GameSentenceMiner 2.8.16__tar.gz → 2.8.18__tar.gz

{gamesentenceminer-2.8.16 → gamesentenceminer-2.8.18}/GameSentenceMiner/ocr/owocr_area_selector.py

@@ -713,6 +713,10 @@ class ScreenSelector:
         canvas.bind('<B1-Motion>', on_drag)  # Left drag
         canvas.bind('<ButtonRelease-1>', on_release)  # Left click release
         canvas.bind('<Button-3>', on_right_click)  # Right click delete
+        canvas.bind('<Control-s>', self.save_rects)  # Save
+        canvas.bind('<Control-z>', self.undo_last_rect)  # Undo
+        canvas.bind('<Control-y>', self.redo_last_rect)  # Redo
+
 
         # --- Bind Global Actions to the window (apply to all windows) ---
         # Use lambdas to ensure the correct toggle function is called if needed,

{gamesentenceminer-2.8.16 → gamesentenceminer-2.8.18}/GameSentenceMiner/ocr/owocr_helper.py

@@ -209,7 +209,7 @@ def do_second_ocr(ocr1_text, rectangle_index, time, img):
     global twopassocr, ocr2, last_ocr1_results, last_ocr2_results
     last_result = ([], -1)
     try:
-        orig_text, text = run.process_and_write_results(img, None, None, last_result, TextFiltering(),
+        orig_text, text = run.process_and_write_results(img, None, None, None, None,
                                                         engine=ocr2)
         previous_ocr2_text = last_ocr2_results[rectangle_index]
         if fuzz.ratio(previous_ocr2_text, text) >= 80:

{gamesentenceminer-2.8.16 → gamesentenceminer-2.8.18}/GameSentenceMiner/owocr/owocr/__main__.py

@@ -1,4 +1,4 @@
-from .run import run, init_config
+from GameSentenceMiner.owocr.owocr.run import run, init_config
 
 
 def main():

{gamesentenceminer-2.8.16 → gamesentenceminer-2.8.18}/GameSentenceMiner/owocr/owocr/ocr.py

@@ -1048,3 +1048,79 @@ class GeminiOCR:
 
     def _preprocess(self, img):
         return pil_image_to_bytes(img, png_compression=1)
+
+
+class GroqOCR:
+    name = 'groq'
+    readable_name = 'Groq OCR'
+    key = 'j'
+    available = False
+
+    def __init__(self, config={'api_key': None}):
+        try:
+            import groq
+            self.api_key = config['api_key']
+            if not self.api_key:
+                logger.warning('Groq API key not provided, GroqOCR will not work!')
+            else:
+                self.client = groq.Groq(api_key=self.api_key)
+                self.available = True
+                logger.info('Groq OCR ready')
+        except ImportError:
+            logger.warning('groq module not available, GroqOCR will not work!')
+        except Exception as e:
+            logger.error(f'Error initializing Groq client: {e}')
+
+    def __call__(self, img_or_path):
+        if not self.available:
+            return (False, 'GroqOCR is not available due to missing API key or configuration error.')
+
+        try:
+            if isinstance(img_or_path, str) or isinstance(img_or_path, Path):
+                img = Image.open(img_or_path).convert("RGB")
+            elif isinstance(img_or_path, Image.Image):
+                img = img_or_path.convert("RGB")
+            else:
+                raise ValueError(f'img_or_path must be a path or PIL.Image, instead got: {img_or_path}')
+
+            img_base64 = self._preprocess(img)
+            if not img_base64:
+                return (False, 'Error processing image for Groq.')
+
+            prompt = (
+                "Analyze this image and extract text from it"
+                # "(speech bubbles or panels containing character dialogue). From the extracted dialogue text, "
+                # "filter out any furigana. Ignore and do not include any text found outside of dialogue boxes, "
+                # "including character names, speaker labels, or sound effects. Return *only* the filtered dialogue text. "
+                # "If no text is found within dialogue boxes after applying filters, return an empty string. "
+                # "OR, if there are no text bubbles or dialogue boxes found, return everything."
+                "Do not include any other output, formatting markers, or commentary, only the text from the image."
+            )
+
+            response = self.client.chat.completions.create(
+                model="meta-llama/llama-4-scout-17b-16e-instruct",
+                messages=[
+                    {
+                        "role": "user",
+                        "content": [
+                            {"type": "text", "text": prompt},
+                            {"type": "image_url", "image_url": {"url": f"data:image/png;base64,{img_base64}"}},
+                        ],
+                    }
+                ],
+                max_tokens=300
+            )
+
+            if response.choices and response.choices[0].message.content:
+                text_output = response.choices[0].message.content.strip()
+                return (True, text_output)
+            else:
+                return (True, "")
+
+        except FileNotFoundError:
+            return (False, f'File not found: {img_or_path}')
+        except Exception as e:
+            return (False, f'Groq API request failed: {e}')
+
+    def _preprocess(self, img):
+        return base64.b64encode(pil_image_to_bytes(img, png_compression=1)).decode('utf-8')

{gamesentenceminer-2.8.16 → gamesentenceminer-2.8.18}/GameSentenceMiner/owocr/owocr/run.py

@@ -585,7 +585,8 @@ def process_and_write_results(img_or_path, write_to, notifications, last_result,
         for i, instance in enumerate(engine_instances):
             if instance.name.lower() in engine.lower():
                 engine_instance = instance
-                last_result = (last_result[0], i)
+                if last_result:
+                    last_result = (last_result[0], i)
                 break
     else:
         engine_instance = engine_instances[engine_index]

{gamesentenceminer-2.8.16 → gamesentenceminer-2.8.18}/GameSentenceMiner.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: GameSentenceMiner
-Version: 2.8.16
+Version: 2.8.18
 Summary: A tool for mining sentences from games. Update: Multi-Line Mining! Fixed!
 Author-email: Beangate <bpwhelan95@gmail.com>
 License: MIT License

{gamesentenceminer-2.8.16 → gamesentenceminer-2.8.18}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: GameSentenceMiner
-Version: 2.8.16
+Version: 2.8.18
 Summary: A tool for mining sentences from games. Update: Multi-Line Mining! Fixed!
 Author-email: Beangate <bpwhelan95@gmail.com>
 License: MIT License

{gamesentenceminer-2.8.16 → gamesentenceminer-2.8.18}/pyproject.toml

@@ -7,7 +7,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "GameSentenceMiner"
-version = "2.8.16"
+version = "2.8.18"
 description = "A tool for mining sentences from games. Update: Multi-Line Mining! Fixed!"
 readme = "README.md"
 requires-python = ">=3.10"