajim.common.app import Any
from gajim.plugins.plugins_i18n import _
from .models import openai_whisper
from .stt_voice_messages import STTVoiceMessagesPlugin
log = logging.getLogger('gajim.p.stt_voice_messages_config')
SUPPORTED_MODELS: dict[str, dict[str, typing.Union[list[str], Any, str]]] = {
'model_openaiwhisper': {
'moduls': ['whisper'],
'class': openai_whisper.WhisperModel,
'name': 'OpenAI Whisper'
},
'model_ctranslate2': {
'moduls': ['ctranslate2'],
'class': None,
'name': _('CTranslate2')
},
'model_faster-whisper': {
'moduls': ['faster-whisper'],
'class': None,
'name:': _('Faster-Whisper')
},
'model_distill': {
'moduls': ['transformers', 'accelerate', 'datasets[audio]'],
'class': None,
'name': _('Distill')
}
}
class Configuration:
def __init__(self, plugin: STTVoiceMessagesPlugin):
self.plugin = plugin
self._available_models: dict[
str, dict[str, typing.Union[list[str], Any, str]]] = {}
self.check_available_moduls()
@property
def available_models(self) -> dict[str, dict[str, typing.Union[list[str], Any, str]]]:
return self._available_models
def on_setting(self, value: Any, data: Any) -> None:
if isinstance(value, str):
value.strip()
log.debug('plugin config before:\n %s', self.plugin.config.data)
self.plugin.config[data] = value
log.debug('plugin config after:\n %s', self.plugin.config.data)
def on_set_model(self, value: Any, data: Any) -> None:
if isinstance(value, str):
value.strip()
log.debug('plugin config before:\n %s', self.plugin.config.data)
self.plugin.config['model_class'] = self._available_models[value][
'class']
self.on_setting(value, data)
log.debug('plugin config after:\n %s', self.plugin.config.data)
@staticmethod
def is_module_available(module: str) -> bool:
try:
__import__(module)
return True
except ModuleNotFoundError:
log.debug('Could not find module %s', module)
return False
except ImportError as ex:
log.debug(str(ex))
return False
def check_available_moduls(self):
for model in SUPPORTED_MODELS:
available = True
for modul in SUPPORTED_MODELS[model]['moduls']:
if not self.is_module_available(modul):
available = False
continue
if available:
self._available_models[model] = SUPPORTED_MODELS[model]
if (self.plugin.config.data['model_class'] is None
and len(self._available_models) > 0):
model = list(self._available_models)[0]
self.on_set_model(model, 'model')
log.debug('Choose first available model!')
else:
log.debug('Available model already chosen!')
log.debug('models = %s', self._available_models)

127
stt_voice_messages/gtk/config_dialog.py
View File

@@ -16,6 +16,7 @@
from __future__ import annotations
import logging
import typing
from pathlib import Path
from typing import TYPE_CHECKING
@@ -23,6 +24,7 @@ import whisper
from gi.repository import Gtk
from gajim.common import app
from gajim.common.app import Any
from gajim.gtk.builder import get_builder
from gajim.gtk.const import Setting, SettingKind, SettingType
from gajim.gtk.settings import SettingsBox
@@ -30,7 +32,7 @@ from gajim.gtk.sidebar_switcher import SideBarSwitcher
from gajim.plugins.helpers import get_builder
from gajim.plugins.plugins_i18n import _
from ..configs import Configuration
from ..models import openai_whisper
if TYPE_CHECKING:
from ..stt_voice_messages import STTVoiceMessagesPlugin
@@ -38,6 +40,95 @@ if TYPE_CHECKING:
log = logging.getLogger('gajim.p.stt_voice_messages_config_dialog')
SUPPORTED_MODELS: dict[str, dict[str, typing.Union[list[str], Any, str]]] = {
'model_openaiwhisper': {
'moduls': ['whisper'],
'class': openai_whisper.WhisperModel,
'name': 'OpenAI Whisper'
},
'model_ctranslate2': {
'moduls': ['ctranslate2'],
'class': None,
'name': _('CTranslate2')
},
'model_faster-whisper': {
'moduls': ['faster-whisper'],
'class': None,
'name:': _('Faster-Whisper')
},
'model_distill': {
'moduls': ['transformers', 'accelerate', 'datasets[audio]'],
'class': None,
'name': _('Distill')
}
}
class Configuration:
def __init__(self, plugin: STTVoiceMessagesPlugin):
self._plugin = plugin
self._available_models: dict[
str, dict[str, typing.Union[list[str], Any, str]]] = {}
self.check_available_moduls()
@property
def plugin(self) -> STTVoiceMessagesPlugin:
return self._plugin
@property
def available_models(self) -> dict[str, dict[str, typing.Union[list[str], Any, str]]]:
return self._available_models
def on_setting(self, value: Any, data: Any) -> None:
if isinstance(value, str):
value.strip()
log.debug('plugin config before:\n %s', self.plugin.config.data)
self.plugin.config[data] = value
log.debug('plugin config after:\n %s', self.plugin.config.data)
def on_set_model(self, value: Any, data: Any) -> None:
if isinstance(value, str):
value.strip()
log.debug('plugin config before:\n %s', self.plugin.config.data)
self.plugin.config['model_class'] = self._available_models[value][
'class']
self.on_setting(value, data)
log.debug('plugin config after:\n %s', self.plugin.config.data)
@staticmethod
def is_module_available(module: str) -> bool:
try:
__import__(module)
return True
except ModuleNotFoundError:
log.debug('Could not find module %s', module)
return False
except ImportError as ex:
log.debug(str(ex))
return False
def check_available_moduls(self):
for model in SUPPORTED_MODELS:
available = True
for modul in SUPPORTED_MODELS[model]['moduls']:
if not self.is_module_available(modul):
available = False
continue
if available:
self._available_models[model] = SUPPORTED_MODELS[model]
if (self.plugin.config.data['model_class'] is None
and len(self._available_models) > 0):
model = list(self._available_models)[0]
self.on_set_model(model, 'model')
log.debug('Choose first available model!')
else:
log.debug('Available model already chosen!')
log.debug('models = %s', self._available_models)
class PreferenceBox(SettingsBox):
def __init__(self, settings: list[Setting]) -> None:
SettingsBox.__init__(self, None)
@@ -52,7 +143,7 @@ class PreferenceBox(SettingsBox):
class STTVoiceMessagesConfigDialog(Gtk.ApplicationWindow):
def __init__(self, plugin: STTVoiceMessagesPlugin, parent: Gtk.Window) -> None:
def __init__(self, config: Configuration, parent: Gtk.Window) -> None:
Gtk.ApplicationWindow.__init__(self)
self.set_application(app.app)
@@ -72,45 +163,47 @@ class STTVoiceMessagesConfigDialog(Gtk.ApplicationWindow):
self._ui.grid.attach(side_bar_switcher, 0, 0, 1, 1)
self.add(self._ui.grid)
self.config = config
self.plugin = self.config.plugin
prefs: list[tuple[str, type[PreferenceBox]]] = [
('stt_behaviour', self.STTBehaviour),
('models', self.Models),
('whisper_general', self.OpenAIWhisperGeneral),
]
self._add_prefs(prefs)
self.show_all()
class STTBehaviour(PreferenceBox):
def __init__(self, config: Configuration) -> None:
def __init__(self, config_dialog: STTVoiceMessagesConfigDialog) -> None:
settings = [
Setting(SettingKind.SWITCH,
_('Auto Transcribe'),
SettingType.VALUE,
value=config.plugin.config['auto_transcribe'],
value=config_dialog.plugin.config['auto_transcribe'],
data='auto_transcribe',
callback=config.on_setting)
callback=config_dialog.config.on_setting)
]
PreferenceBox.__init__(self, settings)
class Models(PreferenceBox):
def __init__(self, config: Configuration) -> None:
def __init__(self, config_dialog: STTVoiceMessagesConfigDialog) -> None:
models: list[tuple[str, str]] = []
for key, value in config.available_models.items():
assert value['name'] is str
for key, value in config_dialog.config.available_models.items():
models.append(
(key, value['name'])
(key, str(value['name']))
)
settings = [
Setting(SettingKind.COMBO,
_('Speech To Text Model'),
SettingType.VALUE,
value=config.plugin.config['model'],
value=config_dialog.plugin.config['model'],
data='model',
callback=config.on_set_model,
callback=config_dialog.config.on_set_model,
props={'combo_items': models},
desc=_('Choose Model to use')),
]
@@ -118,23 +211,23 @@ class STTVoiceMessagesConfigDialog(Gtk.ApplicationWindow):
PreferenceBox.__init__(self, settings)
class OpenAIWhisperGeneral(PreferenceBox):
def __init__(self, config: Configuration) -> None:
def __init__(self, config_dialog: STTVoiceMessagesConfigDialog) -> None:
settings = [
Setting(SettingKind.POPOVER,
_('Language Model Size'),
SettingType.VALUE,
value=config.plugin.config['whisperai_model_size'],
value=config_dialog.plugin.config['whisperai_model_size'],
data='whisperai_model_size',
callback=config.on_setting,
callback=config_dialog.config.on_setting,
props={'entries': whisper.available_models()}),
Setting(SettingKind.SWITCH,
_('Translate'),
SettingType.VALUE,
value=config.plugin.config['whisperai_translate'],
value=config_dialog.plugin.config['whisperai_translate'],
data='whisperai_translate',
callback=config.on_setting)
callback=config_dialog.config.on_setting)
]
PreferenceBox.__init__(self, settings)

8
stt_voice_messages/gtk/sttbox.py
View File

@@ -13,14 +13,15 @@
# You should have received a copy of the GNU General Public License
# along with Gajim. If not, see <http://www.gnu.org/licenses/>.
from gi.repository import Gtk
import logging
from .. import helper
from gi.repository import Gtk
from gajim.plugins.gajimplugin import GajimPluginConfig
from gajim.plugins.plugins_i18n import _
from .. import helper
log = logging.getLogger('gajim.p.stt_voice_messages_sttbox')
class STTBox(Gtk.Box):
@@ -52,9 +53,6 @@ class STTBox(Gtk.Box):
self.show_all()
#def update_config(self, config: GajimPluginConfig):
# self._model = config.data['class']()
def _on_transcribe_clicked(self, _button: Gtk.Button) -> None:
log.debug('config.data = %s', self._config.data)
model_class = self._config.data['model_class']

10
stt_voice_messages/stt_voice_messages.py
View File

@@ -15,17 +15,13 @@
from __future__ import annotations
import logging
from functools import partial
from pathlib import Path
from gi.repository import Gtk
from gajim.plugins import GajimPlugin
from gajim.plugins.plugins_i18n import _
from .gtk.config_dialog import *
from .gtk.sttbox import STTBox
from .gtk.config_dialog import STTVoiceMessagesConfigDialog
log = logging.getLogger('gajim.p.stt_voice_messages')
@@ -33,7 +29,9 @@ log = logging.getLogger('gajim.p.stt_voice_messages')
class STTVoiceMessagesPlugin(GajimPlugin):
def init(self) -> None:
self.description = _('Transcribes voice messages to text.')
self.config_dialog = partial(STTVoiceMessagesConfigDialog, self)
self._config = Configuration(self)
self._config.check_available_moduls()
self.config_dialog = partial(STTVoiceMessagesConfigDialog, self._config)
self.gui_extension_points = {
'preview_audio': (self._on_preview_audio_created, None),