pygpt-net 2.4.49__py3-none-any.whl → 2.4.51__py3-none-any.whl

CHANGELOG.md

@@ -1,5 +1,15 @@
 # CHANGELOG
 
+## 2.4.51 (2025-01-17)
+
+- Added a "Continuous recording" mode under Audio Input in the Notepad tab, allowing for recording long voice notes and real-time auto-transcription. (beta)
+- A new option has been added in Settings -> Audio -> Continuous recording auto-transcribe interval.
+
+## 2.4.50 (2025-01-16)
+
+- Refactored audio input core.
+- Added audio input volume progress bar.
+
 ## 2.4.49 (2025-01-16)
 
 - Fix: stream render in Assistants mode.

README.md

@@ -2,7 +2,7 @@
 
 [![pygpt](https://snapcraft.io/pygpt/badge.svg)](https://snapcraft.io/pygpt)
 
-Release: **2.4.49** | build: **2025.01.16** | Python: **>=3.10, <3.13**
+Release: **2.4.51** | build: **2025.01.17** | Python: **>=3.10, <3.13**
 
 > Official website: https://pygpt.net | Documentation: https://pygpt.readthedocs.io
 > 
@@ -118,6 +118,7 @@ sudo snap connect pygpt:camera
 
 ```commandline
 sudo snap connect pygpt:audio-record :audio-record
+sudo snap connect pygpt:alsa
 ```
 
 **Connecting IPython in Docker in Snap version**:
@@ -3952,6 +3953,16 @@ may consume additional tokens that are not displayed in the main window.
 
 ## Recent changes:
 
+**2.4.51 (2025-01-17)**
+
+- Added a "Continuous recording" mode under Audio Input in the Notepad tab, allowing for recording long voice notes and real-time auto-transcription. (beta)
+- A new option has been added in Settings -> Audio -> Continuous recording auto-transcribe interval.
+
+**2.4.50 (2025-01-16)**
+
+- Refactored audio input core.
+- Added audio input volume progress bar.
+
 **2.4.49 (2025-01-16)**
 
 - Fix: stream render in Assistants mode.
@@ -3977,67 +3988,6 @@ may consume additional tokens that are not displayed in the main window.
 - Introduced a new mode in "Chat with Files": "Retrieve Only", which allows for retrieving raw documents from the index.
 - Fixed a bug related to tool calls in the Gemini provider when using Chat with Files mode.
 
-**2.4.45 (2024-12-16)**
-
-- Enhanced web data loaders UI.
-
-**2.4.44 (2024-12-16)**
-
-- Enhanced web data loaders.
-- Web loaders have been added to attachments, allowing external web content to be attached to context via the "+Web" button in the Attachments tab.
-- Improved handling of attachments in groups and added an attachment icon when a group contains attachments.
-
-**2.4.43 (2024-12-15)**
-
-- Fix: Bug on attachment upload.
-- Added: Attachments uploaded in groups are now available for all contexts in the group (beta).
-
-**2.4.42 (2024-12-15)**
-
-- Added Mailer plugin, which allows sending and retrieving emails from the server, and reading them. It currently supports only SMTP.
-- Added 'web_request' command to the Web Search plugin, enabling GET/POST/PUT and other connections to any address and API endpoint. It also supports sending POST data, files, headers, cookies, and more.
-- Improved audio output.
-- Enhanced visibility of the Video menu.
-- Other fixes.
-
-**2.4.41 (2024-12-14)**
-
-- Improved switching between columns on a split screen.
-- Added visual identification of the active column.
-
-**2.4.40 (2024-12-13)**
-
-- Enhanced Split Screen mode, now promoted from beta to stable.
-- Python Code Interpreter tool added to the Tabs.
-- HTML/JS Canvas tool added to the Tabs.
-- Added attachment icon to the context list if context has attachments.
-- Improved audio playback.
-- Improved web search.
-- Added a thumbnail image to web search results.
-- Added a new commands to web search: "extract_images" and "extract_links".
-- Added the option "Use raw content (without summarization)" to the web search plugin, which provides a more detailed result to the main model.
-- Extended the default maximum result characters to 50,000 in the web search plugin.
-
-**2.4.39 (2024-12-09)**
-
-- Added "Split Screen" mode (accessible via the switch in the bottom-right corner of the screen), which allows you to work in two windows simultaneously. It is currently experimental (beta). Future updates will include Code Interpreter and Canvas running in tabs.
-
-- Fixed: Language switch.
-
-**2.4.38 (2024-12-08)**
-
-- Added the ability to select a style for chat display between: Blocks, ChatGPT-like, and ChatGPT-like Wide. New option in the menu: Config -> Theme -> Style...
-- Added configuration options for audio input in Settings -> Audio -> Audio Input Device, Channels, and Sampling rate.
-
-**2.4.37 (2024-11-30)**
-
-- The `Query only` mode in `Uploaded` tab has been renamed to `RAG`.
-- New options have been added under `Settings -> Files and Attachments`: 
-  - `Use history in RAG query`: When enabled, the content of the entire conversation will be used when preparing a query if the mode is set to RAG or Summary.
-  - `RAG limit`: This option is applicable only if 'Use history in RAG query' is enabled. It specifies the limit on how many recent entries in the conversation will be used when generating a query for RAG. A value of 0 indicates no limit.
-- Cache: dynamic parts of the system prompt (from plugins) have been moved to the very end of the prompt stack to enable the use of prompt cache mechanisms in OpenAI.
-
-
 # Credits and links
 
 **Official website:** <https://pygpt.net>

pygpt_net/CHANGELOG.txt

@@ -1,3 +1,13 @@
+2.4.51 (2025-01-17)
+
+- Added a "Continuous recording" mode under Audio Input in the Notepad tab, allowing for recording long voice notes and real-time auto-transcription. (beta)
+- A new option has been added in Settings -> Audio -> Continuous recording auto-transcribe interval.
+
+2.4.50 (2025-01-16)
+
+- Refactored audio input core.
+- Added audio input volume progress bar.
+
 2.4.49 (2025-01-16)
 
 - Fix: stream render in Assistants mode.

pygpt_net/__init__.py

@@ -6,15 +6,15 @@
 # GitHub:  https://github.com/szczyglis-dev/py-gpt   #
 # MIT License                                        #
 # Created By  : Marcin Szczygliński                  #
-# Updated Date: 2025.01.16 01:00:00                  #
+# Updated Date: 2025.01.17 03:00:00                  #
 # ================================================== #
 
 __author__ = "Marcin Szczygliński"
 __copyright__ = "Copyright 2025, Marcin Szczygliński"
 __credits__ = ["Marcin Szczygliński"]
 __license__ = "MIT"
-__version__ = "2.4.49"
-__build__ = "2025.01.16"
+__version__ = "2.4.51"
+__build__ = "2025.01.17"
 __maintainer__ = "Marcin Szczygliński"
 __github__ = "https://github.com/szczyglis-dev/py-gpt"
 __report__ = "https://github.com/szczyglis-dev/py-gpt/issues"

pygpt_net/controller/access/voice.py

@@ -6,13 +6,11 @@
 # GitHub:  https://github.com/szczyglis-dev/py-gpt   #
 # MIT License                                        #
 # Created By  : Marcin Szczygliński                  #
-# Updated Date: 2024.12.14 08:00:00                  #
+# Updated Date: 2025.01.16 17:00:00                  #
 # ================================================== #
 
 from typing import Optional, List, Dict, Any
 
-import pyaudio
-import wave
 import os
 
 from PySide6.QtCore import QTimer, Slot
@@ -38,9 +36,6 @@ class Voice:
         """
         self.window = window
         self.is_recording = False
-        self.frames = []
-        self.p = None
-        self.stream = None
         self.timer = None
         self.input_file = "voice_control.wav"
         self.thread_started = False
@@ -218,15 +213,6 @@ class Voice:
 
     def start_recording(self):
         """Start recording"""
-        self.frames = []  # clear audio frames
-
-        def callback(in_data, frame_count, time_info, status):
-            self.frames.append(in_data)
-            if self.is_recording:
-                return (in_data, pyaudio.paContinue)
-            else:
-                return (in_data, pyaudio.paComplete)
-
         try:
             self.is_recording = True
             self.switch_btn_stop()
@@ -235,23 +221,26 @@ class Voice:
             if self.window.controller.audio.is_playing():
                 self.window.controller.audio.stop_output()
 
+            # set audio volume bar
+            self.window.core.audio.capture.set_bar(
+                self.window.ui.nodes['voice.control.btn'].bar
+            )
+
             # start timeout timer to prevent infinite recording
             if self.timer is None:
                 self.timer = QTimer()
                 self.timer.timeout.connect(self.stop_timeout)
                 self.timer.start(self.TIMEOUT_SECONDS * 1000)
 
-            self.p = pyaudio.PyAudio()
-            self.stream = self.p.open(format=pyaudio.paInt16,
-                                      channels=1,
-                                      rate=44100,
-                                      input=True,
-                                      frames_per_buffer=1024,
-                                      stream_callback=callback)
+            if not self.window.core.audio.capture.check_audio_input():
+                raise Exception("Audio input not working.")
+                # IMPORTANT!!!!
+                # Stop here if audio input not working!
+                # This prevents the app from freezing when audio input is not working!
 
+            self.window.core.audio.capture.start()  # start recording if audio is OK
             self.window.update_status(trans('audio.speak.now'))
             self.window.dispatch(AppEvent(AppEvent.VOICE_CONTROL_STARTED))  # app event
-            self.stream.start_stream()
         except Exception as e:
             self.is_recording = False
             self.window.core.debug.log(e)
@@ -270,35 +259,29 @@ class Voice:
 
         :param timeout: True if stopped due to timeout
         """
+        self.window.core.audio.capture.reset_audio_level()
         self.is_recording = False
         if self.timer:
             self.timer.stop()
             self.timer = None
         self.switch_btn_start()  # switch button to start
         path = os.path.join(self.window.core.config.path, self.input_file)
+        self.window.core.audio.capture.set_path(path)
 
-        if self.stream is not None:
-            self.stream.stop_stream()
-            self.stream.close()
-            self.p.terminate()
-
+        if self.window.core.audio.capture.has_source():
+            self.window.core.audio.capture.stop()  # stop recording
             # abort if timeout
             if timeout:
                 self.window.dispatch(AppEvent(AppEvent.VOICE_CONTROL_STOPPED))  # app event
                 self.window.update_status("Aborted.".format(self.TIMEOUT_SECONDS))
                 return
 
-            if self.frames:
-                if len(self.frames) < self.MIN_FRAMES:
+            if self.window.core.audio.capture.has_frames():
+                frames = self.window.core.audio.capture.get_frames()
+                if len(frames) < self.MIN_FRAMES:
                     self.window.update_status(trans("status.audio.too_short"))
                     self.window.dispatch(AppEvent(AppEvent.VOICE_CONTROL_STOPPED))  # app event
                     return
-                wf = wave.open(path, 'wb')
-                wf.setnchannels(1)
-                wf.setsampwidth(self.p.get_sample_size(pyaudio.paInt16))
-                wf.setframerate(44100)
-                wf.writeframes(b''.join(self.frames))
-                wf.close()
                 self.window.dispatch(AppEvent(AppEvent.VOICE_CONTROL_SENT))  # app event
                 self.handle_thread(True)  # handle transcription in simple mode
         else:

pygpt_net/controller/audio/__init__.py

@@ -6,7 +6,7 @@
 # GitHub:  https://github.com/szczyglis-dev/py-gpt   #
 # MIT License                                        #
 # Created By  : Marcin Szczygliński                  #
-# Updated Date: 2024.12.14 18:00:00                  #
+# Updated Date: 2025.01.17 02:00:00                  #
 # ================================================== #
 
 import os
@@ -29,6 +29,8 @@ class Audio:
     def setup(self):
         """Setup controller"""
         self.update()
+        if self.window.core.config.get("audio.input.continuous", False):
+            self.window.ui.plugin_addon['audio.input.btn'].continuous.setChecked(True)
 
     def toggle_input(
             self,
@@ -54,6 +56,18 @@ class Audio:
         else:
             self.enable_output()
 
+    def toggle_continuous(self, state: bool):
+        """
+        Toggle continuous audio input
+
+        :param state: True to enable, False to disable
+        """
+        if state:
+            self.window.core.config.set("audio.input.continuous", True)
+        else:
+            self.window.core.config.set("audio.input.continuous", False)
+        self.window.core.config.save()
+
     def enable_output(self):
         """Enable audio output"""
         self.toggle_output_icon(True)

pygpt_net/controller/lang/custom.py

@@ -6,7 +6,7 @@
 # GitHub:  https://github.com/szczyglis-dev/py-gpt   #
 # MIT License                                        #
 # Created By  : Marcin Szczygliński                  #
-# Updated Date: 2024.12.09 00:00:00                  #
+# Updated Date: 2025.01.17 02:00:00                  #
 # ================================================== #
 
 from PySide6.QtCore import Qt
@@ -46,6 +46,7 @@ class Custom:
 
         # checkboxes
         self.window.ui.plugin_addon['audio.input'].btn_toggle.setText(trans('audio.speak.btn'))
+        self.window.ui.plugin_addon['audio.input.btn'].continuous.setText(trans('audio.speak.btn.continuous'))
         self.window.ui.config['assistant']['tool.file_search'].box.setText(trans('assistant.tool.file_search'))
         self.window.ui.config['assistant']['tool.code_interpreter'].box.setText(
             trans('assistant.tool.code_interpreter')

pygpt_net/controller/ui/tabs.py

@@ -6,7 +6,7 @@
 # GitHub:  https://github.com/szczyglis-dev/py-gpt   #
 # MIT License                                        #
 # Created By  : Marcin Szczygliński                  #
-# Updated Date: 2024.12.14 07:00:00                  #
+# Updated Date: 2025.01.17 02:00:00                  #
 # ================================================== #
 
 from typing import Any, Optional
@@ -182,6 +182,12 @@ class Tabs:
         if prev_tab != idx or prev_column != column_idx:
             self.window.dispatch(AppEvent(AppEvent.TAB_SELECTED))  # app event
 
+        # show/hide audio record footer
+        if tab.type == Tab.TAB_NOTEPAD:
+            self.window.ui.plugin_addon['audio.input.btn'].notepad_footer.setVisible(True)
+        else:
+            self.window.ui.plugin_addon['audio.input.btn'].notepad_footer.setVisible(False)
+
         self.window.controller.ui.update()
         self.update_current()
 

pygpt_net/core/audio/__init__.py

@@ -6,17 +6,19 @@
 # GitHub:  https://github.com/szczyglis-dev/py-gpt   #
 # MIT License                                        #
 # Created By  : Marcin Szczygliński                  #
-# Updated Date: 2024.12.14 18:00:00                  #
+# Updated Date: 2025.01.16 17:00:00                  #
 # ================================================== #
 
 import re
 from typing import Union, Optional, Tuple, List
 
+from PySide6.QtMultimedia import QMediaDevices
 from bs4 import UnicodeDammit
 
 from pygpt_net.provider.audio_input.base import BaseProvider as InputBaseProvider
 from pygpt_net.provider.audio_output.base import BaseProvider as OutputBaseProvider
 
+from .capture import Capture
 from .whisper import Whisper
 
 
@@ -28,6 +30,7 @@ class Audio:
         :param window: Window instance
         """
         self.window = window
+        self.capture = Capture(window)
         self.whisper = Whisper(window)
         self.providers = {
             "input": {},
@@ -41,21 +44,12 @@ class Audio:
 
         :return devices list: [(id, name)]
         """
-        import pyaudio
-        devices = []
-        try:
-            p = pyaudio.PyAudio()
-            num_devices = p.get_device_count()
-            for i in range(num_devices):
-                info = p.get_device_info_by_index(i)
-                if info["maxInputChannels"] > 0:
-                    dammit = UnicodeDammit(info["name"])
-                    devices.append((i, dammit.unicode_markup))
-                    # print(f"Device ID {i}: {info['name']}")
-            p.terminate()
-        except Exception as e:
-            print(f"Audio input devices receive error: {e}")
-        return devices
+        devices = QMediaDevices.audioInputs()
+        devices_list = []
+        for index, device in enumerate(devices):
+            dammit = UnicodeDammit(device.description())
+            devices_list.append((index, dammit.unicode_markup))
+        return devices_list
 
     def is_device_compatible(self, device_index: int) -> bool:
         """
@@ -69,7 +63,6 @@ class Audio:
         channels = int(self.window.core.config.get('audio.input.channels', 1))
         p = pyaudio.PyAudio()
         info = p.get_device_info_by_index(device_index)
-        supported = False
         try:
             p.is_format_supported(
                 rate=rate,