zulip/zerver/data_import/mattermost.py

"""
spec:
https://docs.mattermost.com/administration/bulk-export.html
"""
import os
import logging
import subprocess
import ujson
import re
import shutil

from typing import Any, Callable, Dict, List, Set

from django.conf import settings
from django.utils.timezone import now as timezone_now
from django.forms.models import model_to_dict

from zerver.models import Recipient, RealmEmoji, Reaction
from zerver.lib.utils import (
    process_list_in_batches,
)
from zerver.lib.emoji import NAME_TO_CODEPOINT_PATH
from zerver.data_import.import_util import ZerverFieldsT, build_zerver_realm, \
    build_stream, build_realm, build_message, create_converted_data_files, \
    make_subscriber_map, build_recipients, build_user_profile, \
    build_stream_subscriptions, build_personal_subscriptions, SubscriberHandler, \
    build_realm_emoji, make_user_messages

from zerver.data_import.mattermost_user import UserHandler
from zerver.data_import.sequencer import NEXT_ID, IdMapper

def make_realm(realm_id: int, team: Dict[str, Any]) -> ZerverFieldsT:
    # set correct realm details
    NOW = float(timezone_now().timestamp())
    domain_name = settings.EXTERNAL_HOST
    realm_subdomain = team["name"]

    zerver_realm = build_zerver_realm(realm_id, realm_subdomain, NOW, 'Mattermost')
    realm = build_realm(zerver_realm, realm_id, domain_name)

    # We may override these later.
    realm['zerver_defaultstream'] = []

    return realm

def process_user(user_dict: Dict[str, Any], realm_id: int, team_name: str,
                 user_id_mapper: IdMapper) -> ZerverFieldsT:
    def is_team_admin(user_dict: Dict[str, Any]) -> bool:
        if user_dict["teams"] is None:
            return False
        for team in user_dict["teams"]:
            if team["name"] == team_name and "team_admin" in team["roles"]:
                return True
        return False

    def get_full_name(user_dict: Dict[str, Any]) -> str:
        full_name = "{} {}".format(user_dict["first_name"], user_dict["last_name"])
        if full_name.strip():
            return full_name
        return user_dict['username']

    avatar_source = 'G'
    full_name = get_full_name(user_dict)
    id = user_id_mapper.get(user_dict['username'])
    delivery_email = user_dict['email']
    email = user_dict['email']
    is_realm_admin = is_team_admin(user_dict)
    is_guest = False
    short_name = user_dict['username']
    date_joined = int(timezone_now().timestamp())
    timezone = 'UTC'

    if user_dict["is_mirror_dummy"]:
        is_active = False
        is_mirror_dummy = True
    else:
        is_active = True
        is_mirror_dummy = False

    return build_user_profile(
        avatar_source=avatar_source,
        date_joined=date_joined,
        delivery_email=delivery_email,
        email=email,
        full_name=full_name,
        id=id,
        is_active=is_active,
        is_realm_admin=is_realm_admin,
        is_guest=is_guest,
        is_mirror_dummy=is_mirror_dummy,
        realm_id=realm_id,
        short_name=short_name,
        timezone=timezone,
    )

def convert_user_data(user_handler: UserHandler,
                      user_id_mapper: IdMapper,
                      user_data_map: Dict[str, Dict[str, Any]],
                      realm_id: int,
                      team_name: str) -> None:

    user_data_list = []
    for username in user_data_map:
        user = user_data_map[username]
        if check_user_in_team(user, team_name) or user["is_mirror_dummy"]:
            user_data_list.append(user)

    for raw_item in user_data_list:
        user = process_user(raw_item, realm_id, team_name, user_id_mapper)
        user_handler.add_user(user)

def convert_channel_data(channel_data: List[ZerverFieldsT],
                         user_data_map: Dict[str, Dict[str, Any]],
                         subscriber_handler: SubscriberHandler,
                         stream_id_mapper: IdMapper,
                         user_id_mapper: IdMapper,
                         realm_id: int,
                         team_name: str) -> List[ZerverFieldsT]:
    channel_data_list = [
        d
        for d in channel_data
        if d['team'] == team_name
    ]

    channel_members_map = {}  # type: Dict[str, List[str]]
    channel_admins_map = {}  # type: Dict[str, List[str]]

    def initialize_stream_membership_dicts() -> None:
        for channel in channel_data:
            channel_name = channel["name"]
            channel_members_map[channel_name] = []
            channel_admins_map[channel_name] = []

        for username in user_data_map:
            user_dict = user_data_map[username]
            teams = user_dict["teams"]
            if user_dict["teams"] is None:
                continue

            for team in teams:
                if team["name"] != team_name:
                    continue
                for channel in team["channels"]:
                    channel_roles = channel["roles"]
                    channel_name = channel["name"]
                    if "channel_admin" in channel_roles:
                        channel_admins_map[channel_name].append(username)
                    elif "channel_user" in channel_roles:
                        channel_members_map[channel_name].append(username)

    def get_invite_only_value_from_channel_type(channel_type: str) -> bool:
        # Channel can have two types in Mattermost
        # "O" for a public channel.
        # "P" for a private channel.
        if channel_type == 'O':
            return False
        elif channel_type == 'P':
            return True
        else:  # nocoverage
            raise Exception('unexpected value')

    streams = []
    initialize_stream_membership_dicts()

    for channel_dict in channel_data_list:
        now = int(timezone_now().timestamp())
        stream_id = stream_id_mapper.get(channel_dict['name'])
        stream_name = channel_dict["name"]
        invite_only = get_invite_only_value_from_channel_type(channel_dict['type'])

        stream = build_stream(
            date_created=now,
            realm_id=realm_id,
            name=channel_dict['display_name'],
            # Purpose describes how the channel should be used. It is similar to
            # stream description and is shown in channel list to help others decide
            # whether to join.
            # Header text always appears right next to channel name in channel header.
            # Can be used for advertising the purpose of stream, making announcements as
            # well as including frequently used links. So probably not a bad idea to use
            # this as description if the channel purpose is empty.
            description=channel_dict["purpose"] or channel_dict['header'],
            stream_id=stream_id,
            # Mattermost export don't include data of archived(~ deactivated) channels.
            deactivated=False,
            invite_only=invite_only,
        )

        channel_users = set()
        for username in channel_admins_map[stream_name]:
            channel_users.add(user_id_mapper.get(username))

        for username in channel_members_map[stream_name]:
            channel_users.add(user_id_mapper.get(username))

        if channel_users:
            subscriber_handler.set_info(
                stream_id=stream_id,
                users=channel_users,
            )
        streams.append(stream)

    return streams

def get_name_to_codepoint_dict() -> Dict[str, str]:
    with open(NAME_TO_CODEPOINT_PATH) as fp:
        return ujson.load(fp)

def build_reactions(realm_id: int, total_reactions: List[ZerverFieldsT], reactions: List[ZerverFieldsT],
                    message_id: int, name_to_codepoint: ZerverFieldsT,
                    user_id_mapper: IdMapper, zerver_realmemoji: List[ZerverFieldsT]) -> None:
    realmemoji = {}
    for realm_emoji in zerver_realmemoji:
        realmemoji[realm_emoji['name']] = realm_emoji['id']

    # For the unicode emoji codes, we use equivalent of
    # function 'emoji_name_to_emoji_code' in 'zerver/lib/emoji' here
    for mattermost_reaction in reactions:
        emoji_name = mattermost_reaction['emoji_name']
        username = mattermost_reaction["user"]
        # Check in unicode emoji
        if emoji_name in name_to_codepoint:
            emoji_code = name_to_codepoint[emoji_name]
            reaction_type = Reaction.UNICODE_EMOJI
        # Check in realm emoji
        elif emoji_name in realmemoji:
            emoji_code = realmemoji[emoji_name]
            reaction_type = Reaction.REALM_EMOJI
        else:  # nocoverage
            continue

        if not user_id_mapper.has(username):
            continue

        reaction_id = NEXT_ID('reaction')
        reaction = Reaction(
            id=reaction_id,
            emoji_code=emoji_code,
            emoji_name=emoji_name,
            reaction_type=reaction_type)

        reaction_dict = model_to_dict(reaction, exclude=['message', 'user_profile'])
        reaction_dict['message'] = message_id
        reaction_dict['user_profile'] = user_id_mapper.get(username)
        total_reactions.append(reaction_dict)

def get_mentioned_user_ids(raw_message: Dict[str, Any], user_id_mapper: IdMapper) -> Set[int]:
    user_ids = set()
    content = raw_message["content"]

    # usernames can be of the form user.name, user_name, username., username_, user.name_ etc
    matches = re.findall("(?<=^|(?<=[^a-zA-Z0-9-_.]))@(([A-Za-z0-9]+[_.]?)+)", content)

    for match in matches:
        possible_username = match[0]
        if user_id_mapper.has(possible_username):
            user_ids.add(user_id_mapper.get(possible_username))
    return user_ids

def process_raw_message_batch(realm_id: int,
                              raw_messages: List[Dict[str, Any]],
                              subscriber_map: Dict[int, Set[int]],
                              user_id_mapper: IdMapper,
                              user_handler: UserHandler,
                              get_recipient_id_from_receiver_name: Callable[[str, int], int],
                              is_pm_data: bool,
                              output_dir: str,
                              zerver_realmemoji: List[Dict[str, Any]],
                              total_reactions: List[Dict[str, Any]],
                              ) -> None:

    def fix_mentions(content: str, mention_user_ids: Set[int]) -> str:
        for user_id in mention_user_ids:
            user = user_handler.get_user(user_id=user_id)
            mattermost_mention = '@{short_name}'.format(**user)
            zulip_mention = '@**{full_name}**'.format(**user)
            content = content.replace(mattermost_mention, zulip_mention)

        content = content.replace('@channel', '@**all**')
        content = content.replace('@all', '@**all**')
        # We don't have an equivalent for Mattermost's @here mention which mentions all users
        # online in the channel.
        content = content.replace('@here', '@**all**')
        return content

    mention_map = dict()  # type: Dict[int, Set[int]]
    zerver_message = []

    import html2text
    h = html2text.HTML2Text()

    name_to_codepoint = get_name_to_codepoint_dict()

    for raw_message in raw_messages:
        message_id = NEXT_ID('message')
        mention_user_ids = get_mentioned_user_ids(raw_message, user_id_mapper)
        mention_map[message_id] = mention_user_ids

        content = fix_mentions(
            content=raw_message['content'],
            mention_user_ids=mention_user_ids,
        )
        content = h.handle(content)

        if len(content) > 10000:  # nocoverage
            logging.info('skipping too-long message of length %s' % (len(content),))
            continue

        pub_date = raw_message['pub_date']
        sender_user_id = raw_message['sender_id']
        try:
            recipient_id = get_recipient_id_from_receiver_name(raw_message["receiver_id"], Recipient.STREAM)
        except KeyError:
            logging.debug("Could not find recipient_id for a message, skipping.")
            continue
        rendered_content = None

        topic_name = 'imported from mattermost'

        message = build_message(
            content=content,
            message_id=message_id,
            pub_date=pub_date,
            recipient_id=recipient_id,
            rendered_content=rendered_content,
            topic_name=topic_name,
            user_id=sender_user_id,
            has_attachment=False,
        )
        zerver_message.append(message)
        build_reactions(realm_id, total_reactions, raw_message["reactions"], message_id,
                        name_to_codepoint, user_id_mapper, zerver_realmemoji)

    zerver_usermessage = make_user_messages(
        zerver_message=zerver_message,
        subscriber_map=subscriber_map,
        is_pm_data=is_pm_data,
        mention_map=mention_map,
    )

    message_json = dict(
        zerver_message=zerver_message,
        zerver_usermessage=zerver_usermessage,
    )

    dump_file_id = NEXT_ID('dump_file_id' + str(realm_id))
    message_file = "/messages-%06d.json" % (dump_file_id,)
    create_converted_data_files(message_json, output_dir, message_file)

def process_posts(num_teams: int,
                  team_name: str,
                  realm_id: int,
                  post_data: List[Dict[str, Any]],
                  get_recipient_id_from_receiver_name: Callable[[str, int], int],
                  subscriber_map: Dict[int, Set[int]],
                  output_dir: str,
                  is_pm_data: bool,
                  masking_content: bool,
                  user_id_mapper: IdMapper,
                  user_handler: UserHandler,
                  username_to_user: Dict[str, Dict[str, Any]],
                  zerver_realmemoji: List[Dict[str, Any]],
                  total_reactions: List[Dict[str, Any]]) -> None:

    post_data_list = []
    for post in post_data:
        if "team" not in post:
            # Mattermost doesn't specify a team for private messages
            # in its export format.  This line of code requires that
            # we only be importing data from a single team (checked
            # elsewhere) -- we just assume it's the target team.
            post_team = team_name
        else:
            post_team = post["team"]
        if post_team == team_name:
            post_data_list.append(post)

    def message_to_dict(post_dict: Dict[str, Any]) -> Dict[str, Any]:
        sender_id = user_id_mapper.get(post_dict["user"])
        content = post_dict['message']

        if masking_content:
            content = re.sub('[a-z]', 'x', content)
            content = re.sub('[A-Z]', 'X', content)

        if "reactions" in post_dict:
            reactions = post_dict["reactions"] or []
        else:
            reactions = []

        return dict(
            sender_id=sender_id,
            receiver_id=post_dict["channel"],
            content=content,
            pub_date=int(post_dict['create_at'] / 1000),
            reactions=reactions
        )

    raw_messages = []
    for post_dict in post_data_list:
        raw_messages.append(message_to_dict(post_dict))
        message_replies = post_dict["replies"]
        # Replies to a message in Mattermost are stored in the main message object.
        # For now, we just append the replies immediately after the original message.
        if message_replies is not None:
            for reply in message_replies:
                reply["channel"] = post_dict["channel"]
                raw_messages.append(message_to_dict(reply))

    def process_batch(lst: List[Dict[str, Any]]) -> None:
        process_raw_message_batch(
            realm_id=realm_id,
            raw_messages=lst,
            subscriber_map=subscriber_map,
            user_id_mapper=user_id_mapper,
            user_handler=user_handler,
            get_recipient_id_from_receiver_name=get_recipient_id_from_receiver_name,
            is_pm_data=is_pm_data,
            output_dir=output_dir,
            zerver_realmemoji=zerver_realmemoji,
            total_reactions=total_reactions,
        )

    chunk_size = 1000

    process_list_in_batches(
        lst=raw_messages,
        chunk_size=chunk_size,
        process_batch=process_batch,
    )

def write_message_data(num_teams: int,
                       team_name: str,
                       realm_id: int,
                       post_data: List[Dict[str, Any]],
                       zerver_recipient: List[ZerverFieldsT],
                       subscriber_map: Dict[int, Set[int]],
                       output_dir: str,
                       masking_content: bool,
                       stream_id_mapper: IdMapper,
                       user_id_mapper: IdMapper,
                       user_handler: UserHandler,
                       username_to_user: Dict[str, Dict[str, Any]],
                       zerver_realmemoji: List[Dict[str, Any]],
                       total_reactions: List[Dict[str, Any]]) -> None:
    stream_id_to_recipient_id = {}

    for d in zerver_recipient:
        if d['type'] == Recipient.STREAM:
            stream_id_to_recipient_id[d['type_id']] = d['id']

    def get_recipient_id_from_receiver_name(receiver_name: str, recipient_type: int) -> int:
        if recipient_type == Recipient.STREAM:
            receiver_id = stream_id_mapper.get(receiver_name)
            recipient_id = stream_id_to_recipient_id[receiver_id]
        return recipient_id

    process_posts(
        num_teams=num_teams,
        team_name=team_name,
        realm_id=realm_id,
        post_data=post_data,
        get_recipient_id_from_receiver_name=get_recipient_id_from_receiver_name,
        subscriber_map=subscriber_map,
        output_dir=output_dir,
        is_pm_data=False,
        masking_content=masking_content,
        user_id_mapper=user_id_mapper,
        user_handler=user_handler,
        username_to_user=username_to_user,
        zerver_realmemoji=zerver_realmemoji,
        total_reactions=total_reactions,
    )

def write_emoticon_data(realm_id: int,
                        custom_emoji_data: List[Dict[str, Any]],
                        data_dir: str,
                        output_dir: str) -> List[ZerverFieldsT]:
    '''
    This function does most of the work for processing emoticons, the bulk
    of which is copying files.  We also write a json file with metadata.
    Finally, we return a list of RealmEmoji dicts to our caller.

    In our data_dir we have a pretty simple setup:

        The exported JSON file will have emoji rows if it contains any custom emoji
            {
                "type": "emoji",
                "emoji": {"name": "peerdium", "image": "exported_emoji/h15ni7kf1bnj7jeua4qhmctsdo/image"}
            }
            {
                "type": "emoji",
                "emoji": {"name": "tick", "image": "exported_emoji/7u7x8ytgp78q8jir81o9ejwwnr/image"}
            }

        exported_emoji/ - contains a bunch of image files:
            exported_emoji/7u7x8ytgp78q8jir81o9ejwwnr/image
            exported_emoji/h15ni7kf1bnj7jeua4qhmctsdo/image

    We move all the relevant files to Zulip's more nested
    directory structure.
    '''

    logging.info('Starting to process emoticons')

    flat_data = [
        dict(
            path=d['image'],
            name=d['name'],
        )
        for d in custom_emoji_data
    ]

    emoji_folder = os.path.join(output_dir, 'emoji')
    os.makedirs(emoji_folder, exist_ok=True)

    def process(data: ZerverFieldsT) -> ZerverFieldsT:
        source_sub_path = data['path']
        source_path = os.path.join(data_dir, source_sub_path)

        target_fn = data["name"]
        target_sub_path = RealmEmoji.PATH_ID_TEMPLATE.format(
            realm_id=realm_id,
            emoji_file_name=target_fn,
        )
        target_path = os.path.join(emoji_folder, target_sub_path)

        os.makedirs(os.path.dirname(target_path), exist_ok=True)

        source_path = os.path.abspath(source_path)
        target_path = os.path.abspath(target_path)

        shutil.copyfile(source_path, target_path)

        return dict(
            path=target_path,
            s3_path=target_path,
            file_name=target_fn,
            realm_id=realm_id,
            name=data['name'],
        )

    emoji_records = list(map(process, flat_data))
    create_converted_data_files(emoji_records, output_dir, '/emoji/records.json')

    realmemoji = [
        build_realm_emoji(
            realm_id=realm_id,
            name=rec['name'],
            id=NEXT_ID('realmemoji'),
            file_name=rec['file_name'],
        )
        for rec in emoji_records
    ]
    logging.info('Done processing emoticons')

    return realmemoji

def create_username_to_user_mapping(user_data_list: List[Dict[str, Any]]) -> Dict[str, Dict[str, Any]]:
    username_to_user = {}
    for user in user_data_list:
        username_to_user[user["username"]] = user
    return username_to_user

def check_user_in_team(user: Dict[str, Any], team_name: str) -> bool:
    if user["teams"] is None:
        # This is null for users not on any team
        return False
    for team in user["teams"]:
        if team["name"] == team_name:
            return True
    return False

def label_mirror_dummy_users(num_teams: int, team_name: str,
                             mattermost_data: Dict[str, List[Dict[str, Any]]],
                             username_to_user: Dict[str, Dict[str, Any]]) -> None:
    # This function might looks like a great place to label admin users. But
    # that won't be fully correct since we are iterating only though posts and
    # it covers only users that has sent atleast one message.
    for post in mattermost_data["post"]:
        if "team" not in post:
            if num_teams > 1:
                raise AssertionError("Cannot import private messages in export with multiple teams.")
            elif num_teams == 0:
                raise AssertionError("Export with 0 teams")
            else:
                post_team = team_name
        else:
            post_team = post["team"]

        if post_team == team_name:
            user = username_to_user[post["user"]]
            if not check_user_in_team(user, team_name):
                user["is_mirror_dummy"] = True

def reset_mirror_dummy_users(username_to_user: Dict[str, Dict[str, Any]]) -> None:
    for username in username_to_user:
        user = username_to_user[username]
        user["is_mirror_dummy"] = False

def mattermost_data_file_to_dict(mattermost_data_file: str) -> Dict[str, List[Dict[str, Any]]]:
    mattermost_data = {}  # type: Dict[str, List[Dict[str, Any]]]
    mattermost_data["version"] = []
    mattermost_data["team"] = []
    mattermost_data["channel"] = []
    mattermost_data["user"] = []
    mattermost_data["post"] = []
    mattermost_data["emoji"] = []

    with open(mattermost_data_file, "r") as fp:
        for line in fp:
            row = ujson.loads(line.rstrip("\n"))
            data_type = row["type"]
            mattermost_data[data_type].append(row[data_type])
    return mattermost_data

def do_convert_data(mattermost_data_dir: str, output_dir: str, masking_content: bool) -> None:
    username_to_user = {}  # type: Dict[str, Dict[str, Any]]

    os.makedirs(output_dir, exist_ok=True)
    if os.listdir(output_dir):  # nocoverage
        raise Exception("Output directory should be empty!")

    mattermost_data_file = os.path.join(mattermost_data_dir, "export.json")
    mattermost_data = mattermost_data_file_to_dict(mattermost_data_file)

    username_to_user = create_username_to_user_mapping(mattermost_data["user"])

    for team in mattermost_data["team"]:
        realm_id = NEXT_ID("realm_id")
        team_name = team["name"]

        user_handler = UserHandler()
        subscriber_handler = SubscriberHandler()
        user_id_mapper = IdMapper()
        stream_id_mapper = IdMapper()

        print("Generating data for", team_name)
        realm = make_realm(realm_id, team)
        realm_output_dir = os.path.join(output_dir, team_name)

        reset_mirror_dummy_users(username_to_user)
        label_mirror_dummy_users(len(mattermost_data["team"]), team_name, mattermost_data, username_to_user)

        convert_user_data(
            user_handler=user_handler,
            user_id_mapper=user_id_mapper,
            user_data_map=username_to_user,
            realm_id=realm_id,
            team_name=team_name,
        )

        zerver_stream = convert_channel_data(
            channel_data=mattermost_data["channel"],
            user_data_map=username_to_user,
            subscriber_handler=subscriber_handler,
            stream_id_mapper=stream_id_mapper,
            user_id_mapper=user_id_mapper,
            realm_id=realm_id,
            team_name=team_name,
        )

        realm['zerver_stream'] = zerver_stream

        all_users = user_handler.get_all_users()

        zerver_recipient = build_recipients(
            zerver_userprofile=all_users,
            zerver_stream=zerver_stream,
        )
        realm['zerver_recipient'] = zerver_recipient

        stream_subscriptions = build_stream_subscriptions(
            get_users=subscriber_handler.get_users,
            zerver_recipient=zerver_recipient,
            zerver_stream=zerver_stream,
        )

        personal_subscriptions = build_personal_subscriptions(
            zerver_recipient=zerver_recipient,
        )

        # Mattermost currently supports only exporting messages from channels.
        # Personal messages and huddles are not exported.
        zerver_subscription = personal_subscriptions + stream_subscriptions
        realm['zerver_subscription'] = zerver_subscription

        zerver_realmemoji = write_emoticon_data(
            realm_id=realm_id,
            custom_emoji_data=mattermost_data["emoji"],
            data_dir=mattermost_data_dir,
            output_dir=realm_output_dir,
        )
        realm['zerver_realmemoji'] = zerver_realmemoji

        subscriber_map = make_subscriber_map(
            zerver_subscription=zerver_subscription,
        )

        total_reactions = []  # type: List[Dict[str, Any]]
        write_message_data(
            num_teams=len(mattermost_data["team"]),
            team_name=team_name,
            realm_id=realm_id,
            post_data=mattermost_data["post"],
            zerver_recipient=zerver_recipient,
            subscriber_map=subscriber_map,
            output_dir=realm_output_dir,
            masking_content=masking_content,
            stream_id_mapper=stream_id_mapper,
            user_id_mapper=user_id_mapper,
            user_handler=user_handler,
            username_to_user=username_to_user,
            zerver_realmemoji=zerver_realmemoji,
            total_reactions=total_reactions,
        )
        realm['zerver_reaction'] = total_reactions
        realm['zerver_userprofile'] = user_handler.get_all_users()
        realm['sort_by_date'] = True

        create_converted_data_files(realm, realm_output_dir, '/realm.json')
        # Mattermost currently doesn't support exporting avatars
        create_converted_data_files([], realm_output_dir, '/avatars/records.json')
        # Mattermost currently doesn't support exporting uploads
        create_converted_data_files([], realm_output_dir, '/uploads/records.json')

        # Mattermost currently doesn't support exporting attachments
        attachment = {"zerver_attachment": []}  # type: Dict[str, List[Any]]
        create_converted_data_files(attachment, realm_output_dir, '/attachment.json')

        logging.info('Start making tarball')
        subprocess.check_call(["tar", "-czf", realm_output_dir + '.tar.gz', realm_output_dir, '-P'])
        logging.info('Done making tarball')