GameSentenceMiner 2.10.5__py3-none-any.whl → 2.10.7__py3-none-any.whl

GameSentenceMiner/config_gui.py

@@ -271,9 +271,10 @@ class ConfigApp:
                 video_player_path=self.video_player_path.get(),
                 multi_line_line_break=self.multi_line_line_break.get(),
                 multi_line_sentence_storage_field=self.multi_line_sentence_storage_field.get(),
-                use_anki_note_creation_time=self.use_anki_note_creation_time.get(),
+                # use_anki_note_creation_time=self.use_anki_note_creation_time.get(),
                 ocr_websocket_port=int(self.ocr_websocket_port.get()),
                 texthooker_communication_websocket_port=int(self.texthooker_communication_websocket_port.get()),
+                plaintext_websocket_port=int(self.plaintext_websocket_export_port.get()),
             ),
             ai=Ai(
                 enabled=self.ai_enabled.get(),
@@ -1329,14 +1330,23 @@ class ConfigApp:
         self.texthooker_communication_websocket_port.grid(row=self.current_row, column=1, sticky='EW', pady=2)
         self.current_row += 1
 
-        HoverInfoLabelWidget(advanced_frame, text="Use Anki Creation Date for Audio Timing:",
-                             tooltip="Use the Anki note creation date for audio timing instead of the OBS replay time.",
+        HoverInfoLabelWidget(advanced_frame, text="Plaintext Websocket Export Port:",
+                             tooltip="Port for GSM Plaintext WebSocket Export communication. Does nothing right now, hardcoded to 55002",
                              row=self.current_row, column=0)
-        self.use_anki_note_creation_time = tk.BooleanVar(value=self.settings.advanced.use_anki_note_creation_time)
-        ttk.Checkbutton(advanced_frame, variable=self.use_anki_note_creation_time, bootstyle="round-toggle").grid(
-            row=self.current_row, column=1, sticky='W', pady=2)
+        self.plaintext_websocket_export_port = ttk.Entry(advanced_frame)
+        self.plaintext_websocket_export_port.insert(0, str(self.settings.advanced.plaintext_websocket_port))
+        self.plaintext_websocket_export_port.grid(row=self.current_row, column=1, sticky='EW', pady=2)
         self.current_row += 1
 
+
+        # HoverInfoLabelWidget(advanced_frame, text="Use Anki Creation Date for Audio Timing:",
+        #                      tooltip="Use the Anki note creation date for audio timing instead of the OBS replay time.",
+        #                      row=self.current_row, column=0)
+        # self.use_anki_note_creation_time = tk.BooleanVar(value=self.settings.advanced.use_anki_note_creation_time)
+        # ttk.Checkbutton(advanced_frame, variable=self.use_anki_note_creation_time, bootstyle="round-toggle").grid(
+        #     row=self.current_row, column=1, sticky='W', pady=2)
+        # self.current_row += 1
+
         HoverInfoLabelWidget(advanced_frame, text="Reset Line Hotkey:",
                              tooltip="Hotkey to reset the current line of dialogue.", row=self.current_row, column=0)
         self.reset_line_hotkey = ttk.Entry(advanced_frame)

GameSentenceMiner/gsm.py

@@ -198,6 +198,7 @@ class VideoToAudioHandler(FileSystemEventHandler):
         vad_result = vad_processor.trim_audio_with_vad(trimmed_audio, vad_trimmed_audio, game_line)
         if timing_only:
             return vad_result
+        vad_trimmed_audio = vad_result.output_audio
         if get_config().audio.ffmpeg_reencode_options_to_use and os.path.exists(vad_trimmed_audio):
             ffmpeg.reencode_file_with_user_config(vad_trimmed_audio, final_audio_output,
                                                   get_config().audio.ffmpeg_reencode_options_to_use)

GameSentenceMiner/obs.py

@@ -334,7 +334,7 @@ def get_screenshot(compression=-1):
         logger.error(f"Error getting screenshot: {e}")
         return None
 
-def get_screenshot_base64():
+def get_screenshot_base64(compression=0, width=None, height=None):
     try:
         # update_current_game()
         current_game = get_current_game()
@@ -346,10 +346,8 @@ def get_screenshot_base64():
         if not current_source_name:
             logger.error("No active source found in the current scene.")
             return None
-        response = client.get_source_screenshot(name=current_source_name, img_format='png', quality=0, width=None, height=None)
+        response = client.get_source_screenshot(name=current_source_name, img_format='png', quality=compression, width=width, height=height)
         if response and response.image_data:
-            with open('screenshot_response.txt', 'wb') as f:
-                f.write(str(response).encode())
             return response.image_data
         else:
             logger.error(f"Error getting base64 screenshot: {response}")

GameSentenceMiner/ocr/gsm_ocr_config.py

@@ -91,7 +91,7 @@ def get_window(title):
     import pygetwindow as gw
     all_windows = gw.getWindowsWithTitle(title)
     if not all_windows:
-        raise ValueError(f"No windows found with title '{title}'.")
+        return None
 
     filtered_windows = []
     for window in all_windows:
@@ -101,7 +101,7 @@ def get_window(title):
         filtered_windows.append(window)
 
     if not filtered_windows:
-        raise ValueError(f"No non-cmd.exe windows found with title '{title}'.")
+        return None
 
     ret = None
     for window in filtered_windows:

GameSentenceMiner/ocr/owocr_helper.py

@@ -182,11 +182,11 @@ all_cords = None
 rectangles = None
 last_ocr2_result = ""
 
-def do_second_ocr(ocr1_text, time, img, filtering):
+def do_second_ocr(ocr1_text, time, img, filtering, ignore_furigana_filter=False):
     global twopassocr, ocr2, last_ocr2_result
     try:
         orig_text, text = run.process_and_write_results(img, None, last_ocr2_result, filtering, None,
-                                                        engine=ocr2, furigana_filter_sensitivity=furigana_filter_sensitivity)
+                                                        engine=ocr2, furigana_filter_sensitivity=furigana_filter_sensitivity if not ignore_furigana_filter else 0)
         if fuzz.ratio(last_ocr2_result, orig_text) >= 90:
             logger.info("Seems like the same text from previous ocr2 result, not sending")
             return
@@ -313,7 +313,7 @@ def run_oneocr(ocr_config: OCRConfig, rectangles):
 
     run.init_config(False)
     try:
-        run.run(read_from="screencapture",
+        run.run(read_from="screencapture" if window else "",
                 read_from_secondary="clipboard" if ss_clipboard else None,
                 write_to="callback",
                 screen_capture_area=screen_area,
@@ -345,14 +345,14 @@ def add_ss_hotkey(ss_hotkey="ctrl+shift+g"):
     def capture():
         print("Taking screenshot...")
         img = cropper.run()
-        do_second_ocr("", datetime.now(), img, filtering)
+        do_second_ocr("", datetime.now(), img, filtering, ignore_furigana_filter=True)
     def capture_main_monitor():
         print("Taking screenshot of main monitor...")
         with mss.mss() as sct:
             main_monitor = sct.monitors[1] if len(sct.monitors) > 1 else sct.monitors[0]
             img = sct.grab(main_monitor)
             img_bytes = mss.tools.to_png(img.rgb, img.size)
-            do_second_ocr("", datetime.now(), img_bytes, filtering)
+            do_second_ocr("", datetime.now(), img_bytes, filtering, ignore_furigana_filter=True)
     hotkey_reg = None
     try:
         hotkey_reg = keyboard.add_hotkey(ss_hotkey, capture)
@@ -389,7 +389,7 @@ def set_force_stable_hotkey():
     print("Press Ctrl+Shift+F to toggle force stable mode.")
 
 if __name__ == "__main__":
-    global ocr1, ocr2, twopassocr, language, ss_clipboard, ss, ocr_config, furigana_filter_sensitivity, area_select_ocr_hotkey
+    global ocr1, ocr2, twopassocr, language, ss_clipboard, ss, ocr_config, furigana_filter_sensitivity, area_select_ocr_hotkey, window
     import sys
 
     import argparse
@@ -428,7 +428,8 @@ if __name__ == "__main__":
         if ocr_config.window:
             start_time = time.time()
             while time.time() - start_time < 30:
-                if get_window(ocr_config.window):
+                window = get_window(ocr_config.window)
+                if window or manual:
                     break
                 logger.info(f"Window: {ocr_config.window} Could not be found, retrying in 1 second...")
                 time.sleep(1)

GameSentenceMiner/owocr/owocr/run.py

@@ -406,7 +406,8 @@ class TextFiltering:
                         break
         else:
             for block in new_blocks:
-                if lang not in ["ja", "zh"] or self.classify(block)[0] == lang:
+                # This only filters out NON JA/ZH from text when lang is JA/ZH
+                if lang not in ["ja", "zh"] or self.classify(block)[0] in ['ja', 'zh']:
                     final_blocks.append(block)
 
         text = '\n'.join(final_blocks)

GameSentenceMiner/util/configuration.py

@@ -243,6 +243,7 @@ class VAD:
 @dataclass_json
 @dataclass
 class Advanced:
+    plaintext_websocket_port: int = -1
     audio_player_path: str = ''
     video_player_path: str = ''
     show_screenshot_buttons: bool = False
@@ -252,6 +253,11 @@ class Advanced:
     texthooker_communication_websocket_port: int = 55001
     use_anki_note_creation_time: bool = True
 
+    def __post_init__(self):
+        if self.plaintext_websocket_port == -1:
+            self.plaintext_websocket_port = self.texthooker_communication_websocket_port + 1
+
+
 @dataclass_json
 @dataclass
 class Ai:

GameSentenceMiner/util/downloader/oneocr_dl.py

@@ -1,4 +1,5 @@
 import os
+import time
 import zipfile
 import shutil
 from os.path import expanduser
@@ -61,12 +62,13 @@ class Downloader:
         Main function to attempt download and extraction.
         Tries official source first, then a fallback URL.
         """
-        # if checkdir(self.oneocr_dir):
-        #     print("Files already exist in cache.")
-        #     return True
+        if checkdir(self.oneocr_dir):
+            print("Files already exist in cache.")
+            return True
 
         try:
             print("Attempting to download from official source...")
+            # raise Exception("")
             self.downloadofficial()
             print("Download and extraction from official source successful.")
             return True

GameSentenceMiner/util/model.py

@@ -184,7 +184,7 @@ class VADResult:
         self.end = end
         self.model = model
         self.segments = segments if segments is not None else []
-        self.output_audio = None
+        self.output_audio = output_audio
 
     def __repr__(self):
         return f"VADResult(success={self.success}, start={self.start}, end={self.end}, model={self.model}, output_audio={self.output_audio})"