GameSentenceMiner 2.5.14__py3-none-any.whl → 2.6.1__py3-none-any.whl

GameSentenceMiner/ai/gemini.py

@@ -0,0 +1,143 @@
+import google.generativeai as genai
+
+from GameSentenceMiner.configuration import get_config, logger
+
+MODEL = "gemini-2.0-flash"  # or "gemini-pro-vision" if you need image support
+
+genai.configure(api_key=get_config().ai.api_key)
+model = genai.GenerativeModel(MODEL)
+
+def translate_with_context(lines, sentence, current_line_index, game_title=""):
+    """
+    Translates a line of dialogue with context from surrounding lines.
+
+    Args:
+        lines: A list of strings representing the dialogue lines.
+        sentence: Sentence to get translation for
+        current_line_index: The index of the line to translate.
+        game_title: Optional title of the game for added context.
+
+    Returns:
+        A string containing the translated sentence with context.
+    """
+
+    if not lines or current_line_index < 0 or current_line_index >= len(lines):
+        return "Invalid input."
+
+    context_lines = []
+
+    # Get the previous 10 lines (or fewer if at the beginning)
+    for i in range(max(0, current_line_index - 10), current_line_index):
+        context_lines.append(lines[i].text)
+
+    # Get the current line
+    current_line = lines[current_line_index]
+    context_lines.append(current_line.text)
+
+    #Get the next 10 lines (or fewer if at the end)
+    for i in range(current_line_index + 1, min(current_line_index + 11, len(lines))):
+        context_lines.append(lines[i].text)
+
+    ai_config = get_config().ai
+
+#this is ugly, but prettier in the output... so idk
+    if ai_config.use_canned_translation_prompt:
+        prompt_to_use = \
+f"""
+Translate the following Japanese dialogue from the game {game_title} into natural, context-aware English. Focus on preserving the tone, intent, and emotional nuance of the original text, paying close attention to the context provided by surrounding lines. The dialogue may include slang, idioms, implied meanings, or game-specific terminology that should be adapted naturally for English-speaking players. Ensure the translation feels immersive and aligns with the game's narrative style and character voices.
+Translate only the specified line below, providing a single result. Do not include additional text, explanations, or other lines unless explicitly requested. Allow expletives if more natural. Allow HTML tags for emphasis, italics, and other formatting as needed. Please also try to preserve existing HTML tags from the specified sentence if appropriate.
+
+Line to Translate:
+"""
+    elif ai_config.use_canned_context_prompt:
+        prompt_to_use = \
+f"""
+Provide a very brief summary of the scene in English based on the provided Japanese dialogue and context. Focus on the characters' actions and the immediate situation being described.
+
+Current Sentence:
+"""
+    else:
+        prompt_to_use = ai_config.custom_prompt
+
+
+    prompt = \
+f"""
+Dialogue Context:
+
+{chr(10).join(context_lines)}
+
+I am playing the game {game_title}. With that, and the above dialogue context in mind, answer the following prompt.
+
+{prompt_to_use}
+
+{sentence}
+"""
+
+    logger.debug(prompt)
+    try:
+        response = model.generate_content(prompt)
+        return response.text.strip()
+    except Exception as e:
+        return f"Translation failed: {e}"
+
+# Example Usage: Zero Escape: 999 examples
+
+# zero_escape_dialogue1 = [
+#     "扉は開いた…？",
+#     "まさか、こんな仕掛けが…",
+#     "一体、何が起こっているんだ？",
+#     "この数字の意味は…？",
+#     "落ち着いて、考えるんだ。",
+#     "でも、時間が…",
+#     "まだ、諦めるな！",
+#     "一体、誰が…？",
+#     "まさか、あの人が…？",
+#     "もう、ダメだ…",
+#     "まだ、希望はある！",
+#     "この部屋から、脱出するんだ！",
+#     "でも、どうやって…？",
+#     "何か、手がかりがあるはずだ！",
+#     "早く、見つけないと…",
+# ]
+#
+#
+# current_line_index = 3
+# translation = translate_with_context(zero_escape_dialogue1, current_line_index, "Zero Escape: 999")
+# print(f"Original: {zero_escape_dialogue1[current_line_index]}")
+# print(f"Translation: {translation}")
+#
+# # Example with fewer context lines at the beginning.
+# zero_escape_dialogue2 = [
+#     "このアミュレット…",
+#     "何かを感じる…",
+#     "この数字は…",
+#     "９…？",
+#     "まさか、これは…",
+#     "何かの手がかり…？",
+#     "急がないと…",
+#     "時間がない…",
+#     "早く、脱出を…",
+# ]
+#
+# current_line_index = 3
+# translation = translate_with_context(zero_escape_dialogue2, current_line_index, "Zero Escape: 999")
+# print(f"Original: {zero_escape_dialogue2[current_line_index]}")
+# print(f"Translation: {translation}")
+#
+# #example with fewer context lines at the end.
+# zero_escape_dialogue3 = [
+#     "この状況、理解できない。",
+#     "誰かが、私たちを閉じ込めたのか？",
+#     "なぜ、こんなことを…？",
+#     "このゲームの目的は…？",
+#     "一体、何が真実なんだ？",
+#     "信じられるのは、誰…？",
+#     "疑心暗鬼になるな。",
+#     "でも、どうすれば…？",
+#     "とにかく、進むしかない。"
+# ]
+#
+# current_line_index = 4
+# translation = translate_with_context(zero_escape_dialogue3, current_line_index, "Zero Escape: 999")
+# print(f"Original: {zero_escape_dialogue3[current_line_index]}")
+# print(f"Translation: {translation}")

GameSentenceMiner/anki.py

@@ -8,9 +8,10 @@ from datetime import datetime, timedelta
 from requests import post
 
 from GameSentenceMiner import obs, util, notification, ffmpeg
+from GameSentenceMiner.ai.gemini import translate_with_context
 from GameSentenceMiner.configuration import *
 from GameSentenceMiner.configuration import get_config
-from GameSentenceMiner.gametext import get_text_event
+from GameSentenceMiner.gametext import get_text_event, get_all_lines
 from GameSentenceMiner.model import AnkiCard
 from GameSentenceMiner.utility_gui import get_utility_window
 from GameSentenceMiner.obs import get_current_game
@@ -63,6 +64,17 @@ def update_anki_card(last_note: AnkiCard, note=None, audio_path='', video_path='
     if not get_config().screenshot.enabled:
         logger.info("Skipping Adding Screenshot to Anki, Screenshot is disabled in settings")
 
+    if note and 'fields' in note and get_config().ai.enabled:
+        sentence_field = note['fields'].get(get_config().anki.sentence_field, {})
+        sentence_to_translate = sentence_field if sentence_field else last_note.get_field(
+            get_config().anki.sentence_field)
+        translation = translate_with_context(get_all_lines(), sentence_to_translate,
+                                                                   game_line.index, get_current_game())
+        logger.info(translation)
+        note['fields']['SentenceMeaning'] = translation
+    else:
+        logger.error("Invalid note object. Cannot update SentenceMeaning.")
+
     if prev_screenshot_in_anki:
         note['fields'][get_config().anki.previous_image_field] = prev_screenshot_html
 

GameSentenceMiner/config_gui.py

@@ -1,5 +1,5 @@
 import tkinter as tk
-from tkinter import filedialog, messagebox, simpledialog
+from tkinter import filedialog, messagebox, simpledialog, scrolledtext
 
 import ttkbootstrap as ttk
 
@@ -77,6 +77,7 @@ class ConfigApp:
         self.create_hotkeys_tab()
         self.create_profiles_tab()
         self.create_advanced_tab()
+        self.create_ai_tab()
 
         ttk.Button(self.window, text="Save Settings", command=self.save_settings).pack(pady=20)
 
@@ -155,7 +156,8 @@ class ConfigApp:
                 custom_ffmpeg_settings=self.screenshot_custom_ffmpeg_settings.get(),
                 screenshot_hotkey_updates_anki=self.screenshot_hotkey_update_anki.get(),
                 seconds_after_line = self.seconds_after_line.get(),
-                use_beginning_of_line_as_screenshot=self.use_beginning_of_line_as_screenshot.get()
+                use_beginning_of_line_as_screenshot=self.use_beginning_of_line_as_screenshot.get(),
+                use_new_screenshot_logic=self.use_new_screenshot_logic.get()
             ),
             audio=Audio(
                 enabled=self.audio_enabled.get(),
@@ -199,6 +201,15 @@ class ConfigApp:
                 show_screenshot_buttons=self.show_screenshot_button.get(),
                 multi_line_line_break=self.multi_line_line_break.get(),
                 multi_line_sentence_storage_field=self.multi_line_sentence_storage_field.get(),
+            ),
+            ai=Ai(
+                enabled=self.ai_enabled.get(),
+                # provider=self.provider.get(),
+                anki_field=self.ai_anki_field.get(),
+                api_key=self.ai_api_key.get(),
+                use_canned_translation_prompt=self.use_canned_translation_prompt.get(),
+                use_canned_context_prompt=self.use_canned_context_prompt.get(),
+                custom_prompt=self.custom_prompt.get("1.0", tk.END)
             )
         )
 
@@ -262,6 +273,7 @@ class ConfigApp:
             self.create_hotkeys_tab()
             self.create_profiles_tab()
             self.create_advanced_tab()
+            self.create_ai_tab()
 
 
     def increment_row(self):
@@ -739,6 +751,11 @@ class ConfigApp:
         ttk.Checkbutton(screenshot_frame, variable=self.use_beginning_of_line_as_screenshot).grid(row=self.current_row, column=1, sticky='W')
         self.add_label_and_increment_row(screenshot_frame, "Enable to use the beginning of the line as the screenshot point. Adjust the above setting to fine-tine timing.", row=self.current_row, column=2)
 
+        ttk.Label(screenshot_frame, text="Use alternative screenshot logic:").grid(row=self.current_row, column=0, sticky='W')
+        self.use_new_screenshot_logic = tk.BooleanVar(value=self.settings.screenshot.use_new_screenshot_logic)
+        ttk.Checkbutton(screenshot_frame, variable=self.use_new_screenshot_logic).grid(row=self.current_row, column=1, sticky='W')
+        self.add_label_and_increment_row(screenshot_frame, "Enable to use the new screenshot logic. This will try to take the screenshot in the middle of the voiceline, or middle of the line if no audio/vad.", row=self.current_row, column=2)
+
     @new_tab
     def create_audio_tab(self):
         audio_frame = ttk.Frame(self.notebook)
@@ -953,6 +970,63 @@ class ConfigApp:
         self.multi_line_sentence_storage_field.grid(row=self.current_row, column=1)
         self.add_label_and_increment_row(advanced_frame, "Field in Anki for storing the multi-line sentence temporarily.", row=self.current_row, column=2)
 
+
+    @new_tab
+    def create_ai_tab(self):
+        ai_frame = ttk.Frame(self.notebook)
+        self.notebook.add(ai_frame, text='AI')
+
+        ttk.Label(ai_frame, text="Enabled:").grid(row=self.current_row, column=0, sticky='W')
+        self.ai_enabled = tk.BooleanVar(value=self.settings.ai.enabled)
+        ttk.Checkbutton(ai_frame, variable=self.ai_enabled).grid(row=self.current_row, column=1, sticky='W')
+        self.add_label_and_increment_row(ai_frame, "Enable or disable AI integration.", row=self.current_row, column=2)
+
+        ttk.Label(ai_frame, text="Anki Field:").grid(row=self.current_row, column=0, sticky='W')
+        self.ai_anki_field = ttk.Entry(ai_frame)
+        self.ai_anki_field.insert(0, self.settings.ai.anki_field)
+        self.ai_anki_field.grid(row=self.current_row, column=1)
+        self.add_label_and_increment_row(ai_frame, "Field in Anki for AI-generated content.", row=self.current_row,
+                                         column=2)
+
+        # ttk.Label(ai_frame, text="Provider:").grid(row=self.current_row, column=0, sticky='W')
+        # self.provider = ttk.Combobox(ai_frame,
+        #                              values=[AI_GEMINI])
+        # self.provider.set(self.settings.ai.provider)
+        # self.provider.grid(row=self.current_row, column=1)
+        # self.add_label_and_increment_row(ai_frame, "Select the AI provider. Currently only Gemini is supported.", row=self.current_row, column=2)
+
+        ttk.Label(ai_frame, text="API Key:").grid(row=self.current_row, column=0, sticky='W')
+        self.ai_api_key = ttk.Entry(ai_frame, show="*")  # Mask the API key for security
+        self.ai_api_key.insert(0, self.settings.ai.api_key)
+        self.ai_api_key.grid(row=self.current_row, column=1)
+        self.add_label_and_increment_row(ai_frame, "API key for the selected AI provider (Gemini only currently).", row=self.current_row,
+                                         column=2)
+
+        ttk.Label(ai_frame, text="Use Canned Translation Prompt:").grid(row=self.current_row, column=0, sticky='W')
+        self.use_canned_translation_prompt = tk.BooleanVar(value=self.settings.ai.use_canned_translation_prompt)
+        ttk.Checkbutton(ai_frame, variable=self.use_canned_translation_prompt).grid(row=self.current_row, column=1,
+                                                                                    sticky='W')
+        self.add_label_and_increment_row(ai_frame, "Use a pre-defined translation prompt for AI.", row=self.current_row,
+                                         column=2)
+
+        ttk.Label(ai_frame, text="Use Canned Context Prompt:").grid(row=self.current_row, column=0, sticky='W')
+        self.use_canned_context_prompt = tk.BooleanVar(value=self.settings.ai.use_canned_context_prompt)
+        ttk.Checkbutton(ai_frame, variable=self.use_canned_context_prompt).grid(row=self.current_row, column=1,
+                                                                                sticky='W')
+        self.add_label_and_increment_row(ai_frame, "Use a pre-defined context prompt for AI.", row=self.current_row,
+                                         column=2)
+
+        ttk.Label(ai_frame, text="Custom Prompt:").grid(row=self.current_row, column=0, sticky='W')
+
+        self.custom_prompt = scrolledtext.ScrolledText(ai_frame, width=50, height=5)  # Adjust height as needed
+        self.custom_prompt.insert(tk.END, self.settings.ai.custom_prompt)
+        self.custom_prompt.grid(row=self.current_row, column=1)
+
+        self.add_label_and_increment_row(ai_frame, "Custom prompt for AI processing.", row=self.current_row, column=2)
+
+        return ai_frame
+
+
     def on_profile_change(self, event):
         print("profile Changed!")
         self.save_settings(profile_change=True)

GameSentenceMiner/configuration.py

@@ -27,6 +27,8 @@ WHISPER_SMALL = 'small'
 WHISPER_MEDIUM = 'medium'
 WHSIPER_LARGE = 'large'
 
+AI_GEMINI = 'gemini'
+
 INFO = 'INFO'
 DEBUG = 'DEBUG'
 
@@ -107,6 +109,7 @@ class Screenshot:
     screenshot_hotkey_updates_anki: bool = False
     seconds_after_line: float = 1.0
     use_beginning_of_line_as_screenshot: bool = True
+    use_new_screenshot_logic: bool = False
 
 
 @dataclass_json
@@ -175,6 +178,16 @@ class Advanced:
     multi_line_line_break: str = '<br>'
     multi_line_sentence_storage_field: str = ''
 
+@dataclass_json
+@dataclass
+class Ai:
+    enabled: bool = False
+    anki_field: str = ''
+    provider: str = AI_GEMINI
+    api_key: str = ''
+    use_canned_translation_prompt: bool = True
+    use_canned_context_prompt: bool = False
+    custom_prompt: str = ''
 
 
 @dataclass_json
@@ -191,6 +204,7 @@ class ProfileConfig:
     hotkeys: Hotkeys = field(default_factory=Hotkeys)
     vad: VAD = field(default_factory=VAD)
     advanced: Advanced = field(default_factory=Advanced)
+    ai: Ai = field(default_factory=Ai)
 
 
     # This is just for legacy support

GameSentenceMiner/ffmpeg.py

@@ -17,13 +17,13 @@ def get_ffprobe_path():
 ffmpeg_base_command_list = [get_ffmpeg_path(), "-hide_banner", "-loglevel", "error", '-nostdin']
 
 
-def get_screenshot(video_file, time_from_end):
-    time_from_end_to_capture = -time_from_end if time_from_end else -1
+def get_screenshot(video_file, screenshot_timing):
+    screenshot_timing = screenshot_timing if screenshot_timing else 1
     output_image = make_unique_file_name(os.path.join(
         get_config().paths.screenshot_destination, f"{obs.get_current_game(sanitize=True)}.{get_config().screenshot.extension}"))
     # FFmpeg command to extract the last frame of the video
     ffmpeg_command = ffmpeg_base_command_list + [
-        "-sseof", f"{time_from_end_to_capture}",  # Seek to 1 second before the end of the video
+        "-ss", f"{screenshot_timing}",  # Seek to 1 second after the beginning
         "-i", f"{video_file}",
         "-vframes", "1"  # Extract only one frame
     ]
@@ -52,31 +52,51 @@ def get_screenshot_for_line(video_file, game_line):
     return get_screenshot(video_file, get_screenshot_time(video_file, game_line))
 
 
-
-def get_screenshot_time(video_path, game_line, default_beginning=False):
+def get_screenshot_time(video_path, game_line, default_beginning=False, vad_beginning=None, vad_end=None):
     if game_line:
         line_time = game_line.time
     else:
+        # Assuming initial_time is defined elsewhere if game_line is None
         line_time = initial_time
 
     file_length = get_video_duration(video_path)
     file_mod_time = get_file_modification_time(video_path)
 
+    # Calculate when the line occurred within the video file (seconds from start)
     time_delta = file_mod_time - line_time
-    total_seconds = file_length - time_delta.total_seconds()
-
-    time_from_end = file_length - total_seconds - get_config().screenshot.seconds_after_line
+    line_timestamp_in_video = file_length - time_delta.total_seconds()
+    screenshot_offset = get_config().screenshot.seconds_after_line
+
+    # Calculate screenshot time from the beginning by adding the offset
+    if vad_beginning and vad_end:
+        logger.debug("Using VAD to determine screenshot time")
+        screenshot_time_from_beginning = line_timestamp_in_video + vad_end - screenshot_offset
+    elif get_config().screenshot.use_new_screenshot_logic:
+        if game_line.next:
+            logger.debug("Finding time between lines for screenshot")
+            screenshot_time_from_beginning = line_timestamp_in_video + ((game_line.next.time - game_line.time).total_seconds() / 2)
+        else:
+            logger.debug("Using end of line for screenshot")
+            screenshot_time_from_beginning = file_length - screenshot_offset
+    else:
+        screenshot_time_from_beginning = line_timestamp_in_video + screenshot_offset
 
-    if time_from_end < 0 or time_from_end > file_length:
+    # Check if the calculated time is out of bounds
+    if screenshot_time_from_beginning < 0 or screenshot_time_from_beginning > file_length:
         logger.error(
-            "Calculated screenshot time is out of bounds for trimmed video")
+             f"Calculated screenshot time ({screenshot_time_from_beginning:.2f}s) is out of bounds for video (length {file_length:.2f}s)."
+        )
         if default_beginning:
-            logger.info("Defaulting to using the beginning of the Replay Buffer")
-            return file_length - 1.0
-        logger.info("Defaulting to using the end of the Replay Buffer")
-        return 0
+            logger.info("Defaulting to using the beginning of the video (1.0s)")
+            # Return time for the start of the video
+            return 1.0
+        logger.info(f"Defaulting to using the end of the video ({file_length:.2f}s)")
+        return file_length - screenshot_offset
+
+    logger.info("Screenshot time from beginning: " + str(screenshot_time_from_beginning))
 
-    return time_from_end
+    # Return the calculated time from the beginning
+    return screenshot_time_from_beginning
 
 
 def process_image(image_file):

GameSentenceMiner/gametext.py

@@ -6,6 +6,7 @@ from datetime import datetime
 
 import pyperclip
 import websockets
+from websockets import InvalidStatusCode
 
 from GameSentenceMiner import util
 from GameSentenceMiner.model import AnkiCard
@@ -30,6 +31,7 @@ class GameLine:
     time: datetime
     prev: 'GameLine'
     next: 'GameLine'
+    index: int = 0
 
     def get_previous_time(self):
         if self.prev:
@@ -44,6 +46,7 @@ class GameLine:
 @dataclass
 class GameText:
     values: list[GameLine]
+    game_line_index = 0
 
     def __init__(self):
         self.values = []
@@ -64,7 +67,8 @@ class GameText:
         return None
 
     def add_line(self, line_text):
-        new_line = GameLine(line_text, datetime.now(), self.values[-1] if self.values else None, None)
+        new_line = GameLine(line_text, datetime.now(), self.values[-1] if self.values else None, None, self.game_line_index)
+        self.game_line_index += 1
         if self.values:
             self.values[-1].next = new_line
         self.values.append(new_line)
@@ -106,25 +110,38 @@ class ClipboardMonitor(threading.Thread):
 
 async def listen_websocket():
     global current_line, current_line_time, line_history, reconnecting, websocket_connected
+    try_other = False
+    websocket_url = f'ws://{get_config().general.websocket_uri}'
     while True:
+        if try_other:
+            websocket_url = f'ws://{get_config().general.websocket_uri}/api/ws/text/origin'
         try:
-            async with websockets.connect(f'ws://{get_config().general.websocket_uri}', ping_interval=None) as websocket:
+            async with websockets.connect(websocket_url, ping_interval=None) as websocket:
+                logger.info("TextHooker Websocket Connected!")
                 if reconnecting:
                     logger.info(f"Texthooker WebSocket connected Successfully!" + " Disabling Clipboard Monitor." if get_config().general.use_clipboard else "")
                     reconnecting = False
                 websocket_connected = True
+                try_other = True
                 while True:
                     message = await websocket.recv()
-
+                    logger.debug(message)
                     try:
                         data = json.loads(message)
                         if "sentence" in data:
                             current_clipboard = data["sentence"]
-                    except json.JSONDecodeError:
+                    except json.JSONDecodeError or TypeError:
                         current_clipboard = message
                     if current_clipboard != current_line:
                         handle_new_text_event(current_clipboard)
-        except (websockets.ConnectionClosed, ConnectionError) as e:
+        except (websockets.ConnectionClosed, ConnectionError, InvalidStatusCode) as e:
+            if isinstance(e, InvalidStatusCode):
+                e: InvalidStatusCode
+                if e.status_code == 404:
+                    logger.info("Texthooker WebSocket connection failed. Attempting some fixes...")
+                    try_other = True
+
+                logger.error(f"Texthooker WebSocket connection failed. Please check if the Texthooker is running and the WebSocket URI is correct.")
             websocket_connected = False
             if not reconnecting:
                 logger.warning(f"Texthooker WebSocket connection lost, Defaulting to clipboard if enabled. Attempting to Reconnect...")
@@ -225,3 +242,11 @@ def get_mined_line(last_note: AnkiCard, lines):
 
 def get_time_of_line(line):
     return line_history.get_time(line)
+
+
+def get_all_lines():
+    return line_history.values
+
+
+def get_line_history():
+    return line_history

GameSentenceMiner/gsm.py

@@ -116,9 +116,6 @@ class VideoToAudioHandler(FileSystemEventHandler):
                     mined_line = get_mined_line(last_note, lines)
                     line_cutoff = get_utility_window().get_next_line_timing()
 
-                ss_timing = 0
-                if mined_line and line_cutoff or mined_line and get_config().screenshot.use_beginning_of_line_as_screenshot:
-                    ss_timing = ffmpeg.get_screenshot_time(video_path, mined_line)
                 if last_note:
                     logger.debug(last_note.to_json())
                 selected_lines = get_utility_window().get_selected_lines()
@@ -128,7 +125,7 @@ class VideoToAudioHandler(FileSystemEventHandler):
 
                 if get_config().anki.sentence_audio_field and get_config().audio.enabled:
                     logger.debug("Attempting to get audio from video")
-                    final_audio_output, should_update_audio, vad_trimmed_audio = VideoToAudioHandler.get_audio(
+                    final_audio_output, should_update_audio, vad_trimmed_audio, vad_beginning, vad_end = VideoToAudioHandler.get_audio(
                         start_line,
                         line_cutoff,
                         video_path)
@@ -136,10 +133,17 @@ class VideoToAudioHandler(FileSystemEventHandler):
                     final_audio_output = ""
                     should_update_audio = False
                     vad_trimmed_audio = ""
+                    vad_beginning = 0
+                    vad_end = 0
                     if not get_config().audio.enabled:
                         logger.info("Audio is disabled in config, skipping audio processing!")
                     elif not get_config().anki.sentence_audio_field:
                         logger.info("No SentenceAudio Field in config, skipping audio processing!")
+
+                ss_timing = 1
+                if mined_line and line_cutoff or mined_line and get_config().screenshot.use_beginning_of_line_as_screenshot:
+                    ss_timing = ffmpeg.get_screenshot_time(video_path, mined_line, vad_beginning, vad_end)
+
                 if get_config().anki.update_anki and last_note:
                     anki.update_anki_card(last_note, note, audio_path=final_audio_output, video_path=video_path,
                                           tango=tango,
@@ -170,10 +174,11 @@ class VideoToAudioHandler(FileSystemEventHandler):
         final_audio_output = make_unique_file_name(os.path.join(get_config().paths.audio_destination,
                                                                 f"{obs.get_current_game(sanitize=True)}.{get_config().audio.extension}"))
         should_update_audio = True
+        vad_beginning, vad_end = 0, 0
         if get_config().vad.do_vad_postprocessing:
-            should_update_audio = do_vad_processing(get_config().vad.selected_vad_model, trimmed_audio, vad_trimmed_audio)
+            should_update_audio, vad_beginning, vad_end = do_vad_processing(get_config().vad.selected_vad_model, trimmed_audio, vad_trimmed_audio)
             if not should_update_audio:
-                should_update_audio = do_vad_processing(get_config().vad.selected_vad_model, trimmed_audio,
+                should_update_audio, vad_beginning, vad_end = do_vad_processing(get_config().vad.selected_vad_model, trimmed_audio,
                                                         vad_trimmed_audio)
             if not should_update_audio and get_config().vad.add_audio_on_no_results:
                 logger.info("No voice activity detected, using full audio.")
@@ -184,7 +189,7 @@ class VideoToAudioHandler(FileSystemEventHandler):
                                                   get_config().audio.ffmpeg_reencode_options)
         elif os.path.exists(vad_trimmed_audio):
             shutil.move(vad_trimmed_audio, final_audio_output)
-        return final_audio_output, should_update_audio, vad_trimmed_audio
+        return final_audio_output, should_update_audio, vad_trimmed_audio, vad_beginning, vad_end
 
 
 def do_vad_processing(model, trimmed_audio, vad_trimmed_audio, second_pass=False):
@@ -226,10 +231,9 @@ def play_video_in_external(line, filepath):
 
     if start:
         if "vlc" in get_config().advanced.video_player_path:
-            command.append("--start-time")
+            command.extend(["--start-time", convert_to_vlc_seconds(start), '--one-instance'])
         else:
-            command.append("--start")
-        command.append(convert_to_vlc_seconds(start))
+            command.extend(["--start", convert_to_vlc_seconds(start)])
     command.append(os.path.normpath(filepath))
 
     logger.info(" ".join(command))

GameSentenceMiner/obs.py

@@ -118,7 +118,6 @@ def do_obs_call(request, from_dict = None, retry=10):
             return from_dict(response.datain)
         return None
     except Exception as e:
-        logger.error(e)
         if "socket is already closed" in str(e) or "object has no attribute" in str(e):
             if retry > 0:
                 time.sleep(1)

GameSentenceMiner/util.py

@@ -59,7 +59,7 @@ def timedelta_to_ffmpeg_friendly_format(td_obj):
 
 
 def get_file_modification_time(file_path):
-    mod_time_epoch = os.path.getmtime(file_path)
+    mod_time_epoch = os.path.getctime(file_path)
     mod_time = datetime.fromtimestamp(mod_time_epoch)
     return mod_time
 

GameSentenceMiner/vad/silero_trim.py

@@ -31,7 +31,7 @@ def process_audio_with_silero(input_audio, output_audio):
 
     if not voice_activity:
         logger.info("No voice activity detected in the audio.")
-        return False
+        return False, 0, 0
 
     # Trim based on the first and last speech detected
     start_time = voice_activity[0]['start'] if voice_activity else 0
@@ -40,4 +40,4 @@ def process_audio_with_silero(input_audio, output_audio):
     # Trim the audio using FFmpeg
     ffmpeg.trim_audio(input_audio, start_time + get_config().vad.beginning_offset, end_time + get_config().audio.end_offset, output_audio)
     logger.info(f"Trimmed audio saved to: {output_audio}")
-    return True
+    return True, start_time + get_config().vad.beginning_offset, end_time + get_config().audio.end_offset

GameSentenceMiner/vad/vosk_helper.py

@@ -127,7 +127,7 @@ def process_audio_with_vosk(input_audio, output_audio):
 
     if not voice_activity:
         logger.info("No voice activity detected in the audio.")
-        return False
+        return False, 0, 0
 
     # Trim based on the first and last speech detected
     start_time = voice_activity[0]['start'] if voice_activity else 0
@@ -142,7 +142,7 @@ def process_audio_with_vosk(input_audio, output_audio):
     # Trim the audio using FFmpeg
     ffmpeg.trim_audio(input_audio, start_time + get_config().vad.beginning_offset, end_time + get_config().audio.end_offset, output_audio)
     logger.info(f"Trimmed audio saved to: {output_audio}")
-    return True
+    return True, start_time + get_config().vad.beginning_offset, end_time + get_config().audio.end_offset
 
 
 def get_vosk_model():

GameSentenceMiner/vad/whisper_helper.py

@@ -74,7 +74,7 @@ def process_audio_with_whisper(input_audio, output_audio):
 
     if not voice_activity:
         logger.info("No voice activity detected in the audio.")
-        return False
+        return False, 0, 0
 
     # Trim based on the first and last speech detected
     start_time = voice_activity[0]['start']
@@ -89,7 +89,7 @@ def process_audio_with_whisper(input_audio, output_audio):
     # Trim the audio using FFmpeg
     ffmpeg.trim_audio(input_audio, start_time + get_config().vad.beginning_offset, end_time + get_config().audio.end_offset, output_audio)
     logger.info(f"Trimmed audio saved to: {output_audio}")
-    return True
+    return True, start_time + get_config().vad.beginning_offset, end_time + get_config().audio.end_offset
 
 
 # Load Whisper model initially

{gamesentenceminer-2.5.14.dist-info → gamesentenceminer-2.6.1.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: GameSentenceMiner
-Version: 2.5.14
+Version: 2.6.1
 Summary: A tool for mining sentences from games. Update: Multi-Line Mining! Fixed!
 Author-email: Beangate <bpwhelan95@gmail.com>
 License: MIT License
@@ -33,6 +33,7 @@ Requires-Dist: win10toast
 Requires-Dist: numpy
 Requires-Dist: pystray
 Requires-Dist: pywin32; sys_platform == "win32"
+Requires-Dist: google-generativeai
 Dynamic: license-file
 
 # Game Sentence Miner

gamesentenceminer-2.6.1.dist-info/RECORD

@@ -0,0 +1,31 @@
+GameSentenceMiner/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+GameSentenceMiner/anki.py,sha256=6cXCPJRp_en0ZopRCaH7vXJkfdRDFftsX15yLvCvNmI,14082
+GameSentenceMiner/config_gui.py,sha256=AifBgNaeWcBhuFoQz8QJ4srM1Ck8KjHmnOmI-aRcFHg,66965
+GameSentenceMiner/configuration.py,sha256=F-WOOjofV-GrBX2SCpArw5LMqpMGf_KCOzjvQeZZlhM,20287
+GameSentenceMiner/ffmpeg.py,sha256=iUj-1uLLJla6jjGDKuc1FbcXjMogYpoDhrQqCbQFcWA,13359
+GameSentenceMiner/gametext.py,sha256=C8cOMKgKF3AdcaNOKSEYkdg-DAlZ8yXzhiqXJKXIgeA,8365
+GameSentenceMiner/gsm.py,sha256=l4g2qmW-3Wgzd-TJ6YS8qZsN0dIiky20vNaAgnoJJnc,24716
+GameSentenceMiner/model.py,sha256=JdnkT4VoPOXmOpRgFdvERZ09c9wLN6tUJxdrKlGZcqo,5305
+GameSentenceMiner/notification.py,sha256=FY39ChSRK0Y8TQ6lBGsLnpZUFPtFpSy2tweeXVoV7kc,2809
+GameSentenceMiner/obs.py,sha256=xaySBuh-xnShqCPB3T8Lxr_PErNOGRMM4ISolNqmXDI,7618
+GameSentenceMiner/package.py,sha256=YlS6QRMuVlm6mdXx0rlXv9_3erTGS21jaP3PNNWfAH0,1250
+GameSentenceMiner/util.py,sha256=W49gqYlhbzZe-17zHMFcAjNeeteXrv_USHT7aBDKSqM,6825
+GameSentenceMiner/utility_gui.py,sha256=H4aOddlsrVR768RwbMzYScCziuOz1JeySUigNrPlaac,7692
+GameSentenceMiner/ai/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+GameSentenceMiner/ai/gemini.py,sha256=6kTQPuRH16D-1srhrWa5uPGIy-jBqNm9eRdKBSbpiXA,5423
+GameSentenceMiner/communication/__init__.py,sha256=_jGn9PJxtOAOPtJ2rI-Qu9hEHVZVpIvWlxKvqk91_zI,638
+GameSentenceMiner/communication/send.py,sha256=oOJdCS6-LNX90amkRn5FL2xqx6THGm56zHR2ntVIFTE,229
+GameSentenceMiner/communication/websocket.py,sha256=pTcUe_ZZRp9REdSU4qalhPmbT_1DKa7w18j6RfFLELA,3074
+GameSentenceMiner/downloader/Untitled_json.py,sha256=RUUl2bbbCpUDUUS0fP0tdvf5FngZ7ILdA_J5TFYAXUQ,15272
+GameSentenceMiner/downloader/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+GameSentenceMiner/downloader/download_tools.py,sha256=mI1u_FGBmBqDIpCH3jOv8DOoZ3obgP5pIf9o9SVfX2Q,8131
+GameSentenceMiner/vad/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+GameSentenceMiner/vad/silero_trim.py,sha256=ULf3zwS-JMsY82cKF7gZxREHw8L6lgpWF2U1YqgE9Oc,1681
+GameSentenceMiner/vad/vosk_helper.py,sha256=125X8C9NxFPlWWpoNsbOnEqKx8RCjXN109zNx_QXhyg,6070
+GameSentenceMiner/vad/whisper_helper.py,sha256=JJ-iltCh813XdjyEw0Wn5DaErf6PDqfH0Efu1Md8cIY,3543
+gamesentenceminer-2.6.1.dist-info/licenses/LICENSE,sha256=OXLcl0T2SZ8Pmy2_dmlvKuetivmyPd5m1q-Gyd-zaYY,35149
+gamesentenceminer-2.6.1.dist-info/METADATA,sha256=d9eL6-5Xs0DRGNJPFY3wol_7u3eInO1T5obTwdWhvxQ,5467
+gamesentenceminer-2.6.1.dist-info/WHEEL,sha256=CmyFI0kx5cdEMTLiONQRbGQwjIoR1aIYB7eCAQ4KPJ0,91
+gamesentenceminer-2.6.1.dist-info/entry_points.txt,sha256=2APEP25DbfjSxGeHtwBstMH8mulVhLkqF_b9bqzU6vQ,65
+gamesentenceminer-2.6.1.dist-info/top_level.txt,sha256=V1hUY6xVSyUEohb0uDoN4UIE6rUZ_JYx8yMyPGX4PgQ,18
+gamesentenceminer-2.6.1.dist-info/RECORD,,

gamesentenceminer-2.5.14.dist-info/RECORD

@@ -1,29 +0,0 @@
-GameSentenceMiner/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-GameSentenceMiner/anki.py,sha256=FHm76i7WYEs--szhzZjz0B6DW4JuDQopmjR3RjYBogs,13358
-GameSentenceMiner/config_gui.py,sha256=WVSN239lyVm5GwKzE1HXiCSCu2CNpkcDm8TGKlTJpMc,62093
-GameSentenceMiner/configuration.py,sha256=obBhz7aAFkhQAhIqrocv2jRTOqIJ0tFEI_saUCyMAMc,19924
-GameSentenceMiner/ffmpeg.py,sha256=fzOxrn-a4KqFdUY2oove164CTDOSsdPQtzqRW5f1P7c,12002
-GameSentenceMiner/gametext.py,sha256=LORVdE2WEo1CDI8gonc7qxrhbS4KFKXFQVKjhlkpLbc,7368
-GameSentenceMiner/gsm.py,sha256=ZdYVJKilDFEGdC3g5V5pjSr4G2JkNNDIt7BQX6yc28A,24456
-GameSentenceMiner/model.py,sha256=JdnkT4VoPOXmOpRgFdvERZ09c9wLN6tUJxdrKlGZcqo,5305
-GameSentenceMiner/notification.py,sha256=FY39ChSRK0Y8TQ6lBGsLnpZUFPtFpSy2tweeXVoV7kc,2809
-GameSentenceMiner/obs.py,sha256=MlxRToq5wALPI1XrD8rxEU-N8mWII91cNJWY7rUa5uI,7642
-GameSentenceMiner/package.py,sha256=YlS6QRMuVlm6mdXx0rlXv9_3erTGS21jaP3PNNWfAH0,1250
-GameSentenceMiner/util.py,sha256=Awhy57vX4NgQzygqKaGQn2EJ75T0uiXlhmINFOWlQkU,6825
-GameSentenceMiner/utility_gui.py,sha256=H4aOddlsrVR768RwbMzYScCziuOz1JeySUigNrPlaac,7692
-GameSentenceMiner/communication/__init__.py,sha256=_jGn9PJxtOAOPtJ2rI-Qu9hEHVZVpIvWlxKvqk91_zI,638
-GameSentenceMiner/communication/send.py,sha256=oOJdCS6-LNX90amkRn5FL2xqx6THGm56zHR2ntVIFTE,229
-GameSentenceMiner/communication/websocket.py,sha256=pTcUe_ZZRp9REdSU4qalhPmbT_1DKa7w18j6RfFLELA,3074
-GameSentenceMiner/downloader/Untitled_json.py,sha256=RUUl2bbbCpUDUUS0fP0tdvf5FngZ7ILdA_J5TFYAXUQ,15272
-GameSentenceMiner/downloader/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-GameSentenceMiner/downloader/download_tools.py,sha256=mI1u_FGBmBqDIpCH3jOv8DOoZ3obgP5pIf9o9SVfX2Q,8131
-GameSentenceMiner/vad/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-GameSentenceMiner/vad/silero_trim.py,sha256=-thDIZLuTLra3YBj7WR16Z6JeDgSpge2YuahprBvD8I,1585
-GameSentenceMiner/vad/vosk_helper.py,sha256=BI_mg_qyrjNbuEJjXSUDoV0FWEtQtEOAPmrrNixnZ_8,5974
-GameSentenceMiner/vad/whisper_helper.py,sha256=OF4J8TPPoKPJR1uFwrWAZ2Q7v0HJkVvNGmF8l1tACX0,3447
-gamesentenceminer-2.5.14.dist-info/licenses/LICENSE,sha256=OXLcl0T2SZ8Pmy2_dmlvKuetivmyPd5m1q-Gyd-zaYY,35149
-gamesentenceminer-2.5.14.dist-info/METADATA,sha256=sm1RIt4O0bGoTOY3gMeVxcj7qbm_whFLk_6Jgf_blVk,5433
-gamesentenceminer-2.5.14.dist-info/WHEEL,sha256=CmyFI0kx5cdEMTLiONQRbGQwjIoR1aIYB7eCAQ4KPJ0,91
-gamesentenceminer-2.5.14.dist-info/entry_points.txt,sha256=2APEP25DbfjSxGeHtwBstMH8mulVhLkqF_b9bqzU6vQ,65
-gamesentenceminer-2.5.14.dist-info/top_level.txt,sha256=V1hUY6xVSyUEohb0uDoN4UIE6rUZ_JYx8yMyPGX4PgQ,18
-gamesentenceminer-2.5.14.dist-info/RECORD,,