Merge branch 'main' of https://github.com/mush42/piper-nvda

.pre-commit-config.yaml

@@ -0,0 +1,7 @@
+repos:
+  - repo: https://github.com/pre-commit/pre-commit-hooks
+    rev: v4.3.0
+    hooks:
+      - id: check-ast
+      - id: check-case-conflict
+      - id: check-yaml

addon/globalPlugins/piper_voices_globalPlugin/__init__.py

@@ -45,7 +45,7 @@ def __init__(self, *args, **kwargs):
             # Translators: label of a menu item
             _("Piper voice &manager..."),
             # Translators: Piper's voice manager menu item help
-            _("Open the voice manager to preview and download piper voices"),
+            _("Open the voice manager to preview, install or download piper voices"),
         )
         gui.mainFrame.sysTrayIcon.menu.Bind(wx.EVT_MENU, self.on_manager, self.itemHandle)
 

addon/globalPlugins/piper_voices_globalPlugin/voice_manager.py

@@ -14,12 +14,13 @@
 import threading
 
 import wx
+from wx.adv import CommandLinkButton
 import gui
 import nvwave
 import synthDriverHandler
 from logHandler import log
 
-from . import PiperTextToSpeechSystem
+from . import PiperTextToSpeechSystem, PIPER_VOICES_DIR
 from . import voice_download
 from . import helpers
 from .components import AsyncSnakDialog, ColumnDefn, ImmutableObjectListView, SimpleDialog, make_sized_static_box
@@ -35,7 +36,7 @@ def __init__(self, parent):
         super().__init__(parent, -1)
         self.__already_populated = threading.Event()
         # Add controls
-        # Translators: lable for a list of installed voices
+        # Translators: label for a list of installed voices
         voices_label = wx.StaticText(self, -1, _("Installed voices"))
         self.voices_list = ImmutableObjectListView(
             self,
@@ -53,12 +54,25 @@ def __init__(self, parent):
         )
         self.buttons_panel = SizedPanel(self, -1)
         self.buttons_panel.SetSizerType("horizontal")
-        # Translators: lable for a button for showing voice model card
+        # Translators: label for a button for showing voice model card
         self.model_card_button = wx.Button(self.buttons_panel, -1, _("&Voice model card..."))
-        # Translators: lable for a button for removing a voice
+        # Translators: label for a button for removing a voice
         self.remove_voice_button = wx.Button(self.buttons_panel, -1, _("&Remove voice..."))
+        add_voice_button = CommandLinkButton(
+            self,
+            -1,
+            # Translators: the main label of the install button
+            _("Install from local file"),
+            # Translators: the note for this button
+            _(
+                "Install a voice from a local archive.\n"
+                "The archive contains the voice model and configuration.\n"
+                "The archive should have a (.tar.gz) file extension."
+            )
+        )
         self.Bind(wx.EVT_BUTTON, self.on_model_card, self.model_card_button)
         self.Bind(wx.EVT_BUTTON, self.on_remove_voice, self.remove_voice_button)
+        self.Bind(wx.EVT_BUTTON, self._on_install_voice_from_tar, add_voice_button)
 
     def update_voices_list(self, set_focus=False, invalidate_synth_voices_cache=False):
         voices = list(PiperTextToSpeechSystem.load_piper_voices_from_nvda_config_dir())
@@ -162,6 +176,50 @@ def on_remove_voice(self, event):
                 )
                 self.update_voices_list(set_focus=True, invalidate_synth_voices_cache=True)
 
+    def _on_install_voice_from_tar(self, event):
+        openFileDialog = wx.FileDialog(
+            parent=gui.mainFrame,
+            # Translators: title for a dialog for opening a file
+            message=_("Choose voice archive file "),
+            defaultDir=wx.GetUserHome(),
+            wildcard="Tar archives *.tar.gz | *.tar.gz",
+            style=wx.FD_OPEN,
+        )
+        gui.runScriptModalDialog(
+            openFileDialog,
+            functools.partial(self._get_process_tar_archive, openFileDialog),
+        )
+
+    def _get_process_tar_archive(self, dialog, res):
+        if res != wx.ID_OK:
+            return
+        filepath = dialog.GetPath().strip()
+        if not filepath:
+            return
+        try:
+            voice_key = PiperTextToSpeechSystem.install_voice(
+                filepath, PIPER_VOICES_DIR
+            )
+        except:
+            log.error("Failed to install voice from archive", exc_info=True)
+            gui.messageBox(
+                # Translators: message telling the user that installing the voice has failed
+                _(
+                    "Failed to install voice from archive. See NVDA's log for more details."
+                ),
+                _("Voice installation failed"),
+                style=wx.ICON_ERROR,
+            )
+        else:
+            gui.messageBox(
+                # Translators: message telling the user that installing the voice is successful
+                _(
+                    "Voice {voice} has been installed successfully."
+                ).format(voice=voice_key),
+                _("Voice installed successfully"),
+                style=wx.ICON_INFORMATION,
+            )
+
 
 class OnlinePiperVoicesPanel(SizedPanel):
     def __init__(self, parent):

addon/synthDrivers/piper_neural_voices/tts_system.py

@@ -7,8 +7,10 @@
 import io
 import operator
 import os
+import re
 import string
 import sys
+import tarfile
 import typing
 import wave
 from abc import ABC, abstractmethod
@@ -223,6 +225,10 @@ async def speak_text(self, text):
 
 
 class PiperTextToSpeechSystem:
+
+    VOICE_NAME_REGEX = re.compile(        r"voice(-|_)(?P<language>[a-z]+[_]?([a-z]+)?)(-|_)(?P<name>[a-z]+)(-|_)(?P<quality>(high|medium|low|x-low))"
+    )
+
     def __init__(
         self, voices: Sequence[PiperVoice], speech_options: SpeechOptions = None
     ):
@@ -362,3 +368,37 @@ def load_voices_from_directory(
                 continue
             rv.append(voice)
         return rv
+
+    @classmethod
+    def install_voice(cls, voice_archive_path, dest_dir):
+        """Uniform handleing of voice tar archives."""
+        archive_path = Path(voice_archive_path)
+        voice_name = archive_path.name.rstrip("".join(archive_path.suffixes))
+        match = cls.VOICE_NAME_REGEX.match(voice_name)
+        if match is None:
+            raise ValueError(f"Invalid voice archive: `{archive_path}`")
+        info = match.groupdict()
+        language = info["language"]
+        name = info["name"]
+        quality = info["quality"]
+        voice_key = f"{language}-{name}-{quality}"
+        with tarfile.open(os.fspath(archive_path), "r:gz") as tar:
+            members = tar.getmembers()
+            try:
+                m_onnx_model = next(m for m in members if m.name.endswith(".onnx"))
+                m_model_config = next(
+                    m for m in members if m.name.endswith(".onnx.json")
+                )
+            except StopIteration:
+                raise ValueError(f"Invalid voice archive: `{archive_path}`")
+            dst = Path(dest_dir).joinpath(voice_key)
+            dst.mkdir(parents=True, exist_ok=True)
+            tar.extract(m_onnx_model, path=os.fspath(dst), set_attrs=False)
+            tar.extract(m_model_config, path=os.fspath(dst), set_attrs=False)
+            try:
+                m_model_card = next(m for m in members if m.name.endswith("MODEL_CARD"))
+            except StopIteration:
+                pass
+            else:
+                tar.extract(m_model_card, path=os.fspath(dst), set_attrs=False)
+            return voice_key

buildVars.py

@@ -41,7 +41,7 @@ def _(arg):
 	# Add-on update channel (default is None, denoting stable releases,
 	# and for development releases, use "dev".)
 	# Do not change unless you know what you are doing!
-	"addon_updateChannel": None,
+	"addon_updateChannel": "beta",
 	# Add-on license such as GPL 2
 	"addon_license": "GPL 2",
 	# URL for the license document the ad-on is licensed under