Added OCR processor

src/core/containers/right_box.py

@@ -3,14 +3,14 @@
 # Lib imports
 import gi
 gi.require_version('Gtk', '3.0')
-gi.require_version('Gdk', '3.0')
 from gi.repository import Gtk
-from gi.repository import Gdk
 
 # Application imports
+from .image_view_scroll import ImageViewScroll
+
 from ..widgets.button_controls import ButtonControls
 from ..widgets.path_label import PathLabel
-from .image_view_scroll import ImageViewScroll
+from ..widgets.ocr_window import OCRWindow
 
 
 
@@ -38,6 +38,8 @@ class RightBox(Gtk.Box):
         event_system.subscribe("background_fill", self._toggle_background)
 
     def _load_widgets(self):
+        window = OCRWindow()
+
         self.add(ButtonControls())
         self.add(PathLabel())
         self.add(ImageViewScroll())

src/core/widgets/button_controls.py

@@ -40,6 +40,7 @@ class ButtonControls(Gtk.ButtonBox):
         hflip_button   = Gtk.Button()
         rrotate_button = Gtk.Button()
         zoomin_button  = Gtk.Button()
+        ocr_button     = Gtk.Button()
 
         self._set_class(self.fit_button)
 
@@ -51,6 +52,7 @@ class ButtonControls(Gtk.ButtonBox):
         hflip_button.set_tooltip_text("Flip Horizontal")
         rrotate_button.set_tooltip_text("Rotate Right")
         zoomin_button.set_tooltip_text("Zoom In")
+        ocr_button.set_tooltip_text("OCR")
 
         zoomout_button.set_image( Gtk.Image.new_from_file(f"{icons_path}/zoom-out.png") )
         lrotate_button.set_image( Gtk.Image.new_from_file(f"{icons_path}/rotate-left.png") )
@@ -60,6 +62,7 @@ class ButtonControls(Gtk.ButtonBox):
         hflip_button.set_image( Gtk.Image.new_from_file(f"{icons_path}/flip-horizontal.png") )
         rrotate_button.set_image( Gtk.Image.new_from_file(f"{icons_path}/rotate-right.png") )
         zoomin_button.set_image( Gtk.Image.new_from_file(f"{icons_path}/zoom-in.png") )
+        ocr_button.set_image( Gtk.Image.new_from_file(f"{icons_path}/ocr.png") )
 
         zoomout_button.set_always_show_image(True)
         lrotate_button.set_always_show_image(True)
@@ -69,6 +72,7 @@ class ButtonControls(Gtk.ButtonBox):
         hflip_button.set_always_show_image(True)
         rrotate_button.set_always_show_image(True)
         zoomin_button.set_always_show_image(True)
+        ocr_button.set_always_show_image(True)
 
         zoomout_button.connect("clicked", self._zoom_out)
         lrotate_button.connect("clicked", self._rotate_left)
@@ -78,6 +82,7 @@ class ButtonControls(Gtk.ButtonBox):
         hflip_button.connect("clicked", self._horizontal_flip)
         rrotate_button.connect("clicked", self._rotate_right)
         zoomin_button.connect("clicked", self._zoom_in)
+        ocr_button.connect("clicked", self._show_ocr)
 
         center_widget.add(zoomout_button)
         center_widget.add(lrotate_button)
@@ -87,6 +92,7 @@ class ButtonControls(Gtk.ButtonBox):
         center_widget.add(hflip_button)
         center_widget.add(rrotate_button)
         center_widget.add(zoomin_button)
+        center_widget.add(ocr_button)
 
         self.set_center_widget(center_widget)
 
@@ -125,3 +131,6 @@ class ButtonControls(Gtk.ButtonBox):
     def _unset_class(self, target):
         ctx = target.get_style_context()
         ctx.remove_class("button-highlighted")
+
+    def _show_ocr(self, widget):
+        event_system.emit("show_ocr")

src/core/widgets/image_view.py

@@ -50,6 +50,7 @@ class ImageView(ImageViewMixin, Gtk.Image):
         event_system.subscribe("size_allocate", self._size_allocate)
         event_system.subscribe("handle_file_from_dnd", self._handle_file_from_dnd)
 
+        event_system.subscribe("get_active_image_path", self._get_active_image_path)
         event_system.subscribe("zoom_out", self._zoom_out)
         event_system.subscribe("rotate_left", self._rotate_left)
         event_system.subscribe("vertical_flip", self._vertical_flip)
@@ -86,6 +87,8 @@ class ImageView(ImageViewMixin, Gtk.Image):
             self.set_as_static(path)
 
         self.pixbuff      = self.work_pixbuff.copy()
+        self.pixbuff.path = path
+
         width        = self.pixbuff.get_width()
         height       = self.pixbuff.get_height()
         size         = sizeof_fmt( getsize(path) )

src/core/widgets/image_view_mixin.py

@@ -10,6 +10,12 @@ from gi.repository import GdkPixbuf
 
 
 class ImageViewMixin:
+    def _get_active_image_path(self):
+        if self.pixbuff and self.pixbuff.path:
+            return self.pixbuff.path
+
+        return None
+
     def _zoom_out(self):
         if self.work_pixbuff and self.pixbuff:
             # TODO: Setup scale factor setting to pull from settings...

src/core/widgets/ocr_window.py

@@ -0,0 +1,163 @@
+# Python imports
+import os
+import requests
+import subprocess
+
+# Lib imports
+import gi
+gi.require_version('Gtk', '3.0')
+gi.require_version('GtkSource', '4')
+from gi.repository import Gtk
+from gi.repository import GLib
+from gi.repository import GtkSource
+
+# Application imports
+
+
+
+class OCRWindow(Gtk.Window):
+    def __init__(self):
+        super(OCRWindow, self).__init__()
+
+        self.tesseract_path = f"{settings.get_home_config_path()}/tesseract-ocr.AppImage"
+        self.download_url   = "https://github.com/AlexanderP/tesseract-appimage/releases/download/v5.3.3/tesseract-5.3.3-x86_64.AppImage"
+
+        self._setup_styling()
+        self._setup_signals()
+        self._subscribe_to_events()
+        self._load_widgets()
+
+
+    def _setup_styling(self):
+        self.set_title(f"Tesseract OCR")
+        self.set_icon_from_file( settings.get_window_icon() )
+        self.set_gravity(5)  # 5 = CENTER
+        self.set_position(1) # 1 = CENTER, 4 = CENTER_ALWAYS
+
+        self.set_default_size(480, 600)
+        self.set_size_request(480, 600)
+
+
+    def _setup_signals(self):
+        self.connect("delete-event", self._tear_down)
+
+    def _subscribe_to_events(self):
+        event_system.subscribe("show_ocr", self._show_ocr)
+
+    def _load_widgets(self):
+        scrolled_window = Gtk.ScrolledWindow()
+        box             = Gtk.Box()
+        download_button = Gtk.Button(label = "Download Tesseract OCR")
+        run_ocr_button  = Gtk.Button(label = "Run OCR")
+        text_view       = GtkSource.View()
+
+        download_button.set_tooltip_text("Download Tesseract OCR")
+        run_ocr_button.set_tooltip_text("Run OCR")
+
+        download_button.connect("clicked", self._download_tesseract_ocr)
+        run_ocr_button.connect("clicked", self._run_ocr_button)
+
+        box.set_orientation(Gtk.Orientation.VERTICAL)
+        text_view.set_vexpand(True)
+        text_view.set_show_line_numbers(True)
+        ctx = scrolled_window.get_style_context()
+        ctx.add_class("container-padding-5px")
+
+        scrolled_window.add(text_view)
+        box.add(download_button)
+        box.add(run_ocr_button)
+        box.add(scrolled_window)
+        self.add(box)
+
+        box.show_all()
+
+        if os.path.exists(self.tesseract_path):
+            download_button.hide()
+        else:
+            run_ocr_button.hide()
+            text_view.hide()
+
+
+    def _show_ocr(self):
+        self.show()
+
+    def _tear_down(self, widget = None, eve = None):
+        self.hide()
+
+        # Return True to NOT propigate call (would actually destroy window n children)
+        return True
+
+    def _download_tesseract_ocr(self, button):
+        parent           = button.get_parent()
+        spinner          = Gtk.Spinner.new()
+
+        parent.add(spinner)
+        GLib.idle_add(self.__download_file, self.download_url, spinner)
+
+    @daemon_threaded
+    def __download_file(self, url = None, spinner = None):
+        if not url: return
+        if not spinner: return
+        if not url == self.download_url: return
+
+        spinner.show()
+        spinner.start()
+        
+        # NOTE the stream = True parameter below
+        with requests.get(url, stream = True) as r:
+            r.raise_for_status()
+
+            with open(self.tesseract_path, 'wb') as f:
+                for chunk in r.iter_content(chunk_size = 8192):
+                    # If chunk encoded response uncomment and set chunk_size parameter to None.
+                    # if chunk:
+                    f.write(chunk)
+
+        spinner.stop()
+
+        GLib.idle_add(self.__process_dl_finished, spinner)
+
+    def __process_dl_finished(self, spinner = None):
+        if not spinner: return
+        
+        parent           = spinner.get_parent()
+        download_button, \
+        run_ocr_button,  \
+        text_view,       \
+        spinner          = parent.get_children()
+
+        parent.remove(spinner)
+        if os.path.exists(self.tesseract_path):
+            self.__set_as_executable(self.tesseract_path)
+            run_ocr_button.show()
+            text_view.show()
+            download_button.hide()
+
+    def __set_as_executable(self, tesseract_path):
+        if not tesseract_path: return
+
+        os.access(tesseract_path, os.X_OK)
+        try:
+            command = ["chmod", "544", tesseract_path]
+            with subprocess.Popen(command, stdout = subprocess.PIPE) as proc:
+                result = proc.stdout.read().decode("UTF-8").strip()
+        except Exception as e:
+            logger.error(f"Couldn't chmod\nFile:  {properties.file_uri}")
+            logger.error( repr(e) )
+
+
+    def _run_ocr_button(self, button):
+        active_image = event_system.emit_and_await("get_active_image_path")
+
+        if not active_image: return
+
+        scrolled_window = button.get_parent().get_children()[2]
+        text_view = scrolled_window.get_children()[0]
+
+        command   = [self.tesseract_path, active_image, "stdout"]
+        result    = subprocess.run(command, stdout = subprocess.PIPE)
+        data      = result.stdout.decode('utf-8')
+
+        logger.debug(command)
+        logger.debug(data)
+        text_view.get_buffer().set_text(data, -1)