Bikarhêner:Balyozxane/skrîpt/py/removegiredan.py

Ji Wîkîpediya, ensîklopediya azad.
#!/usr/bin/env python3
"""
Şablona Girêdan ji rûpelan radike eger gotara kurdî hebe.

Bikaranîn

python pwb.py removegiredan -ns:0 -transcludes:"Girêdan" -async -always

python pwb.py removegiredan -ns:0 -transcludes:"Girêdan" -showdiff

The following parameters are supported:

-always           The bot won't ask for confirmation when putting a page.

-showdiff         The bot will show the differences in the console.

-async            Edits will be performed asynchronously.

Use global -simulate option for test purposes. No changes to live wiki
will be done.

"""
#
# (C) Balyozxane
#
# Distributed under the terms of the MIT license.
#

import pywikibot
from pywikibot import pagegenerators
from pywikibot.bot import (
    AutomaticTWSummaryBot,
    ConfigParserBot,
    ExistingPageBot,
    SingleSiteBot,
)
import mwparserfromhell
from kucosmetics import CANCEL, CosmeticChangesToolkit
import urllib.parse
import requests

VERBOSE = False
TESTING = False

# This is required for the text that is shown when you run this script
# with the parameter -help.
docuReplacements = {'&params;': pagegenerators.parameterHelp}  # noqa: N816


class XebatBot(
    # Refer pywikobot.bot for generic bot classes
    SingleSiteBot,  # A bot only working on one site
    ConfigParserBot,  # A bot which reads options from scripts.ini setting file
    # CurrentPageBot,  # Sets 'current_page'. Process it in treat_page method.
    #                  # Not needed here because we have subclasses
    ExistingPageBot,  # CurrentPageBot which only treats existing pages
    AutomaticTWSummaryBot,  # Automatically defines summary; needs summary_key
):
    use_redirects = False  # treats non-redirects only
    summary_key = 'basic-changing'

    update_options = {
        'async': False,
        'showdiff': False,
        'ignore': CANCEL.MATCH,
    }

    def __init__(self, *args, **kwargs):
        super().__init__(*args, **kwargs)
        # Retrieve redirects for templates
        self.bot_name = "User:Balyozxane/skrîpt/py/removegiredan.py"

        if TESTING:
            self.giredan_redirects = ['Ill', 'Girêdana hevziman', 'Illm', 'G', 'Girêdan']
        else:
            self.giredan_redirects = self.get_template_redirects("Girêdan")

    def get_template_redirects(self, template_title):
        template_title = "Şablon:" + template_title
        template_page = pywikibot.Page(self.site, template_title)
        redirects = template_page.backlinks(filter_redirects=True, namespaces=[10])
        redirect_titles = [redirect.title(with_ns=False) for redirect in redirects]
        redirect_titles.append(template_title.split(":")[-1])

        if VERBOSE:
            print(f"{template_title} redirects:\n{redirect_titles}")
        return redirect_titles

    @staticmethod
    def get_kuwiki_title(lang_code, title):
        encoded_title = urllib.parse.quote(title)
        url = f"https://www.wikidata.org/w/api.php?action=wbgetentities&sites={lang_code}wiki&titles={encoded_title}&props=sitelinks&format=json"
        response = requests.get(url)
        data = response.json()
        entity = next(iter(data["entities"].values()))  # Get the first (and only) entity
        if "sitelinks" in entity and "kuwiki" in entity["sitelinks"]:
            kuwiki_title = entity["sitelinks"]["kuwiki"]["title"]

            if VERBOSE:
                print(f"Kurdish title for '{lang_code}:{title}' fetched: '{kuwiki_title}'")

            kuwiki_title = kuwiki_title.replace('Kategorî:', '')
            return kuwiki_title
        else:
            if VERBOSE:
                print(f"Kurdish title for '{lang_code}:{title}' not found.")
            return None

    def do_kozmetik(self, old_text):
        kozmetik_cebu = ""
        cc_toolkit = CosmeticChangesToolkit(self.current_page,
                                            ignore=self.opt.ignore)
        new_text, summaries = cc_toolkit.change(old_text)
        applied_summaries = ', '.join(summaries.values())
        if new_text is not False and new_text != old_text:
            kozmetik_cebu = "; paqijiyên kozmetîk"
            if applied_summaries:
                kozmetik_cebu += f' ({applied_summaries}.)'

        return new_text, kozmetik_cebu

    def replace_template(self, parsed):
        lang_code = ""
        en_title = ""
        for template in parsed.filter_templates():
            template_name = template.name.strip()
            template_name = template_name[0].upper() + template_name[1:]
            if template_name in self.giredan_redirects:

                if template.has(3):
                    en_title = str(template.get(3).value.strip())

                    if template.has(2):
                        lang_code = str(template.get(2).value.strip())

                    if template.has("lt"):
                        ku_title = str(template.get("lt").value.strip())
                    else:
                        ku_title = str(template.get(1).value.strip())
                else:
                    if template.has(1):
                        en_title = str(template.get(1).value.strip())
                    if template.has(2):
                        lang_code = str(template.get(2).value.strip())

                    if template.has("lt"):
                        ku_title = str(template.get("lt").value.strip())
                    else:
                        ku_title = str(template.get(1).value.strip())

                if ku_title and lang_code and en_title:
                    ku_page = self.get_kuwiki_title(lang_code, en_title)
                    if ku_page:
                        # Construct the replacement text

                        replacement_text = f"[[{ku_page}|{ku_title}]]"

                        parsed.replace(template, replacement_text)

                        if VERBOSE:
                            print("Template replaced successfully!")
                    else:
                        if VERBOSE:
                            print("Kurdish page not found.")
                else:
                    if VERBOSE:
                        print("Incomplete information to replace template.")
        return str(parsed)

    def treat_page(self) -> None:
        
        if self.current_page.namespace() != 0:
            if VERBOSE:
                print("Skipping Namespace not 0.")
            return
        
        text = self.current_page.text
        wikicode = mwparserfromhell.parse(text)

        updated_text = self.replace_template(wikicode)
        if text != updated_text:
            cleaned_new_text, kozmetik_cebu = self.do_kozmetik(updated_text)

            summary = f'[[{self.bot_name}|Bot]]: Şablona {{{{[[Şablon:Girêdan|girêdan]]}}}} hat rakirin{kozmetik_cebu}'

            self.put_current(
                cleaned_new_text,
                summary=summary,
                asynchronous=self.opt['async'],
                show_diff=self.opt['showdiff']
            )


def main(*args: str) -> None:
    """
    Process command line arguments and invoke bot.

    If args is an empty list, sys.argv is used.

    :param args: command line arguments
    """
    options = {}
    # Process global arguments to determine desired site
    local_args = pywikibot.handle_args(args)

    # This factory is responsible for processing command line arguments
    # that are also used by other scripts and that determine on which pages
    # to work on.
    gen_factory = pagegenerators.GeneratorFactory()

    # Process pagegenerators arguments
    local_args = gen_factory.handle_args(local_args)

    # Parse your own command line arguments
    for arg in local_args:
        arg, _, value = arg.partition(':')
        option = arg[1:]
        if option in ('-always', '-async', '-showdiff'):
            options[option[1:]] = True
        elif option == '-ignore':
            value = value.upper()
            try:
                options['ignore'] = getattr(CANCEL, value)
            except AttributeError:
                raise ValueError(f'Unknown ignore mode {value!r}!')
        # take the remaining options as booleans.
        # You will get a hint if they aren't pre-defined in your bot class
        else:
            options[option] = True
    # The preloading option is responsible for downloading multiple
    # pages from the wiki simultaneously.
    gen = gen_factory.getCombinedGenerator(preload=True)

    # check if further help is needed
    if not pywikibot.bot.suggest_help(missing_generator=not gen):
        # pass generator and private options to the bot
        bot = XebatBot(generator=gen, **options)
        bot.run()  # guess what it does


if __name__ == '__main__':
    main()