Added basic selection rendering (incl. testing/demo code)

MainWindow.glade

@@ -3,6 +3,8 @@
 <interface>
   <requires lib="gtk+" version="3.24"/>
   <template class="main_window" parent="GtkApplicationWindow">
+    <property name="width-request">800</property>
+    <property name="height-request">500</property>
     <property name="can-focus">False</property>
     <property name="title" translatable="yes">PDF Table Extractor</property>
     <property name="icon-name">document-page-setup</property>

experiments/draw_test.py

@@ -1,67 +0,0 @@
-import sys
-import traceback
-from typing import Dict, List, Optional, Sequence, Tuple, TypeAlias
-
-import gi
-gi.require_version("Gtk", "3.0")
-from gi.repository import GLib, Gio, Gtk, GObject
-gi.require_foreign("cairo")
-import cairo
-
-
-class CustomDrawingArea(Gtk.DrawingArea):
-    def __init__(self, *args, **kwargs):
-        super().__init__(*args, **kwargs)
-
-        self.set_size_request(200, 200)
-        self.set_app_paintable(True)
-        self.connect_after("draw", self.on_draw, {})
-
-    def on_draw(self, widget: Gtk.DrawingArea, cr: cairo.Context, data: GObject.GPointer):
-        width = widget.get_allocated_width()
-        height = widget.get_allocated_height()
-        print(f"w: {width}, h: {height}")
-
-        sctx = widget.get_style_context()
-        Gtk.render_background(sctx, cr, 0, 0, width, height)
-
-        cr.set_source_rgba(1.0, 0.0, 0.0, 1.0)
-        cr.rectangle(0, 0, width, height)
-        cr.fill()
-
-
-class MainWindow(Gtk.ApplicationWindow):
-    def __init__(self, *args, **kwargs):
-        super().__init__(*args, **kwargs)
-
-        self.app: Application = self.get_application() # type: ignore
-        assert self.app is not None
-
-        box = Gtk.Box(orientation=Gtk.Orientation.VERTICAL)
-        self.add(box)
-
-        self.area = CustomDrawingArea()
-        frame = Gtk.Frame(label="DrawingArea")
-        frame.add(self.area)
-        box.pack_start(frame, expand=True, fill=True, padding=20)
-        box.add(Gtk.Button(label=":)"))
-
-
-class Application(Gtk.Application):
-    def __init__(self, *args, **kwargs):
-        super().__init__(
-            *args,
-            application_id="zone.lunareclipse.draw_test",
-            flags=Gio.ApplicationFlags.FLAGS_NONE,
-            **kwargs
-        )
-        self.window = None
-
-    def do_activate(self):
-        self.window = self.window or MainWindow(application=self)
-        self.window.show_all()
-
-
-if __name__ == "__main__":
-    app = Application()
-    app.run(sys.argv)

experiments/main.py

@@ -1,17 +0,0 @@
-import argparse
-import tabula
-
-
-if __name__ == "__main__":
-    #parser = argparse.ArgumentParser()
-    #parser.add_argument("filename")
-    #args = parser.parse_args()
-
-    table = tabula.io.read_pdf(
-        "../sources/2018_Torres-Benitez_Metabolomic analysis Parmotrema.pdf", 
-        pages=[5],
-        lattice=False,
-        multiple_tables=True,
-    )
-    print(table)
-    print("test")

experiments/old_pdf_table_extractor_qt.py

@@ -1,54 +0,0 @@
-import sys
-from typing import Dict, List, Optional, Sequence, Tuple, TypeAlias
-
-from PySide6.QtWidgets import QApplication, QDialog, QHBoxLayout, QLabel, QMainWindow, QPushButton, QVBoxLayout, QWidget
-from PySide6.QtCore import Slot
-from PySide6.QtPdf import QPdfDocument
-from PySide6.QtPdfWidgets import QPdfView
-import fitz
-
-TEST_FILENAME = "/home/luna/Documents/Resources/Praca Licencjacka/sources/2018_Torres-Benitez_Metabolomic analysis Parmotrema.pdf"
-
-Coords: TypeAlias = Tuple[float, float]
-
-class Selection:
-    def __init__(self, bounds: Tuple[Coords, Coords], columns: Optional[Sequence[float]] = None):
-        self.bounds = bounds
-        self.columns = columns
-
-class Document:
-    def __init__(self, filename: str):
-        self.filename = filename
-        self.document = fitz.Document(filename)
-        self.selections: Dict[int, List[Selection]] = {}
-
-class SelectablePdfView(QPdfView):
-    pass
-
-class State():
-    pass
-
-class Frontend():
-    def __init__(self, argv: Sequence[str]):
-        self.app = QApplication(argv)
-        self.window = QMainWindow()
-        self.state = State()
-
-        self.thumbnails = QWidget()
-        self.pdfDocument = QPdfDocument()
-        self.optionsPanel = QWidget()
-
-        central_widget = QWidget()
-        cw_layout = QHBoxLayout(central_widget)
-        cw_layout.addWidget(self.thumbnails)
-        cw_layout.addWidget(self.optionsPanel)
-        self.window.setCentralWidget(central_widget)
-
-
-    def exec(self):
-        self.window.show()
-        self.app.exec()
-
-if __name__ == "__main__":
-    app = Frontend(sys.argv)
-    app.exec()

experiments/qt6_test.py

@@ -1,41 +0,0 @@
-import sys
-from typing import Sequence
-
-from PySide6.QtWidgets import QApplication, QDialog, QLabel, QPushButton, QVBoxLayout
-from PySide6.QtCore import Slot
-
-class Application():
-    def __init__(self, argv: Sequence[str]):
-        self.app = QApplication(argv)
-        self.counter = 0
-        self.window = QDialog()
-        self.layout = QVBoxLayout(self.window)
-        self.label = QLabel("0")
-        self.button_increment = QPushButton("Increment counter!")
-        self.button_decrement = QPushButton("Decrement counter!")
-
-        self.window.setWindowTitle("PDF Table Extractor")
-        self.layout.addWidget(self.label)
-        self.layout.addWidget(self.button_increment)
-        self.layout.addWidget(self.button_decrement)
-        self.button_increment.clicked.connect(self.increment)
-        self.button_decrement.clicked.connect(self.decrement)
-        print(self.window.layout())
-
-    @Slot()
-    def increment(self):
-        self.counter += 1
-        self.label.setText(f"{self.counter}")
-
-    @Slot()
-    def decrement(self):
-        self.counter -= 1
-        self.label.setText(f"{self.counter}")
-
-    def exec(self):
-        self.window.show()
-        self.app.exec()
-
-if __name__ == "__main__":
-    app = Application(sys.argv)
-    app.exec()

experiments/test.py

@@ -1,84 +0,0 @@
-import gi
-
-gi.require_version("Gtk", "3.0")
-from gi.repository import Gtk
-
-
-class FileChooserWindow(Gtk.Window):
-    def __init__(self):
-        super().__init__(title="FileChooser Example")
-
-        box = Gtk.Box(spacing=6)
-        self.add(box)
-
-        button1 = Gtk.Button(label="Choose File")
-        button1.connect("clicked", self.on_file_clicked)
-        box.add(button1)
-
-        button2 = Gtk.Button(label="Choose Folder")
-        button2.connect("clicked", self.on_folder_clicked)
-        box.add(button2)
-
-    def on_file_clicked(self, widget):
-        dialog = Gtk.FileChooserDialog(
-            title="Please choose a file", parent=self, action=Gtk.FileChooserAction.OPEN
-        )
-        dialog.add_buttons(
-            Gtk.STOCK_CANCEL,
-            Gtk.ResponseType.CANCEL,
-            Gtk.STOCK_OPEN,
-            Gtk.ResponseType.OK,
-        )
-
-        self.add_filters(dialog)
-
-        response = dialog.run()
-        if response == Gtk.ResponseType.OK:
-            print("Open clicked")
-            print("File selected: " + dialog.get_filename())
-        elif response == Gtk.ResponseType.CANCEL:
-            print("Cancel clicked")
-
-        dialog.destroy()
-
-    def add_filters(self, dialog):
-        filter_text = Gtk.FileFilter()
-        filter_text.set_name("Text files")
-        filter_text.add_mime_type("text/plain")
-        dialog.add_filter(filter_text)
-
-        filter_py = Gtk.FileFilter()
-        filter_py.set_name("Python files")
-        filter_py.add_mime_type("text/x-python")
-        dialog.add_filter(filter_py)
-
-        filter_any = Gtk.FileFilter()
-        filter_any.set_name("Any files")
-        filter_any.add_pattern("*")
-        dialog.add_filter(filter_any)
-
-    def on_folder_clicked(self, widget):
-        dialog = Gtk.FileChooserDialog(
-            title="Please choose a folder",
-            parent=self,
-            action=Gtk.FileChooserAction.SELECT_FOLDER,
-        )
-        dialog.add_buttons(
-            Gtk.STOCK_CANCEL, Gtk.ResponseType.CANCEL, "Select", Gtk.ResponseType.OK
-        )
-        dialog.set_default_size(800, 400)
-
-        response = dialog.run()
-        if response == Gtk.ResponseType.OK:
-            print("Select clicked")
-            print("Folder selected: " + dialog.get_filename())
-        elif response == Gtk.ResponseType.CANCEL:
-            print("Cancel clicked")
-
-        dialog.destroy()
-
-
-win = FileChooserWindow()
-win.connect("destroy", Gtk.main_quit)
-win.show_all()
-Gtk.main()

pdf_table_extractor.py

@@ -2,6 +2,7 @@ import sys
 import traceback
 from typing import Dict, List, Optional, Sequence, Tuple, TypeAlias
 
+import PIL.Image
 import gi
 
 gi.require_version("Gtk", "3.0")
@@ -12,54 +13,94 @@ import cairo
 
 import fitz
 
-TEST_FILENAME = "/home/luna/Documents/Resources/Praca Licencjacka/sources/2018_Torres-Benitez_Metabolomic analysis Parmotrema.pdf"
-
 Coords: TypeAlias = Tuple[float, float]
 
 
 class Selection:
     def __init__(self, bounds: Tuple[Coords, Coords], columns: Optional[Sequence[float]] = None):
         self.bounds = bounds
-        self.columns = columns
+        self.columns: list[float] = list(columns or [])
 
+class Page:
+    def __init__(self, index: int, raw: fitz.Page):
+        self.index = index
+        self.raw = raw
+        self.selections: list[Selection] = []
 
 class Document:
     def __init__(self, filename: str):
         self.filename = filename
-        self.document = fitz.Document(filename)
+        self.raw = fitz.Document(filename)
-        self.selections: Dict[int, List[Selection]] = {}
+        self.pages = []
 
-
+        for i, p in enumerate(self.raw.pages()):  # type: ignore
-class State():
+            page = Page(index=i, raw=p)
-    pass
+            self.pages.append(page)
 
 
 class PdfPage(Gtk.DrawingArea):
     def __init__(self, page, *args, **kwargs):
         super().__init__(*args, **kwargs)
 
-        self.page: fitz.Page = page
+        self.page: Page = page
-        pix = self.page.get_pixmap(dpi=96)  # type: ignore
+        pix: fitz.Pixmap = self.page.raw.get_pixmap(dpi=96)  # type: ignore
         self.set_size_request(pix.width, pix.height)
-        self.set_app_paintable(True)  # type: ignore
         self.connect("draw", self.on_draw, {})
 
+        if self.page.index == 4:
+            self.page.selections.append(Selection(
+                ((0.1, 0.18), (0.9, 0.72)),
+                columns=[0.08, 0.24, 0.34, 0.42, 0.51, 0.59, 0.67, 0.73, 0.91]
+            ))
+
     def on_draw(self, widget: Gtk.DrawingArea, cr: cairo.Context, data: GObject.GPointer):
-        # app: Application =  widget.get_window().get_application() # type: ignore
         width = widget.get_allocated_width()
         height = widget.get_allocated_height()
 
-        cr.set_source_rgba(255, 255, 255)
-        cr.paint()
-
         sctx = widget.get_style_context()
         Gtk.render_background(sctx, cr, 0, 0, width, height)
-        pix = self.page.get_pixmap(dpi=96, alpha=True)  # type: ignore
+
-        mv: memoryview = pix.samples_mv
+        pix: fitz.Pixmap = self.page.raw.get_pixmap(dpi=96)  # type: ignore
+        img = PIL.Image.frombytes("RGBA" if pix.alpha else "RGB", [pix.width, pix.height], pix.samples)
+        img.putalpha(1)
+        img = PIL.Image.merge("RGBA", (lambda r, g, b, a: (b, g, r, a))(*img.split()))  # type: ignore
+        mv: memoryview = memoryview(bytearray(img.tobytes()))
         ims = cairo.ImageSurface.create_for_data(mv, cairo.Format.RGB24, pix.width, pix.height)
         cr.set_source_surface(ims, 0, 0)
         cr.paint()
 
+        for sel in self.page.selections:
+            sel_x1 = sel.bounds[0][0]*pix.width
+            sel_y1 = sel.bounds[0][1]*pix.height
+            sel_x2 = sel.bounds[1][0]*pix.width
+            sel_y2 = sel.bounds[1][1]*pix.height
+
+            # Base settings
+            cr.set_line_cap(cairo.LINE_CAP_BUTT)
+            cr.set_line_width(2)
+
+            # Columns (draw first - below selection)
+            cr.set_dash([5])
+            cr.set_source_rgba(1, 0, 0)
+            for col in sel.columns:
+                col_x = sel_x1 + (sel_x2 - sel_x1) * col
+                cr.move_to(col_x, sel_y1)
+                cr.line_to(col_x, sel_y2)
+                cr.stroke()
+
+            # Selection
+            cr.rectangle(sel_x1, sel_y1, sel_x2 - sel_x1, sel_y2 - sel_y1)
+
+            # White part of the pattern
+            cr.set_source_rgba(1, 1, 1)
+            cr.set_dash([5], 5)
+            cr.stroke_preserve() # important preserve - reuse rectangle
+
+            # Black part of the pattern
+            cr.set_source_rgba(0, 0, 0)
+            cr.set_dash([5])
+            cr.stroke()
+
 
 @Gtk.Template.from_file("MainWindow.glade")
 class MainWindow(Gtk.ApplicationWindow):
@@ -81,8 +122,6 @@ class MainWindow(Gtk.ApplicationWindow):
 
         self.app.connect("notify::document", self.on_document_updated)
 
-        # self.pdf_list_box.add(PdfPage())
-
     # @Gtk.Template.Callback()
     # def example_button_released_cb(self, widget: Gtk.Button, **kwargs):
     #    assert self.example_button == widget
@@ -124,20 +163,35 @@ class MainWindow(Gtk.ApplicationWindow):
     #       editing an existing one won't trigger it
     def on_document_updated(self, recvobj, gparamstring):
         document: Document = self.app.get_property("document")
-        self.header_bar.set_title(document.filename.split("/")[-1])  # type: ignore
+        self.header_bar.set_title(document.filename.split("/")[-1])
-        self.header_bar.set_subtitle(document.filename)  # type: ignore
+        self.header_bar.set_subtitle(document.filename)
 
-        for i in document.document.pages():  # type: ignore
+        for child in self.pdf_list_box.get_children():
-            row = Gtk.ListBoxRow()
+            if type(child) is Gtk.ListBoxRow:
-            page = PdfPage(i)
+                child.destroy()
-            row.add(page)
+
-            self.pdf_list_box.add(row)  # type: ignore
+        for i in document.pages:
+            self.pdf_list_box.add(PdfPage(i))
 
         self.pdf_list_box.show_all()
 
     @Gtk.Template.Callback()
     def on_open_button_small_clicked(self, widget, **kwargs):
-        pass
+        TEST_FILENAME = "/home/luna/Documents/Resources/Praca Licencjacka/sources/2018_Torres-Benitez_Metabolomic analysis Parmotrema.pdf"
+        try:
+            self.app.set_property("document", Document(TEST_FILENAME))
+        except Exception as e:
+            message_dialog = Gtk.MessageDialog(
+                title="An error has occured.",
+                transient_for=self,
+                modal=True,
+                message_type=Gtk.MessageType.ERROR,
+                text=repr(e),
+                secondary_text=traceback.format_exc(),
+                buttons=Gtk.ButtonsType.OK,
+            )
+            message_dialog.run()  # type: ignore
+            message_dialog.destroy()
 
 
 class Application(Gtk.Application):

pyproject.toml

@@ -11,11 +11,14 @@ python = ">=3.11,<3.13"
 pygobject = "^3.48.2"
 tabula-py = "^2.9.0"
 PyMuPDF = "^1.24.0"
+pillow = "^10.3.0"
 #pdfplumber = "^0.11.0"
 #PySide6 = "^6.6.3" # qt
 
 [tool.poetry.group.dev.dependencies]
-pygobject-stubs = "^2.11.0" # this is actually broken lol
+# this is actually broken lol, fix by running following in venv:
+# pip uninstall pygobject-stubs && pip install pygobject-stubs --no-cache-dir --config-settings=config=Gtk3,Gdk3,Soup2
+pygobject-stubs = "^2.11.0" 
 
 [build-system]
 requires = ["poetry-core"]