Refactor: Update MainWindow to set default box display style and load…

… external OCR model

- Updated the `MainWindow` class in `mainwindow.py` to set the default box display style for the OCR model combo box.
- Added the ability to load a custom OCR model file in the `ocrModelChanged` method of `MainWindow`.
- Updated the `setOcrModel` method in `TextDetector` class of `tesseract.py` to handle loading external OCR models.

src/mainwindow.py

@@ -243,11 +243,14 @@ def __init__(self, translator: QTranslator, parent: QObject):
             "General Scoreboard",
             "General Fonts (English)",
             "General Scoreboard Large",
+            "Load External OCR Model",
         ]
         self.ui.comboBox_ocrModel.addItems(ocr_models)
-        self.ui.comboBox_ocrModel.setCurrentIndex(
-            fetch_data("scoresight.json", "ocr_model", 1)
-        )  # default to General Scoreboard
+        # default to General Scoreboard
+        ocr_model_from_storage = fetch_data("scoresight.json", "ocr_model", 1)
+        if type(ocr_model_from_storage) == str:
+            ocr_model_from_storage = 4
+        self.ui.comboBox_ocrModel.setCurrentIndex(ocr_model_from_storage)
 
         self.ui.frame_source_view.setEnabled(False)
         self.ui.groupBox_target_settings.setEnabled(False)
@@ -539,15 +542,35 @@ def detectionCadenceChanged(self, detections_per_second):
                 1000 / detections_per_second
             )
 
-    def ocrModelChanged(self, index):
-        self.globalSettingsChanged("ocr_model", index)
+    def ocrModelChanged(self, index: int | str):
+        ocrModel = None
+        if index == 4:
+            # load a custom OCR model
+            file, _ = QFileDialog.getOpenFileName(
+                self, "Open OCR Model File", "", "OCR Model Files (*.traineddata)"
+            )
+            if not file:
+                return
+            # get absolute file path
+            file = path.abspath(file)
+            self.globalSettingsChanged("ocr_model", file)
+            ocrModel = file
+        elif type(index) == int:
+            self.globalSettingsChanged("ocr_model", index)
+            ocrModel = index
+        elif type(index) == str:
+            # check if the index is a valid existing file
+            if path.exists(index):
+                self.globalSettingsChanged("ocr_model", index)
+                ocrModel = index
+
         # update the ocr model in the text detector
         if (
             self.image_viewer
             and self.image_viewer.timerThread
             and self.image_viewer.timerThread.textDetector
         ):
-            self.image_viewer.timerThread.textDetector.setOcrModel(index)
+            self.image_viewer.timerThread.textDetector.setOcrModel(ocrModel)
 
     def openLogsDialog(self):
         if self.log_dialog is None:

src/tesseract.py

@@ -15,6 +15,7 @@
     TextDetectionTarget,
     TextDetectionTargetWithResult,
 )
+from sc_logging import logger
 
 
 def autocrop(image_in):
@@ -112,25 +113,40 @@ def __init__(self):
         else:
             self.setOcrModel(TextDetector.OcrModelIndex.DAKTRONICS)
 
-    def setOcrModel(self, ocrModelIndex):
+    def setOcrModel(self, ocrModelIndex: OcrModelIndex | str | None = None):
         ocr_model = None
-        if ocrModelIndex == self.OcrModelIndex.DAKTRONICS:
+        external_folder = None
+        if ocrModelIndex == TextDetector.OcrModelIndex.DAKTRONICS:
             ocr_model = "daktronics"
-        if ocrModelIndex == self.OcrModelIndex.SCOREBOARD_GENERAL:
+        elif ocrModelIndex == TextDetector.OcrModelIndex.SCOREBOARD_GENERAL:
             ocr_model = "scoreboard_general"
-        if ocrModelIndex == self.OcrModelIndex.GENERAL_ENGLISH:
+        elif ocrModelIndex == TextDetector.OcrModelIndex.GENERAL_ENGLISH:
             ocr_model = "eng"
-        if ocrModelIndex == self.OcrModelIndex.SCOREBOARD_GENERAL_LARGE:
+        elif ocrModelIndex == TextDetector.OcrModelIndex.SCOREBOARD_GENERAL_LARGE:
             ocr_model = "scoreboard_general_large"
-        if ocr_model is None:
+        elif isinstance(ocrModelIndex, str):
+            # check the model file exists at the path
+            if path.exists(ocrModelIndex):
+                # Take the folder as the tessdata folder
+                external_folder = path.dirname(ocrModelIndex)
+                # Take the model name without extension as the "language"
+                ocr_model = path.basename(ocrModelIndex)
+                ocr_model = path.splitext(ocr_model)[0]
+        elif ocr_model is None:
             return
 
+        logger.info(f"Setting OCR model to {ocr_model}")
+
         with self.api_lock:
             if self.api is not None:
                 self.api.End()
                 self.api = None
             self.api = PyTessBaseAPI(
-                path=resource_path("tesseract", "tessdata"),
+                path=(
+                    resource_path("tesseract", "tessdata")
+                    if external_folder is None
+                    else external_folder
+                ),
                 lang=ocr_model,
             )
             # single word PSM