GameSentenceMiner 2.13.0__py3-none-any.whl → 2.13.1__py3-none-any.whl

GameSentenceMiner/owocr/owocr/ocr.py

@@ -11,7 +11,6 @@ import json
 import base64
 from urllib.parse import urlparse, parse_qs
 
-import jaconv
 import numpy as np
 import rapidfuzz.fuzz
 from PIL import Image
@@ -95,6 +94,7 @@ def empty_post_process(text):
 
 
 def post_process(text, keep_blank_lines=False):
+    import jaconv
     if keep_blank_lines:
         text = '\n'.join([''.join(i.split()) for i in text.splitlines()])
     else:
@@ -436,7 +436,7 @@ class GoogleLens:
         #             res += '\n'
         
         if return_coords:
-            x = (True, res, lines)
+            x = (True, res, response_dict)
         else:
             x = (True, res)
 
@@ -887,7 +887,28 @@ class OneOCR:
             except:
                 logger.warning('Error reading URL from config, OneOCR will not work!')
 
-    def __call__(self, img, furigana_filter_sensitivity=0, sentence_to_check=None, return_coords=False):
+    def get_regex(self, lang):
+        if lang == "ja":
+            self.regex = re.compile(r'[\u3041-\u3096\u30A1-\u30FA\u4E00-\u9FFF]')
+        elif lang == "zh":
+            self.regex = re.compile(r'[\u4E00-\u9FFF]')
+        elif lang == "ko":
+            self.regex = re.compile(r'[\uAC00-\uD7AF]')
+        elif lang == "ar":
+            self.regex = re.compile(r'[\u0600-\u06FF\u0750-\u077F\u08A0-\u08FF\uFB50-\uFDFF\uFE70-\uFEFF]')
+        elif lang == "ru":
+            self.regex = re.compile(r'[\u0400-\u04FF\u0500-\u052F\u2DE0-\u2DFF\uA640-\uA69F\u1C80-\u1C8F]')
+        elif lang == "el":
+            self.regex = re.compile(r'[\u0370-\u03FF\u1F00-\u1FFF]')
+        elif lang == "he":
+            self.regex = re.compile(r'[\u0590-\u05FF\uFB1D-\uFB4F]')
+        elif lang == "th":
+            self.regex = re.compile(r'[\u0E00-\u0E7F]')
+        else:
+            self.regex = re.compile(
+            r'[a-zA-Z\u00C0-\u00FF\u0100-\u017F\u0180-\u024F\u0250-\u02AF\u1D00-\u1D7F\u1D80-\u1DBF\u1E00-\u1EFF\u2C60-\u2C7F\uA720-\uA7FF\uAB30-\uAB6F]')
+
+    def __call__(self, img, furigana_filter_sensitivity=0, return_coords=False):
         lang = get_ocr_language()
         if lang != self.initial_lang:
             self.initial_lang = lang
@@ -911,6 +932,10 @@ class OneOCR:
                         json.dump(ocr_resp, f, indent=4, ensure_ascii=False)
                 # print(json.dumps(ocr_resp))
                 filtered_lines = [line for line in ocr_resp['lines'] if self.regex.search(line['text'])]
+                x_coords = [line['bounding_rect'][f'x{i}'] for line in filtered_lines for i in range(1, 5)]
+                y_coords = [line['bounding_rect'][f'y{i}'] for line in filtered_lines for i in range(1, 5)]
+                if x_coords and y_coords:
+                    crop_coords = (min(x_coords) - 5, min(y_coords) - 5, max(x_coords) + 5, max(y_coords) + 5)
                 # logger.info(filtered_lines)
                 res = ''
                 skipped = []
@@ -964,30 +989,6 @@ class OneOCR:
                     #         else:
                     #             continue
                     #     res += '\n'
-                elif sentence_to_check:
-                    lines_to_build_area = []
-                    widths = []
-                    heights = []
-                    for line in ocr_resp['lines']:
-                        print(line['text'])
-                        if sentence_to_check in line['text'] or line['text'] in sentence_to_check or rapidfuzz.fuzz.partial_ratio(sentence_to_check, line['text']) > 50:
-                            lines_to_build_area.append(line)
-                            res += line['text']
-                            for word in line['words']:
-                                widths.append(word['bounding_rect']['x2'] - word['bounding_rect']['x1'])
-                                heights.append(word['bounding_rect']['y3'] - word['bounding_rect']['y1'])
-
-                    x_coords = [line['bounding_rect'][f'x{i}'] for line in lines_to_build_area for i in
-                                range(1, 5)]
-                    y_coords = [line['bounding_rect'][f'y{i}'] for line in lines_to_build_area for i in
-                                range(1, 5)]
-                    if widths:
-                        avg_width = sum(widths) / len(widths)
-                    if heights:
-                        avg_height = sum(heights) / len(heights)
-                    if x_coords and y_coords:
-                        crop_coords = (
-                            min(x_coords) - 5, min(y_coords) - 5, max(x_coords) + 5, max(y_coords) + 5)
                 elif return_coords:
                     for line in filtered_lines:
                         for word in line['words']:
@@ -998,10 +999,6 @@ class OneOCR:
                             boxes.append(box)
                     res = ocr_resp['text']
                 else:
-                    x_coords = [line['bounding_rect'][f'x{i}'] for line in filtered_lines for i in range(1, 5)]
-                    y_coords = [line['bounding_rect'][f'y{i}'] for line in filtered_lines for i in range(1, 5)]
-                    if x_coords and y_coords:
-                        crop_coords = (min(x_coords) - 5, min(y_coords) - 5, max(x_coords) + 5, max(y_coords) + 5)
                     res = ocr_resp['text']
 
             except RuntimeError as e:
@@ -1019,7 +1016,7 @@ class OneOCR:
 
             res = res.json()['text']
         if return_coords:
-            x = (True, res, boxes)
+            x = (True, res, filtered_lines)
         else:
             x = (True, res, crop_coords)
         if is_path:

GameSentenceMiner/owocr/owocr/run.py

@@ -44,7 +44,6 @@ import queue
 from datetime import datetime
 from PIL import Image, ImageDraw, UnidentifiedImageError
 from loguru import logger
-from pynput import keyboard
 from desktop_notifier import DesktopNotifierSync
 import psutil
 
@@ -384,6 +383,7 @@ class TextFiltering:
                 block_filtered = self.latin_extended_regex.findall(block)
             else:
                 block_filtered = self.latin_extended_regex.findall(block)
+
             if block_filtered:
                 orig_text_filtered.append(''.join(block_filtered))
             else:
@@ -547,39 +547,6 @@ class ScreenshotThread(threading.Thread):
             else:
                 raise ValueError('Window capture is only currently supported on Windows and macOS')
 
-    def __del__(self):
-        if self.macos_window_tracker_instance:
-            self.macos_window_tracker_instance.join()
-        elif self.windows_window_tracker_instance:
-            self.windows_window_tracker_instance.join()
-
-    def setup_persistent_windows_window_tracker(self):
-        global window_open
-        window_open = False
-        def setup_tracker():
-            global window_open
-            self.window_handle, window_title = self.get_windows_window_handle(self.screen_capture_window)
-
-            if not self.window_handle:
-                # print(f"Window '{screen_capture_window}' not found.")
-                return
-
-            set_dpi_awareness()
-            window_open = True
-            self.windows_window_tracker_instance = threading.Thread(target=self.windows_window_tracker)
-            self.windows_window_tracker_instance.start()
-            logger.opt(ansi=True).info(f'Selected window: {window_title}')
-
-        while not terminated:
-            if not window_open:
-                try:
-                    setup_tracker()
-                except ValueError as e:
-                    logger.error(f"Error setting up persistent windows window tracker: {e}")
-                    break
-            time.sleep(5)
-
-
     def get_windows_window_handle(self, window_title):
         def callback(hwnd, window_title_part):
             window_title = win32gui.GetWindowText(hwnd)
@@ -602,7 +569,7 @@ class ScreenshotThread(threading.Thread):
 
     def windows_window_tracker(self):
         found = True
-        while not terminated or window_open:
+        while not terminated:
             found = win32gui.IsWindow(self.window_handle)
             if not found:
                 break
@@ -872,18 +839,9 @@ class OBSScreenshotThread(threading.Thread):
             image_queue.put((result, True))
 
     def connect_obs(self):
-        try:
-            import obsws_python as obs
-            self.obs_client = obs.ReqClient(
-                host=get_config().obs.host,
-                port=get_config().obs.port,
-                password=get_config().obs.password,
-                timeout=10
-            )
-            logger.info("Connected to OBS WebSocket.")
-        except Exception as e:
-            logger.error(f"Failed to connect to OBS: {e}")
-            self.obs_client = None
+        import GameSentenceMiner.obs as obs
+        obs.connect_to_obs_sync()
+
 
     def run(self):
         global last_image
@@ -895,9 +853,12 @@ class OBSScreenshotThread(threading.Thread):
         def init_config(source=None, scene=None):
             obs.update_current_game()
             self.current_source = source if source else obs.get_active_source()
-            self.current_source_name = self.current_source.get('sourceName') if isinstance(self.current_source, dict) else None
+            self.current_source_name = self.current_source.get("sourceName") or None
             self.current_scene = scene if scene else obs.get_current_game()
             self.ocr_config = get_scene_ocr_config()
+            if not self.ocr_config:
+                logger.error("No OCR config found for the current scene.")
+                return
             self.ocr_config.scale_to_custom_size(self.width, self.height)
 
         # Register a scene switch callback in obsws
@@ -925,22 +886,23 @@ class OBSScreenshotThread(threading.Thread):
                 continue
 
             if not self.ocr_config:
+                logger.info("No OCR config found for the current scene. Waiting for scene switch.")
                 time.sleep(1)
                 continue
+            
+            if not self.current_source_name:
+                obs.update_current_game()
+                self.current_source = obs.get_active_source()
+                self.current_source_name = self.current_source.get("sourceName") or None
 
             try:
-                response = self.obs_client.get_source_screenshot(
-                    name=self.current_source_name,
-                    img_format='png',
-                    quality=75,
-                    width=self.width,
-                    height=self.height,
-                )
+                if not self.current_source_name:
+                    logger.error("No active source found in the current scene.")
+                    time.sleep(1)
+                    continue
+                img = obs.get_screenshot_PIL(source_name=self.current_source_name, width=self.width, height=self.height, img_format='png', compression=80)
 
-                if response.image_data:
-                    image_data = base64.b64decode(response.image_data.split(",")[1])
-                    img = Image.open(io.BytesIO(image_data)).convert("RGBA")
-                    
+                if img is not None:
                     if not img.getbbox():
                         logger.info("OBS Not Capturing anything, sleeping.")
                         time.sleep(1)
@@ -1118,11 +1080,10 @@ def signal_handler(sig, frame):
 
 
 def on_window_closed(alive):
-    global terminated, window_open
+    global terminated
     if not (alive or terminated):
         logger.info('Window closed or error occurred, terminated!')
-        window_open = False
-        # terminated = True
+        terminated = True
 
 
 def on_screenshot_combo():
@@ -1464,8 +1425,12 @@ def run(read_from=None,
         read_from_readable.append(f'directory {read_from_path}')
 
     if len(key_combos) > 0:
-        key_combo_listener = keyboard.GlobalHotKeys(key_combos)
-        key_combo_listener.start()
+        try:
+            from pynput import keyboard
+            key_combo_listener = keyboard.GlobalHotKeys(key_combos)
+            key_combo_listener.start()
+        except ImportError:
+            pass
 
     if write_to in ('clipboard', 'websocket', 'callback'):
         write_to_readable = write_to
@@ -1513,6 +1478,7 @@ def run(read_from=None,
                 filter_img = True
                 notify = False
                 last_screenshot_time = time.time()
+                ocr_start_time = datetime.now()
 
         if img == 0:
             on_window_closed(False)

{gamesentenceminer-2.13.0.dist-info → gamesentenceminer-2.13.1.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: GameSentenceMiner
-Version: 2.13.0
+Version: 2.13.1
 Summary: A tool for mining sentences from games. Update: Overlay?
 Author-email: Beangate <bpwhelan95@gmail.com>
 License: MIT License

{gamesentenceminer-2.13.0.dist-info → gamesentenceminer-2.13.1.dist-info}/RECORD

@@ -15,6 +15,9 @@ GameSentenceMiner/assets/icon32.png,sha256=Kww0hU_qke9_22wBuO_Nq0Dv2SfnOLwMhCyGg
 GameSentenceMiner/assets/icon512.png,sha256=HxUj2GHjyQsk8NV433256UxU9phPhtjCY-YB_7W4sqs,192487
 GameSentenceMiner/assets/icon64.png,sha256=N8xgdZXvhqVQP9QUK3wX5iqxX9LxHljD7c-Bmgim6tM,9301
 GameSentenceMiner/assets/pickaxe.png,sha256=VfIGyXyIZdzEnVcc4PmG3wszPMO1W4KCT7Q_nFK6eSE,1403829
+GameSentenceMiner/locales/en_us.json,sha256=zv7qfr23G-os29yXfIE61_tYsTOwUjxF8ADH2YyJ0NA,24881
+GameSentenceMiner/locales/ja_jp.json,sha256=PgAkkN7qmbtuRj5a3jHd9R9FNztamhmhRZysY6TwETU,26283
+GameSentenceMiner/locales/zh_cn.json,sha256=h-YRSz2XN_TkymjkAH1MQ9KqGLE9_7Ru-N4yTqpm8_8,23615
 GameSentenceMiner/ocr/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 GameSentenceMiner/ocr/gsm_ocr_config.py,sha256=Ezj-0k6Wo-una91FvYhMp6KGkRhWYihXzLAoh_Wu2xY,5329
 GameSentenceMiner/ocr/ocrconfig.py,sha256=_tY8mjnzHMJrLS8E5pHqYXZjMuLoGKYgJwdhYgN-ny4,6466
@@ -25,8 +28,8 @@ GameSentenceMiner/owocr/owocr/__init__.py,sha256=87hfN5u_PbL_onLfMACbc0F5j4KyIK9
 GameSentenceMiner/owocr/owocr/__main__.py,sha256=XQaqZY99EKoCpU-gWQjNbTs7Kg17HvBVE7JY8LqIE0o,157
 GameSentenceMiner/owocr/owocr/config.py,sha256=qM7kISHdUhuygGXOxmgU6Ef2nwBShrZtdqu4InDCViE,8103
 GameSentenceMiner/owocr/owocr/lens_betterproto.py,sha256=oNoISsPilVVRBBPVDtb4-roJtAhp8ZAuFTci3TGXtMc,39141
-GameSentenceMiner/owocr/owocr/ocr.py,sha256=6ArGr0xd-Fhkw9uPn4MH3urxbLBwZ-UmxfwoKUUgxio,63459
-GameSentenceMiner/owocr/owocr/run.py,sha256=nkDpXICJCTKgJTS4MYRnaz-GYqAS-GskcSg1ZkGIRuE,67285
+GameSentenceMiner/owocr/owocr/ocr.py,sha256=jjm7TGIOaR-7JfoP_4B24uW7tLsxJL36s2kFndJy_SA,63025
+GameSentenceMiner/owocr/owocr/run.py,sha256=Qm4srtzqj6tZhgicMME4SyjP6NP_2IQRN-AOzVzE828,66011
 GameSentenceMiner/owocr/owocr/screen_coordinate_picker.py,sha256=Na6XStbQBtpQUSdbN3QhEswtKuU1JjReFk_K8t5ezQE,3395
 GameSentenceMiner/util/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 GameSentenceMiner/util/audio_offset_selector.py,sha256=8Stk3BP-XVIuzRv9nl9Eqd2D-1yD3JrgU-CamBywJmY,8542
@@ -64,9 +67,9 @@ GameSentenceMiner/web/templates/index.html,sha256=Gv3CJvNnhAzIVV_QxhNq4OD-pXDt1v
 GameSentenceMiner/web/templates/text_replacements.html,sha256=tV5c8mCaWSt_vKuUpbdbLAzXZ3ATZeDvQ9PnnAfqY0M,8598
 GameSentenceMiner/web/templates/utility.html,sha256=3flZinKNqUJ7pvrZk6xu__v67z44rXnaK7UTZ303R-8,16946
 GameSentenceMiner/wip/get_overlay_coords.py,sha256=pxTuOicSsMMmOLRQH0-3FPoQqsolbncvIMgX2q8ArHc,19787
-gamesentenceminer-2.13.0.dist-info/licenses/LICENSE,sha256=OXLcl0T2SZ8Pmy2_dmlvKuetivmyPd5m1q-Gyd-zaYY,35149
-gamesentenceminer-2.13.0.dist-info/METADATA,sha256=vstCSpe07K_nuP8m0Ur3Zn4Dbfgk8ttM2xxjsTtmVd4,1463
-gamesentenceminer-2.13.0.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-gamesentenceminer-2.13.0.dist-info/entry_points.txt,sha256=2APEP25DbfjSxGeHtwBstMH8mulVhLkqF_b9bqzU6vQ,65
-gamesentenceminer-2.13.0.dist-info/top_level.txt,sha256=V1hUY6xVSyUEohb0uDoN4UIE6rUZ_JYx8yMyPGX4PgQ,18
-gamesentenceminer-2.13.0.dist-info/RECORD,,
+gamesentenceminer-2.13.1.dist-info/licenses/LICENSE,sha256=OXLcl0T2SZ8Pmy2_dmlvKuetivmyPd5m1q-Gyd-zaYY,35149
+gamesentenceminer-2.13.1.dist-info/METADATA,sha256=plYR4MOiOcRPSsEHvWAJ6HNOWz6tFxHyfBdrqxUHCnQ,1463
+gamesentenceminer-2.13.1.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+gamesentenceminer-2.13.1.dist-info/entry_points.txt,sha256=2APEP25DbfjSxGeHtwBstMH8mulVhLkqF_b9bqzU6vQ,65
+gamesentenceminer-2.13.1.dist-info/top_level.txt,sha256=V1hUY6xVSyUEohb0uDoN4UIE6rUZ_JYx8yMyPGX4PgQ,18
+gamesentenceminer-2.13.1.dist-info/RECORD,,