docoptimizer

#!/usr/bin/env python
"""
Sometimes ODT files get slow after pasting many things from the internet
and no matter where else you paste the document's contents, even
AbiWord, the file takes minutes to load (or paste). This program uses
pandoc which so far is the only program I've discovered that can remove
the invisible junk and make the file fast again.

NOTE: It will potentially lose the margins and fonts, make the entire
document italicized, and make the entire document a bullet in a
bulleted list. However, that was only known to happen with one of the
slow documents so maybe it will not happen in every case.
"""
from __future__ import print_function

import os
import platform
import subprocess
import sys

if platform.system() == "Windows":
    HOME = os.environ['USERPROFILE']
else:
    HOME = os.environ['HOME']

PANDOC = "pandoc"

tkdephelp = "sudo apt-get install python3-tk"
dephelp = "sudo apt-get install python3-pil python3-pil.imagetk"
if sys.version_info.major >= 3:
    try:
        import tkinter as tk
        from tkinter import ttk
        from tkinter import messagebox
        from tkinter import filedialog
    except ImportError:
        for sitepackages in site.getsitepackages():
            try_sub = os.path.join(sitepackages, "tkinter")
            if os.path.isdir(try_sub):
                print(venv_error_fmt % try_sub,
                      file=sys.stderr)
            elif os.path.exists(try_sub):
                print("Error: %s exists but is not a directory"
                      "" % repr(try_sub),
                      file=sys.stderr)
            else:
                print("Error: %s is not present" % repr(try_sub),
                      file=sys.stderr)
        raise
else:  # Python 2
    import Tkinter as tk  # type: ignore
    import ttk  # type: ignore
    import tkMessageBox as messagebox
    import tkFileDialog as filedialog
    tkdephelp = "sudo apt-get install python-tk"
    dephelp = "sudo apt-get install python-imaging python-pil.imagetk"


class Tkclass(tk.Tk):
    def __init__(self):
        tk.Tk.__init__(self)
        root = self
        app = Application(self)
        app.master.title("Document Optimizer")
        root.after(1, self.loaded)
        app.mainloop()

    def loaded(self):
        root = self
        # Hide hidden "." folders as per
        # <https://stackoverflow.com/a/54068050>
        # call a dummy dialog with an impossible option to initialize the file
        # dialog without really getting a dialog window; this will throw a
        # TclError, so we need a try...except :
        try:
            root.tk.call('tk_getOpenFile', '-foobarbaz')
        except tk.TclError:
            # _tkinter.TclError: can't invoke "image" command: application has been destroyed
            pass
        # now set the magic variables accordingly
        root.tk.call('set', '::tk::dialog::file::showHiddenBtn', '1')
        root.tk.call('set', '::tk::dialog::file::showHiddenVar', '0')
        screen_w = root.winfo_screenwidth()
        screen_h = root.winfo_screenheight()
        root.geometry("{}x{}".format(int(screen_w/4), int(screen_h/12)))


class Application(tk.Frame):

    def __init__(self, parent):
        self.root = parent
        tk.Frame.__init__(self, parent)
        self.grid(sticky=tk.N+tk.S+tk.E+tk.W)
        self.statusSV = tk.StringVar(master=parent)
        self.pathSV = tk.StringVar(master=parent)
        self.statusSV.set("Choose a file to convert.")
        self.createWidgets()

    def createWidgets(self):
        top = self.winfo_toplevel()
        top.rowconfigure(0, weight=1)
        top.columnconfigure(0, weight=1)
        self.rowconfigure(0, weight=1)
        self.columnconfigure(0, weight=6)
        self.columnconfigure(1, weight=1)
        self.columnconfigure(2, weight=1)
        self.columnconfigure(3, weight=1)
        self.fileEntry = tk.Entry(self, textvariable=self.pathSV)
        self.fileEntry.grid(row=0, column=0, sticky=tk.W+tk.E)

        self.browseButton = tk.Button(self, text='Open...',
                                      command=self.choose_file)
        self.browseButton.grid(row=0, column=1, sticky=tk.W+tk.E)
        self.convertButton = tk.Button(self, text='Convert',
                                       command=self.convert_selected)
        self.convertButton.grid(row=1, column=0,  # sticky=tk.W+tk.E,
                                columnspan=2)
        self.statusLabel = tk.Label(self)
        self.statusLabel.configure(textvariable=self.statusSV)
        self.statusLabel.grid(row=2, column=0, columnspan=2)
        # self.quit = tk.Button(self, text='Quit',command=self.QuitApp)
        # self.quit.grid(row=0, column=3)

    def choose_file(self):
        self.statusSV.set("Choose a file...")
        self.cmd_parts = None
        name = filedialog.askopenfilename(
            parent=self,
            initialdir=HOME,
            filetypes=(('CSV files', '*.odt'),),  # ('all files', '*.*')),
        )
        if not name:
            self.statusSV.set("")
            return
        self.pathSV.set(name)

        # self.statusSV.set(name)
        self.statusSV.set("")

    def convert_selected(self):
        self.statusSV.set("")
        self.cmd_parts = None
        path = self.pathSV.get()
        if not path:
            self.statusSV.set("You must select a file first.")
        if not os.path.isfile(path):
            self.statusSV.set("The selected file doesn't exist.")
        no_ext, dot_ext = os.path.splitext(path)
        self.dst = no_ext + ".docx"
        # working example:
        # pandoc 16\ Links\ slow\ Apr\ 1-\ 2024.odt -f odt -t docx -s \
        #   -o 16\ Links\ slow\ Apr\ 1-\ 2024.docx
        # ^ as per Amanda on <https://stackoverflow.com/a/34299311>
        dst_dir, dst_name = os.path.split(self.dst)
        if os.path.isfile(self.dst):
            answer = messagebox.askyesno(
                title="Confirm Overwrite",
                message=("Already converted.\nDo you want to overwrite\n\"{}\"?"
                         .format(dst_name)),
            )
            if answer != tk.YES:
                self.statusSV.set("Cancelled.")
                return
            os.remove(self.dst)

        self.cmd_parts = [
            PANDOC,
            path,
            "-f",
            "odt",
            "-t",
            "docx",
            "-s",
            "-o",
            self.dst,
        ]

        self.statusSV.set("Please wait...")
        self.root.after(1, self.run_job)

    def run_job(self):
        if not self.cmd_parts:
            self.statusSV.set("The command was interrupted.")
            return
        result = subprocess.Popen(self.cmd_parts)
        text = result.communicate()[0]
        return_code = result.returncode

        dst_dir, dst_name = os.path.split(self.dst)
        if return_code == 0:
            self.statusSV.set("Done (saved \"{}\")".format(dst_name))
        else:
            self.statusSV.set("Failed ({})".format(dst_name))

    def QuitApp(self):
        top=self.winfo_toplevel()
        top.quit()


main = Tkclass()