#!/usr/bin/python
# -*- coding: utf-8 -*-
#
# Published by zhuyifei1999 (https://wikitech.wikimedia.org/wiki/User:Zhuyifei1999)
# under the terms of Creative Commons Attribution-ShareAlike 3.0 Unported (CC BY-SA 3.0)
# https://creativecommons.org/licenses/by-sa/3.0/

from __future__ import unicode_literals

import calendar
import collections
import itertools
import re
import weakref

import pywikibot

pywikibot.handle_args()

SITE = pywikibot.Site('en', 'wikipedia')

linenos = weakref.WeakSet()


class Lineno(object):
    def __init__(self, lineno):
        self.lineno = lineno
        linenos.add(self)

    def get(self):
        assert self.lineno is not None
        return self.lineno

    def rm_lineno(self, lineno):
        if self.lineno is None:
            return
        if lineno == self.lineno:
            self.lineno = None
        elif lineno < self.lineno:
            self.lineno -= 1

    def __repr__(self):
        return 'Lineno(%d)' % self.lineno


H2 = collections.namedtuple('H2', 'lineno year month')
H3 = collections.namedtuple('H2', 'lineno pagename')

reqpage = pywikibot.Page(
    SITE, 'Wikipedia:WikiProject_Guild_of_Copy_Editors/Requests')

reqpage.origtext = reqpage.text


def rm_lineno(lineno):
    for linenoobj in linenos:
        linenoobj.rm_lineno(lineno)

    text = reqpage.text.split('\n')
    text.pop(lineno)
    reqpage.text = '\n'.join(text)


def rm_empty_lineno(lineno):
    text = reqpage.text.split('\n')
    if lineno < len(text) and not text[lineno]:
        rm_lineno(lineno)


sections = collections.defaultdict(
    lambda: collections.defaultdict(lambda: lambda: None))
sections_flat = []

h3 = h2 = None
for i, line in enumerate(reqpage.text.split('\n')):
    if not line:
        continue

    i = Lineno(i)
    reobj = re.match(r'^== *(?P<month>%s) +(?P<year>\d{4,}) *== *$' % ('|'.join(
        filter(None, set(calendar.month_name) | set(calendar.month_abbr)))),
        line)
    if reobj:
        month = reobj.group('month')
        try:
            month = list(calendar.month_name).index(month)
        except ValueError:
            month = list(calendar.month_abbr).index(month)

        year = int(reobj.group('year'))
        h2 = H2(i, year, month)
        h3 = None

    if not h2:
        continue

    reobj = re.match(r"^=== *(?:<s>)?('*\[\[[^\[\]]+\]\]'*)(?:</s>)? *=== *$", line)
    if reobj:
        h3 = H3(i, reobj.group(1))
        obj = sections[h2][h3]
        sections_flat.append(obj)
        obj.h2 = h2
        obj.h3 = h3
        obj.begin = obj.end = i
        obj.acronym = set()
        obj.requester = None
        obj.copyeditors = []
        obj.status = None
        continue

    if not h3:
        continue

    obj = sections[h2][h3]
    obj.end = i

    obj.acronym.update(re.findall(r'\b(?:ACN|ACR|BCN|BCR|CCN|DYK|FAC|FAR|FARC'
                                  r'|FLC|FLRC|GAN|GAR|PR|TFA|OTD)\b', line))

    user = re.findall(r'\[\[\s*user(?:[ _]talk)?:[^\[\]]+\]\]', line, re.I)
    if user:
        user = pywikibot.link_regex.match(user[-1])
        user = pywikibot.Link(user.group('title'), source=SITE)
        user = pywikibot.Page(user)
        if user.isTalkPage():
            user = user.toggleTalkPage()
    else:
        user = None

    timestamp = re.findall(
        r'\d{2}:\d{2}, \d{1,2} [A-Z][a-z]+ \d{4,} \([A-Z]+\)', line)
    if timestamp:
        timestamp = pywikibot.Timestamp.strptime(
            timestamp[-1], '%H:%M, %d %B %Y (%Z)')
    else:
        timestamp = None

    if timestamp:
        # Poor man's goto
        for _ in range(1):
            reobj = re.search(r'\{\{(declined|withdrawn)\}\}', line, re.I)
            if reobj:
                obj.status = reobj.group(1).capitalize(), timestamp
                break

            reobj = re.search(r'\{\{((?:partly[ _])?done)\}\}', line, re.I)
            if reobj and user:
                if user not in obj.copyeditors:
                    obj.copyeditors.append(user)

                if reobj.group(1).lower() == 'done':
                    obj.status = True, timestamp
                break

            if obj.status is None and obj.requester is None and user:
                obj.requester = user, timestamp

archived = []
archives = {}

for obj in sections_flat:
    if not obj.status:
        continue

    if not obj.requester:
        continue

    requester, request_time = obj.requester
    purpose, complete_time = obj.status

    if (pywikibot.Timestamp.utcnow() - complete_time).days < 1:
        continue

    def format_username(user):
        user = pywikibot.User(user)
        if not user.isAnonymous():
            return user.title(
                with_ns=False, as_link=True, allow_interwiki=False)
        else:
            displayname = user.username

            if len(displayname) > 25:
                reobj = re.match(r'^((?:[0-9a-fA-F]{1,4}:){4})'
                                 r'((?:[0-9a-fA-F]{1,4}:){3}'
                                 r'[0-9a-fA-F]{1,4})$', displayname)
                if reobj:
                    displayname = ' '.join(reobj.groups())
            return '[[Special:Contributions/%s|%s]]' % (
                user.username, displayname)

    if purpose is True:
        purpose = ' / '.join(sorted(obj.acronym))
        editors = ', '.join(map(format_username, obj.copyeditors))
    else:
        editors = 'n/a'

    tr = '''|-
| %s
| %s
| %s
| %s
| %s
| %s
''' % (
        request_time.strftime('%Y-%m-%d'),
        obj.h3.pagename,
        format_username(requester),
        complete_time.strftime('%Y-%m-%d'),
        editors,
        purpose
    )

    if request_time.year in archives:
        archive = archives[request_time.year][0]
    else:
        archive = pywikibot.Page(
            SITE, 'Wikipedia:WikiProject_Guild_of_Copy_Editors/'
            'Requests/Archives/%d' % request_time.year)
        archive.origtext = archive.text
        archives[request_time.year] = [archive, []]

    partition = (request_time.month - 1) // 3 * 3 + 1
    thispart_title = '==%s–%s==\n' % (
        calendar.month_name[partition], calendar.month_name[partition + 2])
    nextpart_title = None if partition + 3 > 12 else '==%s–%s==\n' % (
        calendar.month_name[partition + 3], calendar.month_name[partition + 5])
    try:
        startpos = archive.text.index(thispart_title)
        endpos = archive.text.index(nextpart_title) if nextpart_title else None
    except ValueError:
        continue

    def findnext(substr):
        if endpos:
            return archive.text.index(substr, startpos, endpos)
        else:
            return archive.text.index(substr, startpos)

    class Fail(Exception):
        pass

    try:
        for i in itertools.count():
            try:
                startpos = min(findnext('\n|-\n'), findnext('\n|}\n')) + 1
            except ValueError:
                raise Fail

            if archive.text[startpos:startpos + 2] == '|}':
                break
            if archive.text[startpos:startpos + 5] == '|-\n|}':
                break
            elif archive.text[startpos:startpos + 5] != '|-\n| ':
                print(repr(archive.text[startpos:startpos + 6]))
                raise Fail

            reobj = re.match(
                r'^\d{4,}-\d{2}-\d{2}', archive.text[startpos + 5:])
            if not reobj:
                raise Fail

            row_time = pywikibot.Timestamp.strptime(reobj.group(0), '%Y-%m-%d')
            if request_time >= row_time:
                continue

            break
    except Fail:
        __import__('traceback').print_exc()
        continue

    archive.text = archive.text[:startpos] + tr + archive.text[startpos:]

    begin = obj.begin.get()
    for i in range(obj.end.get(), obj.begin.get() - 1, -1):
        rm_lineno(i)
    rm_empty_lineno(begin)

    del sections[obj.h2][obj.h3]

    if not sections[obj.h2]:
        lineno = obj.h2.lineno.get()
        rm_lineno(lineno)
        rm_empty_lineno(lineno)

    archived.append((obj.h3.pagename, purpose))
    archives[request_time.year][1].append((obj.h3.pagename, purpose))


def _mk_summary(archived_lst):
    summ = [
        '%s (%s)' % (pagename, purpose) if purpose else pagename
        for pagename, purpose in archived_lst]
    if len(summ) > 1:
        summ[-1] = 'and ' + summ[-1]
    # Special case: [[pagename]] (purpose) and [[pagename]] (purpose)
    if len(summ) == 2 and archived_lst[0][1]:
        return ' '.join(summ)
    return ', '.join(summ)


def mk_summary(archived_lst):
    if len(archived_lst) == 1:
        return 'Bot: Archived request for ' + _mk_summary(archived_lst)
    return 'Bot: Archived requests for ' + _mk_summary(archived_lst)

for archive, archived_inner in archives.values():
    if archived_inner:
        archive.save(mk_summary(archived_inner))

if archived:
    reqpage.save(mk_summary(archived))