#!/usr/bin/env python3
# Someone's Mattermost scripts.
#  Copyright (c) 2016-2020 by Someone <someone@somenet.org> (aka. Jan Vales <jan@jvales.net>)
#  published under MIT-License
#

import json
import os
import pprint
import random
import re
import signal
import sys
import time
import traceback
import mwclient
import mattermost

import config


########
# Util #
########
# global regexes
REXtuwien = re.compile(r'^TU Wien\:(.*) \w\w,?\w?\w? \(.*\).*$', re.IGNORECASE)
REXdaten = re.compile(r'==\s*Daten\s*==', re.MULTILINE|re.DOTALL|re.IGNORECASE)
REXwdata_mmchan = re.compile(r'\*??\s*\{\{mattermost-channel\|([a-zA-Z0-9_ -]+)\}\}\s*\n{0,2}', re.IGNORECASE)

def get_smw_properties(subject, sobj_dict={'sobj': {}}):
    ret = dict()
    for prop in subject:
        if prop['property'] not in ret:
            ret[prop['property']] = []
        for dataitem in prop['dataitem']:
            if dataitem['type'] == 9 and dataitem['item'] in sobj_dict['sobj']:
                ret[prop['property']] += [sobj_dict['sobj'][dataitem['item']]]
            else:
                ret[prop['property']] += [dataitem['item']]
    return ret

def query_smw_properties(site, pagename):
    res = site.api('browsebysubject', subject=pagename)

    # gather all sub objects first
    ret = {'sobj': {}}
    if 'sobj' in res['query']:
        for sobj in res['query']['sobj']:
            ret['sobj'][sobj['subject']] = get_smw_properties(sobj['data'])

    ret = get_smw_properties(res['query']['data'], ret)

    if '_ASK' in ret:
        del ret['_ASK']
    return ret


def mw_pagename_to_mm_chan_mapping(pagename):
    try:
        mmchdisplayname = REXtuwien.search(pagename).group(1)
        vowiurlname = mmchdisplayname.replace(' ', '_')
        mmchname = mmchdisplayname.lower().replace('ä', 'ae').replace('ö', 'oe').replace('ü', 'ue')
        mmchdisplayname = mmchdisplayname[:63]

        mmchname = re.sub(r'[^a-zA-Z0-9_]', r'-', mmchname)
        mmchname = re.sub(r'-+', r'-', mmchname)
        mmchname = mmchname[:63]
        mmchname = mmchname.strip('-')

        # now even group-channels (len == 40) do weird shit. Thanks MM :/
        # Heisen-Issue: for some reason MM uses sometimes the channel name as id if len == 26 gives and when switching between teams.
        if len(mmchname) == 54 or len(mmchname) == 40 or len(mmchname) == 26:
            mmchname = mmchname+'0'

        return (mmchname, (mmchname, mmchdisplayname, vowiurlname, pagename))
    except:
        print('Failed to process pagemapping for: '+pagename)
        return ("", ("", "", "", ""))

#########
# /Util #
#########



def create_update_spam_mm_channel(mws, mma, basepage_name, changes_to_notify_about=dict()):
    print('\ncreate_update_spam_mm_channel(): https://vowi.fsinf.at/wiki/'+basepage_name.replace(' ', '_').replace(')', '%29'))

    # really purge cache :/
    sp = mws.Pages[basepage_name]
    sp.purge()

    if not sp.exists or sp.redirect:
        print('Page was deleted or moved. ignoring.')
        return

    # get smw data of basepage
    semantic_data = query_smw_properties(mws, sp.name)
    sp_text_daten = sp_text_daten_orig = sp.text(section=1)
    if REXdaten.match(sp_text_daten) is None:
        raise Exception('"== Daten ==" not section=1')

    # abort if veraltet
    if 'Ist_veraltet' in semantic_data and semantic_data['Ist_veraltet'][0] == 't':
        print('Page is veraltet. ignoring.')
        return

    # generate metadata
    basepage_name_mapping = mw_pagename_to_mm_chan_mapping(basepage_name)
    metadata = dict()
    metadata['display_name'] = basepage_name_mapping[1][1]
    metadata['purpose'] = 'LVA(s) im VoWi: https://vowi.fsinf.at/LVA/'+basepage_name_mapping[1][2]+'?mm'
    metadata['header'] = 'Links: [LVA(s) im VoWi](https://vowi.fsinf.at/LVA/'+basepage_name_mapping[1][2]+'?mm)'

    if 'Hat_Kurs-ID' in semantic_data:
        metadata['header'] += ' - [LVA in TISS](https://tiss.tuwien.ac.at/course/courseDetails.xhtml?courseNr='+semantic_data['Hat_Kurs-ID'][0]+')'
    if 'Hat_Homepage' in semantic_data:
        metadata['header'] += ' - [LVA-HP]('+semantic_data['Hat_Homepage'][0]+')'

    # create channel, if not exist
    mmchan = mma.get_channel_by_name(config.mm_autochannels_team, basepage_name_mapping[0])
    if 'status_code' in mmchan and mmchan['status_code'] == 404:
        print(mma.create_channel(config.mm_autochannels_team, basepage_name_mapping[0], metadata['display_name'], metadata['purpose'], metadata['header']))

        # really get channel
        mmchan = mma.get_channel_by_name(config.mm_autochannels_team, basepage_name_mapping[0])


    # update wiki:data-section. always.
    sp_text_daten = re.sub(REXwdata_mmchan, '', sp_text_daten)
    sp_text_daten = sp_text_daten.strip()
    sp_text_daten += '\n{{mattermost-channel|'+basepage_name_mapping[0]+'}}'
    if sp_text_daten != sp_text_daten_orig:
        print('Updating MW page: '+str(sp.save(sp_text_daten, section=1, summary='mw_vowi')))


    # update channel
    if 'type' in mmchan and mmchan['type'] == 'O' and ((sp_text_daten != sp_text_daten_orig) or (len(changes_to_notify_about) > 0)):
        mma.patch_channel(mmchan['id'], metadata)

    # spam channel if there is anything to spam about
    if len(changes_to_notify_about) > 0:
        msg = '``BOT-AUTODELETE-SLOW``\nFolgende Seiten wurden in den letzten 24 Stunden im VoWi editiert:\n'
        for v in changes_to_notify_about.values():
            msg += '\nhttps://vowi.fsinf.at/wiki/'+v['title'].replace(' ', '_')
        msg += '\n\n``footer``\n'+str(random.choice(config.mm_footerlist))
        print('Posting to channel:' +str(mma.create_post(mmchan['id'], msg)))



def process_recent_vowi_changes(mws, mma):
    lastdata = dict()
    try:
        with open('data/last.json') as fh:
            lastdata = json.load(fh)
    except:
        print('Failed to load last.json')

    changes = dict()
    if 'failed' in lastdata:
        changes = lastdata['failed']

    if 'delme' in lastdata:
        del lastdata['delme']

    if 'timestamp' in lastdata:
        rc = mws.recentchanges(start=lastdata['timestamp'], dir='newer', namespace=3000, show='!bot|!redirect|!minor', toponly=True)
    else:
        rc = mws.recentchanges(dir='newer', namespace=3000, show='!bot|!redirect|!minor', toponly=False)
        lastdata['timestamp'] = 0
    lastdata['last_timestamp'] = lastdata['timestamp']

    if 'rcid' not in lastdata:
        lastdata['rcid'] = 0

    for c in rc:
        if c['rcid'] == lastdata['rcid']:
            continue

        basepage = c['title'].split('/', 1)[0]
        if basepage not in changes:
            changes[basepage] = dict()
        changes[basepage][c['title']] = c
        changes[basepage][c['title']]['timestamp'] = int(time.strftime('%Y%m%d%H%M%S', changes[basepage][c['title']]['timestamp']))

        if changes[basepage][c['title']]['timestamp'] > lastdata['timestamp']:
            lastdata['timestamp'] = changes[basepage][c['title']]['timestamp']
        if changes[basepage][c['title']]['rcid'] > lastdata['rcid']:
            lastdata['rcid'] = changes[basepage][c['title']]['rcid']

    lastdata['failed'] = changes

    # process all pending changes
    for ch in list(changes):
        semantic_data = query_smw_properties(mws, ch)

        # never mind, its outdated anyway ...
        if 'Ist_veraltet' in semantic_data and semantic_data['Ist_veraltet'][0] == 't':
            del lastdata['failed'][ch]
            continue

        try:
            print("trying: "+ch)
            create_update_spam_mm_channel(mws, mma, ch, changes[ch])
            del lastdata['failed'][ch]
        except Exception as ex:
            print('Exception, skipping: '+ch+' - '+str(sys.exc_info()))
            traceback.print_exc()

        with open('data/last.json', 'w') as fh:
            json.dump(lastdata, fh)

    if len(lastdata['failed']) != 0:
        print('Some changes failed to get posted!')
        pprint.pprint(lastdata)



def process_all_LVAs(mws, mma):
    for p in mws.ask('[[TU Wien:+]][[Ist veraltet::0]]'):
        create_update_spam_mm_channel(mws, mma, ''.join(p))



# remove "has mm-channel" info from outdated lva pages.
def process_outdated_LVAs(mws):
    for p in mws.ask('[[Hat Mattermost-Channel::+]][[Ist veraltet::1]]'):
        sp = mws.Pages[(''.join(p.keys()))]
        print('\nProcessing: https://vowi.fsinf.at/wiki/'+sp.name.replace(' ', '_').replace(')', '%29'))

        sp_text = sp.text(section=1)

        sp_text = re.sub(REXwdata_mmchan, '', sp_text)

        print(sp.save(sp_text, section=1, summary='mw_vowi'))
        sp.purge()



def process_outdated_MMChannels(mws, mma, spam_channel=False):
    # get a dict of all lva-pages that have a mm-channel saved and map them to all the possible channel-names.
    # Legitimate (not outdated LVAs)
    mw_page_names_with_mm_channelname_mappings = dict([mw_pagename_to_mm_chan_mapping(c) for c in [''.join(p) for p in mws.ask('[[Hat Mattermost-Channel::+]]')]])

    # All existing mm channels
    mm_all_channels_with_infos = {mmchan['name']:mmchan for mmchan in mma.get_team_channels(config.mm_autochannels_team)}

    # outdated candidates.
    diff_channel_names = set(mm_all_channels_with_infos.keys()) - set(mw_page_names_with_mm_channelname_mappings.keys())
    diff_channel_names.discard('town-square')
    diff_channel_names.discard('off-topic')

    # get replacements
    # dict of all outdated LVA-pages.
    mw_outdated_page_names_mappings = dict([mw_pagename_to_mm_chan_mapping(c) for c in [''.join(p) for p in mws.ask('[[TU Wien:+]][[Kategorie:LVAs]][[Ist veraltet::1]]')]])
# TODO: seems b0rked.
#    mw_lva_ersetzt_durch = dict()
#    for p in mws.ask('[[Ersetzt durch::+]]|?Ersetzt durch'):
#        for k,v in p.items():
#            mw_lva_ersetzt_durch[k] = v['printouts']['Ersetzt durch'][0]
#
#    mw_lva_wirklich_ersetzt_durch = mw_lva_ersetzt_durch.copy()
#    pprint.pprint(mw_lva_wirklich_ersetzt_durch)
#    return
#    for k,v in mw_lva_ersetzt_durch.items():
#         ersatz = k
#         while ersatz in mw_lva_ersetzt_durch:
#             ersatz = mw_lva_ersetzt_durch[ersatz]
#         mw_lva_wirklich_ersetzt_durch[k] = ersatz
#

    # notify admins
    msg = '``BOT-AUTODELETE-SLOW``\n#### ``Likely outdated MM channels found``\nClick ``Next...`` to load all channels, to make these links work\nCheck via vowi if there is really no not-outdated LVA-page.\n'
    msg_inconsistent = ''
    real_diff_channel_names = diff_channel_names.copy()
    for d in diff_channel_names:
        try:
            msg += ' + ~'+d+' :arrow_right: https://vowi.fsinf.at/wiki/Spezial:FlexiblePrefix/'+mw_outdated_page_names_mappings[d][2]+'\n'
        except:
            msg_inconsistent += ' + ~'+d+' :warning: ``No known VoWi URL - SWM-Consistency issue?``\n'
            real_diff_channel_names.discard(d)

    if len(diff_channel_names) == 0:
        msg = '``BOT-AUTODELETE-SLOW``\n#### `` No outdated MM channels found`` :)\n'

    if len(msg_inconsistent) > 0:
        msg += '\n\nThese ('+str(len(diff_channel_names)-len(real_diff_channel_names))+') are likely SMW inconsistencies and were skipped:\n'+msg_inconsistent
        msg += 'If there are many of these, consider running ``SemanticMediaWiki/maintenance/rebuildData.php``'

    # make real_ the real deal
    diff_channel_names = real_diff_channel_names
    del real_diff_channel_names

    print('Posting '+str(len(msg))+' chars to channel:' +str(mma.create_post(config.mm_admin_channel, msg)))


    # notify channels
    if not spam_channel:
        return

    for mmchan in diff_mmchan_names:
        print('notifying: '+mmchan)
        metadata = dict()
        try:
            metadata['purpose'] = 'Outdated? Dieser LVA-Channel wurde als outdated markiert, da alle seine VoWi-LVAs outdated sind. LVA(s) im VoWi: https://vowi.fsinf.at/LVA/'+mw_mmchans_full_all[mmchan][2].replace(' ', '_')+'?mm'
            metadata['header'] = ':warning: Outdated? :warning: Links: [Outdated LVA(s) im VoWi](https://vowi.fsinf.at/LVA/'+mw_mmchans_full_all[mmchan][2].replace(' ', '_')+'?mm)'
        except:
            metadata['purpose'] = 'Outdated? Dieser LVA-Channel wurde als outdated markiert, da alle seine VoWi-LVAs outdated sind. Sollte dies ein Fehler sein, korrigiere bitte die entsprechende LVA im VoWi.'
            metadata['header'] = ':warning: Outdated? :warning: Dieser LVA-Channel wurde als outdated markiert, da alle seine VoWi-LVAs outdated sind.'

        # update channel info.
        print(mma.patch_channel(mm_mmchans_full[mmchan]['id'], metadata))

        # post "something changed" info
        msg = ('``BOT-AUTODELETE-SLOW``\n### Dieser LVA-Channel wurde als outdated markiert, da alle seine VoWi-LVAs outdated sind.\n'
               'Sollte dies ein Fehler sein, korrigiere bitte die entsprechende LVA-Seite im VoWi.\n'
               '``Falls die LVA von einer neuen Person übernommen wurde und sich dabei die Durchführung grundlegend geändert hat, oder dies zu erwarten ist, lege bitte eine neue LVA-Seite an.``\n\n'
               '## :warning: :warning: :warning: :warning: :warning: :warning: :warning:\n\n'
               )

        try:
            msg += ':arrow_right: Die Nachfolge-LVA laut VoWi ist: https://vowi.fsinf.at/LVA/'+mw_lva_ersetzt_durch[mmchan].replace(' ', '_')+' \n'
        except:
            msg += ':warning: Eine Nachfolge-LVA ist nicht bekannt, im VoWi nicht vermerkt, oder konnte nicht ermittelt werden :(\n'
        try:
            msg += ':arrow_right: Der Nachfolge-LVA-MM-Channel dürfte sein: ~'+mw_pagename_to_mm_chan_mapping(mw_lva_ersetzt_durch[mmchan])[0]
        except:
            msg += ':warning: Ein Nachfolge-LVA-MM-Channel konnte nicht ermittelt werden :('

        print('Posting to channel:' +str(mma.create_post(mm_mmchans_full[mmchan]['id'], msg)))



if __name__ == '__main__':
    def signal_handler(signal, frame):
        print('SIG received. exitting!')
        os._exit(0)
    signal.signal(signal.SIGINT, signal_handler)

    mws = mwclient.Site(config.mw_name, path='/', retry_timeout=120)
    mws.login(config.mw_user, config.mw_user_pw)
    mma = mattermost.MMApi(config.mm_api_url)
    mma.login(config.mm_user, config.mm_user_pw)

    # Use recent changes to create, update and spam channels.
    if len(sys.argv) == 1:
        process_recent_vowi_changes(mws, mma)

    # maint stuff
    # "full" run modes - We cant run the whole script at once, due to out db-backup, likely killing our run mid-excecution. :(
    # "new" should actually not really be needed, as it should be covered by incremental changes.
    if len(sys.argv) > 1 and sys.argv[1] == 'all-pages':
        print(sys.argv[1])
        process_all_LVAs(mws, mma)

    if len(sys.argv) > 1 and sys.argv[1] == 'outdated-pages':
        print(sys.argv[1])
        process_outdated_LVAs(mws)

    if len(sys.argv) > 1 and sys.argv[1] == 'outdated-chans':
        print(sys.argv[1])
        process_outdated_MMChannels(mws, mma, False)

    if len(sys.argv) > 1 and sys.argv[1] == 'outdated-chans-spam':
        print(sys.argv[1])
        process_outdated_MMChannels(mws, mma, True)

    # logout
    mma.revoke_user_session()