fpc.py

#!/usr/bin/env python3
"""This script runs as FPCBot on Wikimedia Commons.

It counts the votes in featured picture nominations,
closes and archives finished nominations,
adds freshly promoted featured pictures to the gallery pages,
and informs uploaders and nominators about the success.
It also processes delisting nominations.

Programmed by Daniel78 at Commons.
Updated by Eatcha, KTC, Aristeas and other Commons users.

The script is based on Pywikibot.  Therefore you can use it
with Pywikibot options (so-called global options);
to list them, use '-help:global' or run 'pwb.py -help'.

In addition, the script understands the following
command line arguments for tasks and (local) options:

Tasks:

-help           Print this help and exit.
-info           Print status and vote count info about current nominations.
-close          Close, count votes and add results to finished nominations.
-park           Park closed and verified nominations.
-test           Test vote counting against an old log.
-checkgallery   Test the gallery links in current nominations.

Options:

-auto           Do not ask before committing edits to pages.
-dry            Do not commit any edits, just print them.
-threads        Use threads to speed things up
                (must be used with '-auto' and/or '-dry').
-fpc            Process candidates for featured picture status.
-delist         Process candidates for delisting from featured picture status
                (if neither '-fpc' nor '-delist' is used,
                all candidates are processed).
-notime         Avoid displaying timestamps in log output.
-match pattern  Only operate on nominations matching this pattern.
"""

# Standard library imports
from __future__ import annotations
import sys
import abc
from collections.abc import Callable
from types import FrameType
from typing import Final, Literal, Type, ClassVar, NamedTuple
import signal
import datetime
import time
import re
import unicodedata
import urllib.parse
import threading
import traceback

# Third-party imports
import pywikibot


# CONSTANTS

# Namespaces, prefixes, page names, etc.
BOT_NAME: Final[str] = "FPCBot"
FILE_NAMESPACE: Final[str] = "File:"
USER_NAMESPACE: Final[str] = "User:"
USER_TALK_NAMESPACE: Final[str] = "User talk:"
FP_PREFIX: Final[str] = "Commons:Featured pictures/"
FPC_PAGE: Final[str] = "Commons:Featured picture candidates"
CAND_PREFIX: Final[str] = f"{FPC_PAGE}/"
CAND_LOG_PREFIX: Final[str] = f"{CAND_PREFIX}Log/"
CHRONO_ARCHIVE_PREFIX: Final[str] = f"{FP_PREFIX}chronological/"
CAND_LIST_PAGE_NAME: Final[str] = f"{CAND_PREFIX}candidate list"
TEST_LOG_PAGE_NAME: Final[str] = f"{CAND_LOG_PREFIX}January 2025"
GALLERY_LIST_PAGE_NAME: Final[str] = "Commons:Featured pictures, list"
FP_TALK_PAGE_NAME: Final[str] = "Commons talk:Featured picture candidates"
UNSORTED_HEADING: Final[str] = "Unsorted"
MAX_ENTRIES_PER_LOG_PART: Final[int] = 100


# Valid voting templates
# Taken from Commons:Polling_templates, including some common redirects
SUPPORT_TEMPLATES: Final[tuple[str, ...]] = (
    "[Ss]upport",
    "[Ss]upp?",
    "[Ss]",
    "[Vv]ote[ _]support",
    "[Ss]trong[ _]support",
    "[Ss]Support",
    "[Ww]eak[ _]support",
    "[Ww]eak[ _][Ss]",
    "[Ww]Support",
    # Variants and translations
    "[Aa][ _]favore?",
    "[Aa][ _]favuri",
    "[Aa]poio",
    "[AaÁá][ _]fabor",
    "[Dd]estek",
    "[Kk]yllä",
    "[Oo]ui",
    "[Pp]our",
    "[Pp]ro",
    "[Pp]RO",
    "[Ss][Ff]or",
    "[Ss]amþykkt",
    "[Ss]im",
    "[Ss]tödjer",
    "[Ss]ubteno",
    "[Ss]up",
    "[Ss]í",
    "[Tt]acaíocht",
    "[Tt]ak",
    "[Tt]aurä",
    "[Vv]oor",
    "[Yy]es",
    "[Υυ]πέρ",
    "[Зз]а",
    "[Пп]адтрымліваю",
    "דעב",
    "เห็นด้วย",
    "ჰო",
    "支持",
    "賛成",
    "찬성",
)
OPPOSE_TEMPLATES: Final[tuple[str, ...]] = (
    "[Oo]ppose",
    "[Oo]pp",
    "[Oo]",
    "[Ss]trong[ _]oppose",
    "[Ss]Oppose",
    "[Ww]eak[ _]oppose",
    "[Ww]eak[ _][Oo]",
    "[Ww]eako",
    # {{FPX contested}} is counted like a normal {{oppose}} vote
    "FPX[ _]contested",
    # Variants and translations
    "[Cc]ontr[aeo]",
    "[Cc]ontrario",
    "[Cc]untrariu",
    "[Dd]iscordo",
    "[Dd]íliostaigh",
    "[Ee]i",
    "[Ee]n[ _]contra",
    "[Ii][ _]gcoinne",
    "[Kk]ar[sş]i",
    "[Kk]ontra",
    "[Kk]ontraŭi",
    "[Ll]ivari",
    "[Mm]autohe",
    "[Mm]ot",
    "[Nn]ein",
    "[Nn]ie",
    "[Nn]o",
    "[Nn]ão",
    "[Oo]pponera",
    "[Ss]tödjer[ _]ej",
    "[Tt]egen",
    "[Áá][ _]móti",
    "[Пп]ротив",
    "[Сс]упраць",
    "נגד",
    "ไม่เห็นด้วย",
    "反対",
    "除外",
    "반대",
)
NEUTRAL_TEMPLATES: Final[tuple[str, ...]] = (
    "[Nn]eutral",
    "[Nn]eu",
    "[Nn]",
    "[Vv]n",
    # Variants and translations
    "[Hh]lutlaus",
    "[Nn]eodrach",
    "[Nn]eutr[aeo]",
    "[Nn]eutrale",
    "[Nn]eŭtrala",
    "[Nn]iutrali",
    "[Nn]øytral",
    "[Oo]partisk",
    "[Tt]arafs[iı]z",
    "[Вв]оздерживаюсь",
    "[Вв]оздржан",
    "[Нн]эўтральна",
    "נמנע",
    "เป็นกลาง",
    "中立",
    "중립",
)
DELIST_TEMPLATES: Final[tuple[str, ...]] = (
    "[Dd]elist",
    # There are no translated versions of this template.
    # Don't add {{Remove}} or {{Del}}, they are for deletion discussions.
)
DELIST_AND_REPLACE_TEMPLATES: Final[tuple[str, ...]] = (
    "[Dd]elistandreplace",
    "[Dd]elist[ _]and[ _][Rr]eplace",
    "[Dd]ar",
    "[Dd]&R",
    # There are no translated versions.
)
KEEP_TEMPLATES: Final[tuple[str, ...]] = (
    "[Kk]eep",
    "[Kk]",
    "[Vv]ote[ _]keep",
    "[Vv]keep",
    "[Vv]k",
    "[Ss]trong[ _]keep",
    "[Ww]eak[ _]keep",
    "[Ww]k",
    "[Vv]wk",
    # Variants and translations
    "[Bb]ehalten",
    "[Bb]ehold",
    "[Bb]ehåll",
    "[Cc]onserver",
    "[Gg]arder",
    "[Mm]antenere?",
    "[Mm]anter",
    "[Ss]avi",
    "[Зз]адржи",
    "เก็บ",
    "保留",
    "維持",
)


# Shared messages

NOMINATION_SUBPAGE_RENAMED: Final[str] = (
    "Fixed nomination subpage name because it could cause problems "
    "with [[Template:Assessments]] and other tools."
)
PLEASE_FIX_HINT: Final[str] = (
    "Please check and fix this so that the bot can process the nomination."
)
PLEASE_RENAME_HINT: Final[str] = (
    "Please rename [[{subpage}]] to a suitable name "
    "so that the bot can process the nomination."
)
SERIOUS_PROBLEM_CHECK_PAGE: Final[str] = (
    "This is a serious problem, please check that page."
)
PLEASE_CHECK_GALLERY_AND_SORT_FPS: Final[str] = (
    "Please check that gallery page and add the new featured picture(s) "
    "from the nomination to the appropriate gallery page."
)
LIST_INCLUDES_MISSING_SUBPAGE: Final[str] = (
    "The candidate list [[{list}]] includes the nomination [[{subpage}]], "
    "but that page does not exist. Perhaps the page has been renamed "
    f"and the list needs to be updated. {PLEASE_FIX_HINT}"
)
COULD_NOT_READ_RECENT_FPS_LIST: Final[str] = (
    f"The bot could not read [[{GALLERY_LIST_PAGE_NAME}|the list]] "
    "of recent featured pictures: {exception}. "
    "Please check the list page and fix it."
)
ADDING_FPS_TO_UNSORTED_SECTION: Final[str] = (
    "Therefore one or more new featured pictures are added "
    "to the ''Unsorted'' section at the bottom of [[{page}]]. "
    "Please sort these images into the correct section."
)
CHECK_DELISTED_FP_AND_UPDATE_ASSESSMENTS: Final[str] = (
    "Please check that page and update the parameters "
    "of the {{{{tl|Assessments}}}} template: "
    "replace <code>featured=1</code> by <code>featured=2</code> "
    "and add or update the <code>com-nom</code> parameter "
    "with the value <code>{subpage}</code>."
)
PLEASE_ADD_FP_ASSESSMENT_CLAIM: Final[str] = (
    "Please add the claim [[:wikidata:Special:EntityPage/P6731|"
    "Commons quality assessment (P6731)]]: "
    "[[:wikidata:Special:EntityPage/Q63348049|"
    "Wikimedia Commons featured picture (Q63348049)]] "
    "to the Structured data of {files}."
)


# Regular expressions

# Building patterns
CAND_PREFIX_PATTERN: Final[str] = CAND_PREFIX.replace(" ", r"[ _]")
MIDDLE_NOMINATION_NAME_PATTERN: Final[str] = (
    r" *(?:[Rr]emoval */ *)?(?:[Ss]et */|(?:[Ff]ile|[Ii]mage) *:) *"
)

# Identify reasonably valid FP nomination subpage names
VALID_NOMINATION_NAME_START_REGEX: Final[re.Pattern] = re.compile(
    f"^{CAND_PREFIX_PATTERN}{MIDDLE_NOMINATION_NAME_PATTERN}"
)
# Find or remove the candidate prefix
CAND_PREFIX_REGEX: Final[re.Pattern] = re.compile(CAND_PREFIX_PATTERN)
# Remove the middle part of nomination names (after deleting the prefix)
MIDDLE_NOMINATION_NAME_REGEX: Final[re.Pattern] = re.compile(
    MIDDLE_NOMINATION_NAME_PATTERN
)
# Remove or replace the candidate prefix and the 'File:'/'Image:' namespace,
# plus any possible crap between the prefix and the namespace
# and faulty spaces before and after the namespace.
FULL_FILE_PREFIX_REGEX: Final[re.Pattern] = re.compile(
    f"^{CAND_PREFIX_PATTERN}" + r".*?(?:[Ff]ile|[Ii]mage) *: *"
)

# Look for results using the old, text-based results format
# which was in use until August 2009.  An example of such a line is:
# '''result:''' 3 support, 2 oppose, 0 neutral => not featured.
OBSOLETE_RESULT_REGEX: Final[re.Pattern] = re.compile(
    r"'''[Rr]esult:'''\s+(\d+)\s+support,\s+(\d+)\s+oppose,\s+(\d+)\s+neutral"
    r"\s*=>\s*((?:not )?featured)"
)
OBSOLETE_DELIST_RESULT_REGEX: Final[re.Pattern] = re.compile(
    r"'''[Rr]esult:'''\s+(\d+)\s+delist,\s+(\d+)\s+keep,\s+(\d+)\s+neutral"
    r"\s*=>\s*((?:not )?delisted)"
)

# Look for verified results using the new, template-based format
VERIFIED_RESULT_REGEX: Final[re.Pattern] = re.compile(
    r"""
    \{\{\s*FPC-results-reviewed\s*\|
    \s*support\s*=\s*(\d+)\s*\|            # (1) Support votes
    \s*oppose\s*=\s*(\d+)\s*\|             # (2) Oppose votes
    \s*neutral\s*=\s*(\d+)\s*\|            # (3) Neutral votes
    \s*featured\s*=\s*(\w+)\s*\|           # (4) Featured, should be 'yes'/'no'
    \s*gallery\s*=\s*([^|\n]*)             # (5) Gallery link (if featured)
    (?:\|\s*alternative\s*=\s*([^|\n]*))?  # (6) For candidates with alternatives:
    .*?\}\}                                #     name of the winning image
    """,
    flags=re.VERBOSE,
)
VERIFIED_DELIST_RESULT_REGEX: Final[re.Pattern] = re.compile(
    r"""
    \{\{\s*FPC-delist-results-reviewed\s*\|
    \s*delist\s*=\s*(\d+)\s*\|   # (1) Delist votes
    \s*keep\s*=\s*(\d+)\s*\|     # (2) Keep votes
    \s*neutral\s*=\s*(\d+)\s*\|  # (3) Neutral votes
    \s*delisted\s*=\s*(\w+)      # (4) Delisted, should be 'yes'/'no'
    .*?\}\}
    """,
    flags=re.VERBOSE,
)

# Simple regexes which check just whether a certain template is present or not
COUNTED_TEMPLATE_REGEX: Final[re.Pattern] = re.compile(
    r"\{\{\s*FPC-results-unreviewed.*?\}\}"
)
REVIEWED_TEMPLATE_REGEX: Final[re.Pattern] = re.compile(
    r"\{\{\s*FPC-results-reviewed.*?\}\}"
)
DELIST_COUNTED_TEMPLATE_REGEX: Final[re.Pattern] = re.compile(
    r"\{\{\s*FPC-delist-results-unreviewed.*?\}\}"
)
DELIST_REVIEWED_TEMPLATE_REGEX: Final[re.Pattern] = re.compile(
    r"\{\{\s*FPC-delist-results-reviewed.*?\}\}"
)

# Find voting templates
VOTING_TEMPLATE_MODEL: Final[str] = r"\{\{\s*(?:%s)\s*(\|.*?)?\s*\}\}"
SUPPORT_VOTE_REGEX: Final[re.Pattern] = re.compile(
    VOTING_TEMPLATE_MODEL % "|".join(SUPPORT_TEMPLATES)
)
OPPOSE_VOTE_REGEX: Final[re.Pattern] = re.compile(
    VOTING_TEMPLATE_MODEL % "|".join(OPPOSE_TEMPLATES)
)
NEUTRAL_VOTE_REGEX: Final[re.Pattern] = re.compile(
    VOTING_TEMPLATE_MODEL % "|".join(NEUTRAL_TEMPLATES)
)
DELIST_VOTE_REGEX: Final[re.Pattern] = re.compile(
    VOTING_TEMPLATE_MODEL % "|".join(DELIST_TEMPLATES)
)
DELIST_AND_REPLACE_VOTE_REGEX: Final[re.Pattern] = re.compile(
    VOTING_TEMPLATE_MODEL % "|".join(DELIST_AND_REPLACE_TEMPLATES)
)
KEEP_VOTE_REGEX: Final[re.Pattern] = re.compile(
    VOTING_TEMPLATE_MODEL % "|".join(KEEP_TEMPLATES)
)

# Does the nomination contain a {{Withdraw(n)}}, {{FPX}} or {{FPD}} template?
WITHDRAWN_FPX_FPD_REGEX: Final[re.Pattern] = re.compile(
    r"\{\{\s*([Ww](?:ithdrawn?|dn)|[Ff]P[XD])\s*(\|.*?)?\}\}"
)
# Does the nomination contain subheadings = subsections?
SECTION_REGEX: Final[re.Pattern] = re.compile(
    r"^={1,4}.+={1,4}\s*$", flags=re.MULTILINE
)
# Find or count displayed images
IMAGES_REGEX: Final[re.Pattern] = re.compile(
    # Because Mediawiki page names can't contain '|', '[', ']', etc.,
    # capturing group 2 stops at the end of the actual image filename,
    # regardless whether there is a '|...' or not.
    r"(\[\[((?:[Ff]ile|[Ii]mage):[^|\]]+).*?\]\])"
)
# Look for a size specification in the image link
IMAGE_SIZE_REGEX: Final[re.Pattern] = re.compile(r"\|.*?(\d+)\s*px")
# Check if there is a 'thumb' parameter in the image link
IMAGE_THUMB_REGEX: Final[re.Pattern] = re.compile(r"\|\s*thumb\b")
# Search nomination for the username of the original creator
CREATOR_NAME_REGEX: Final[re.Pattern] = re.compile(
    r"\{\{[Ii]nfo\}\}.+?"
    r"[Cc]reated +(?:(?:and|\&) +(?:[a-z]+ +)?uploaded +)?by +"
    r"\[\[[Uu]ser:([^|\]\n]+)[|\]]"
)
# Search for first entry in (or end of) <gallery>...</gallery> contents
GALLERY_ENTRY_START_REGEX: Final[re.Pattern] = re.compile(r">(?: *\n)*")

# Find the {{Assessments}} template on an image description page
# (sometimes people break it into several lines, so use '\s' and re.DOTALL)
ASSESSMENTS_TEMPLATE_REGEX: Final[re.Pattern] = re.compile(
    r"\{\{\s*[Aa]ssessments\s*(\|.*?)\}\}", flags=re.DOTALL
)

# Find the associated <gallery> element right after the target subheading.
# Between subheading and <gallery> we allow only:
# 1. a single line (e.g. a 'See also' hint) which must not be a subheading,
# <gallery>, etc.; or
# 2. a comment which can span several lines.
ASSOC_GALLERY_ELEMENT_REGEX: Final[re.Pattern] = re.compile(
    r"((?:[^<= \n][^\n]+\n\s*| *<!--.+?-->\s*)?<gallery\b[^>]*>)\s*", flags=re.DOTALL
)

# Find 'Featured pictures of/from/by ...' categories which must be removed
# if a FP is delisted.  If the category is followed by a NL, remove it, too.
# NB: We must not touch project-specific categories like 'Featured pictures
# on Wikipedia, <language>', hence the negative lookahead assertion '(?!on )'.
TOPICAL_FP_CATEGORY_REGEX: Final[re.Pattern] = re.compile(
    r"\[\[[Cc]ategory: *Featured (?:"
    r"pictures (?!on ).+?"
    r"|(?:[a-z -]+)?photo(?:graphs|graphy|s).*?"
    r"|(?:diagrams|maps).*?"
    r")\]\] *\n?"
)


# Dealing with gallery links

VALID_BASIC_GALLERY_LINKS: Final[set[str]] = {
    # Valid first elements of gallery links
    # All entries have been converted to lowercase to ignore case variants.
    # Keep this list in sync with the actual FP gallery pages!
    "animals",
    # "animated",  # Deprecated, must not be used for new nominations.
    "astronomy",
    "food and drink",
    "fungi",
    "historical",
    "natural phenomena",
    "non-photographic media",
    "objects",
    "other lifeforms",
    "people",
    "photo techniques",
    "places",
    "plants",
    # "sets",  # Deprecated, must not be used for new nominations.
    "space exploration",
    "sports",
}
OBJECT_SECTS_IN_ARCH_ELEMENTS_REGEX: Final[re.Pattern] = re.compile(
    # When we want to find the matching candidate archive category by subject,
    # the gallery 'Objects/Architectural elements' is very difficult --
    # most entries clearly belong to architecture (e.g. the ceilings),
    # others are traditionally sorted under 'Objects', e.g. stained-glass
    # windows, some are mixed (e.g., most FPs of doors and windows belong
    # to architecture, but some show details of wood carving etc.).
    # We do our best by analyzing the section, if present.
    r"(?:fences|floors|ground|ornaments|railings|stained[ _-]glass)",
    flags=re.IGNORECASE,
)


# Creating sort keys for use with candidate archive categories

AUXILIARY_SORT_KEY_TRSL_TABLE: Final[dict[int, str]] = str.maketrans(
    # Some frequent auxiliary character translations for sort keys
    {
        "Æ": "AE", "æ": "ae",
        "Œ": "OE", "œ": "oe",
        "Đ": "D", "đ": "d",  # U+0110, != U+00D0 (eth)
        "Ł": "L", "ł": "l",
        "Ø": "O", "ø": "o",
        "ß": "ss",
    }
)
RUN_OF_WHITESPACE_REGEX: Final[re.Pattern] = re.compile(r"\s+")
REMOVE_FROM_SORT_KEY_REGEX: Final[re.Pattern] = re.compile(
    # Common punctuation marks etc. which should be removed from the sort key.
    r"['\"`´„“”‚‘’«»‹›\u00AD\u200B\u200C\u200D]+"
)
REPLACE_BY_SPACE_IN_SORT_KEY_REGEX: Final[re.Pattern] = re.compile(
    # Common punctuation marks which should be replaced by a simple space
    # in the sort key because they usually have separating character.
    r"[,:;?!\-()/+*^~¿¡–—…·•†\u00A0\u037E\u0387]+"
)
REMOVE_FROM_START_OF_SORT_KEY_REGEX: Final[re.Pattern] = re.compile(
    # Characters which should be removed completely only from the beginning
    # of the sort key, in any count and order, including any whitespace
    # before/between/after them.  Here we also remove '.'.
    r"^[\s.]+"
)
REMOVE_SPACE_BEFORE_INTERPT_REGEX: Final[re.Pattern] = re.compile(
    # Remove some irritating spaces before remaining punctuation marks.
    # Because this regex should be used only *after* the preceding ones,
    # we handle only the remaining punctuation mark '.'.
    r" (?=\.[^.])"
)


# GLOBALS

# Apply all changes without asking the user
_g_auto: bool = False
# Dry run, do not save any changes
_g_dry: bool = False
# Use threads
_g_threads: bool = False
# Don't print timestamps in CLI/log output
_g_log_no_time: bool = False
# Process only nominations with a name matched by this pattern
_g_match_pattern: str = ""
# Flag that will be set to True if CTRL-C was pressed
_g_abort: bool = False
# Pywikibot Site object
_g_site: pywikibot.site.BaseSite | None = None
# Number of the current log part
_g_log_part_no: int | None = None


# CLASSES

class DataAlreadyPresentError(Exception):
    """The data we wanted to add are already present on the page.

    This can happen if the process has previously been interrupted.
    """
    pass


class CouldNotAddDataError(Exception):
    """Error during data insertion on a Commons page."""
    pass


class CandidateTypesToProcess(NamedTuple):
    """Class keeping track of the types of nominations we want to process.

    Attributes:
        fp: Boolean indicating whether to process FP candidates or not.
        delist: Boolean indicating whether to process delisting candidates.
    """

    # Declare types of instance variables
    fp: bool
    delist: bool

    def candidate_class(self, subpage_name: str) -> Type[Candidate] | None:
        """Find out if should we process a nomination, and with which class.

        Args:
            subpage_name: Full name of a nomination subpage.

        Returns:
            The correct Candidate subclass to be used for that nomination,
            or None if we should not process that type of nomination.
        """
        if re.search(r"/ *[Rr]emoval */", subpage_name):
            return DelistCandidate if self.delist else None
        # Default:
        return FPCandidate if self.fp else None

    def describe(self) -> str:
        """Summarize the types of candidates we want to process."""
        types = [
            entry[1] for entry in zip(self, self._fields)  # pylint: disable=no-member
            if entry[0]
        ]
        return f"{', '.join(types)} candidates"


class ThreadCheckCandidate(threading.Thread):
    """A thread executing one of the bot's task on a single candidate."""

    # Declare types of instance variables
    _candidate: Candidate  # The candidate/nomination handled in this thread
    _check: Callable[[Candidate], None]  # The method for the desired task

    def __init__(
        self,
        candidate: Candidate,
        check: Callable[[Candidate], None],
    ) -> None:
        """Initialize the thread.

        Args:
            candidate: The Candidate subclass object to process.
            check: The Candidate class method which should be called.
        """
        super().__init__()
        self._candidate = candidate
        self._check = check

    def run(self) -> None:
        """Execute the desired task for the candidate."""
        self._check(self._candidate)
        self._candidate.clear_cache()


class Candidate(abc.ABC):
    """A featured picture candidate (nomination)

    This abstract base class bundles all common attributes and methods.
    All individual candidates (nominations) are represented by instances
    of the concrete subclasses.
    """

    # Define class constants
    # (these are the values for a normal FP nomination,
    # subclasses must adapt them as needed)
    # Keyword for the title etc. of a successful nomination:
    _SUCCESS_KEYWORD: ClassVar[str] = "featured"
    # Keyword for the title etc. of a failed nomination:
    _FAIL_KEYWORD: ClassVar[str] = "not featured"
    # Compiled regex to find positive votes in the nomination:
    _PRO_VOTE_REGEX: ClassVar[re.Pattern] = SUPPORT_VOTE_REGEX
    # Compiled regex to find negative votes in the nomination:
    _CONTRA_VOTE_REGEX: ClassVar[re.Pattern] = OPPOSE_VOTE_REGEX
    # Compiled regex to find neutral votes in the nomination:
    _NEUTRAL_VOTE_REGEX: ClassVar[re.Pattern] = NEUTRAL_VOTE_REGEX
    # Compiled regex to find templates containing unreviewed results:
    _COUNTED_RES_REGEX: ClassVar[re.Pattern] = COUNTED_TEMPLATE_REGEX
    # Compiled regex to find templates containing reviewed results:
    _REVIEWED_RES_REGEX: ClassVar[re.Pattern] = REVIEWED_TEMPLATE_REGEX
    # Compiled regex to analyse the template with reviewed results:
    _VERIFIED_RES_REGEX: ClassVar[re.Pattern] = VERIFIED_RESULT_REGEX
    # Compiled regex to analyse the obsolete format for reviewed results:
    _OBSOLETE_RES_REGEX: ClassVar[re.Pattern] = OBSOLETE_RESULT_REGEX

    # Declare types of instance variables
    _list_page_name: str  # Name of the candidates list page on Commons
    _page: pywikibot.Page  # The nomination subpage for this candidate
    _filtered_content: str | None  # Text of the nomination w/o comments etc.
    _creation_time: datetime.datetime | None  # Creation time of the nomination
    _days_old: int  # How many days passed since the nomination was created?
    _days_since_last_edit: int  # How many days passed since it was edited?
    _creator: str | None  # Username of the creator of the nominated image
    _uploader: dict[str, str]  # Mapping: filename -> username of uploader
    _nominator: str | None  # Username of the creator of the nomination
    _image_count: int | None  # Count of (full-size) images in the nomination
    _filename: str | None  # Name of the nominated image, empty if not found
    _alternative: str | None  # If there are alternatives: selected image name
    _set_files: list[str] | None  # Names of nominated images (for set noms)
    _pro: int  # Count of pro votes
    _con: int  # Count of contra votes
    _neu: int  # Count of neutral votes

    def __init__(self, page: pywikibot.Page, list_name: str) -> None:
        """Initialize the candidate object.

        Although this is an abstract base class, the initializer is called
        by the concrete subclasses.

        Args:
            page: A pywikibot.Page object for the nomination subpage.
            list_name: A string with the name of the candidate list page.
        """
        # Save passed values
        self._list_page_name = list_name
        self._page = page
        # Set other instance variables to default values
        self._filtered_content = None
        self._creation_time = None
        self._days_old = -1
        self._days_since_last_edit = -1
        self._creator = None
        self._uploader: dict[str, str] = {}
        self._nominator = None
        self._image_count = None
        self._filename = None
        self._alternative = None
        self._set_files = None
        self._pro = -1
        self._con = -1
        self._neu = -1

    @property
    def page(self) -> pywikibot.Page:
        """Return the nomination subpage for this candidate."""
        return self._page

    def print_all_info(self) -> None:
        """Print a line with current status information about this candidate.

        This method is used to generate the '-info' overview of all
        open nominations.
        """
        try:
            self.count_votes()
            out(
                f"{self.cut_title()}: "
                f"{self._type_code()} "
                f"P:{self._pro:02d} "
                f"C:{self._con:02d} "
                f"N:{self._neu:02d} "
                f"Do:{self.days_old():02d} "
                f"De:{self.days_since_last_edit():02d} "
                f"Se:{self.section_count():02d} "
                f"Im:{self.image_count():02d} "
                f"W:{y_n(bool(self.was_cancelled()))} "
                f"S:{'?' if self.is_ignored() else y_n(self.is_passed())} "
                f"({self.status_string()})"
            )
        except pywikibot.exceptions.NoPageError:
            error(f"{self.cut_title()}: -- No such page --")

    def filtered_content(self) -> str:
        """Return a filtered version of the wikitext of the nomination subpage.

        The filtered version omits all comments, stricken text, code examples,
        etc.; it is used for counting the votes etc., and cached because
        we need it often.  If you change the text of the nomination subpage,
        call reset_filtered_content() to flag the cached version as outdated.
        """
        if self._filtered_content is None:
            self._filtered_content = filter_content(self._page.get(get_redirect=False))
        return self._filtered_content

    def reset_filtered_content(self) -> None:
        """Flag filtered content as outdated after changing page contents."""
        self._filtered_content = None

    def clear_cache(self) -> None:
        """Clear all cached page contents."""
        self._page.clear_cache()
        self._filtered_content = None

    def creator(self, link: bool) -> str:
        """Return the name of the user who has originally created the image(s).

        There is no generally applicable way to determine the creator.
        Therefore nominators should use the phrase
            '{{Info}} ... created by [[User:...]]'
        on the nomination subpage in order to identify the original creator
        (we also support some common variants, see the regex constant).

        Args:
            link: Pass True to get a link to the user page, False to get
                just the plain username.

        Returns:
            If the expected phrase is found, the function returns the username
            resp. a link to the user page, else just ''.
        """
        if self._creator is None:
            if match := CREATOR_NAME_REGEX.search(self.filtered_content()):
                self._creator = match.group(1).replace("_", " ").strip()
            else:
                self._creator = ""
        if self._creator and link:
            return user_page_link(self._creator)
        return self._creator

    def uploader(self, filename: str, link: bool) -> str:
        """Return the name of the user who has originally uploaded the image.

        This method works differently than nominator() because all files of
        a set must have the same nominator, but can have different uploaders,
        therefore we need to specify the individual filename.

        Args:
            filename: The name of the image on Commons.
            link: Pass True to get a link to the user page, False to get
                just the plain username.

        Returns:
            The username resp. a link to the user page; on errors 'Unknown'.
        """
        try:
            username = self._uploader[filename]
        except KeyError:
            username = oldest_revision_user(pywikibot.Page(_g_site, filename))
            self._uploader[filename] = username
        if username:
            return user_page_link(username) if link else username
        return "Unknown"

    def nominator(self, link: bool) -> str:
        """Return the name of the user who has created the nomination subpage.

        Args:
            link: Pass True to get a link to the user page, False to get
                just the plain username.

        Returns:
            The username resp. a link to the user page; on errors 'Unknown'.
        """
        if self._nominator is None:
            self._nominator = oldest_revision_user(self._page)
        if self._nominator:
            return user_page_link(self._nominator) if link else self._nominator
        return "Unknown"

    def is_set(self) -> bool:
        """Find out if this candidate is a set nomination."""
        return re.search(r"/ *[Ss]et */", self._page.title()) is not None

    def set_files(self) -> list[str]:
        """Return the names of all nominated images from a set nomination.

        The method tries to retrieve the filenames from a <gallery> element
        in the nomination subpage, checks if the individual files exist
        and resolves any redirects.
        Problems are reported on the FPC talk page.
        For single-file nominations, use filename() instead.

        Returns:
            A list with the names of all nominated image files;
            on errors the result is an empty list.
        """
        # Use cached result if possible
        if self._set_files is not None:
            return self._set_files
        # Change default value from None (meaning 'not retrieved yet')
        # to an empty list (meaning 'files not OK'; used if we return early)
        self._set_files = []

        # Extract contents of the first <gallery>...</gallery> element
        subpage_name = self._page.title()
        cut_title = self.cut_title()
        match = re.search(
            r"<gallery[^>]*>(.+?)</gallery>",
            self.filtered_content(),
            flags=re.DOTALL,
        )
        if not match:
            error(f"{cut_title}: (Error: no <gallery> in set nomination)")
            ask_for_help(
                f"In the set nomination [[{subpage_name}]], the bot "
                "did not find the <code><nowiki><gallery></nowiki></code> "
                "element with the nominated images. "
                f"Perhaps the formatting is damaged. {PLEASE_FIX_HINT}"
            )
            return []
        text_inside_gallery = match.group(1)

        # First try to find files which are properly listed with 'File:'
        # or 'Image:' prefix; they must be the first element on their line,
        # but leading whitespace is tolerated:
        files_list = re.findall(
            r"^ *(?:[Ff]ile|[Ii]mage) *:([^\n|]+)",
            text_inside_gallery,
            flags=re.MULTILINE,
        )
        if not files_list:
            # If we did not find a single file, let's try a casual search
            # for lines which, ahem, seem to start with an image filename:
            files_list = re.findall(
                r"^ *([^|\n:<>\[\]]+\.(?:jpe?g|tiff?|png|svg|webp|xcf))",
                text_inside_gallery,
                flags=re.MULTILINE | re.IGNORECASE,
            )
        if not files_list:
            # Still no files found, so we must skip this candidate
            error(f"{cut_title}: (Error: found no images in set)")
            ask_for_help(
                f"The set nomination [[{subpage_name}]] seems to contain no images. "
                f"Perhaps the formatting is damaged. {PLEASE_FIX_HINT}"
            )
            return []

        # Format and check filenames, resolve any redirects
        for i, filename in enumerate(files_list, start=0):
            # Add (back) the 'File:' prefix (our search omits the prefixes)
            filename = f"{FILE_NAMESPACE}{filename.replace('_', ' ').strip()}"
            page = pywikibot.Page(_g_site, filename)
            if not page.exists():
                # File not found, skip this candidate
                error(f"{cut_title}: (Error: can't find set image '{filename}')")
                ask_for_help(
                    f"The set nomination [[{subpage_name}]] lists the image "
                    f"[[:{filename}]], but that image file does not exist. "
                    f"Perhaps it has been renamed or deleted. {PLEASE_FIX_HINT}"
                )
                return []
            if page.isRedirectPage():
                try:
                    page = page.getRedirectTarget()
                except pywikibot.exceptions.PageRelatedError:
                    # Circular or invalid redirect etc., skip candidate
                    error(
                        f"{cut_title}: (Error: invalid redirect "
                        f"in set image '{filename}')"
                    )
                    ask_for_help(
                        f"The set nomination [[{subpage_name}]] lists the image "
                        f"[[:{filename}]], but the image page contains "
                        f"a circular or invalid redirect. {PLEASE_FIX_HINT}"
                    )
                    return []
                if not page.exists():
                    # Broken redirect, skip candidate
                    error(
                        f"{cut_title}: (Error: broken redirect "
                        f"in set image '{filename}')"
                    )
                    ask_for_help(
                        f"The set nomination [[{subpage_name}]] lists the image "
                        f"[[:{filename}]], but the image page redirects to a file "
                        f"or page which does not exist. {PLEASE_FIX_HINT}"
                    )
                    return []
                out(f"Resolved redirect: '{filename}' -> '{page.title()}'.")
                filename = page.title()  # Update filename.
            files_list[i] = filename

        # Save and return files list
        self._set_files = files_list
        return files_list

    def find_gallery_of_file(self) -> str:
        """Find and polish the gallery link in the nomination subpage."""
        if match := re.search(
            r"Gallery[^\n]+?\[\[[Cc]ommons: ?[Ff]eatured[_ ]pictures ?\/([^\n\]]+)",
            self.filtered_content(),
        ):
            return clean_gallery_link(match.group(1))
        return ""

    def count_votes(self) -> None:
        """Count all votes in this nomination."""
        if self._pro > -1:
            return  # Votes are already counted.
        if text := self.filtered_content():
            self._pro = len(self._PRO_VOTE_REGEX.findall(text))
            self._con = len(self._CONTRA_VOTE_REGEX.findall(text))
            self._neu = len(self._NEUTRAL_VOTE_REGEX.findall(text))
        else:
            error(f"Error - '{self._page.title()}' has no real content")

    def was_cancelled(self, uppercase: bool = False) -> str | Literal[False]:
        """Find out if the nomination has been withdrawn, FPXed or FPDed.

        Nominators can withdraw any of their nominations by adding
        the template {{Withdraw}} to it.
        Users can mark a nomination as hopeless with the {{FPX}} template
        or deny a nomination (because the nominator has exceeded the limit
        for simultaneous nominations) with the {{FPD}} template.

        Args:
            uppercase: Should the return value start with an uppercase letter
                if it is a string?

        Returns:
            A short string which denotes that the nomination has been
            withdrawn, FPXed or FPDed; or False if this is not the case.
        """
        if result := WITHDRAWN_FPX_FPD_REGEX.search(self.filtered_content()):
            # Distinguish templates on the basis of the name's last letter
            match result.group(1)[-1]:
                case "X":
                    return "FPXed"
                case "D":
                    return "FPDed"
                case _:
                    return "Withdrawn" if uppercase else "withdrawn"
        return False

    def rules_of_fifth_day(self) -> bool:
        """Find out if the rules of the 5th day apply to this nomination.

        The rules of the 5th day allow to close a nomination already
        at the 5th day after its creation if it is hopeless (less than two
        support votes) or a clear winner (10 or more support votes,
        no oppose votes).  They do not apply to nominations with alternatives
        because with them the voters' favour can change at any time.
        """
        if self.days_old() < 5:
            return False
        # Rules of the fifth day don't apply to nominations with alternatives
        if self.image_count() > 1:
            return False
        self.count_votes()
        # First rule of the fifth day
        if self._pro <= 1:
            return True
        # Second rule of the fifth day
        if self._pro >= 10 and self._con == 0:
            return True
        # If we arrive here, no rule applies
        return False

    def close(self) -> None:
        """Close the nomination if it is finished.

        If the nomination is finished, the function adds a provisional result
        to the nomination subpage.
        Withdrawn, FPXed and FPDed nominations are just moved to the FPC log.
        """
        subpage_name = self._page.title()
        cut_title = self.cut_title()

        # First make sure that the page actually exists
        if not self._page.exists():
            error(f"{cut_title}: (Error: no such page?!)")
            ask_for_help(
                LIST_INCLUDES_MISSING_SUBPAGE.format(
                    list=self._list_page_name, subpage=subpage_name
                )
            )
            return

        # Close a withdrawn or FPXed/FPDed nomination if at least one full day
        # has passed since the last edit
        if reason := self.was_cancelled():
            old_enough = self.days_since_last_edit() > 0
            action = "closing" if old_enough else "but waiting a day"
            out(f"{cut_title}: {reason}, {action}")
            if old_enough:
                self.move_to_log(reason)
            return

        # Is the nomination still active?
        fifth_day = self.rules_of_fifth_day()
        if not self.is_done() and not fifth_day:
            out(f"{cut_title}: (still active, ignoring)")
            return

        # Is there any other reason not to close the nomination?
        try:
            filtered_text = self.filtered_content()
        except pywikibot.exceptions.PageRelatedError as exc:
            error(f"{cut_title}: (Error: is not readable)")
            ask_for_help(
                f"The bot could not read the nomination subpage [[{subpage_name}]]: "
                f"{format_exception(exc)}. {PLEASE_FIX_HINT}"
            )
            return
        if not filtered_text:
            error(f"{cut_title}: (Error: has no real content)")
            ask_for_help(
                f"The nomination subpage [[{subpage_name}]] seems to be empty. "
                f"{PLEASE_FIX_HINT}"
            )
            return
        if re.search(r"\{\{\s*FPC-closed-ignored.*\}\}", filtered_text):
            out(f"{cut_title}: (marked as ignored, so ignoring)")
            return
        if self._COUNTED_RES_REGEX.search(filtered_text):
            out(f"{cut_title}: (needs review, ignoring)")
            return
        if self._REVIEWED_RES_REGEX.search(filtered_text):
            out(f"{cut_title}: (already closed and reviewed, ignoring)")
            return

        # OK, we should close the nomination
        if self.image_count() <= 1:
            self.count_votes()
        old_text = self._page.get(get_redirect=False)
        new_text = old_text.rstrip() + "\n\n" + self.get_result_string()
        if self.image_count() <= 1:
            new_text = self.fix_heading(new_text)

        # Save the new text of the nomination subpage
        summary = self.get_close_edit_summary(fifth_day)
        commit(old_text, new_text, self._page, summary)
        self.reset_filtered_content()

    def fix_heading(self, text: str, value: str | None = None) -> str:
        """Append a result keyword to the heading of the nomination subpage.

        The function appends the keyword '(not) featured' or '(not) delisted'
        to the heading of the nomination subpage, depending on the result.
        It reports if the nomination does not start correctly with a heading.

        Args:
            text: The complete wikitext of the nomination subpage.
            value: If specified as 'yes' or 'no' (the value of the 'featured'
                or 'delisted' parameter from the reviewed results template),
                the keyword is based on that value, otherwise it is based
                on the result of the method is_passed().

        Returns:
            The modified wikitext of the nomination subpage.
        """
        # Determine the keyword
        match value:
            case "yes":
                success = True
            case "no":
                success = False
            case _:
                success = self.is_passed()
        keyword = self._SUCCESS_KEYWORD if success else self._FAIL_KEYWORD
        # Check if the nomination correctly starts with a level 3+ heading
        text = text.lstrip()  # Silently remove irritating whitespace.
        match = re.match(r"===(.+?)===", text)
        if not match:
            warn(
                f"Nomination '{self._page.title()}' does not start "
                f"with a heading; can't add '{keyword}' to the title."
            )
            ask_for_help(
                f"The nomination [[{self._page.title()}]] does not start with "
                "the usual <code><nowiki>===...===</nowiki></code> heading. "
                "Please check if there is any rubbish at the beginning "
                "and remove it, fix the heading if necessary, "
                f"and add <code>, {keyword}</code> at the end of the heading."
            )
            return text
        # Check whether the heading already contains the keyword or not
        heading = match.group(1).strip()
        if heading.endswith(keyword):
            return text
        # Add the keyword to the heading
        return text.replace(heading, f"{heading}, {keyword}", 1)

    @abc.abstractmethod
    def _type_code(self) -> str:
        """Return a three-letter code of the nomination type for '-info'.

        Abstract method, must be implemented by the subclasses.
        """
        pass

    @abc.abstractmethod
    def get_result_string(self) -> str:
        """Return the results template to be added when closing the nomination.

        Abstract method, must be implemented by the subclasses.
        """
        pass

    @abc.abstractmethod
    def get_close_edit_summary(self, fifth_day: bool) -> str:
        """Return the edit summary to be used when closing the nomination.

        Abstract method, must be implemented by the subclasses.

        Args:
            fifth_day: Is the nomination closed early because one of the
                rules of the 5th day applies to it?
        """
        pass

    @abc.abstractmethod
    def _archive_cat_name_base(self) -> str:
        """Return the basic component of a candidate archive category name.

        Abstract method, must be implemented by the subclasses.
        """
        pass

    def creation_time(self) -> datetime.datetime:
        """Return the time at which the nomination subpage was created.

        Returns:
            A datetime.datetime object with the creation time of the subpage.
            On errors the function returns the current time; therefore
            the nomination will not be closed because it appears as too young.
        """
        if self._creation_time:
            return self._creation_time

        try:
            timestamp = self._page.oldest_revision["timestamp"]
        except pywikibot.exceptions.PageRelatedError:
            error(
                f"Couldn't determine creation time of '{self._page.title()}', "
                "returning now()"
            )
            return datetime.datetime.now(datetime.UTC)
        # MediaWiki timestamps are always stored in UTC,
        # but querying a revision timestamp still returns an offset-naive
        # pywikibot.Timestamp object.  Therefore we convert it right away
        # to an offset-aware datetime object in order to compare it
        # easily and correctly to offset-aware datetime objects:
        self._creation_time = timestamp.replace(tzinfo=datetime.UTC)

        # print "C:" + self._creation_time.isoformat()
        # print "N:" + datetime.datetime.now(datetime.UTC).isoformat()
        return self._creation_time

    def status_string(self) -> str:
        """Return a short string describing the status of the candidate."""
        if cancelled := self.was_cancelled(uppercase=True):
            return cancelled
        if reviewed := self.is_reviewed():
            return reviewed
        if self.is_ignored():
            return "Ignored"
        if self.is_done() or self.rules_of_fifth_day():
            text = self._SUCCESS_KEYWORD if self.is_passed() else self._FAIL_KEYWORD
            return text.capitalize()
        return "Active"

    def _candidate_archive_status_cats(
        self,
        year: int,
        status: str,
    ) -> tuple[str, str]:
        """Return the names of the candidate archive categories by status.

        Args:
            year: The current year.
            status: A keyword for the final status of this nomination,
                i.e. '(not) featured', '(not) delisted', 'withdrawn',
                'FPXed', 'FPDed'.

        Returns:
            A tuple, containing:
            [0] The complete name of the candidate archive category by status
                for this nomination.
            [1] The name of the status supercategory for that category.
        """
        match status:
            case "featured" | "delisted":
                status = "successful"
            case "not featured" | "not delisted":
                status = "unsuccessful"
            case _:
                # Don't change the keyword, just check it to catch bugs:
                assert status in {"withdrawn", "FPXed", "FPDed"}
        superstat = f"{status[0].upper()}{status[1:]}"  # Capitalize only 1st letter.
        cat_name_base = self._archive_cat_name_base()
        return (
            f"Category:{year} {status} {cat_name_base}",
            f"Category:{superstat} {cat_name_base}",
        )

    def _candidate_archive_subject(
        self,
        gallery_link: str | None = None,
    ) -> tuple[str, str]:
        """Return the subject for candidate archive categories by subject.

        The subject keywords of these categories are based on the first part
        of the gallery link, like 'Animals' or 'Natural phenomena'.
        Therefore the categories also correspond to the sections of our list
        of recent FPs which is also used on the FP landing page.
        We just replace some subject names for grammatical reasons
        (all subjects must go with 'FPC of ...'), we split the giant
        'Places' subject in order to separate at least architecture and
        landscape photography (they are independent genres of photography),
        and we separate 'vehicles' and 'architectural elements', too.
        The 'architectural elements' are split further because some sections
        belong to architecture, some rather to other objects.

        Args:
            gallery_link: The gallery link from the nomination (incl. section),
                if already known.  Pass an empty string for delisting
                nominations (they don't use a gallery link).

        Returns:
            A tuple, containing:
            [0] A phrase describing the subject (like 'of plants') for use
                in category names.  We return a phrase (not just the subject)
                because sometimes we must vary the grammatical construction.
            [1] A sort key for the category.
        """
        if gallery_link is None:
            gallery_link = self.find_gallery_of_file()
        if gallery_link == "":  # Delisting nomination, or link not found.
            # For such nominations we use a maintenance category:
            return ("without subject", "!Without")
        # From here on we handle the link as lowercase to ignore case variants
        gallery_link = gallery_link.lower()
        gallery_link, section = split_gallery_link(gallery_link)
        link_parts = re.split(r"\s*/\s*", gallery_link)
        basic_gallery = link_parts[0]
        if basic_gallery not in VALID_BASIC_GALLERY_LINKS:
            warn(
                f"{self.cut_title()}: Invalid basic gallery '{basic_gallery}', "
                "adding to 'without subject' category."
            )
            return ("without subject", "!Without")
        next_part = link_parts[1] if len(link_parts) > 1 else ""
        match basic_gallery:
            case "historical":
                subject = "historical images"
            case "places":
                match next_part:
                    case "architecture":
                        if "cityscapes" in gallery_link:
                            subject = "cityscapes and settlements"
                        else:
                            subject = "architecture"
                    case "interiors":
                        subject = "architecture"
                    case "natural":
                        subject = "natural scenes"
                    case "settlements":
                        subject = "cityscapes and settlements"
                    case _:
                        subject = "places"
            case "objects":
                match next_part:
                    case "architectural elements":
                        if OBJECT_SECTS_IN_ARCH_ELEMENTS_REGEX.search(section):
                            subject = "objects"
                        else:
                            subject = "architecture"
                    case "vehicles":
                        subject = "vehicles"
                    case _:
                        subject = "objects"
            case _:
                subject = basic_gallery
        return (f"of {subject}", f"{subject[0].upper()}{subject[1:]}")

    def days_old(self) -> int:
        """Return the number of days since this nomination was created."""
        if self._days_old != -1:
            return self._days_old

        delta = datetime.datetime.now(datetime.UTC) - self.creation_time()
        self._days_old = delta.days
        return self._days_old

    def days_since_last_edit(self) -> int:
        """Return the number of days since this nomination was last edited.

        Returns:
            The number of whole days since the nomination was last edited;
            if the value cannot be found, the function returns -1.
        """
        if self._days_since_last_edit != -1:
            return self._days_since_last_edit

        try:
            timestamp = self._page.latest_revision["timestamp"]
        except pywikibot.exceptions.PageRelatedError:
            return -1
        # MediaWiki timestamps are always stored in UTC,
        # but querying a revision timestamp still returns an offset-naive
        # pywikibot.Timestamp object.  Therefore we convert it right away
        # to an offset-aware datetime object in order to compare it
        # easily and correctly to offset-aware datetime objects:
        last_edit = timestamp.replace(tzinfo=datetime.UTC)

        delta = datetime.datetime.now(datetime.UTC) - last_edit
        self._days_since_last_edit = delta.days
        return self._days_since_last_edit

    def is_done(self) -> bool:
        """Find out if the voting period for the nomination is over.

        NB: This method does not consider the rules of the fifth day,
        please use rules_of_fifth_day() for that purpose.
        """
        return self.days_old() >= 9

    def is_passed(self) -> bool:
        """Find out if the nomination is successful acc. to the current votes.

        NB: This method does not consider the age of the nomination,
        please test that with is_done() and rules_of_fifth_day().
        NB: This method cannot properly handle nominations with alternatives.
        """
        if self.was_cancelled():
            return False
        self.count_votes()
        return self._pro >= 7 and (self._pro >= 2 * self._con)

    def is_reviewed(self) -> str | Literal[False]:
        """Return a keyword for the review status of this candidate.

        Returns:
            A keyword for use with status_string(), indicating whether
            the nomination has already been closed and reviewed,
            or has been counted, but is still waiting for the review.
            If neither the one nor the other applies, it returns False.
        """
        text = self.filtered_content()
        if self._REVIEWED_RES_REGEX.search(text):
            return "Reviewed"
        if self._COUNTED_RES_REGEX.search(text):
            return "Counted"
        return False

    def is_ignored(self) -> bool:
        """Find out if the nomination needs manual vote counting.

        Nominations with alternative images require manual counting,
        therefore this method returns True for any nomination with more
        than one full-size image.
        """
        return self.image_count() > 1

    def section_count(self) -> int:
        """Count the number of sections in this nomination."""
        return len(SECTION_REGEX.findall(self.filtered_content()))

    def image_count(self) -> int:
        """Count the number of images in this nomination.

        This method is used to detect nominations with alternative images.
        Therefore it ignores the images from set nominations (which are
        formatted with a <gallery> element) and reports 0 for them.
        Small images and thumbnails are not counted, too, because they are
        probably just inline icons and not alternative images.
        """
        if self._image_count is not None:
            return self._image_count
        self._image_count = 0
        if images := IMAGES_REGEX.findall(self.filtered_content()):
            for image_link, _ in images:
                if not is_just_thumbnail(image_link):
                    self._image_count += 1
        return self._image_count

    def existing_results(self) -> list[tuple[str, ...]]:
        """Search and parse verified (reviewed) results template(s).

        The method scans the text of the nomination subpage for this candidate
        for any verified (reviewed) results template(s) and parses them.

        Returns:
            A list of tuples of strings.
            That list is empty if the nomination was not closed,
            if the results were not verified (reviewed) yet,
            or if the results do not use one of the usual formats.
            The list contains one or more (normally: exactly one) tuple
            if the nomination has been verified correctly.
            The length of the tuple(s) varies, depending on the results format,
            but only the first four values are important for a comparison:
            [0] count of support votes,
            [1] count of oppose votes,
            [2] count of neutral votes,
            [3] ('yes'|'no'|'(not )?featured'|'(not )?delisted').
        """
        text = self.filtered_content()
        # Search first for result(s) using the new template-base format,
        # and if this fails for result(s) in the old text-based format:
        results = self._VERIFIED_RES_REGEX.findall(text)
        if not results:
            results = self._OBSOLETE_RES_REGEX.findall(text)
        return results

    def compare_result_to_count(self) -> None:
        """Compare an existing verified result with a fresh vote counting.

        If there is an existing result the method compares it with a fresh
        counting of the votes and checks whether the results match or not.
        This is useful to test the vote counting code of the bot
        and to find possibly incorrect old results.
        """
        # Check status and get old result(s)
        if cancelled := self.was_cancelled():
            out(f"{self.cut_title()}: (ignoring, was {cancelled})")
            return
        if self.image_count() > 1:
            out(f"{self.cut_title()}: (ignoring, contains alternatives)")
            return
        results = self.existing_results()
        if not results:
            out(f"{self.cut_title()}: (ignoring, has no results)")
            return
        if len(results) > 1:
            out(f"{self.cut_title()}: (ignoring, has several results)")
            return

        # We have exactly one old result, so recount the votes and compare
        old_result = results[0]
        old_success = old_result[3].lower() in {"yes", "featured", "delisted"}
        old_pro = int(old_result[0])
        old_con = int(old_result[1])
        old_neu = int(old_result[2])
        self.count_votes()
        if (
            self._pro == old_pro
            and self._con == old_con
            and self._neu == old_neu
            and old_success == self.is_passed()
        ):
            status = "OK"
        else:
            status = "FAIL"

        # Print result as list entry to console
        out(
            f"{self.cut_title()}: "
            f"P:{self._pro:02d}/{old_pro:02d} "
            f"C:{self._con:02d}/{old_con:02d} "
            f"N:{self._neu:02d}/{old_neu:02d} "
            f"S:{y_n(self.is_passed())}/{y_n(old_success)} "
            f"({status})"
        )

    def cut_title(self) -> str:
        """Return a fixed width title for the nomination."""
        title = self.subpage_name(keep_prefix=False, keep_number=True)
        # We skip 'removal/', 'File:' etc., but 'Set/' is informative
        if self.is_set():
            title = f"Set/{title}"
        return title[0:50].ljust(50)

    def all_filenames(self) -> list[str]:
        """List all nominated image files.

        Returns:
            A list with the cleaned names of all nominated image files.
            If this is a single-file nomination, it has just one entry.
            If there were errors in determining the file names,
            the list is empty.
        """
        if self.is_set():
            files = self.set_files()
        else:
            filename = self.filename()
            files = [filename] if filename else []
        return files

    def filename(self) -> str:
        """Return the name of the image nominated in this candidate.

        The method tries to retrieve the filename, checks if the file exists
        and resolves any redirects.
        Problems are reported on the FPC talk page.
        For set nominations, use set_files() instead.

        Returns:
            The filename of the nominated image; on errors an empty string.
        """
        # Try the selected alternative or a cached result first
        if self._alternative is not None:
            return self._alternative
        if self._filename is not None:
            return self._filename
        # Change default value from None (meaning 'not retrieved yet')
        # to an empty string (meaning 'file not OK'; used if we return early)
        self._filename = ""

        # Try to derive the filename from the name of the nomination subpage
        page: pywikibot.Page | None  # Help typecheckers.
        subpage_name = self._page.title()
        cut_title = self.cut_title()
        if match := FULL_FILE_PREFIX_REGEX.search(subpage_name):
            filename = subpage_name[match.end(0):]
            # Use standard 'File:' namespace and remove '/2' etc.
            filename = FILE_NAMESPACE + re.sub(r" */ *\d+ *$", "", filename, count=1)
            page = pywikibot.Page(_g_site, filename)
            if not page.exists():
                # Image page not found; try the 1st image in the nomination
                warn(f"{cut_title}: (Did not find '{filename}', trying first image...)")
                page = self._first_real_image_in_nomination()
                if page is None:
                    error(f"{cut_title}: (Error: can't find image page)")
                    ask_for_help(
                        f"The nomination [[{subpage_name}]] is about the image "
                        f"[[:{filename}]], but that image file does not exist. "
                        f"Perhaps the file has been renamed. {PLEASE_FIX_HINT}"
                    )
                    return ""
        else:
            # Bad nomination subpage name; try the 1st image in the nomination
            warn(f"{cut_title}: (Could not derive filename, trying first image...)")
            page = self._first_real_image_in_nomination()
            if page is None:
                error(f"{cut_title}: (Error: bad nomination subpage name)")
                ask_for_help(
                    f"The name of the nomination subpage [[{subpage_name}]] "
                    "is irregular, therefore the bot cannot identify "
                    f"the nominated image. {PLEASE_FIX_HINT}"
                )
                return ""
        # If we arrive here, 'page' should point to a valid page.

        # Check if the image was renamed and try to resolve the redirect
        if page.isRedirectPage():
            filename = page.title()
            try:
                page = page.getRedirectTarget()
            except pywikibot.exceptions.PageRelatedError:
                # Circular or invalid redirect etc., skip candidate
                error(f"{cut_title}: (Error: invalid redirect)")
                ask_for_help(
                    f"The nomination [[{subpage_name}]] is about the image "
                    f"[[:{filename}]], but the image page contains a circular "
                    f"or invalid redirect. {PLEASE_FIX_HINT}"
                )
                return ""
            if not page.exists():
                # Broken redirect, skip candidate
                error(f"{cut_title}: (Error: broken redirect)")
                ask_for_help(
                    f"The nomination [[{subpage_name}]] is about the image "
                    f"[[:{filename}]], but the image page redirects to a file "
                    f"or page which does not exist. {PLEASE_FIX_HINT}"
                )
                return ""
            out(f"Resolved redirect: '{filename}' -> '{page.title()}'.")

        # Use the official spelling from the file page
        self._filename = page.title()
        return self._filename

    def _first_real_image_in_nomination(self) -> pywikibot.Page | None:
        """Return the page of the first image in the nomination.

        Returns:
            A pywikibot.Page object for the first image file
            which is linked in reasonable size (not as a mere thumbnail)
            in the text of the nomination subpage and actually exists.
            If no such image file is found, returns None.
        """
        images = IMAGES_REGEX.findall(self.filtered_content())
        for image_link, image_name in images:
            if not is_just_thumbnail(image_link):
                page = pywikibot.Page(_g_site, image_name)
                if page.exists():
                    return page
        return None

    def subpage_name(
        self,
        keep_prefix: bool = True,
        keep_number: bool = True,
    ) -> str:
        """Return the name of the nomination subpage.

        The method returns the name of the nomination subpage for this
        candidate without the leading 'Commons:Featured picture candidates/'
        (if you want to include it, just call 'self._page.title()' instead).

        Args:
            keep_prefix: Keep the 'removal/' and/or 'Set' prefixes
                and/or the 'File:' namespace prefix, if present.
            keep_number: Keep the trailing '/2', '/3', etc.

        Use 'keep_number=True' and adjust the 'keep_prefix' parameter
        to get tailor-made values for the 'com-nom' and 'subpage' parameters
        of the {{Assessments}} and user notification templates:
        for {{Assessments}}, pass 'keep_prefix=False' to remove the 'Set/',
        'removal/', and 'File:'/'Image:' prefixes (plus their combinations);
        for the user notification templates, pass 'keep_prefix=True' to keep
        these parts without any normalization.

        Use 'keep_prefix=False, keep_number=False' to get a clean title
        for the nomination, e.g. as title for a set nomination.
        """
        name = self._page.title()
        name = name.replace("_", " ")
        name = CAND_PREFIX_REGEX.sub("", name, count=1).strip()
        if not keep_prefix:
            name = MIDDLE_NOMINATION_NAME_REGEX.sub("", name, count=1)
        if not keep_number:  # Remove trailing '.../2' etc. of repeated noms.
            name = re.sub(r" */ *\d+ *$", "", name, count=1)
        return name

    def _get_current_log_page(
        self,
        year: int,
        month: str,  # Full English month name
        subpage_name: str,
    ) -> tuple[pywikibot.Page, int, str]:
        """Find and read the current FPC log page, creating it if necessary.

        Until November 2005, we just used one FPC log page per month.
        But long log pages were not rendered correctly because they contain
        too many template transclusions.  Therefore now we split
        the monthly log into parts containing at most MAX_ENTRIES_PER_LOG_PART
        nominations each.  This has made the search for the current log page
        more complicated, it includes creating a new part if the old one
        has reached the threshold.  We save the current part number of the log
        in a global variable to speed up the handling of the next nomination.

        Args:
            year: The current year as an integer, e.g. 2025.
            month: The English name of the current month, e.g. 'January'.
            subpage_name: The full name of the nomination subpage.

        Returns:
            A tuple, containing:
            [0] a Pyikibot page object for the current part of the log;
            [1] an integer with the number of the current part of the log;
            [2] a string with the current text of the current part of the log,
                empty if we have just started a new part of the log.

        Raises:
            DataAlreadyPresentError: The nomination is already in the log;
                that can happen if the previous run of the bot was interrupted.
                We still need to remove the nomination from the candidate list.
            CouldNotAddDataError: Due to an error the nomimation can't be added
                to the log; an error description is passed in the exception,
                we have to report it and must not remove the nomination.
        """
        global _g_log_part_no

        # Find the name of the current part of the log for the current month
        if _g_log_part_no is not None:
            part_no = _g_log_part_no
            log_page_name = build_log_page_name(month, year, _g_log_part_no)
        else:
            part_no = 1
            log_page_name = build_log_page_name(month, year, 1)
            if pywikibot.Page(_g_site, log_page_name).exists():
                # There is already at least one log part for the month,
                # find the last existing part
                for i in range(2, 1001):
                    next_page_name = build_log_page_name(month, year, i)
                    if not pywikibot.Page(_g_site, next_page_name).exists():
                        break
                    log_page_name = next_page_name
                    part_no = i
                else:
                    error("Error - no free log part page name?")
                    raise CouldNotAddDataError(
                        "The bot tried to generate a new part of the FPC log "
                        f"for {month} {year}, but did not find a free page "
                        f"from [[{build_log_page_name(month, year, 1)}]] "
                        f"up to [[{next_page_name}]]. "
                        "''Something is rotten in the state of Commons'', "
                        "or in the code of the bot."
                    )
            # Implict else: create a new log for the new month, part 1.

        # Read the current part of the log
        log_page = pywikibot.Page(_g_site, log_page_name)
        try:
            log_text = log_page.get(get_redirect=False).strip()
        except pywikibot.exceptions.NoPageError:
            out(f"Starting new log page '{log_page_name}'...")
            log_text = ""
        except pywikibot.exceptions.IsRedirectPageError:
            # Try to resolve the redirect
            try:
                log_page = log_page.getRedirectTarget()
                log_text = log_page.get(get_redirect=False).strip()
            except pywikibot.exceptions.PageRelatedError as exc:
                # Circular, nested or invalid redirect etc.
                error(f"Log page '{log_page_name}' was moved, redirect is invalid.")
                raise CouldNotAddDataError(
                    f"The log page [[{log_page_name}]] has been renamed, "
                    "but the bot could not resolve the redirect: "
                    f"{format_exception(exc)}."
                ) from exc
            out(f"Resolved redirect: '{log_page_name}' -> '{log_page.title()}'")
            log_page_name = log_page.title()

        # Check the existing text of the current log part
        if log_text:
            # Is the nomination already in the log?
            # This can happen if the process has previously been interrupted.
            if re.search(wikipattern(subpage_name), log_text):
                raise DataAlreadyPresentError(subpage_name)
            # Is the log page already full?
            count = len(CAND_PREFIX_REGEX.findall(log_text))
            if count >= MAX_ENTRIES_PER_LOG_PART:
                # We have to start a new log part; it must not exist yet
                part_no += 1
                log_page_name = build_log_page_name(month, year, part_no)
                log_page = pywikibot.Page(_g_site, log_page_name)
                if log_page.exists():
                    error(f"Error - log page '{log_page_name}' already exists")
                    raise CouldNotAddDataError(
                        "The bot tried to generate a new part of the FPC log "
                        f"for {month} {year}, but the page [[{log_page_name}]] "
                        "already exists. Either there is a bug in the bot, "
                        "or the log pages are in a muddle."
                    )
                out(f"Moving on to new log part '{log_page_name}'...")
                log_text = ""
                count = 0
        else:
            count = 0

        # Update the global value to handle the next candidate more easily;
        # if this entry reaches the limit, the next one must start a new part
        _g_log_part_no = (
            part_no if (count + 1) < MAX_ENTRIES_PER_LOG_PART
            else part_no + 1
        )
        return (log_page, part_no, log_text)

    def move_to_log(self, status: str, gallery_link: str | None = None) -> None:
        """Move the nomination from the candidate list to the log.

        This is the last step of the parking procedure for FP candidates
        as well as for delisting candidates.
        Call this method only for closed and verified candidates.
        It removes the nomination from the list of current candidates
        and adds it to the log for the current month; since February 2026,
        it also adds candidate archive categories to the nomination subpage.

        Args:
            status: A keyword for the final status, like 'featured'.
            gallery_link: The gallery link from the nomination (incl. section),
                if already known.  Pass an empty string for delisting
                nominations (they don't use a gallery link).
        """
        subpage_name = self._page.title()
        now = datetime.datetime.now(datetime.UTC)

        # Add candidate archive categories to the nomination subpage
        self._add_archive_categories(status, now, gallery_link)

        # Append nomination to the current log page
        try:
            log_page, part_no, old_log_text = self._get_current_log_page(
                now.year, now.strftime("%B"), subpage_name
            )
        except DataAlreadyPresentError:
            # Great, we are already done here and can remove the nomination.
            out(
                f"Skipping add in move_to_log() for '{subpage_name}', "
                "candidate is already in the log."
            )
        except CouldNotAddDataError as exc:
            ask_for_help(
                f"{exc} Please add [[{subpage_name}]] to the correct log page "
                "to ensure the log is complete, and search for the cause "
                "of this error and fix it."
            )
            # So we must not remove the nomination from the candidate list!
            return
        else:
            if old_log_text:
                new_log_text = f"{old_log_text}\n{{{{{subpage_name}}}}}"
            else:
                new_log_text = (
                    "{{FPC log page header"
                    f"|month={now:%m}|year={now.year}|part={part_no}"
                    "}}\n\n"  # Empty line
                    f"{{{{{subpage_name}}}}}"
                )
            job = "Added" if old_log_text else "Started new log page, added"
            summary = f"{job} [[{subpage_name}]] ({status})"
            commit(old_log_text, new_log_text, log_page, summary)

        # Remove nomination from the list of current nominations
        candidates_list_page = pywikibot.Page(_g_site, self._list_page_name)
        old_cand_text = candidates_list_page.get(get_redirect=False)
        pattern = r" *\{\{\s*" + wikipattern(subpage_name) + r"\s*\}\} *\n?"
        new_cand_text = re.sub(pattern, "", old_cand_text, count=1)
        if old_cand_text == new_cand_text:
            # This can happen if the process has previously been interrupted.
            out(
                f"Skipping remove in move_to_log() for '{subpage_name}', "
                "candidate not found in list."
            )
        else:
            summary = f"Removed [[{subpage_name}]] ({status})"
            commit(old_cand_text, new_cand_text, candidates_list_page, summary)

    def _add_archive_categories(
        self,
        status: str,
        now: datetime.datetime,
        gallery_link: str | None = None,
    ) -> None:
        """Add candidate archive categories to the nomination subpage.

        Args:
            status: A keyword for the final status, like 'featured'.
            now: The current date and time.
            gallery_link: The gallery link from the nomination (incl. section),
                if already known.  Pass an empty string for delisting
                nominations (they don't use a gallery link).
        """
        # Add archive categories to nomination subpage
        old_text = self._page.get(get_redirect=False)
        year = now.year
        month_cat = f"Category:{now:%B} {year} featured picture candidates"
        if month_cat in old_text:
            out(
                f"Skipping _add_archive_categories() for '{self._page.title()}', "
                "archive categories already present."
            )
            return
        status_cat, status_supercat = self._candidate_archive_status_cats(year, status)
        subject_phrase, subject_key = self._candidate_archive_subject(gallery_link)
        subject_cat = f"Category:{year} featured picture candidates {subject_phrase}"
        if subject_phrase.startswith("without"):
            ask_for_help(
                "The bot program could not infer the subject of the nomination "
                f"[[{self._page.title()}]], therefore it adds the nomination "
                f"to the maintenance category [[:{subject_cat}]]. "
                "Please replace that category by the appropriate annual archive "
                "category for the subject of the nomination."
            )
        categories = [month_cat, status_cat, subject_cat]  # Also defines order
        if self.is_set():
            set_cat = f"Category:{year} featured picture set nominations"
            categories.append(set_cat)
        else:
            set_cat = ""
        key = name_to_sort_key(self.subpage_name(keep_prefix=False, keep_number=False))
        new_text = (
            f"{old_text.rstrip()}\n\n"
            "<noinclude>\n"
            f"{{{{DEFAULTSORT:{key}}}}}\n"
            + "".join(f"[[{cat}]]\n" for cat in categories)
            + "</noinclude>"
        )
        summary = (
            "Added candidate archive categories "
            + ", ".join(f"[[{cat}]]" for cat in categories)
        )
        commit(old_text, new_text, self._page, summary)
        self.reset_filtered_content()

        # Create the candidate archive categories if necessary
        self._create_month_and_year_categories(now, month_cat)
        self._create_type_and_status_categories(year, status_cat, status_supercat)
        self._create_subject_categories(year, subject_cat, subject_phrase, subject_key)
        if set_cat:
            self._create_set_category(year, set_cat)

    @staticmethod
    def _create_month_and_year_categories(
        now: datetime.datetime,
        month_cat: str,
    ) -> None:
        """Create candidate archive categories for month and year, if necessary.

        Args:
            now: The current date and time.
            month_cat: The name of the archive category for the current month.
        """
        # Create the category for the current month if necessary
        category_page = pywikibot.Page(_g_site, month_cat)
        if category_page.exists():
            # This means the following categories exist, too, so we can just ...
            return
        year = now.year
        header = f"{{{{FPC archive category header|year={year}}}}}"
        month_supercat = f"Category:{year} featured picture candidates by month"
        new_text = f"{header}\n\n[[{month_supercat}| {now:%m}]]"
        summary = "Created new candidate archive category for the month"
        commit("", new_text, category_page, summary)

        # Do we also need to create the 'by month' supercategory?
        category_page = pywikibot.Page(_g_site, month_supercat)
        if not category_page.exists():
            year_supercat = f"Category:{year} featured picture candidates"
            new_text = f"{header}\n\n[[{year_supercat}| Month]]"
            summary = "Created new candidate archive supercategory 'by month'"
            commit("", new_text, category_page, summary)

            # Do we also need to create the supercategory for the year?
            category_page = pywikibot.Page(_g_site, year_supercat)
            if not category_page.exists():
                new_text = (
                    f"{header}\n\n"
                    f"[[Category:Featured picture candidates by year| {year}]]"
                )
                summary = "Created new candidate archive supercategory for the year"
                commit("", new_text, category_page, summary)

    def _create_type_and_status_categories(
        self,
        year: int,
        status_category: str,
        status_supercat: str,
    ) -> None:
        """Create candidate archive categories for type and status, if necessary.

        Args:
            year: The current year.
            status_category: Name of the category for nominations from the current
                year which have the same type and status as this nomination.
            status_supercat: Name of the supercategory for all nominations
                with that type and status.
        """
        # Create the status category if necessary
        category_page = pywikibot.Page(_g_site, status_category)
        if category_page.exists():
            # This means the following category exists, too, so we can just ...
            return
        header = f"{{{{FPC archive category header|year={year}}}}}"
        type_supercat = f"Category:{year} {self._archive_cat_name_base()}"
        new_text = (
            f"{header}\n\n"
            f"[[{type_supercat}| ]]\n"
            f"[[{status_supercat}| {year}]]"
        )
        summary = "Created new candidate archive category for the status"
        commit("", new_text, category_page, summary)

        # Do we also need to create the supercategory for the type?
        category_page = pywikibot.Page(_g_site, type_supercat)
        if not category_page.exists():
            by_type_supercat = f"Category:{year} featured picture candidates by type"
            key = (
                " /" if "replacing" in type_supercat
                else (" -" if "delist" in type_supercat else " +")
            )
            new_text = (
                f"{header}\n\n"
                f"[[{by_type_supercat}|{key}]]"
            )
            summary = "Created new candidate archive supercategory for the type"
            commit("", new_text, category_page, summary)

            # Do we also need to create a 'by type' supercategory?
            category_page = pywikibot.Page(_g_site, by_type_supercat)
            if not category_page.exists():
                new_text = (
                    f"{header}\n\n"
                    f"[[Category:{year} featured picture candidates| Type]]"
                )
                summary = "Created new candidate archive supercategory 'by type'"
                commit("", new_text, category_page, summary)
        # We don't need to create the supercategory for the status
        # because it is the same for all years and does already exist.

    @staticmethod
    def _create_subject_categories(
        year: int,
        subject_category: str,
        subject_phrase: str,
        subject_key: str,
    ) -> None:
        """Create candidate archive categories for the subject, if necessary.

        Args:
            year: The current year.
            subject_category: Name of the category for nominations from the current
                year which have the same subject as this nomination.
            subject_phrase: A phrase describing the subject (like 'of plants')
                for use in category names.
            subject_key: A sort key for the categories.
        """
        # Create the subject category if necessary
        category_page = pywikibot.Page(_g_site, subject_category)
        if category_page.exists():
            # This means the following categories exist, too, so we can just ...
            return
        header_with_year = f"{{{{FPC archive category header|year={year}}}}}"
        is_maintenance = subject_key[0] == "!"
        header = (  # Maintenance categories add a special param to the header:
            f"{{{{FPC archive category header|year={year}|maintenance=yes}}}}"
            if is_maintenance
            else header_with_year
        )
        subject_supercat = f"Category:Featured picture candidates {subject_phrase}"
        year_supercat = f"Category:{year} featured picture candidates by subject"
        new_text = (
            f"{header}\n\n"
            f"[[{year_supercat}|{subject_key}]]\n"
            f"[[{subject_supercat}| {year}]]"
        )
        summary = "Created new candidate archive category for the subject"
        commit("", new_text, category_page, summary)

        # Do we also need to create the 'by subject' supercategory for the year?
        category_page = pywikibot.Page(_g_site, year_supercat)
        if not category_page.exists():
            new_text = (
                f"{header_with_year}\n\n"  # Don't add maintenance param here!
                f"[[Category:{year} featured picture candidates| Subject]]"
            )
            summary = "Created new candidate archive supercategory 'by subject'"
            commit("", new_text, category_page, summary)

        # Do we also need to create the supercategory for the subject?
        category_page = pywikibot.Page(_g_site, subject_supercat)
        if not category_page.exists():
            header = (  # Here we need the header template without year:
                "{{FPC archive category header|maintenance=yes}}" if is_maintenance
                else "{{FPC archive category header}}"
            )
            new_text = (
                f"{header}\n\n"
                f"[[Category:Featured picture candidates by subject|{subject_key}]]"
            )
            summary = "Created new candidate archive supercategory for the subject"
            commit("", new_text, category_page, summary)
        # We don't need to create the base category, it is always the same.

    @staticmethod
    def _create_set_category(year: int, set_category: str) -> None:
        """Create candidate archive category for set nominations, if necessary.

        Args:
            year: The current year.
            set_category: The name of the candidate archive category
                for set nominations from the current year.
        """
        category_page = pywikibot.Page(_g_site, set_category)
        if category_page.exists():
            return
        new_text = (
            f"{{{{FPC archive category header|year={year}}}}}\n\n"
            f"[[Category:{year} featured picture candidates by type| >]]\n"
            f"[[Category:Featured picture set nominations| {year}]]"
        )
        summary = "Created new candidate archive category for set nominations"
        commit("", new_text, category_page, summary)
        # We don't need to create the supercategories -- the year category
        # is created elsewhere, the set category is always the same.

    def check_gallery(self) -> None:
        """Check if the gallery link is valid and report any problems.

        Like the '-info' task, '-checkgallery' is not used by the bot.
        It allows users to check the gallery links in all open FP nominations
        in order to find and fix gallery link problems before the bot program
        parks each nomination and has to complain if a link does not work.
        """
        self._check_gallery_link()

    @abc.abstractmethod
    def _check_gallery_link(self) -> None:
        """Check if the gallery link is valid and report any problems.

        We have to define check_gallery() in the base class in order to link
        and call it like all other task methods from _handle_task().
        However the real work must be done in this subroutine which is
        implemented by the individual subclasses because its code needs
        to call subclass-specific methods.
        """
        pass

    def park(self) -> None:
        """Park the nomination, if possible.

        The method checks that the candidate has exactly one valid verified
        result, that the image file(s) exist(s) and that there are no other
        serious obstacles.
        If yes, it executes the parking procedure for the candidate --
        if the nomination was successful, it promotes the new FP(s) or
        delists the former FP respectively;
        if it has failed, it just archives the nomination.
        """
        subpage_name = self._page.title()
        cut_title = self.cut_title()

        # Check that the nomination subpage actually exists
        if not self._page.exists():
            error(f"{cut_title}: (Error: no such page?!)")
            ask_for_help(
                LIST_INCLUDES_MISSING_SUBPAGE.format(
                    list=self._list_page_name, subpage=subpage_name
                )
            )
            return

        # Withdrawn/FPXed/FPDed nominations are handled by close()
        if cancelled := self.was_cancelled():
            out(f"{cut_title}: (ignoring, was {cancelled})")
            return

        # Look for verified results
        # (leaving out stricken or commented results which have been corrected)
        results = self._VERIFIED_RES_REGEX.findall(self.filtered_content())
        # Stop if there is not exactly one valid verified result
        if not results:
            out(f"{cut_title}: (ignoring, no verified results)")
            return
        if len(results) > 1:
            error(f"{cut_title}: (Error: several verified results?)")
            ask_for_help(
                f"The nomination [[{subpage_name}]] seems to contain "
                "more than one verified result. "
                "Please remove (or cross out) all but one of the results."
            )
            return

        # Are the nominated images readily available?
        if not self.all_filenames():
            # Could not find nominated image(s), error was already reported.
            return

        # We should now have a candidate with verified result that we can park
        verified_result = results[0]
        success = verified_result[3].lower()
        if success in {"yes", "no"}:
            # If the keyword has not yet been added to the heading, add it now
            old_text = self._page.get(get_redirect=False)
            new_text = self.fix_heading(old_text, success)
            if new_text != old_text:
                commit(old_text, new_text, self._page, "Fixed heading", minor=True)
                self.reset_filtered_content()
            # Park the candidate
            if success == "yes":
                self.handle_passed_candidate(verified_result)
            else:
                self.move_to_log(self._FAIL_KEYWORD)
        else:
            error(f"{cut_title}: (Error: invalid verified success status '{success}')")
            ask_for_help(
                f"The verified success status <code>{success}</code> "
                f"in the results template of [[{subpage_name}]] "
                f"is invalid. {PLEASE_FIX_HINT}"
            )

    @abc.abstractmethod
    def handle_passed_candidate(self, results: tuple[str, ...]) -> None:
        """Handle the parking procedure for a passed candidate.

        Abstract method, must be implemented by the subclasses.

        Args:
            results: A tuple with strings representing the values which
                have been assigned to the individual parameters of the
                reviewed results template in the nomination.
        """
        pass


class FPCandidate(Candidate):
    """A candidate (nomination) for featured picture status

    Each instance of this class represents a featured picture nomination;
    the class is used both for single-file and set nominations.
    """

    # Define class constants:
    # all class constants are inherited, no changes necessary.

    # Declare types of instance variables:
    # all instance variables are inherited, see superclass.

    # No __init__():
    # the class just uses the initializer of the superclass.

    def _type_code(self) -> str:
        """Return a three-letter code of the nomination type for '-info'.

        Overrides the abstract method from the superclass, implementing it
        for FP candidates.
        """
        # In this class the result doesn't depend on the particular candidate.
        return "New"

    def get_result_string(self) -> str:
        """Return the results template to be added when closing a nomination.

        Overrides the abstract method from the superclass, implementing it
        for FP candidates.
        """
        gallery = self.find_gallery_of_file()
        if self.image_count() > 1:
            return (
                "{{FPC-results-unreviewed"
                "|support=X|oppose=X|neutral=X"
                f"|featured=X|gallery={gallery}|alternative="
                "|sig=<small>'''Note: this candidate has several alternatives. "
                "Thus, if featured, the code <code>alternative=</code> "
                "in this template needs to be followed by the filename "
                "of the chosen alternative.'''</small> "
                "/~~~~}}"
            )
        # A simple FP nomination
        self.count_votes()
        return (
            "{{FPC-results-unreviewed"
            f"|support={self._pro}|oppose={self._con}|neutral={self._neu}"
            f"|featured={yes_no(self.is_passed())}"
            f"|gallery={gallery}"
            "|sig=~~~~}}"
        )

    def get_close_edit_summary(self, fifth_day: bool) -> str:
        """Return the edit summary to be used when closing a nomination.

        Overrides the abstract method from the superclass, implementing it
        for FP candidates.
        """
        if self.image_count() > 1:
            return "Closing for review - contains alternatives, needs manual counting"
        # A simple FP nomination
        self.count_votes()
        return (
            f"Closing for review ({self._pro} support, "
            f"{self._con} oppose, {self._neu} neutral, "
            f"featured: {yes_no(self.is_passed())}, "
            f"5th day: {yes_no(fifth_day)})"
        )

    def _archive_cat_name_base(self) -> str:
        """Return the basic component of a candidate archive category name.

        Overrides the abstract method from the superclass, implementing it
        for FP candidates.
        """
        # In this class the result doesn't depend on the particular candidate.
        return "candidates for featured picture status"

    def _check_gallery_link(self) -> None:
        """Check if the gallery link is valid and report any problems.

        The method overrides the abstract method from the superclass,
        implementing it for FP candidates.
        """
        cut_title = self.cut_title()
        # Find and split the gallery link
        gallery_link = self.find_gallery_of_file()
        if not gallery_link:
            warn(f"{cut_title}: Found no gallery link.")
            return
        gallery_page_name, section = split_gallery_link(gallery_link)
        # Read the gallery page
        full_page_name = f"{FP_PREFIX}{gallery_page_name}"
        page = pywikibot.Page(_g_site, full_page_name)
        try:
            old_text = page.get(get_redirect=False)
        except pywikibot.exceptions.NoPageError:
            warn(f"{cut_title}: Gallery page '{full_page_name}' does not exist.")
            return
        except (
            pywikibot.exceptions.PageRelatedError,
            pywikibot.exceptions.InvalidTitleError,
        ) as exc:
            warn(f"{cut_title}: Can't read gallery page '{full_page_name}': {exc}")
            return
        # Search for the section to which we have to add the new FP(s)
        result = self._find_gallery_insertion_place(
            gallery_page_name, section, old_text, False
        )
        if isinstance(result, tuple):
            section_name, _ = result
            out(f"{cut_title}: OK - '{gallery_page_name}', section '{section_name}'")
        else:  # The result is a warning.
            warn(f"{cut_title}: {result}")

    def handle_passed_candidate(self, results: tuple[str, ...]) -> None:
        """Promote a new featured picture (or set of featured pictures).

        The method overrides the abstract method from the superclass,
        implementing it for FP candidates.  It adds the new FP(s) to
        the list of recently promoted FPs and to the specified gallery page,
        inserts the {{Assessments}} template into the description page(s),
        creates an entry in the chronological archives, notifies nominator,
        uploader and creator, etc.
        """
        subpage_name = self._page.title()
        cut_title = self.cut_title()

        # Some methods need the full gallery link with section anchor,
        # others only the gallery page name or even just the basic gallery.
        full_gallery_link = clean_gallery_link(results[4])
        gallery_page, section = split_gallery_link(full_gallery_link)
        if not gallery_page:
            error(f"{cut_title}: (ignoring, gallery not defined)")
            ask_for_help(
                f"The gallery link in the nomination [[{subpage_name}]] "
                f"is empty or broken. {PLEASE_FIX_HINT}"
            )
            return
        match = re.search(r"^(.*?)(?:/|$)", gallery_page)
        assert match is not None  # Regex matches always, help typecheckers.
        basic_gallery = match.group(1)

        # If there is more than one image, search for the selected alternative
        if self.image_count() > 1:
            if len(results) > 5 and results[5].strip():
                alternative = results[5].replace("_", " ").strip()
                if not pywikibot.Page(_g_site, alternative).exists():
                    error(
                        f"{cut_title}: (ignoring, specified alternative "
                        f"'{alternative}' not found)"
                    )
                    ask_for_help(
                        f"Cannot find the alternative [[:{alternative}]] "
                        f"specified by the nomination [[{subpage_name}]]. "
                        f"{PLEASE_FIX_HINT}"
                    )
                    self._alternative = ""  # Mark as invalid, like _filename.
                    return
                self._alternative = alternative
            else:
                error(f"{cut_title}: (ignoring, alternative not set)")
                ask_for_help(
                    f"The nomination [[{subpage_name}]] contains several "
                    "images, but does not specify the selected alternative. "
                    f"{PLEASE_FIX_HINT}"
                )
                return

        # Promote the new featured picture(s)
        files = self.all_filenames()
        if not files:
            # Could not identify the nominated image(s), error already reported
            return
        self.add_to_featured_list(basic_gallery, files)
        self.add_to_gallery_page(gallery_page, section, files)
        self.add_assessments(files)
        # ! self.add_assessment_to_media_info(files)  # Disabled 2026-04-01
        self.add_to_current_month(files)
        self.notify_nominator(files)
        self.notify_uploader_and_creator(files)
        self.move_to_log(self._SUCCESS_KEYWORD, full_gallery_link)

    def add_to_featured_list(self, section_name: str, files: list[str]) -> None:
        """Add the new featured picture to the list of recent FPs.

        This is STEP 1 of the parking procedure for new featured pictures.
        Call this method only for closed and verified candidates.
        The list of recent FPs is also used on the FP landing page;
        set nominations are represented by the first image from the set.

        Args:
            section_name: The section name, like 'Animals' or 'Places'.
                (The list of recently featured images uses the basic part
                of the gallery links as section names.)
            files: List with filename(s) of the featured picture or set.
        """
        filename = files[0]  # For set nominations just use the first file.

        # Read the list
        page = pywikibot.Page(_g_site, GALLERY_LIST_PAGE_NAME)
        try:
            old_text = page.get(get_redirect=False)
        except pywikibot.exceptions.PageRelatedError as exc:
            error(f"Error - can't read list of recent FPs: {exc}")
            fexc = format_exception(exc)
            ask_for_help(
                COULD_NOT_READ_RECENT_FPS_LIST.format(exception=fexc)
                + f" Then please add the new FP [[:{filename}]] "
                f"to the section ''{section_name}''."
            )
            return

        # Check if the image is already on the page.
        # This can happen if the process has previously been interrupted.
        if re.search(wikipattern(filename), old_text):
            out(
                f"Skipping add_to_featured_list() for '{filename}', "
                "image is already listed."
            )
            return

        # Find the correct section and its <gallery> element;
        # remove the last entry/entries from the <gallery> element,
        # keeping the 3 newest ones, and insert the new FP before them.
        esc_section = re.escape(section_name)
        match = re.search(
            r"\n==\s*\{\{\{\s*\d+\s*\|\s*(" + esc_section + r")\s*\}\}\}\s*==\s*"
            r"<gallery[^\n>]*>(.+?)</gallery>",
            old_text,
            flags=re.DOTALL | re.IGNORECASE,
        )
        if not match:
            error(f"Error - can't find gallery section '{section_name}'.")
            ask_for_help(
                f"The bot could not add the new featured picture [[:{filename}]] "
                f"to the list at [[{GALLERY_LIST_PAGE_NAME}]] "
                f"because it did not find the section ''{section_name}''. "
                "Either there is no subheading with that name, "
                "or it is not followed immediately by a valid "
                "<code><nowiki><gallery></nowiki></code> element. "
                "Please check whether the list page is OK or needs a fix, "
                "and add the new FP by hand to the correct section."
            )
            return
        real_section_name = match.group(1)
        entries = match.group(2).strip().splitlines()
        formatted = "\n".join(entry.strip() for entry in entries[:3])
        new_text = (
            old_text[:match.start(2)]
            + f"\n{filename}|{bare_filename(filename)}\n"
            + f"{formatted}\n"
            + old_text[match.end(2):]
        )

        # Commit the new text
        summary = f"Added [[{filename}]] to section '{real_section_name}'"
        commit(old_text, new_text, page, summary)

    def add_to_gallery_page(
        self,
        gallery_page_name: str,
        section: str,
        files: list[str],
    ) -> None:
        """Add the new featured picture(s) to the FP gallery page.

        This is STEP 2 of the parking procedure for new featured pictures.
        Call this method only for closed and verified candidates.
        For set nominations all files from the set are added at once
        to the same place on the same FP gallery page.

        Args:
            gallery_page_name: The name of the gallery page, without the
                leading 'Commons:Featured pictures/'.
            section: The section anchor which denotes the target section
                on that page (can be empty).
            files: List with filename(s) of the featured picture or set.
        """
        subpage_name = self._page.title()

        # Read the gallery page
        full_page_name = f"{FP_PREFIX}{gallery_page_name}"
        page = pywikibot.Page(_g_site, full_page_name)
        try:
            old_text = page.get(get_redirect=False)
        except pywikibot.exceptions.NoPageError:
            error(f"Error - gallery page '{full_page_name}' does not exist.")
            ask_for_help(
                f"The gallery page [[{full_page_name}]] which was specified "
                f"by the nomination [[{subpage_name}]] does not exist. "
                f"{PLEASE_CHECK_GALLERY_AND_SORT_FPS}"
            )
            return
        except (
            pywikibot.exceptions.PageRelatedError,
            pywikibot.exceptions.InvalidTitleError,
        ) as exc:
            error(f"Error - can't read gallery page '{full_page_name}': {exc}")
            ask_for_help(
                f"The bot could not read the gallery page [[{full_page_name}]] "
                f"which was specified by the nomination [[{subpage_name}]]: "
                f"{format_exception(exc)}. {PLEASE_CHECK_GALLERY_AND_SORT_FPS}"
            )
            return

        # Check if some of the new FPs are already on the page.
        # This can happen if the process has previously been interrupted.
        # We skip these files but handle any file which is not yet present.
        new_files = [
            file for file in files if not re.search(wikipattern(file), old_text)
        ]
        if not new_files:
            # Not a single file needs to be added, so we can stop here.
            out(
                f"Skipping add_to_gallery_page() for '{subpage_name}', "
                "image(s) already listed."
            )
            return
        # Format the new entries and a hint for the edit summary
        new_entries = "".join(
            f"{filename}|{bare_filename(filename)}\n" for filename in new_files
        )
        files_for_summary = f"[[{new_files[0]}]]"
        if len(new_files) > 1:
            files_for_summary += f" and {len(new_files) - 1} more set file(s)"

        # Search for the section to which we have to add the new FPs
        insert_at: slice | None  # Help typecheckers.
        result = self._find_gallery_insertion_place(
            gallery_page_name, section, old_text, True
        )
        if isinstance(result, tuple):
            real_section_name, insert_at = result
            summary = f"Added {files_for_summary} to section '{real_section_name}'"
        elif insert_at := self._find_unsorted_insertion_place(full_page_name, old_text):
            summary = f"Added {files_for_summary} to the 'Unsorted' section"
        else:  # Serious error with gallery page, already reported
            return

        # Add the new FP(s) to the gallery page
        new_text = (
            f"{old_text[:insert_at.start]}\n"
            + new_entries
            + old_text[insert_at.stop:]
        )
        commit(old_text, new_text, page, summary)

    def _find_gallery_insertion_place(
        self,
        gallery_page_name: str,
        section: str,
        old_text: str,
        report_errors: bool,
    ) -> tuple[str, slice] | str:
        """Search for the start of the <gallery>...</gallery> element
        of the section to which we have to add the new featured picture(s).

        Args:
            gallery_page_name: The name of the gallery page, without the
                leading 'Commons:Featured pictures/'.
            section: The section anchor from the gallery link.
            old_text: The complete old text of the gallery page.
            report_errors: If True, warnings are printed to the CLI and
                a request for help is posted on the FPC talk page;
                if False, warnings are just returned.

        Returns:
            If successful, a tuple containing
            [0] the official subheading of the target section and
            [1] a slice object describing the index values of the characters
            which should be replaced by the new entries.
            If not successful, a string with a short warning;
            this means we did not find a valid target section and have
            to use the 'Unsorted' section instead.
        """
        subpage_name = self._page.title()
        full_page_name = f"{FP_PREFIX}{gallery_page_name}"
        unsorted_hint = ADDING_FPS_TO_UNSORTED_SECTION.format(page=full_page_name)

        # Have we got a section anchor?
        if not section:
            # There was no section anchor
            warning = "No section anchor, adding FP(s) to 'Unsorted' section."
            if report_errors:
                warn(warning)
                ask_for_help(
                    f"The gallery link in the nomination [[{subpage_name}]] "
                    f"points to the gallery page [[{full_page_name}]], "
                    "but does not specify the desired section on that page. "
                    f"{unsorted_hint}"
                )
            return warning

        # Search for the subheading matching the section anchor
        # We handle spaces before colons as optional (these spaces are common
        # in subheadings on biology gallery pages, users often forget them).
        esc_section = re.escape(section)
        esc_section = re.sub(r"(?:\\? )*:", r" *:", esc_section)
        match = re.search(
            r"\n=+ *(" + esc_section + r") *=+(?: *\n)+",
            old_text,
            flags=re.IGNORECASE,
        )
        if not match:
            warning = (
                "Found no matching subheading, adding FP(s) "
                "to 'Unsorted' section."
            )
            if report_errors:
                warn(warning)
                ask_for_help(
                    f"The section anchor ''{section}'' in the gallery link "
                    f"of the nomination [[{subpage_name}]] does not match "
                    f"any subheading on the gallery page [[{full_page_name}]] "
                    f"letter for letter. {unsorted_hint}"
                )
            return warning
        real_section_name = match.group(1)

        # Check if that subheading opens a valid target section,
        # i.e., whether it is directly followed by the associated
        # <gallery>...</gallery> element or not
        match = ASSOC_GALLERY_ELEMENT_REGEX.match(old_text, pos=match.end(0))
        if not match:
            warning = (
                "Target subheading not followed immediately by <gallery>, "
                "adding FP(s) to 'Unsorted' section."
            )
            if report_errors:
                warn(warning)
                ask_for_help(
                    f"The gallery link in the nomination [[{subpage_name}]] "
                    f"points to a heading on [[{full_page_name}]], "
                    f"but [[{full_page_name}#{real_section_name}|that heading]] "
                    "is not a valid target because it is not followed "
                    "immediately by an associated "
                    "<code><nowiki><gallery></nowiki></code> element. "
                    "Perhaps this is a superordinate heading and the image "
                    "should be added to one of its subsections; "
                    "but to which one? Or there is a special gallery page "
                    "for featured pictures of that subject and the section "
                    "is just a cross reference; in this case the entry "
                    "for the new featured picture must be moved to "
                    f"the special gallery page. \u2013 {unsorted_hint}"
                )
            return warning

        # If we arrive here, we have found a valid target section.
        # Check if that section is just the 'Unsorted' section
        # (this actually happens; it's valid, but not helpful,
        # so we handle the request, but also ask for help).
        if real_section_name == UNSORTED_HEADING:
            warning = "Gallery link points to 'Unsorted' section."
            if report_errors:
                warn(warning)
                ask_for_help(
                    f"The gallery link in the nomination [[{subpage_name}]] "
                    "instructs the bot to put the new featured picture(s) "
                    f"into the ''Unsorted'' section of [[{full_page_name}]]. "
                    "This is not exactly helpful because that section "
                    "is used only for images which need to be sorted "
                    "into a more specific section. "
                    "So please move the new featured picture(s) "
                    "to a more appropriate place."
                )
            return warning

        # Success
        return (real_section_name, slice(match.end(1), match.end(0)))

    def _find_unsorted_insertion_place(
        self,
        full_page_name: str,
        old_text: str,
    ) -> slice | None:
        """Search for the start of the <gallery>...</gallery> element
        of the 'Unsorted' section in order to insert the new FP(s);
        it should be just the last <gallery> element on the gallery page.

        Args:
            full_page_name: The full name of the gallery page.
            old_text: The complete old text of the gallery page.

        Returns:
            If successful, a slice object describing the index values
            of the characters which should be replaced by the new entries;
            or None if we did not even find a usabale 'Unsorted' section.
        """
        if (start := old_text.rfind("<gallery")) >= 0:
            if match := GALLERY_ENTRY_START_REGEX.search(old_text, pos=start):
                return slice(match.start(0) + 1, match.end(0))
        error(
            "Error - found no 'Unsorted' section on "
            f"'{full_page_name}', can't add new FP(s)."
        )
        ask_for_help(
            f"The gallery page [[{full_page_name}]] which was "
            f"specified by the nomination [[{self._page.title()}]] "
            "seems to be invalid or broken. The bot did not find "
            "a valid <code><nowiki><gallery></nowiki></code> element "
            "for the ''Unsorted'' section on that page. "
            "Either there is no such section or the formatting of its "
            "<code><nowiki><gallery></nowiki></code> element is damaged. "
            f"{PLEASE_CHECK_GALLERY_AND_SORT_FPS}"
        )
        return None

    def add_assessments(self, files: list[str]) -> None:
        """Add the {{Assessments}} template to the image description page(s).

        This is STEP 3 of the parking procedure for new featured pictures.
        Call this method only for closed and verified candidates.
        For set nominations the template is added to the description page
        of every single image from the set.

        Args:
            files: List with filename(s) of the featured picture or set.
        """
        subpage_name = self.subpage_name(keep_prefix=False, keep_number=True)
        for filename in files:
            # Try to get and read the image description page
            page = pywikibot.Page(_g_site, filename)
            try:
                old_text = page.get(get_redirect=False)
            except pywikibot.exceptions.PageRelatedError as exc:
                # If the image has been deleted etc., we must just ignore it
                error(f"Error - can't read '{filename}': {exc}")
                continue

            # Search and (if found) update the {{Assessments}} template
            found, up_to_date, new_text = update_assessments_template(
                old_text, 1, subpage_name
            )
            if found:
                if up_to_date:
                    # Old and new template are identical, so skip this file,
                    # but continue to check other files (for set nominations)
                    out(
                        f"Skipping add_assessments() for '{filename}', "
                        "image is already featured."
                    )
                    continue
                # Else: The {{Assessments}} template was found and updated.
            else:
                # There is no {{Assessments}} template, so just add a new one.
                tmpl = f"{{{{Assessments|featured=1|com-nom={subpage_name}}}}}"
                # Search for the best location, in order of priority:
                # 1) At the very end of the file description stuff by putting
                # it right before the header of the license section;
                # 2) after the location templates (usually they appear after
                # the info templates and are displayed in unity with them);
                # 3) after one of the common information templates.
                if match := re.search(
                    r"\n== *\{\{ *int:license-header *\}\} *==", old_text
                ):
                    end = match.start(0)
                elif template_pos := find_template_pos(
                    old_text, r"(?:[Oo]bject[ _])?[Ll]ocation(?:[ _]dec)?"
                ):
                    end = template_pos.stop
                elif template_pos := find_template_pos(
                    old_text,
                    r"[Ii]nformation|[Aa]rtwork|[Pp]hotograph|[Aa]rt[ _][Pp]hoto",
                ):
                    end = template_pos.stop
                else:
                    end = 0
                if end:
                    # Use no empty line before, 1 empty line after the template
                    new_text = (
                        f"{old_text[:end].rstrip()}\n"
                        f"{tmpl}\n"
                        "\n"
                        f"{old_text[end:].lstrip()}"
                    )
                else:
                    # Searches have failed, just add the template at the top
                    new_text = f"{tmpl}\n\n{old_text.lstrip()}"

            # Commit the new text
            try:
                commit(old_text, new_text, page, "FP promotion")
            except pywikibot.exceptions.PageSaveRelatedError as exc:
                error(
                    f"Error - can't save '{filename}', "
                    f"can't add/update {{{{Assessments}}}}: {exc}"
                )
                ask_for_help(
                    "The bot cannot add or update the {{tl|Assessments}} "
                    f"template for the featured picture [[:{filename}]] "
                    "because there was an error saving the page: "
                    f"{format_exception(exc)}. "
                    "Please add or update the {{tl|Assessments}} manually."
                )

    def add_assessment_to_media_info(self, files: list[str]) -> None:
        """Add a FP assessment claim to the structured data of the new FP(s).

        This is STEP 4 of the parking procedure for new featured pictures.
        Call this method only for closed and verified candidates.
        It adds the 'Commons quality assessment' (P6731) claim
        'Wikimedia Commons featured picture' (Q63348049) to the Media Info
        (structured data) of the new featured picture, resp. of all files
        from a successful set nomination.

        Args:
            files: List with filename(s) of the featured picture or set.
        """
        # As effective date for the FP status we use the last modification
        # of the nomination subpage, normally made by the closing user.
        # To limit the precision to the day/date, we use "'precision': 11"
        # (right now more precise values are not supported on Wikidata:
        # https://phabricator.wikimedia.org/T57755).  Therefore we must
        # literally nullify the time part of the timestamp,
        # else MediaWiki rejects the timestamp as 'Malformed input'.
        try:
            timestamp = self._page.latest_revision["timestamp"]
        except pywikibot.exceptions.PageRelatedError:
            timestamp = datetime.datetime.now(datetime.UTC)
        iso_timestamp = timestamp.strftime("%Y-%m-%dT00:00:00Z")

        # Prepare data for the FP assessment claim
        try:
            fp_claim_site = pywikibot.Site("wikidata", "wikidata")
        except pywikibot.exceptions.Error as exc:
            # Creating a claim requires a Wikidata Site object, but sometimes
            # creating it fails due to connection errors.  In this case we
            # report the error and skip this step of the parking procedure,
            # handing it over to manual handling by adding a request for help.
            error(f"Error - could not create Site object for Wikidata: {exc}")
            file_links = ", ".join(f"[[:{filename}]]" for filename in files)
            ask_for_help(
                "The bot could not add a featured picture assessment claim "
                "to the Structured data of one or more new FP(s) because "
                "creating and connecting a Pywikibot <code>Site</code> object "
                f"for Wikidata has failed: {format_exception(exc)}. "
                + PLEASE_ADD_FP_ASSESSMENT_CLAIM.format(files=file_links)
            )
            return
        fp_claim_data = {
            "mainsnak": {
                "snaktype": "value",
                "property": "P6731",
                "datatype": "wikibase-item",
                "datavalue": {
                    "value": {
                        "entity-type": "item",
                        "numeric-id": 63348049,
                    },
                    "type": "wikibase-entityid",
                },
            },
            "type": "statement",
            "rank": "normal",
            "qualifiers": {
                "P580": [
                    {
                        "snaktype": "value",
                        "property": "P580",
                        "datatype": "time",
                        "datavalue": {
                            "value": {
                                "time": iso_timestamp,
                                "precision": 11,
                                "after": 0,
                                "before": 0,
                                "timezone": 0,
                                "calendarmodel": "http://www.wikidata.org/entity/Q1985727",
                            },
                            "type": "time",
                        },
                    },
                ],
            },
            "qualifiers-order": ["P580"],
        }

        for filename in files:
            # Get the Media Info for the image
            file_page = pywikibot.FilePage(_g_site, title=filename)
            if not file_page.exists():
                error(f"Error - image '{filename}' not found.")
                continue
            media_info = file_page.data_item()
            structured_data = media_info.get(force=True)
            try:
                statements = structured_data["statements"]
            except KeyError:
                error(
                    f"Error - no 'statements' entry in structured data for '{filename}'."
                )
                continue

            # Is there already a FP assessment claim?
            try:
                quality_assessments = statements["P6731"]
            except KeyError:
                # No 'Commons quality assessment' (P6731) claims at all
                claim_already_present = False
            else:
                for claim in quality_assessments:
                    if is_fp_assessment_claim(claim):
                        claim_already_present = True
                        break
                else:
                    # We did not leave the loop via 'break': claim not found
                    claim_already_present = False

            # Add the claim if necessary
            if claim_already_present:
                out(
                    f"Skipping add_assessment_to_media_info() for '{filename}', "
                    "FP assessment claim already present."
                )
            else:
                # We must use a new Claim instance with every file,
                # else Pywikibot raises a ValueError.
                fp_claim = pywikibot.page.Claim.fromJSON(
                    site=fp_claim_site, data=fp_claim_data
                )
                try:
                    commit_media_info_changes(filename, media_info, [], [fp_claim])
                except pywikibot.exceptions.PageSaveRelatedError as exc:
                    error(f"Error - can't save changes to '{filename}': {exc}")
                    file_link = f"[[:{filename}]]"
                    ask_for_help(
                        "The bot could not add a featured picture assessment "
                        f"claim to the Structured data of {file_link} "
                        "because there was an error saving that page: "
                        f"{format_exception(exc)}. "
                        + PLEASE_ADD_FP_ASSESSMENT_CLAIM.format(files=file_link)
                    )

    def add_to_current_month(self, files: list[str]) -> None:
        """Add the candidate to the chronological archive of featured pictures.

        This is STEP 5 of the parking procedure for new featured pictures.
        Call this method only for closed and verified candidates.
        Set nominations are represented by the first image from the set.

        Args:
            files: List with filename(s) of the featured picture or set.
        """
        filename = files[0]  # For set nominations just use the first file.

        # Extract voting results
        if match := VERIFIED_RESULT_REGEX.search(self.filtered_content()):
            ws = match.group(1)
            wo = match.group(2)
            wn = match.group(3)
        else:
            error(f"Error - no verified result in '{self._page.title()}'.")
            ask_for_help(
                f"The nomination [[{self._page.title()}]] is closed, "
                "but does not contain a valid verified result. "
                f"{PLEASE_FIX_HINT}"
            )
            return

        # Get the chronological archive page for the current month
        now = datetime.datetime.now(datetime.UTC)
        year = now.year
        month = now.strftime("%B")  # Full local month name, here: English
        monthpage = f"{CHRONO_ARCHIVE_PREFIX}{month} {year}"
        page = pywikibot.Page(_g_site, monthpage)
        try:
            old_text = page.get(get_redirect=False)
        except pywikibot.exceptions.NoPageError:
            old_text = ""

        if old_text:
            # First check if the image is already on the page.
            # This can happen if the process has previously been interrupted.
            if re.search(wikipattern(filename), old_text):
                out(
                    f"Skipping add_to_current_month() for '{filename}', "
                    "image is already listed."
                )
                return
            # Find the number of entries in the gallery
            if match := re.search(
                r"<gallery\b[^>]*>(\n.*)</gallery>",
                old_text,
                flags=re.DOTALL,
            ):
                # Because of the obligatory NL after '<gallery>' even
                # an empty gallery must yield a count of 1, as we need it.
                count = match.group(1).count("\n")
            else:
                error(f"Error - no valid <gallery> element in '{monthpage}'.")
                ask_for_help(
                    f"The chronological archive page [[{monthpage}]] contains "
                    "no <code><nowiki><gallery></nowiki></code> element. "
                    "Please check the page."
                )
                return
            job = "Added"
        else:
            # The page does not exist yet (new month) or is empty,
            # so create its contents from scratch.
            # Add the category with a sort key and in <noinclude>
            # because archive pages are sometimes included by other pages
            # which should not appear in the archive category.
            old_text = (
                f"== {month} {year} ==\n"
                "<noinclude>[[Category:Featured picture candidate "
                f"chronological archives| {year} {now:%m}]]</noinclude>\n"
                "<gallery>\n</gallery>"
            )
            count = 1
            job = "Started new chronological archive page, added"

        # Assemble the new entry and append it to the end of the gallery
        if self.is_set():
            set_name = self.subpage_name(keep_prefix=False, keep_number=False)
            title = f"Set: {set_name} ({len(files)} files)"
            summary = f"{job} set [[{self._page.title()}|{set_name}]]"
        else:
            title = bare_filename(filename)
            summary = f"{job} [[{filename}]]"
        if creator_link := self.creator(link=True):
            creator_hint = f"created by {creator_link}, "
        else:
            creator_hint = ""
        uploader_link = self.uploader(filename, link=True)
        nominator_link = self.nominator(link=True)
        new_text = old_text.replace(
            "</gallery>",
            f"{filename}|[[{self._page.title()}|{count}]] '''{title}'''<br> "
            f"{creator_hint}"
            f"uploaded by {uploader_link}, "
            f"nominated by {nominator_link},<br> "
            f"{{{{s|{ws}}}}}, {{{{o|{wo}}}}}, {{{{n|{wn}}}}}\n"
            "</gallery>",
            1,
        )
        commit(old_text, new_text, page, summary)

    def notify_nominator(self, files: list[str]) -> None:
        """Notify the nominator of the new FP(s).

        This is STEP 6 of the parking procedure for new featured pictures.
        Call this method only for closed and verified candidates.
        It adds a FP promotion template to the nominator's talk page.

        Args:
            files: List with filename(s) of the featured picture or set.
        """
        # Get and read nominator talk page
        talk_link = f"{USER_TALK_NAMESPACE}{self.nominator(link=False)}"
        talk_page = pywikibot.Page(_g_site, talk_link)
        ignoring = "but ignoring since it's just the nominator notification."
        try:
            old_text = talk_page.get(get_redirect=False)
        except pywikibot.exceptions.NoPageError:
            # Undefined user talk pages are uncommon because every new user
            # is welcomed by an automatic message.  So better stop here.
            warn(f"The user talk page '{talk_link}' is undefined, {ignoring}")
            return
        except pywikibot.exceptions.IsRedirectPageError:
            # Try to resolve the redirect
            try:
                talk_page = talk_page.getRedirectTarget()
                old_text = talk_page.get(get_redirect=False)
            except pywikibot.exceptions.PageRelatedError:
                # Circular, nested or invalid redirect etc.
                warn(
                    f"The user talk page '{talk_link}' was moved "
                    f"and the redirect is invalid, {ignoring}"
                )
                return
            new_talk_link = talk_page.title()
            out(f"User talk page redirect: '{talk_link}' -> '{new_talk_link}'")
            talk_link = new_talk_link  # Update the talk page name.

        subpage_name = self.subpage_name(keep_prefix=True, keep_number=True)
        if self.is_set():
            # Notifications for set nominations add a gallery to the talk page
            # and use a special template with an appropriate message.
            # Since August 2025 we use an improved version of the template.
            nomination_link = self._page.title()
            set_title = self.subpage_name(keep_prefix=False, keep_number=False)
            template = f"{{{{FPpromotionSet2|{set_title}|subpage={subpage_name}}}}}"
            # Check if there already is a promotion template on the talk page.
            # This can happen if the process has previously been interrupted.
            if re.search(wikipattern(template), old_text):
                out(
                    f"Skipping notify_nominator() for set '{set_title}', "
                    f"promotion template is already present at '{talk_link}'."
                )
                return
            entries = "\n".join(
                f"  {filename}|{bare_filename(filename)}" for filename in files
            )
            new_text = (
                f"{old_text.rstrip()}\n"
                "\n"
                "== Set Promoted to FP ==\n"
                '<gallery mode="packed-hover" heights="120px">\n'
                f"{entries}\n"
                "</gallery>\n"
                f"{template} /~~~~"
            )
            summary = f"FP promotion of set [[{nomination_link}|{set_title}]]"

        else:
            # Single FP nomination
            filename = files[0]
            template = f"{{{{FPpromotion|{filename}|subpage={subpage_name}}}}}"
            # Check if there already is a promotion template on the talk page.
            # This can happen if the process has previously been interrupted.
            if re.search(wikipattern(template), old_text):
                out(
                    f"Skipping notify_nominator() for '{filename}', "
                    f"promotion template is already present at '{talk_link}'."
                )
                return
            new_text = (
                f"{old_text.rstrip()}\n"
                "\n"
                "== FP Promotion ==\n"
                f"{template} /~~~~"
            )
            summary = f"FP promotion of [[{filename}]]"

        # Commit the new text
        try:
            commit(old_text, new_text, talk_page, summary)
        except pywikibot.exceptions.LockedPageError:
            warn(f"The user talk page '{talk_link}' is locked, {ignoring}")
        except pywikibot.exceptions.PageSaveRelatedError:
            # Any other error during saving; e.g., that a talk page disallows
            # bot edits by using the {{Bots}} template
            warn(f"Can't save user talk page '{talk_link}', {ignoring}")

    def notify_uploader_and_creator(self, files: list[str]) -> None:
        """Notify the uploader and (optionally) the creator of the new FP(s).

        This is STEP 7 of the parking procedure for new featured pictures.
        Call this method only for closed and verified candidates.
        It adds a FP promotion template to the talk page(s) of the uploader(s)
        and (optionally) of the original creator of the promoted image(s).
        (Sometimes the creator is different from the uploader, e.g. if we
        promote a variant of an image which has been retouched by another user.
        In this case we notify also the original creator, if possible.)

        To understand this method and how it differs from notify_nominator(),
        consider that all files in a set nomination have the same nominator,
        but they may have been uploaded by different users.  Of course
        this is very unusual and discouraged by the current FPC rules,
        but the bot stills supports that special case.  Therefore this method
        handles the files one by one, unlike notify_nominator().
        (Theoretically we would also need to support different creators,
        but at least for now we extract the creator name from the nomination,
        therefore we can handle just a single creator per nomination.)

        Args:
            files: List with filename(s) of the featured picture or set.
        """
        ignored_pages: set[str] = set()
        redirects: dict[str, str] = {}  # Mapping: old page name -> new name
        nominator_name = self.nominator(link=False)
        creator_name = self.creator(link=False)
        for filename in files:
            # Check if nominator, uploader and creator are the same user,
            # to avoid adding two templates to the same talk page
            uploader_name = self.uploader(filename, link=False)
            if uploader_name != nominator_name:
                self._notify_uploader_or_creator(
                    filename, True, uploader_name, ignored_pages, redirects
                )
            else:
                out(
                    f"Skipping uploader notification for '{filename}', "
                    "uploader is identical to nominator."
                )
            if (
                creator_name
                and not is_same_user(creator_name, nominator_name)
                and not is_same_user(creator_name, uploader_name)
            ):
                self._notify_uploader_or_creator(
                    filename, False, creator_name, ignored_pages, redirects
                )
            else:
                out(
                    f"Skipping creator notification for '{filename}', "
                    + (
                        "creator is identical to nominator/uploader."
                        if creator_name
                        else "can't identify the creator."
                    )
                )

    def _notify_uploader_or_creator(
        self,
        filename: str,
        is_uploader: bool,
        username: str,
        ignored_pages: set[str],
        redirects: dict[str, str],
    ) -> None:
        """Notify the uploader or creator of a new featured picture.

        Subroutine which implements the uploader/creator notification.

        Args:
            filename: The complete filename of the new featured picture.
            is_uploader: True if we want to notify the uploader,
                False if we want to notify the creator of the picture.
            username: The username of the uploader/creator.
            ignored_pages: A set with problematic user talk pages which
                must be ignored.
            redirects: A dictionary with resolved user talk page redirects,
                used to avoid repeated page redirect resolutions.
        """
        if is_uploader:
            role = "uploader"
            tmpl_name = "FPpromotedUploader"
        else:
            role = "creator"
            tmpl_name = "FPpromotedCreator"
        ignoring = f"but ignoring since it's just the {role} notification."

        # Find and read the user talk page
        talk_link = f"{USER_TALK_NAMESPACE}{username}"
        if talk_link in ignored_pages:
            # Don't load or report undefined or locked talk pages twice
            return
        talk_link = redirects.get(talk_link, talk_link)
        talk_page = pywikibot.Page(_g_site, talk_link)
        try:
            old_text = talk_page.get(get_redirect=False)
        except pywikibot.exceptions.NoPageError:
            # Undefined user talk pages are uncommon because every new user
            # is welcomed by an automatic message.  So better stop here.
            warn(f"The user talk page '{talk_link}' is undefined, {ignoring}")
            ignored_pages.add(talk_link)
            return
        except pywikibot.exceptions.IsRedirectPageError:
            # Try to resolve the redirect
            try:
                talk_page = talk_page.getRedirectTarget()
                old_text = talk_page.get(get_redirect=False)
            except pywikibot.exceptions.PageRelatedError:
                # Circular, nested or invalid redirect etc.
                warn(
                    f"The user talk page '{talk_link}' was moved "
                    f"and the redirect is invalid, {ignoring}"
                )
                ignored_pages.add(talk_link)
                return
            # Record redirect to avoid repeated resolving, update variable
            new_talk_link = talk_page.title()
            redirects[talk_link] = new_talk_link
            out(f"User talk page redirect: '{talk_link}' -> '{new_talk_link}'")
            talk_link = new_talk_link

        # Assemble the template
        subpage_name = self.subpage_name(keep_prefix=True, keep_number=True)
        template = f"{{{{{tmpl_name}|{filename}|subpage={subpage_name}}}}}"

        # Check if there already is a promotion template for the file
        # on the user talk page.  If yes, we skip that file.
        if re.search(wikipattern(template), old_text):
            out(
                f"Skipping {role} notification for '{filename}', "
                f"promotion template is already present at '{talk_link}'."
            )
            return

        # Update the description and commit the new text
        new_text = (
            f"{old_text.rstrip()}\n"
            "\n"
            "== FP Promotion ==\n"
            f"{template} /~~~~"
        )
        summary = f"FP promotion of [[{filename}]]"
        try:
            commit(old_text, new_text, talk_page, summary)
        except pywikibot.exceptions.LockedPageError:
            warn(f"The user talk page '{talk_link}' is locked, {ignoring}")
            ignored_pages.add(talk_link)
        except pywikibot.exceptions.PageSaveRelatedError:
            # Any other error during saving; e.g., that a talk page disallows
            # bot edits by using the {{Bots}} template
            warn(f"Can't save user talk page '{talk_link}', {ignoring}")
            ignored_pages.add(talk_link)


class DelistCandidate(Candidate):
    """A candidate (nomination) for delisting from featured picture status

    Each instance of this class represents a delisting nomination.
    NB: The class handles only the delisting of individual FPs,
    set delisting nominations will be implemented soon.
    """

    # Define class constants
    # Adapt values for the needs of this class:
    _SUCCESS_KEYWORD = "delisted"
    _FAIL_KEYWORD = "not delisted"
    _PRO_VOTE_REGEX = DELIST_VOTE_REGEX
    _CONTRA_VOTE_REGEX = KEEP_VOTE_REGEX
    _NEUTRAL_VOTE_REGEX = NEUTRAL_VOTE_REGEX
    _COUNTED_RES_REGEX = DELIST_COUNTED_TEMPLATE_REGEX
    _REVIEWED_RES_REGEX = DELIST_REVIEWED_TEMPLATE_REGEX
    _VERIFIED_RES_REGEX = VERIFIED_DELIST_RESULT_REGEX
    _OBSOLETE_RES_REGEX = OBSOLETE_DELIST_RESULT_REGEX

    # Declare types of instance variables:
    # all instance variables are inherited, see superclass.

    # No __init__():
    # the class just uses the initializer of the superclass.

    def is_delist_and_replace(self) -> bool:
        """Find out whether this is a delist-and-replace nomination or not."""
        # Check two criteria because sometimes people use unusual formatting
        # (e.g. a <gallery> element) for D&R nominations.  There could be
        # false-positive results, but these are better that false-negative ones
        # because D&R nominations are rare, so false-positive results will
        # certainly be noticed, unlike false-negative ones.
        return (
            self.image_count() > 1
            or bool(DELIST_AND_REPLACE_VOTE_REGEX.search(self.filtered_content()))
        )

    def _type_code(self) -> str:
        """Return a three-letter code of the nomination type for '-info'.

        Overrides the abstract method from the superclass, implementing it
        for delisting candidates.
        """
        return "D&R" if self.is_delist_and_replace() else "Del"

    def get_result_string(self) -> str:
        """Return the results template to be added when closing a nomination.

        Overrides the abstract method from the superclass, implementing it
        for delisting candidates.
        """
        if self.is_delist_and_replace():
            # A delist-and-replace or other special nomination
            return (
                "{{FPC-delist-results-unreviewed"
                "|delist=X|keep=X|neutral=X|delisted=X"
                "|sig=<small>'''Note: This looks like a delist-and-replace "
                "nomination (or something else special). "
                "Therefore it must be counted and processed manually.'''"
                "</small> ~~~~}}"
            )
        # A delisting or set delisting nomination
        self.count_votes()
        return (
            "{{FPC-delist-results-unreviewed"
            f"|delist={self._pro}|keep={self._con}|neutral={self._neu}"
            f"|delisted={yes_no(self.is_passed())}"
            "|sig=~~~~}}"
        )

    def get_close_edit_summary(self, fifth_day: bool) -> str:
        """Return the edit summary to be used when closing a nomination.

        Overrides the abstract method from the superclass, implementing it
        for delisting candidates.
        """
        if self.is_delist_and_replace():
            # A delist-and-replace or other special nomination
            return (
                "Closing for review - looks like a delist-and-replace "
                "nomination, needs manual counting"
            )
        # A delisting or set delisting nomination
        self.count_votes()
        return (
            "Closing for review "
            f"({self._pro} delist, {self._con} keep, {self._neu} neutral, "
            f"delisted: {yes_no(self.is_passed())}, "
            f"5th day: {yes_no(fifth_day)})"
        )

    def _archive_cat_name_base(self) -> str:
        """Return the basic component of a candidate archive category name.

        Overrides the abstract method from the superclass, implementing it
        for delisting candidates.
        """
        return (
            "candidates for delisting and replacing featured pictures"
            if self.is_delist_and_replace()
            else "candidates for delisting from featured picture status"
        )

    def _check_gallery_link(self) -> None:
        """Check if the gallery link is valid and report any problems.

        The method overrides the abstract method from the superclass,
        implementing it for delisting candidates.
        """
        out(f"{self.cut_title()}: OK - delisting candidate, doesn't need a gallery link")

    def handle_passed_candidate(self, results: tuple[str, ...]) -> None:
        """Handle the parking procedure for a passed delisting candidate.

        The method overrides the abstract method from the superclass,
        implementing it for delisting candidates.  It removes the image(s)
        from FP gallery pages, updates the {{Assessents}} template(s)
        and removes FP categories from the image description page(s),
        marks entries in the chronological archives as delisted, etc.
        """
        if self.is_delist_and_replace():
            # Support for delist-and-replace nominations is not implemented.
            ask_for_help(
                f"The nomination [[{self._page.title()}]] looks like "
                "a delist-and-replace nomination or something else special, "
                "but the bot cannot handle this (yet). "
                "Please take care of the images from that nomination "
                "and remove or replace them manually."
            )
            return
        files = self.all_filenames()
        if not files:
            # Could not identify the nominated image(s), error already reported
            return
        self.remove_from_featured_list(files)
        self.remove_from_gallery_pages(files, results)
        self.remove_assessments(files)
        # ! self.remove_assessment_from_media_info(files)  # Disabled 2026-04-01
        self.move_to_log(self._SUCCESS_KEYWORD, "")  # No gallery link here.

    def remove_from_featured_list(self, files: list[str]) -> None:
        """Remove the delisted FP(s) from the list of recent FPs.

        This is STEP 1 of the parking procedure for delisting candidates.
        It removes the delisted featured picture(s) from the list of new FPs
        which is also used (transcluded) on the FP landing page.

        Usually this step is not necessary.  Until August 2025, a comment
        in the code said:
            'We skip checking the FP landing page with the newest FPs;
            the chance that the image is still there is very small,
            and even then that page will soon be updated anyway.'
        This is correct.  But some sections of the list are updated
        only very rarely (e.g. the 'Other lifeforms' section),
        so a delisted FP could stay there for years.  That would be bad,
        and removing a FP from the list is easy, so we just do it.

        Args:
            files: The filename(s) of the delisted featured picture(s).
        """
        # Read the list
        page = pywikibot.Page(_g_site, GALLERY_LIST_PAGE_NAME)
        try:
            old_text = page.get(get_redirect=False)
        except pywikibot.exceptions.PageRelatedError as exc:
            error(f"Error - can't read list of recent FPs: {exc}")
            fexc = format_exception(exc)
            if len(files) == 1:
                file_desc = f"the delisted featured picture [[:{files[0]}]]"
            else:
                file_desc = (
                    "any of the delisted featured pictures "
                    + ", ".join(f"[[:{filename}]]" for filename in files)
                )
            ask_for_help(
                COULD_NOT_READ_RECENT_FPS_LIST.format(exception=fexc)
                + f" If {file_desc} appears on that page, please remove it."
            )
            return

        # Remove the image(s), if present, from the list
        # (there should be at most one of the images in the list,
        # but we better search for all of them to be on the save side)
        removed_files = []
        new_text = old_text
        for filename in files:
            fn_pattern = wikipattern(filename.replace(FILE_NAMESPACE, ""))
            new_text, n = re.subn(r"\n[^\n]*" + fn_pattern + r"[^\n]*", "", new_text)
            if n > 0:
                removed_files.append(filename)
        if not removed_files:
            out(
                f"Skipping remove_from_featured_list() for '{self._page.title()}', "
                "image(s) not found in list."
            )
            return
        summary = f"Removed [[{removed_files[0]}]] "
        if (n := len(removed_files)) > 1:
            summary += f"and {n - 1} more set file(s) "
        summary += f"per [[{self._page.title()}]]"
        commit(old_text, new_text, page, summary)

    def remove_from_gallery_pages(
        self,
        files: list[str],
        results: tuple[str, ...],
    ) -> None:
        """Remove delisted FP(s) from gallery pages and chronological archives.

        This is STEP 2 of the parking procedure for delisting candidates.
        It removes the delisted FP(s) from all FP gallery pages and marks
        their entries in the chronological archive pages as delisted.

        Args:
            files: The filename(s) of the delisted featured picture(s).
            results: A tuple with strings representing the values which
                have been assigned to the individual parameters of the
                reviewed results template in the nomination.
        """
        images_per_page: dict[pywikibot.Page, list[str]] = {}
        for filename in files:
            file_page = pywikibot.FilePage(_g_site, title=filename)
            if not file_page.exists():
                error(f"Error - image '{filename}' not found.")
                ask_for_help(
                    "The bot tried to remove the delisted featured picture "
                    f"[[:{filename}]] from the gallery pages, but it seems "
                    "the file page does not exist. Please check "
                    "whether it has been renamed or deleted, and remove it "
                    "from all featured picture gallery pages."
                )
                continue
            using_pages = file_page.using_pages(
                namespaces=["Commons"], filterredir=False
            )
            for page in using_pages:
                page_name = page.title()
                if not page_name.startswith(FP_PREFIX):
                    # Any other page -- don't remove the image here, of course.
                    continue
                if page_name.startswith(CHRONO_ARCHIVE_PREFIX):
                    # Chronological archive page: mark the image as delisted
                    self._mark_image_as_delisted(page, filename, results)
                else:
                    # FP gallery page -- if this is a set delisting, the page
                    # probably contains several or all images from the set,
                    # so collect them first and then remove all at once.
                    if page in images_per_page:
                        images_per_page[page].append(filename)
                    else:
                        images_per_page[page] = [filename]
        # Now remove the collected images from the gallery page(s)
        for page, filenames in images_per_page.items():
            self._remove_images_from_gallery_page(page, filenames)

    def _remove_images_from_gallery_page(
        self,
        page: pywikibot.Page,
        files: list[str],
    ) -> None:
        """Remove all delisted images from the gallery page."""
        page_name = page.title()
        out(f"Removing delisted image(s) from '{page_name}'...")
        try:
            old_text = page.get(get_redirect=False)
        except pywikibot.exceptions.PageRelatedError as exc:
            error(f"Error - could not read '{page_name}': {exc}")
            formatted = "\n".join(f"* [[:{filename}]]" for filename in files)
            ask_for_help(
                f"The bot could not read the gallery page [[{page_name}]]: "
                f"{format_exception(exc)}. Please check that page "
                "and remove the following delisted featured picture(s):\n"
                f"{formatted}\n"
            )
            return
        removed_files = []
        remaining_files = []
        new_text = old_text
        for filename in files:
            fn_pattern = wikipattern(filename.replace(FILE_NAMESPACE, ""))
            new_text, n = re.subn(
                r"(\[\[)?([Ff]ile|[Ii]mage):" + fn_pattern + r".*\n", "", new_text
            )
            if n == 0:
                remaining_files.append(filename)
            else:
                removed_files.append(filename)
        if remaining_files:
            for filename in remaining_files:
                error(f"Error - could not remove '{filename}' from '{page_name}'.")
            formatted = "\n".join(f"* [[:{filename}]]" for filename in remaining_files)
            ask_for_help(
                "The bot could not remove the following delisted featured "
                f"picture(s) from the gallery page [[{page_name}]]:\n"
                f"{formatted}\nPlease remove them manually."
            )
        if removed_files:
            summary = f"Removed [[{removed_files[0]}]] "
            if (n := len(removed_files)) > 1:
                summary += f"and {n - 1} more set file(s) "
            summary += f"per [[{self._page.title()}]]"
            commit(old_text, new_text, page, summary)

    def _mark_image_as_delisted(
        self,
        page: pywikibot.Page,
        filename: str,
        results: tuple[str, ...],
    ) -> None:
        """Mark the image on a chronological archive page as delisted."""
        page_name = page.title()
        out(f"Adding 'delisted' note to '{page_name}'...")
        try:
            old_text = page.get(get_redirect=False)
        except pywikibot.exceptions.PageRelatedError as exc:
            error(f"Error - could not read '{page_name}': {exc}")
            ask_for_help(
                f"The bot could not read the archive page [[{page_name}]]: "
                f"{format_exception(exc)}. Please check that page "
                f"and mark the former featured picture [[:{filename}]] "
                "as delisted."
            )
            return
        fn_pattern = wikipattern(filename.replace(FILE_NAMESPACE, ""))
        if match := re.search(
            r"((?:[Ff]ile|[Ii]mage):" + fn_pattern + r"(.*))\n", old_text
        ):
            if re.search(r"[Dd]elisted", match.group(2)):
                out(f"Already marked as delisted on '{page_name}'.")
                return
            now = datetime.datetime.now(datetime.UTC)
            entry = (
                # Entries often end with trailing spaces, strip them
                f"{match.group(1).rstrip()}<br> "
                f"'''[[{self._page.title()}|Delisted]] {now:%Y-%m-%d} "
                f"({results[1]}\u2013{results[0]})'''"
            )
            new_text = (
                f"{old_text[:match.start(1)]}{entry}{old_text[match.end(1):]}"
            )
        else:
            # Did not find the image.  That's OK e.g. for the overview
            # pages which include the archives of each half year,
            # therefore don't print an error here.
            out(
                f"Did not find '{filename}' on '{page_name}'; "
                "that's OK if this is just a transclusion page etc."
            )
            return
        summary = f"Delisted [[{filename}]] per [[{self._page.title()}]]"
        commit(old_text, new_text, page, summary)

    def remove_assessments(self, files: list[str]) -> None:
        """Remove the FP status from the image description page(s).

        This is STEP 3 of the parking procedure for delisting candidates.
        It searches the image description page(s) of the delisted FP(s)
        for the {{Assessments}} template and changes the template's
        parameters in order to mark the FP(s) as delisted.

        Args:
            files: The filename(s) of the delisted featured picture(s).
        """
        subpage_name = self.subpage_name(keep_prefix=False, keep_number=True)
        for filename in files:
            # Get and read image description page
            image_page = pywikibot.Page(_g_site, filename)
            try:
                old_text = image_page.get(get_redirect=False)
            except pywikibot.exceptions.PageRelatedError as exc:
                error(f"Error - can't read '{filename}': {exc}")
                ask_for_help(
                    "The bot could not read the delisted featured picture "
                    f"[[:{filename}]]: {format_exception(exc)}. "
                    + CHECK_DELISTED_FP_AND_UPDATE_ASSESSMENTS.format(
                        subpage=subpage_name
                    )
                )
                continue

            # Search and (if found) update the {{Assessments}} template
            found, up_to_date, new_text = update_assessments_template(
                old_text, 2, subpage_name
            )
            if not found:
                error(f"Error - no {{{{Assessments}}}} found on '{filename}'.")
                ask_for_help(
                    "The bot tried to update the {{tl|Assessments}} template "
                    f"for the delisted featured picture [[:{filename}]], "
                    "but did not find the template on that page. "
                    + CHECK_DELISTED_FP_AND_UPDATE_ASSESSMENTS.format(
                        subpage=subpage_name
                    )
                )
                continue
            if up_to_date:
                # This can happen if the process has previously been interrupted.
                out(
                    f"Skipping add_assessments() for '{filename}', "
                    "image is already delisted."
                )
                continue

            # Remove 'Featured pictures of/from/by ...' categories
            new_text = TOPICAL_FP_CATEGORY_REGEX.sub("", new_text)

            # Commit the new text
            summary = f"Delisted per [[{self._page.title()}]]"
            try:
                commit(old_text, new_text, image_page, summary)
            except pywikibot.exceptions.PageSaveRelatedError as exc:
                error(
                    f"Error - can't save '{filename}', "
                    f"can't update {{{{Assessments}}}} template: {exc}"
                )
                ask_for_help(
                    "The bot cannot update the {{tl|Assessments}} template "
                    f"for the delisted featured picture [[:{filename}]] "
                    "because saving changes to that page has failed: "
                    f"{format_exception(exc)}. "
                    + CHECK_DELISTED_FP_AND_UPDATE_ASSESSMENTS.format(
                        subpage=subpage_name
                    )
                )

    def remove_assessment_from_media_info(self, files: list[str]) -> None:
        """Remove the FP assessment claim from the structured data.

        This is STEP 4 of the parking procedure for delisting candidates.
        It removes the 'Commons quality assessment' (P6731) claim
        'Wikimedia Commons featured picture' (Q63348049)
        from the Media Info (structured data) for the former FP(s).

        Args:
            files: The filename(s) of the delisted featured picture(s).
        """
        for filename in files:
            # Get the Media Info for the image
            file_page = pywikibot.FilePage(_g_site, title=filename)
            if not file_page.exists():
                error(f"Error - image '{filename}' not found.")
                ask_for_help(
                    "The bot tried to remove the featured picture "
                    "assessment claim from the delisted featured picture "
                    f"[[:{filename}]], but it seems the file page "
                    "does not exist. Please check whether it has been "
                    "renamed or deleted, and remove the assessment claim."
                )
                return
            media_info = file_page.data_item()
            structured_data = media_info.get(force=True)
            try:
                quality_assessments = structured_data["statements"]["P6731"]
            except KeyError:
                warn(
                    "Found no 'Commons quality assessment' (P6731) claims "
                    f"for '{filename}'."
                )
                return

            # Search for the claim(s) to be removed
            # (normally there should be at most one FP claim, but I have seen
            # weird things, so handle multiple FP claims to be on the save side)
            claims_to_remove = []
            for claim in quality_assessments:
                if is_fp_assessment_claim(claim):
                    claims_to_remove.append(claim)
            if claims_to_remove:
                try:
                    commit_media_info_changes(
                        filename, media_info, claims_to_remove, []
                    )
                except pywikibot.exceptions.PageSaveRelatedError as exc:
                    error(
                        f"Error - saving '{filename}' failed, "
                        f"can't remove FP claim: {exc}")
                    ask_for_help(
                        "The bot tried to remove the featured picture "
                        "assessment claim from the delisted featured picture "
                        f"[[:{filename}]], but saving changes to that page "
                        f"has failed: {format_exception(exc)}. "
                        "Please remove the assessment claim manually."
                    )
            else:
                warn(
                    "Found no 'Wikimedia Commons featured picture' assessment "
                    f"claim (Q63348049) for '{filename}'."
                )


# FUNCTIONS

def wikipattern(text: str) -> str:
    """Prepare a filename, page name etc. for the use in a regex.

    Args:
        text: A short wikitext string, typically a filename or page name.

    Returns:
        A copy of the string in which all characters that have a special
        regex meaning have been escaped, and in which spaces and underscores
        are treated as interchangeable, as usual in MediaWiki filenames,
        page names, etc.
    """
    return re.sub(r"(?:\\ |_)", r"[ _]", re.escape(text))


# If this assertion ever fails, re.escape() handles spaces differently now,
# so please update the regex in the function above.
assert re.escape(" ") == r"\ "


def out(
    text: str,
    newline: bool = True,
    date: bool = False,
    heading: bool = False,
) -> None:
    """Output information or status messages to the console or log.

    Args:
        text: The message that should be printed or logged.
        newline: Whether to end the line after the message or not.
        date: Whether to output date and time before the message or not.
        heading: Whether this is a heading or other emphasized text, or not;
            emphasized text is printed to the console in blue.
    """
    if heading:
        text = f"<<lightblue>>{text}<<default>>"
    dstr = (
        f"{datetime.datetime.now(datetime.UTC).strftime('%Y-%m-%d %H:%M:%S')}: "
        if date and not _g_log_no_time
        else ""
    )
    pywikibot.stdout(f"{dstr}{text}", newline=newline)


def warn(text: str, newline: bool = True) -> None:
    """Output a warning to the console or log.

    Use this if something does not work as expected, but it is probably
    not necessary to take action.
    The text of the warning is printed to the console in yellow.

    Args:
        text: The warning that should be printed or logged.
        newline: Whether to end the line after the warning or not.
    """
    # TODO: Consider to use pywikibot.warning() instead of pywikibot.stdout(),
    # but first clarify whether any log settings need to be be changed
    # on the server then.
    pywikibot.stdout(f"<<lightyellow>>{text}<<default>>", newline=newline)


def error(text: str, newline: bool = True) -> None:
    """Output an error message to the console or log.

    Use this if something does not work and it is probably necessary
    to take action, e.g. to fix the wikitext of a nomination or page,
    or to improve the code of the bot program.
    The text of the error message is printed to the console in red.

    Args:
        text: The error message that should be printed or logged.
        newline: Whether to end the line after the message or not.
    """
    # TODO: Consider to use pywikibot.error() instead of pywikibot.stdout(),
    # but first clarify whether any log settings need to be be changed
    # on the server then.
    pywikibot.stdout(f"<<lightred>>{text}<<default>>", newline=newline)


def find_candidates(
    list_page_name: str,
    which_types: CandidateTypesToProcess,
) -> list[Candidate]:
    """Build a list with candidate objects for all nominations.

    The list is based either on the FPC list page with the current candidates
    or on a log page with closed nominations.
    It retains the original order of entries and omits faulty entries.
    If we find redirects to renamed nomination subpages, they are resolved
    (so the candidate objects point to the actual nominations),
    and the page with the list of candidates is updated.

    Args:
        list_page_name: Name of the page on Commons which lists
            the candidates (technically: transcludes the nomination subpages).
        which_types: A CandidateTypesToProcess object, specifying
            which types of nominations we should process.

    Returns:
        A list of Candidate objects, in the same order as on the list page.
    """
    # Extract nomination subpage names
    out(f"Extracting {which_types.describe()}, checking for redirects...")
    page = pywikibot.Page(_g_site, list_page_name)
    try:
        old_text = page.get(get_redirect=False)
    except pywikibot.exceptions.PageRelatedError as exc:
        error(f"Error - can't read candidate list '{list_page_name}': {exc}.")
        ask_for_help(
            f"The bot could not read the candidate list [[{list_page_name}]]: "
            f"{format_exception(exc)}. {SERIOUS_PROBLEM_CHECK_PAGE}"
        )
        return []
    without_comments = re.sub(r"<!--.+?-->", "", old_text, flags=re.DOTALL)
    subpage_entries = re.findall(
        r"(\{\{ *(Commons:Featured[ _]picture[ _]candidates */[^\n}]+?)\}\})",
        without_comments,
    )
    if not subpage_entries:
        error(f"Error - no candidates found in '{list_page_name}'.")
        ask_for_help(
            f"The candidate list [[{list_page_name}]] does not appear "
            "to contain a single nomination. That is a bit peculiar. "
            "Please check whether this is correct or not."
        )
        return []
    match_pattern = _g_match_pattern.lower()
    candidates: list[Candidate] = []  # Needs type hint: list is invariant.
    redirects: list[tuple[str, str]] = []

    for _, subpage_name in subpage_entries:
        # Skip nominations which are not of the expected type
        if not (candidate_class := which_types.candidate_class(subpage_name)):
            continue
        # Skip nominations which do not match the '-match' argument
        if match_pattern:
            comparison_name = CAND_PREFIX_REGEX.sub("", subpage_name, count=1).lower()
            if match_pattern not in comparison_name:
                continue
        subpage: pywikibot.Page | None  # Help typecheckers.
        subpage = pywikibot.Page(_g_site, subpage_name)
        # Check if nomination exists (filter out damaged links)
        if not subpage.exists():
            error(f"Error - nomination '{subpage_name}' not found, ignoring.")
            ask_for_help(
                LIST_INCLUDES_MISSING_SUBPAGE.format(
                    list=list_page_name, subpage=subpage_name
                )
            )
            continue
        # Check for redirects and and resolve them
        subpage = _resolve_nomination_subpage_redirect(subpage, redirects)
        if subpage is None:
            # Error during resolving, ignore nomination
            continue
        # Rename subpages with bad title (space after namespace prefix, etc.)
        subpage = _rename_nomination_subpage_with_bad_title(subpage, redirects)
        if subpage is None:
            # Error during renaming, ignore nomination
            continue
        # OK, seems the nomination is fine -- append candidate object
        candidates.append(candidate_class(subpage, list_page_name))

    # If we have found any redirects, update the candidates page
    if redirects:
        new_text = old_text
        for old_name, new_name in redirects:
            new_text = new_text.replace(old_name, new_name, 1)
        summary = f"Resolved {len(redirects)} redirect(s) to renamed nomination(s)"
        commit(old_text, new_text, page, summary)
    return candidates


def _resolve_nomination_subpage_redirect(
    subpage: pywikibot.Page,
    redirects: list[tuple[str, str]],
) -> pywikibot.Page | None:
    """Check if a nomination subpage is a redirect and try to resolve it.

    Sometimes nomination subpages are renamed, leaving a redirect which is
    still listed (transcluded) in the list of candidates.  We have to deal
    with the nomination subpage of every candidate several times,
    therefore we try to find and resolve all such redirects right at the
    beginning of each bot run when we read the list of candidates.

    Args:
        subpage: A pywikibot.Page object for the nomination subpage.
        redirects: This list is used to collect all resolved redirects.
            If the page is found to be redirect that can be resolved,
            the function appends a tuple with the old and the new page name
            to this list.

    Returns:
        If successful, a pywikibot.Page object for the real nomination
        subpage; if an error occurs, just None.
    """
    if subpage.isRedirectPage():
        old_name = subpage.title()
        try:
            subpage = subpage.getRedirectTarget()
        except (
            pywikibot.exceptions.PageRelatedError,
            pywikibot.exceptions.SectionError,
        ):
            # Circular or invalid redirect, redirect to invalid section, etc.
            error(f"Error - invalid nomination redirect page '{old_name}', ignoring.")
            ask_for_help(
                f"The nomination subpage [[{old_name}]] "
                f"contains an invalid redirect. {PLEASE_FIX_HINT}"
            )
            return None
        # A nomination subpage must redirect to a whole page, never to a
        # section.  A malformed "#REDIRECT [[Target#Section]]" yields a
        # target that still carries the section on pywikibot versions before
        # 9.3, which added the 'ignore_section' parameter to
        # getRedirectTarget().  Calling Page.get() on such a target later
        # raises an uncaught SectionError and aborts the whole bot run.
        # Treat a redirect that points to a section as invalid instead.
        if subpage.section():
            error(
                f"Error - nomination redirect '{old_name}' points to a "
                "section, ignoring."
            )
            ask_for_help(
                f"The nomination subpage [[{old_name}]] redirects to a "
                "section of a page rather than to a page itself, "
                f"which the bot cannot process. {PLEASE_FIX_HINT}"
            )
            return None
        new_name = subpage.title()
        out(f"Nomination '{old_name}' has been renamed to '{new_name}'")
        redirects.append((old_name, new_name))
    return subpage


def _rename_nomination_subpage_with_bad_title(
    subpage: pywikibot.Page,
    redirects: list[tuple[str, str]],
) -> pywikibot.Page | None:
    """Rename nomination subpages with problematic titles.

    The function checks if a nomination subpage has a bad title which can
    cause problems.  If yes, it tries to rename that subpage; if this is
    not possible, it reports the problem on the FPC talk page.

    Args:
        subpage: A pywikibot.Page object for the nomination subpage.
        redirects: This list is used to collect all resolved redirects.
            If the page gets renamed, the function appends a tuple
            with the old and the new page name to this list.

    Returns:
        If successful, a pywikibot.Page object for the renamed subpage;
        if an error occurs, returns None.
    """
    old_name = subpage.title()
    # Check for invalid names which do not start with the obligatory prefix
    # or do not even allow to determine the nomination type
    # (we do not try to repair such names because the intent is unclear)
    if not VALID_NOMINATION_NAME_START_REGEX.match(old_name):
        error(f"Error - invalid nomination name: '{old_name}'.")
        ask_for_help(
            f"The nomination subpage [[{old_name}]] has an invalid name: "
            f"it does not begin with the obligatory ''{CAND_PREFIX}'' or "
            "contains neither the <code>File:</code> namespace prefix "
            "nor the signal word <code>/Set/</code>. "
            "Therefore the type of the nomination is undefined. "
            "The bot may fail to handle that nomination correctly, "
            "the {{tl|Assessments}} template will not be able "
            "to link to that nomination, "
            "and editors will be confused, too. "
            f"{PLEASE_RENAME_HINT.format(subpage=old_name)}"
        )
        return None
    # Check for fixable problems: space after namespace prefix, etc.
    new_name = re.sub(r" */ *(?:[Ff]ile|[Ii]mage) *: *", "/File:", old_name, count=1)
    new_name = re.sub(r" */ *[Ss]et */ *", "/Set/", new_name, count=1)
    new_name = re.sub(r" */ *[Rr]emoval */ *", "/removal/", new_name, count=1)
    if new_name != old_name:
        if pywikibot.Page(_g_site, new_name).exists():
            error(
                "Error - could not fix name of nomination subpage "
                f"'{old_name}', ignoring."
            )
            ask_for_help(
                f"The nomination subpage [[{old_name}]] has a tricky name: "
                "it contains irritating spaces around the namespace prefix "
                "etc., or uses the old <code>Image:</code> namespace. "
                "This can confuse the {{tl|Assessments}} template and "
                "similar tools, and does not help the FP maintainers. "
                f"The bot tried to rename the subpage to [[{new_name}]], "
                "but there is already a subpage with that name. "
                f"{PLEASE_RENAME_HINT.format(subpage=old_name)}"
            )
            return None
        out(f"\nAbout to rename nomination '{old_name}' to '{new_name}':")
        if _confirm_changes(old_name, summary=NOMINATION_SUBPAGE_RENAMED):
            subpage = subpage.move(
                new_name,
                reason=NOMINATION_SUBPAGE_RENAMED,
                noredirect=False,  # Avoid problems with any links.
            )
            out(f"Renamed nomination '{old_name}' to '{new_name}'.")
            redirects.append((old_name, new_name))
            # Update all self-links (and all links to the nominated image)
            # in the nomination subpage
            old_text = subpage.get(get_redirect=False)
            new_text = old_text.replace(old_name, new_name)
            new_text = re.sub(
                r"(\[\[:?|/) *(?:[Ff]ile|[Ii]mage) *: *", r"\1File:", new_text
            )
            if new_text != old_text:
                summary = "Updated links after renaming nomination subpage"
                commit(old_text, new_text, subpage, summary)
            # NB: If for any reason this update fails (or is not confirmed
            # in interactive mode) the self-links in the nomination still work
            # thanks to the redirect created above.
        else:
            out(f"Renaming of '{old_name}' ignored.")
    return subpage


def check_candidates(
    check: Callable[[Candidate], None],
    list_page_name: str,
    which_types: CandidateTypesToProcess,
    descending: bool = True,
) -> None:
    """Call a function on each nomination found in the list of candidates.

    Args:
        check: A method of the Candidate class which should be called
            on each candidate.
        list_page_name: The name of the page which includes all nominations;
            i.e. either the page with the list of current candidates
            or a log page that we want to check for test purposes.
        which_types: A CandidateTypesToProcess object, specifying
            which types of nominations we should process.
        descending: Specify True if the page puts the newest entries first,
            False if it runs from the oldest to the newest entry.
            So we can always handle the candidates in chronological order.
    """
    if _g_site is None:  # Test is also necessary to help typecheckers.
        error("Fatal error - _g_site not initialized, call main() first.")
        return

    # Find all current candidates
    candidates = find_candidates(list_page_name, which_types)
    if not candidates:
        out(
            f"Found no {which_types.describe()}"
            f"{' matching the -match argument' if _g_match_pattern else ''}."
        )
        return
    if descending:
        candidates.reverse()

    # Handle each candidate with the specified method
    total = len(candidates)
    for i, candidate in enumerate(candidates, start=1):
        if not _g_threads:
            out(f"({i:03d}/{total:03d}) ", newline=False, date=True)
        try:
            if _g_threads:
                while threading.active_count() >= pywikibot.config.max_external_links:
                    time.sleep(0.1)
                thread = ThreadCheckCandidate(candidate, check)
                thread.start()
            else:
                check(candidate)
                candidate.clear_cache()
        except pywikibot.exceptions.NoPageError as exc:
            error(f"Error - no such page: '{exc}'")
            ask_for_help(
                f"During the processing of [[{candidate.page.title()}]], "
                "the bot could not find a page (perhaps it has been renamed "
                f"without leaving a redirect): {format_exception(exc)}. "
                f"{SERIOUS_PROBLEM_CHECK_PAGE}"
            )
        except pywikibot.exceptions.IsRedirectPageError as exc:
            error(f"Error - unexpected redirect: '{exc}'")
            ask_for_help(
                f"During the processing of [[{candidate.page.title()}]], "
                "the bot found that a page contains an unexpected redirect: "
                f"{format_exception(exc)}. {SERIOUS_PROBLEM_CHECK_PAGE}"
            )
        except pywikibot.exceptions.PageSaveRelatedError as exc:
            error(f"Error - can't save page: '{exc}'")
            ask_for_help(
                f"During the processing of [[{candidate.page.title()}]], "
                "the bot could not save changes to the text of a page: "
                f"{format_exception(exc)}. This can be harmless, but also "
                "a serious error. Please check that page. If the issue "
                "with the page does not get fixed, you may need to process "
                "that nomination manually."
            )
        # These exceptions just stop the candidate, continue with the next one.
        if _g_abort:
            break
        # Take the load off the server and avoid 429 'Too many requests' errors
        time.sleep(1)


def filter_content(text: str) -> str:
    """Filter out content that must be ignored when counting votes etc.

    Currently this includes:
    * comments
    * the <s>, <strike>, <del> tags for striking out text
    * the <nowiki> tag which is just for displaying syntax
    * the {{Strikethrough|...}} template
    * image notes
    * collapse boxes

    Args:
        text: The wikitext string that should be filtered.

    Returns:
        The filtered text.
    """
    text = re.sub(r"<!--.*?-->", "", text, flags=re.DOTALL)
    text = strip_tags(text, r"s(?:trike)?|del|nowiki")
    text = strip_templates(text, r"[Ss]trikethrough")
    text = re.sub(
        r"\{\{\s*[Ii]mageNote\s*\|.*?\}\}.*?\{\{\s*[iI]mageNoteEnd.*?\}\}",
        "",
        text,
        flags=re.DOTALL,
    )
    text = re.sub(
        r"\{\{\s*[Cc](?:ollapse[ _]top|ot)\s*\|.*?\}\}.*?"
        r"\{\{\s*[Cc](?:ollapse[ _]bottom|ob)\s*\}\}",
        "",
        text,
        flags=re.DOTALL,
    )
    return text


def strip_tags(text: str, tags: str) -> str:
    """Remove all instances of some HTML tag(s) (incl. contents) from the text.

    Args:
        text: The wikitext you want to free from the HTML tag(s).
        tags: String with the name(s) of the HTML tags you want to remove;
            handled as a regex fragment, so you can supply several names
            by separating them with '|' or by using '(?:...)?', etc.

    Returns:
        The filtered text.
    """
    return re.sub(
        "<(" + tags + r")(?:\s[^>]*)?>.*?</\1\s*>",
        "",
        text,
        flags=re.DOTALL | re.IGNORECASE,
    )


def strip_templates(text: str, template_names: str) -> str:
    """Remove all instances of the specified template(s) from the text.

    Args:
        text: The wikitext you want to free from the template(s).
        template_names: String with the names of the template(s) to remove;
            handled as a regex fragment, so you can supply several names
            by separating them with '|' or by using '(?:...)?', etc.

    Returns:
        The filtered text.
    """
    pos = 0
    while tmpl_pos := find_template_pos(text, template_names, pos):
        text = f"{text[:tmpl_pos.start]}{text[tmpl_pos.stop:]}"
        pos = tmpl_pos.start
    return text


def clean_gallery_link(gallery_link: str) -> str:
    """Clean and polish a gallery link.

    The function
    * replaces underscores and non-breaking spaces by plain spaces
      (underscores are present if users just copy the link,
      a NBSP can be entered by accident with some keyboard settings);
    * replaces double spaces by a single one;
    * removes leading/trailing whitespace;
    * replaces %-encoded characters by their plain text counterparts.

    Args:
        gallery_link: The gallery link from the nomination.

    Returns:
        The polished gallery link.
    """
    link = re.sub(r"[ _\u00A0]+", " ", gallery_link).strip()
    if "%" in link:
        try:
            link = urllib.parse.unquote(link, errors="strict")
        except UnicodeError:
            # Keep the encoded value of the link, just log the error
            error("Error - invalid %xx escape in gallery link.")
    return link


def split_gallery_link(gallery_link: str) -> tuple[str, str]:
    """Split the gallery link into gallery page name and section anchor.

    Args:
        gallery_link: The gallery link from the nomination.

    Returns:
        A tuple, containing:
        [0] str: The name of the gallery page, without the leading
            'Commons:Featured pictures/'.
        [1] str: The section anchor (can be empty).
    """
    link_parts = gallery_link.split("#", maxsplit=1)
    page_name = link_parts[0].strip()
    section = link_parts[1].strip() if len(link_parts) > 1 else ""
    return (page_name, section)


def bare_filename(filename: str) -> str:
    """Return the bare filename without 'File:' prefix and file extension.

    This is useful for labels, image captions, etc.

    Args:
        filename: The name of a file on Commons, e.g. 'File:Test 1.jpg'.

    Returns:
        The bare name of that file, e.g. 'Test 1'.
    """
    return re.sub(
        r"^(?:[Ff]ile|[Ii]mage):(.+?)\.\w{2,4}$",
        r"\1",
        filename,
        count=1,
    ).strip()


def is_just_thumbnail(image_link: str) -> bool:
    """Find out if a wikitext image link is just a thumbnail or icon.

    Args:
        image_link: A wikitext image link, e.g. '[[File:Test.jpg|thumb]]'.

    Returns:
        True if the link represents just a thumbnail or icon, else False.
    """
    if IMAGE_THUMB_REGEX.search(image_link):
        return True
    size = IMAGE_SIZE_REGEX.search(image_link)
    if size and (int(size.group(1)) <= 150):
        return True
    return False


def yes_no(value: bool) -> str:
    """Translate a boolean value to 'yes' and 'no' resp."""
    return "yes" if value else "no"


def y_n(value: bool) -> str:
    """Translate a boolean value to 'Y' and 'N' resp., for use in tables."""
    return "Y" if value else "N"


def user_page_link(username: str) -> str:
    """Build a link to the user page of the specified user."""
    return f"[[{USER_NAMESPACE}{username}|{username}]]"


def is_same_user(username_1: str, username_2: str) -> bool:
    """Find out if two usernames are effectively identical.

    The MediaWiki software handles the first character of page names
    and user names (after the namespace prefix) case-insensitively,
    but the remaining characters case-sensitively.  We must imitate
    that behaviour in order to compare usernames accurately.

    Args:
        username_1: The first username.
        username_2: The second username.

    Returns:
        True if the two usernames are effectively identical, else False.
    """
    return (
        username_1[0].upper() == username_2[0].upper()
        and username_1[1:] == username_2[1:]
    )


def build_log_page_name(month_name: str, year: int, part: int) -> str:
    """Assemble the page name of a FPC log page.

    Args:
        month_name: The English name of a month, e.g. 'January'.
        year: The year as an integer, e.g. 2025.
        part: The current part of the monthly log, e.g. 1, 2, 3 ...

    Returns:
        The full page name of the current log page part,
        e.g. 'Commons:Featured picture candidates/Log/January 2025-1'.
    """
    return f"{CAND_LOG_PREFIX}{month_name} {year}-{part}"


def name_to_sort_key(name: str) -> str:
    """Convert a page or nomination base name to a sort key.

    It is more or less impossible to sort arbitrary strings in a way which
    is satisfying for all languages and styles at the same time, because
    languages have very different rules for collation etc.
    This function is just a compromise: it is simple enough to be practical
    and at least allows for better sorting of many pages; e.g., it makes sure
    that we sort 'Château.jpg' before 'Chur.jpg' and not after it,
    and sort '"All right".jpg' under 'A', not under '"'.

    The function removes accents and diacritics; it translates some carefully
    selected compound characters acc. to their most common interpretation
    for sorting; it removes common punctuation marks which are usually
    not taken into account when sorting items, or replaces them by spaces;
    and unifies all (runs of) whitespace chars to simple spaces.
    """
    try:
        result = unicodedata.normalize("NFKD", name.strip())
        result = "".join(
            char for char in result if not unicodedata.combining(char)
        )
    except UnicodeError:
        return name
    result = result.translate(AUXILIARY_SORT_KEY_TRSL_TABLE)
    result = REMOVE_FROM_SORT_KEY_REGEX.sub("", result)
    result = REPLACE_BY_SPACE_IN_SORT_KEY_REGEX.sub(" ", result)
    result = REMOVE_FROM_START_OF_SORT_KEY_REGEX.sub("", result, count=1)
    result = RUN_OF_WHITESPACE_REGEX.sub(" ", result)
    result = REMOVE_SPACE_BEFORE_INTERPT_REGEX.sub(r"", result)
    return result.strip()


def format_exception(exc: Exception) -> str:
    """Format an exception nicely in order to use it in requests for help."""
    # Pywikibot exception messages often (but not always) end with '.',
    # therefore we remove any trailing period to allow us to integrate
    # the message nicely into a full sentence.
    message = str(exc).strip().rstrip(".")
    # Some messages contain a NL which breaks our formatting with ''...''
    message = message.replace("\n", " ")
    # Link to files instead of displaying them
    message = re.sub(r"\[\[ *(?:[Ff]ile|[Ii]mage) *: *", "[[:File:", message)
    # Escape quoted template names with {{tl|...}},
    # except the template-linking templates {{tl}}, {{t}}, {{t0}}, etc.
    message = re.sub(
        r"\{\{\s*(?!(?:t[0-9]?|tl[a-z]*)\s*\|)",
        r"{{tl|",
        message,
        flags=re.IGNORECASE,
    )
    name = type(exc).__name__
    return f"''{message}'' (<code>{name}</code>)"


def is_fp_assessment_claim(claim: pywikibot.page.Claim) -> bool:
    """Find out if a Claim object is a FP assessment claim.

    The function tests whether a Pywikibot page.Claim object
    represents a 'Commons quality assessment' (P6731) claim
    with the value 'Wikimedia Commons featured picture' (Q63348049).
    """
    # For now a simple string comparison seems sufficient.
    # If this fails because of format variations etc.,
    # use a regex comparison or explore the nested data values.
    plain = repr(claim)
    return "'property': 'P6731'" in plain and "'numeric-id': 63348049" in plain


def oldest_revision_user(page: pywikibot.Page) -> str:
    """Return the name of the user who has originally created a Commons page.

    Args:
        page: A pywikibot.Page object for a page (e.g. the description page
            of an image or a nomination subpage) on Commons.

    Returns:
        The name of the user who has created the oldest (first) revision
        of the page; on errors just ''.
    """
    try:
        username = page.oldest_revision["user"]
        assert isinstance(username, str)  # Catch errors, help typecheckers.
        return username.strip()
    except (pywikibot.exceptions.PageRelatedError, AttributeError):
        return ""


def find_template_pos(
    text: str,
    template_names: str,
    pos: int = 0,
) -> slice | None:
    """Search for a template to find its start and stop character indices.

    We use a specific function because normal regexes as supported by
    Python's 're' module cannot properly deal with nested templates.

    Args:
        text: The wikitext of the page you want to search.
        template_names: String with the allowed template name(s);
            handled as a regex fragment, so you can supply several names
            by separating them with '|' or by using '(?:...)?', etc.
        pos: Start position of the search (optional).  This allows you
            to search for the next template instance etc.

    Returns:
        A slice with the positions of the first character of the template
        as start and the position of the first character after the template
        as stop value; or None if the template was not found.
    """
    # NB: we need parens around template_names because it may contain '|'.
    pattern = re.compile(r"(\{\{\s*(?:" + template_names + r")\s*)[|{}]")
    match = pattern.search(text, pos=pos)
    if not match:
        return None
    lvl = 0
    cp = match.end(1)

    while cp < len(text):
        ns = text.find("{{", cp)
        ne = text.find("}}", cp)

        # If we see no end tag, we give up
        if ne == -1:
            return None

        # Handle case when there are no more start tags
        if ns == -1:
            if not lvl:
                return slice(match.start(0), ne + 2)
            lvl -= 1
            cp = ne + 2

        elif not lvl and ne < ns:
            return slice(match.start(0), ne + 2)
        elif ne < ns:
            lvl -= 1
            cp = ne + 2
        else:
            lvl += 1
            cp = ns + 2
    # Apparently we never found it
    return None


def update_assessments_template(
    old_text: str,
    featured_value: Literal[1, 2],
    com_nom_value: str,
) -> tuple[bool, bool, str]:
    """Update the {{Assessments}} template on an image description page.

    This function is used both for FP and delisting candidates.
    If the wikitext of an image description page already contains
    an {{Assessments}} template, the function updates it to
    the specified 'featured' and 'com-nom' values.
    It also updates the old 'subpage' parameter name to 'com-nom',
    but preserves the wikitext formatting of the {{Assessments}} template
    because sometimes users format it with spaces, newlines, etc.

    Args:
        old_text: The wikitext of the image description page.
        featured_value: New value for the 'featured' parameter;
            possible values: 1 (featured) or 2 (formerly featured).
        com_nom_value: New value for the 'com-nom' parameter.

    Returns:
        A tuple, containing:
        [0] bool: Did the text contain an {{Assessments}} template?
        [1] bool: If there was a template, was it already up-to-date?
        [2] str: The updated text of the image description page.
    """
    if match := ASSESSMENTS_TEMPLATE_REGEX.search(old_text):
        params = match.group(1)
        # Search and update/append 'featured' parameter
        fstr = str(featured_value)
        if m := re.search(r"\|\s*featured\s*=\s*(\w+)", params):
            if m.group(1) != fstr:
                params = f"{params[:m.start(1)]}{fstr}{params[m.end(1):]}"
                after = m.start(1) + len(fstr)
            else:
                after = m.end(1)
        else:
            params += f"|featured={fstr}"
            after = len(params)
        # Search and update/append 'com-nom' parameter
        # (the end of the regex is so complicated because we want to leave
        # any whitespace after the 'com-nom' value unchanged,
        # therefore it must be excluded from group 2)
        if m := re.search(
            r"\|\s*(com-nom|subpage)\s*=\s*(.+?)\s*(?:$|[{}|\n])", params
        ):
            if m.group(1) == "subpage":
                # We can replace the old name directly because the length
                # of 'subpage' and 'com-nom' is identical
                params = f"{params[:m.start(1)]}com-nom{params[m.end(1):]}"
            if m.group(2) != com_nom_value:
                params = f"{params[:m.start(2)]}{com_nom_value}{params[m.end(2):]}"
        else:
            # Insert new 'com-nom' right after the 'featured' parameter
            params = f"{params[:after]}|com-nom={com_nom_value}{params[after:]}"
        # Check and assemble result
        if params == match.group(1):
            return (True, True, old_text)
        new_text = f"{old_text[:match.start(1)]}{params}{old_text[match.end(1):]}"
        return (True, False, new_text)
    return (False, False, old_text)


def ask_for_help(message: str) -> None:
    """Add a short notice to the FPC talk page, asking for help.

    The request for help feature was introduced in July 2025 to allow
    a simple communication between the bot program and the FPC regulars.
    It allows users to recognize and to fix problems without having
    to search and decipher the log files.  It is especially useful
    if the problem is very probably caused by a broken link, some
    wikitext syntax error, etc. on a Commons page, i.e. simple issues
    a normal human editor can correct easily.

    Args:
        message: A concise description of the problem in one or two
            short, but complete sentences; normally they should end
            with a request to change this or that to help the bot.
    """
    talk_page = pywikibot.Page(_g_site, FP_TALK_PAGE_NAME)
    try:
        old_text = talk_page.get()
    except Exception as exc:  # pylint: disable=broad-exception-caught
        # Catch any exceptions to avoid circular exceptions
        error(f"Error - could not read FPC talk page: {exc}")
        return
    if message in old_text:
        return  # Don't post the same message twice.
    new_text = old_text.rstrip() + (
        f"\n\n== {BOT_NAME} asking for help ==\n"
        f"[[File:Robot icon.svg|64px|left|link={USER_NAMESPACE}{BOT_NAME}]]\n"
        f"{message} Thank you! / ~~~~"
    )
    try:
        commit(old_text, new_text, talk_page, "Added request for help")
    except Exception as exc:  # pylint: disable=broad-exception-caught
        # Catch any exceptions to avoid circular exceptions
        error(f"Error - could not add post to FPC talk page: {exc}")


def _confirm_changes(page_name: str, summary: str | None = None) -> bool:
    """Ask the supervising user whether to save or discard the changes.

    Args:
        page_name: Name (title) of the page on Wikimedia Commons.
        summary: Optional string with the edit summary; omit if there is
            no custom edit summary, e.g. for structured data changes.

    Returns:
        True if changes should be saved, False if changes should be discarded.
        (If the user decides to quit, we quit immediately, no return value.)
    """
    if _g_dry:
        return False
    if _g_auto:
        return True
    choice = pywikibot.bot.input_choice(
        f"Do you want to accept these changes to '{page_name}'"
        + (f" with summary '{summary}'?" if summary else "?"),
        [("yes", "y"), ("no", "n"), ("quit", "q")],
        automatic_quit=False,
    )
    match choice:
        case "y":
            return True
        case "n":
            return False
        case "q":
            out("Aborting.")
            sys.exit()
        case _:
            error("Congrats, you found a bug in pywikibot.bot.input_choice().")
            sys.exit()


def commit(
    old_text: str,
    new_text: str,
    page: pywikibot.Page,
    summary: str,
    minor: bool = False,
) -> None:
    """Commit the new text of the page.

    Unless running in automatic mode the function first shows a diff
    and asks whether the supervising user accepts the changes or not.

    Args:
        old_text: Old text of the page, used to show the diff.
        new_text: New text of the page to be submitted.
        page: Pywikibot Page object for the concerned page on Commons.
        summary: The edit summary for the page history.
        minor (optional): Is this a minor edit?
    """
    # Show the diff
    page_name = page.title()
    out(f"\nAbout to commit changes to '{page_name}':")
    lines_of_context = 0 if (_g_auto and not _g_dry) else 3
    pywikibot.showDiff(
        old_text,
        new_text,
        context=lines_of_context,
    )

    # Decide whether to save the changes
    if _confirm_changes(page_name, summary):
        page.text = new_text
        page.save(summary=summary, watch=None, minor=minor, bot=True)
    else:
        out(f"Changes to '{page_name}' ignored.")


def commit_media_info_changes(
    filename: str,
    media_info: pywikibot.MediaInfo,
    claims_to_remove: list[pywikibot.page.Claim],
    claims_to_add: list[pywikibot.page.Claim],
) -> None:
    """Apply changes to the Media Info (structured data) of an image.

    When changing the Media Info (structured data) for an image, we cannot
    use the normal commit mechanism.  Therefore we print a kind of
    home-made diff.  In interactive mode we ask the supervising user
    whether to save the changes or not; if yes, we apply the changes.

    Args:
        filename: Name of the affected image file.
        media_info: A Pywikibot MediaInfo instance representing
            the Media Info (structured data) for the image.
        claims_to_remove: A list of Pywikibot Claim instances
            representing the statement(s) to be removed; can be empty.
        claims_to_add: A list of Pywikibot Claim instances
            representing the statement(s) to be added; can be empty.
    """
    assert claims_to_remove or claims_to_add

    # Show the diff
    out(f"\nAbout to change the Media Info (structured data) of '{filename}':")
    if claims_to_remove:
        removing = "- " + "\n- ".join(repr(claim) for claim in claims_to_remove)
        pywikibot.stdout(f"<<lightred>>{removing}<<default>>")
    if claims_to_add:
        adding = "+ " + "\n+ ".join(repr(claim) for claim in claims_to_add)
        pywikibot.stdout(f"<<lightgreen>>{adding}<<default>>")

    # Decide whether to save the changes
    if _confirm_changes(filename):
        # TODO: Whenever this is called a 2nd time or after another change,
        # I get a Pywikibot warning:
        #   'WARNING: API error badtoken: Invalid CSRF token.'
        # This appears to be harmless (the change is still saved),
        # but keep an eye on it.  Maybe related to Pywikibot bugs, cf. e.g.:
        #   https://phabricator.wikimedia.org/T261050
        if claims_to_remove:
            media_info.removeClaims(claims_to_remove)
        if claims_to_add:
            for claim in claims_to_add:
                media_info.addClaim(claim, bot=True)
        out(f"Media Info (structured data) of '{filename}' changed.")
    else:
        out(f"Changes to '{filename}' ignored.")


def main(*args: str) -> None:
    """Run the bot program.

    This function is the main entry point of the bot program.
    It parses and checks the CLI arguments, defines global variables,
    selects the desired tasks and calls the appropriate functions.

    Args:
        *args: If you run the script in the usual way as bot program,
            this function is called without any arguments and uses
            the CLI arguments.  However for test purposes etc. you can
            import the script like a module and call this method
            from Python code.  In this case pass strings with the same
            values as the CLI arguments, then '*args' packs all these
            values into a single tuple.
    """
    global _g_site

    # Acquire CLI arguments, let Pywikibot handle the global arguments
    # (including '-help') and get the rest as a simple list
    override_args = args if args else None
    try:
        local_args = pywikibot.handle_args(args=override_args, do_help=True)
    except ConnectionError:
        error("Error - can't connect to the Commons server, aborting.")
        sys.exit()

    # Pywikibot can create the site object only after handling the arguments
    _g_site = pywikibot.Site()
    if not _g_site.logged_in():
        _g_site.login()

    # Inspect local arguments and perform the desired task(s)
    try:
        task_args, which_types = _inspect_local_arguments(local_args)
        for task in task_args:
            _handle_task(task, which_types)
    except Exception as exc:  # pylint: disable=broad-exception-caught
        # 'Last resort' exception handler:
        # report exception with stack trace on the FPC talk page
        stack_trace = traceback.format_exc().rstrip()
        stack_trace = re.sub(  # Abbreviate file paths to filenames
            r'(File ").+?/([^/\n]+\.py")', r"\1\2", stack_trace
        )
        ask_for_help(
            "The bot has stopped because of an uncaught exception: "
            f"{format_exception(exc)}. Abbreviated stack trace:\n"
            f"<pre>{stack_trace}</pre>\n"
            "Developers, please look into this."
        )
        raise exc  # Stop the bot and enable full error logging.
    out("Done.", heading=True)


def _inspect_local_arguments(
    local_args: list[str]
) -> tuple[list[str], CandidateTypesToProcess]:
    """Derive desired task(s) and options from the local arguments.

    Args:
        local_args: A list with local argument strings like '-park'.

    Returns:
        A tuple, containing:
        [0] list[str]: The task arguments in the original order specified
            by the user.
        [1] A CandidateTypesToProcess object specifying which types
            of nominations we want to process.
    """
    global _g_auto
    global _g_dry
    global _g_threads
    global _g_log_no_time
    global _g_match_pattern

    # Define default values
    fpc = delist = False
    task_args = []

    # First look for arguments which act as options for all tasks
    # and extract the task arguments, keeping them in the desired order
    i = 0
    while i < len(local_args):
        arg = local_args[i]
        match arg:
            case "-auto":
                _g_auto = True
            case "-dry":
                _g_dry = True
            case "-threads":
                _g_threads = True
            case "-fpc":
                fpc = True
            case "-delist":
                delist = True
            case "-notime":
                _g_log_no_time = True
            case "-match":
                # So the next argument must be the pattern string
                try:
                    _g_match_pattern = local_args[i + 1]
                except IndexError:
                    error("Error - '-match' must be followed by a pattern, aborting.")
                    sys.exit()
                i += 1  # Skip the pattern argument.
            case _:
                task_args.append(arg)
        i += 1

    # If neither -fpc nor -delist is used we handle all candidates
    if not (fpc or delist):
        fpc = delist = True
    which_types = CandidateTypesToProcess(fpc, delist)

    # We can't use the interactive mode with threads
    if _g_threads and (not _g_dry and not _g_auto):
        error("Error - '-threads' must be used with '-dry' or '-auto'.")
        sys.exit()

    # Check the task arguments
    if not task_args:
        error(
            "Error - you need to specify at least one task "
            "like '-info', '-close', '-park', etc.; see '-help'."
        )
        sys.exit()
    if invalid_args := set(task_args) - {
        "-test", "-info", "-checkgallery", "-close", "-park"
    }:
        # To provide a helpful error message, abort before handling even
        # the first argument and report all invalid arguments at once.
        formatted = ", ".join(f"'{arg}'" for arg in sorted(invalid_args))
        error(f"Error - unknown argument(s) {formatted}; aborting, see '-help'.")
        sys.exit()

    return (task_args, which_types)


def _handle_task(task: str, which_types: CandidateTypesToProcess) -> None:
    """Perform the desired task for all concerned nominations.

    Args:
        task: String describing the task we want to execute; see the code
            of this function for the allowed values.
        which_types: A CandidateTypesToProcess object specifying which types
            of nominations we want to process.
    """
    match task:
        case "-test":
            out("Recounting votes for testing...", heading=True)
            check_candidates(
                Candidate.compare_result_to_count,
                TEST_LOG_PAGE_NAME,
                which_types,
                descending=False,
            )
        case "-info":
            out("Gathering information about candidates...", heading=True)
            check_candidates(Candidate.print_all_info, CAND_LIST_PAGE_NAME, which_types)
        case "-checkgallery":
            if not which_types.fp:
                warn("Checking gallery links makes sense only with FP candidates.")
                return
            out("Checking gallery links...", heading=True)
            check_candidates(Candidate.check_gallery, CAND_LIST_PAGE_NAME, which_types)
        case "-close":
            out("Closing finished candidates...", heading=True)
            check_candidates(Candidate.close, CAND_LIST_PAGE_NAME, which_types)
        case "-park":
            if _g_threads and _g_auto:
                warn("Auto-parking using threads is disabled for now...")
                return
            out("Parking finished candidates...", heading=True)
            check_candidates(Candidate.park, CAND_LIST_PAGE_NAME, which_types)
        case _:
            # This means we have forgotten to update the invalid_args test.
            error(f"Error - unknown argument '{task}'; aborting, see '-help'.")
            sys.exit()


def signal_handler(signal_number: int, frame: FrameType | None) -> None:
    """Handle a SIGINT (keyboard, Ctrl-C) interrupt.

    The parameters of this function must follow the scheme demanded
    by signal.signal() for a custom signal handler.  Therefore we must
    specify parameters even if we do not need them.

    Args:
        signal_number: The number of the signal, one of the constants
            defined in Python's 'signal' module.  Because we use this handler
            only for keyboard interrupts, that value must be signal.SIGINT.
        frame: the current stack frame -- a frame object or None.
    """
    global _g_abort
    print("\n\nReceived SIGINT, will abort...\n")
    _g_abort = True


# PROGRAM SETUP

# Install a custom handler for SIGINT (keyboard, Ctrl-C) interrupts
signal.signal(signal.SIGINT, signal_handler)

# Define the entry point for the bot program with the common idiom
if __name__ == "__main__":
    try:
        main()
    finally:
        pywikibot.stopme()