fix(windows): Bundle tessdata for OCR support out of the box

.github/workflows/release.yml

@@ -81,6 +81,14 @@ jobs:
     - name: Copy files to directory for installer
       run: mkdir installer; cp ./x64/Release-Full/ccextractorwinfull.exe ./installer; cp ./x64/Release-Full/*.dll ./installer
       working-directory: ./windows
+    - name: Download tessdata for OCR support
+      run: |
+        mkdir -p ./installer/tessdata
+        # Download English traineddata from tessdata_fast (smaller, faster, good for most use cases)
+        Invoke-WebRequest -Uri "https://github.com/tesseract-ocr/tessdata_fast/raw/main/eng.traineddata" -OutFile "./installer/tessdata/eng.traineddata"
+        # Download OSD (Orientation and Script Detection) for automatic script detection
+        Invoke-WebRequest -Uri "https://github.com/tesseract-ocr/tessdata_fast/raw/main/osd.traineddata" -OutFile "./installer/tessdata/osd.traineddata"
+      working-directory: ./windows
     - name: install WiX
       run: dotnet tool uninstall --global wix; dotnet tool install --global wix --version 6.0.2 && wix extension add -g WixToolset.UI.wixext/6.0.2
     - name: Make sure WiX works

src/lib_ccx/ocr.c

@@ -8,6 +8,11 @@
 #include <dirent.h>
 #include "ccx_encoders_helpers.h"
 #include "ccx_encoders_spupng.h"
+#ifdef _WIN32
+#include <windows.h>
+#elif defined(__APPLE__)
+#include <mach-o/dyld.h>
+#endif
 #include "ocr.h"
 
 struct ocrCtx
@@ -100,22 +105,87 @@ void delete_ocr(void **arg)
 	freep(arg);
 }
 
+/**
+ * get_executable_directory
+ *
+ * Returns the directory containing the executable.
+ * Returns a pointer to a static buffer, or NULL on failure.
+ */
+static const char *get_executable_directory(void)
+{
+	static char exe_dir[1024] = {0};
+	static int initialized = 0;
+
+	if (initialized)
+		return exe_dir[0] ? exe_dir : NULL;
+
+	initialized = 1;
+
+#ifdef _WIN32
+	char exe_path[MAX_PATH];
+	DWORD len = GetModuleFileNameA(NULL, exe_path, MAX_PATH);
+	if (len == 0 || len >= MAX_PATH)
+		return NULL;
+
+	// Find the last backslash and truncate there
+	char *last_sep = strrchr(exe_path, '\\');
+	if (last_sep)
+	{
+		*last_sep = '\0';
+		strncpy(exe_dir, exe_path, sizeof(exe_dir) - 1);
+		exe_dir[sizeof(exe_dir) - 1] = '\0';
+	}
+#elif defined(__linux__)
+	char exe_path[1024];
+	ssize_t len = readlink("/proc/self/exe", exe_path, sizeof(exe_path) - 1);
+	if (len <= 0)
+		return NULL;
+	exe_path[len] = '\0';
+
+	char *last_sep = strrchr(exe_path, '/');
+	if (last_sep)
+	{
+		*last_sep = '\0';
+		strncpy(exe_dir, exe_path, sizeof(exe_dir) - 1);
+		exe_dir[sizeof(exe_dir) - 1] = '\0';
+	}
+#elif defined(__APPLE__)
+	char exe_path[1024];
+	uint32_t size = sizeof(exe_path);
+	if (_NSGetExecutablePath(exe_path, &size) != 0)
+		return NULL;
+
+	char *last_sep = strrchr(exe_path, '/');
+	if (last_sep)
+	{
+		*last_sep = '\0';
+		strncpy(exe_dir, exe_path, sizeof(exe_dir) - 1);
+		exe_dir[sizeof(exe_dir) - 1] = '\0';
+	}
+#endif
+
+	return exe_dir[0] ? exe_dir : NULL;
+}
+
 /**
  * probe_tessdata_location
  *
  * This function probe tesseract data location
  *
  * Priority of Tesseract traineddata file search paths:-
  * 1. tessdata in TESSDATA_PREFIX, if it is specified. Overrides others
- * 2. tessdata in current working directory
- * 3. tessdata in /usr/share
+ * 2. tessdata in executable directory (for bundled tessdata)
+ * 3. tessdata in current working directory
+ * 4. tessdata in system locations (/usr/share, etc.)
+ * 5. tessdata in default Tesseract install location (Windows)
  */
 char *probe_tessdata_location(const char *lang)
 {
 	int ret = 0;
 
 	const char *paths[] = {
 	    getenv("TESSDATA_PREFIX"),
+	    get_executable_directory(),
 	    "./",
 	    "/usr/share/",
 	    "/usr/local/share/",

windows/installer.wxs

@@ -4,6 +4,7 @@
 		<MediaTemplate EmbedCab="yes"/>
 		<Feature Id="CCX" Title="CCExtractor Setup" Level="1">
 			<ComponentGroupRef Id="CCX_Components_MainFolder"/>
+			<ComponentGroupRef Id="CCX_Components_tessdata"/>
 			<ComponentGroupRef Id="CCX_Components_MainFolder_data"/>
 			<ComponentGroupRef Id="CCX_Components_MainFolder_data_flutter_assets"/>
 			<ComponentGroupRef Id="CCX_Components_MainFolder_data_flutter_assets_assets"/>
@@ -40,6 +41,7 @@
     </StandardDirectory>
 		<StandardDirectory Id="ProgramFiles6432Folder">
 			<Directory Id="INSTALLFOLDER" Name="CCExtractor">
+				<Directory Id="CCX_tessdata" Name="tessdata"/>
 				<Directory Id="CCX_data" Name="data">
 					<Directory Id="CCX_data_flutter_assets" Name="flutter_assets">
 						<Directory Id="CCX_data_flutter_assets_assets" Name="assets"/>
@@ -194,6 +196,15 @@
 				<File Id="filWmUnmaximize" KeyPath="yes" Source="./installer/data/flutter_assets/packages/window_manager/images/ic_chrome_unmaximize.png"/>
 			</Component>
 		</ComponentGroup>
+		<!-- Tesseract OCR data files for HardSubx feature -->
+		<ComponentGroup Id="CCX_Components_tessdata" Directory="CCX_tessdata">
+			<Component Id="cmpTessdataEng" Guid="{A1234567-8901-2345-6789-0123456789AB}">
+				<File Id="filEngTraineddata" KeyPath="yes" Source="./installer/tessdata/eng.traineddata"/>
+			</Component>
+			<Component Id="cmpTessdataOsd" Guid="{B2345678-9012-3456-7890-123456789ABC}">
+				<File Id="filOsdTraineddata" KeyPath="yes" Source="./installer/tessdata/osd.traineddata"/>
+			</Component>
+		</ComponentGroup>
 	</Fragment>
 </Wix>