Support LINE users with no discoverable ID

AKA "strangers". Should only happen to non-friends in rooms (not groups!)

matrix_puppeteer_line/db/__init__.py

@@ -3,6 +3,7 @@ from mautrix.util.async_db import Database
 from .upgrade import upgrade_table
 from .user import User
 from .puppet import Puppet
+from .stranger import Stranger
 from .portal import Portal
 from .message import Message
 from .media import Media
@@ -10,8 +11,8 @@ from .receipt_reaction import ReceiptReaction
 
 
 def init(db: Database) -> None:
-    for table in (User, Puppet, Portal, Message, Media, ReceiptReaction):
+    for table in (User, Puppet, Stranger, Portal, Message, Media, ReceiptReaction):
         table.db = db
 
 
-__all__ = ["upgrade_table", "User", "Puppet", "Portal", "Message", "Media", "ReceiptReaction"]
+__all__ = ["upgrade_table", "User", "Puppet", "Stranger", "Portal", "Message", "Media", "ReceiptReaction"]

matrix_puppeteer_line/db/stranger.py

@@ -0,0 +1,93 @@
+# matrix-puppeteer-line - A very hacky Matrix-LINE bridge based on running LINE's Chrome extension in Puppeteer
+# Copyright (C) 2020-2021 Tulir Asokan, Andrew Ferrazzutti
+#
+# This program is free software: you can redistribute it and/or modify
+# it under the terms of the GNU Affero General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU Affero General Public License for more details.
+#
+# You should have received a copy of the GNU Affero General Public License
+# along with this program.  If not, see <https://www.gnu.org/licenses/>.
+from typing import Optional, ClassVar, TYPE_CHECKING
+
+from attr import dataclass
+from random import randint, seed
+
+from mautrix.util.async_db import Database
+
+fake_db = Database("") if TYPE_CHECKING else None
+
+
+@dataclass
+class Stranger:
+    db: ClassVar[Database] = fake_db
+
+    # Optional properties are ones that should be set by Puppet
+    fake_mid: str
+    name: Optional[str] = None
+    avatar_path: Optional[str] = None
+    available: bool = False
+
+    async def insert(self) -> None:
+        q = ("INSERT INTO stranger (fake_mid, name, avatar_path, available) "
+             "VALUES ($1, $2, $3, $4)")
+        await self.db.execute(q, self.fake_mid, self.name, self.avatar_path, self.available)
+
+    async def update_profile_info(self) -> None:
+        q = ("UPDATE stranger SET name=$2, avatar_path=$3 "
+             "WHERE fake_mid=$1")
+        await self.db.execute(q, self.fake_mid, self.name, self.avatar_path)
+
+    async def make_available(self) -> None:
+        q = ("UPDATE stranger SET available=true "
+             "WHERE name=$1 AND avatar_path=$2")
+        await self.db.execute(q, self.name, self.avatar_path)
+
+    @classmethod
+    async def get_by_mid(cls, mid: str) -> Optional['Stranger']:
+        q = ("SELECT fake_mid, name, avatar_path, available "
+             "FROM stranger WHERE fake_mid=$1")
+        row = await cls.db.fetchrow(q, mid)
+        if not row:
+            return None
+        return cls(**row)
+
+    @classmethod
+    async def get_by_profile(cls, info: 'Participant') -> Optional['Stranger']:
+        q = ("SELECT fake_mid, name, avatar_path, available "
+             "FROM stranger WHERE name=$1 AND avatar_path=$2")
+        row = await cls.db.fetchrow(q, info.name, info.avatar.path)
+        if not row:
+            return None
+        return cls(**row)
+
+    @classmethod
+    async def get_any_available(cls) -> Optional['Stranger']:
+        q = ("SELECT fake_mid, name, avatar_path, available "
+             "FROM stranger WHERE available=true")
+        row = await cls.db.fetchrow(q)
+        if not row:
+            return None
+        return cls(**row)
+
+    @classmethod
+    async def init_available_or_new(cls) -> 'Stranger':
+        stranger = await cls.get_any_available()
+        if not stranger:
+            while True:
+                fake_mid = "_STRANGER_"
+                for _ in range(32):
+                    fake_mid += f"{randint(0,15):x}"
+                if await cls.get_by_mid(fake_mid) != None:
+                    # Extremely unlikely event of a randomly-generated ID colliding with another.
+                    # If it happens, must be not that unlikely after all, so pick a new seed.
+                    seed()
+                else:
+                    stranger = cls(fake_mid)
+                    break
+        return stranger

matrix_puppeteer_line/db/upgrade.py

@@ -109,7 +109,23 @@ async def upgrade_read_receipts(conn: Connection) -> None:
 @upgrade_table.register(description="Media metadata")
 async def upgrade_deduplicate_blob(conn: Connection) -> None:
     await conn.execute("""ALTER TABLE media
-       ADD COLUMN IF NOT EXISTS mime_type  TEXT,
+        ADD COLUMN IF NOT EXISTS mime_type  TEXT,
-       ADD COLUMN IF NOT EXISTS file_name  TEXT,
+        ADD COLUMN IF NOT EXISTS file_name  TEXT,
-       ADD COLUMN IF NOT EXISTS size       INTEGER
+        ADD COLUMN IF NOT EXISTS size       INTEGER
    """)
+
+
+@upgrade_table.register(description="Strangers")
+async def upgrade_strangers(conn: Connection) -> None:
+    await conn.execute("""
+    CREATE TABLE IF NOT EXISTS stranger (
+        fake_mid     TEXT NOT NULL UNIQUE,
+        name         TEXT NOT NULL,
+        avatar_path  TEXT NOT NULL,
+        available    BOOLEAN NOT NULL DEFAULT false,
+
+        PRIMARY KEY (name, avatar_path),
+        FOREIGN KEY (fake_mid)
+            REFERENCES puppet (mid)
+            ON DELETE CASCADE
+   )""")

matrix_puppeteer_line/portal.py

@@ -205,29 +205,34 @@ class Portal(DBPortal, BasePortal):
                 intent = None
         return intent
 
-    async def handle_remote_message(self, source: 'u.User', sender: Optional['p.Puppet'],
+    async def handle_remote_message(self, source: 'u.User', evt: Message) -> None:
-                                    evt: Message) -> None:
+        if await DBMessage.get_by_mid(evt.id):
+            self.log.debug(f"Ignoring duplicate message {evt.id}")
+            return
+
         if evt.is_outgoing:
             if source.intent:
+                sender = None
                 intent = source.intent
             else:
                 if not self.invite_own_puppet_to_pm:
                     self.log.warning(f"Ignoring message {evt.id}: double puppeting isn't enabled")
                     return
+                sender = p.Puppet.get_by_mid(evt.sender.id) if not self.is_direct else None
                 intent = await self._bridge_own_message_pm(source, sender, f"message {evt.id}")
                 if not intent:
                     return
-        elif self.other_user:
-            intent = (await p.Puppet.get_by_mid(self.other_user)).intent
-        elif sender:
-            intent = sender.intent
         else:
-            self.log.warning(f"Ignoring message {evt.id}: sender puppet is unavailable")
+            sender = await p.Puppet.get_by_mid(self.other_user if self.is_direct else evt.sender.id)
-            return
+            # TODO Respond to name/avatar changes of users in a DM
-
+            if not self.is_direct:
-        if await DBMessage.get_by_mid(evt.id):
+                if sender:
-            self.log.debug(f"Ignoring duplicate message {evt.id}")
+                    await sender.update_info(evt.sender, source.client)
-            return
+                else:
+                    self.log.warning(f"Could not find ID of LINE user who sent event {evt.id}")
+                    sender = await p.Puppet.get_by_profile(evt.sender, source.client)
+            intent = sender.intent
+            intent.ensure_joined(self.mxid)
 
         if evt.image and evt.image.url:
             if not evt.image.is_sticker or self.config["bridge.receive_stickers"]:
@@ -318,6 +323,9 @@ class Portal(DBPortal, BasePortal):
                 format=Format.HTML if msg_html else None,
                 body=msg_text, formatted_body=msg_html)
             event_id = await self._send_message(intent, content, timestamp=evt.timestamp)
+        # TODO Joins/leaves/invites/rejects, which are sent as LINE message events after all!
+        #      Also keep track of strangers who leave / get blocked / become friends
+        #      (maybe not here for all of that)
         else:
             content = TextMessageEventContent(
                 msgtype=MessageType.NOTICE,
@@ -349,11 +357,13 @@ class Portal(DBPortal, BasePortal):
             if reaction:
                 await self.main_intent.redact(self.mxid, reaction.mxid)
                 await reaction.delete()
+            # If there are as many receipts as there are chat participants, then everyone
+            # must have read the message, so send real read receipts from each puppet.
             # TODO Not just -1 if there are multiple _OWN_ puppets...
             if receipt_count == len(self._last_participant_update) - 1:
-                for participant in filter(lambda participant: not p.Puppet.is_mid_for_own_puppet(participant), self._last_participant_update):
+                for mid in filter(lambda mid: not p.Puppet.is_mid_for_own_puppet(mid), self._last_participant_update):
-                    puppet = await p.Puppet.get_by_mid(participant)
+                    intent = (await p.Puppet.get_by_mid(mid)).intent
-                    await puppet.intent.send_receipt(self.mxid, event_id)
+                    await intent.send_receipt(self.mxid, event_id)
             else:
                 # TODO Translatable string for "Read by"
                 reaction_mxid = await self.main_intent.react(self.mxid, event_id, f"(Read by {receipt_count})")
@@ -418,8 +428,13 @@ class Portal(DBPortal, BasePortal):
             if self._main_intent is self.az.intent:
                 self._main_intent = (await p.Puppet.get_by_mid(self.other_user)).intent
         for participant in conv.participants:
-            puppet = await p.Puppet.get_by_mid(participant.id)
+            # REMINDER: multi-user chats include your own LINE user in the participant list
-            await puppet.update_info(participant, client)
+            if participant.id != None:
+                puppet = await p.Puppet.get_by_mid(participant.id, client)
+                await puppet.update_info(participant, client)
+            else:
+                self.log.warning(f"Could not find ID of LINE user {participant.name}")
+                puppet = await p.Puppet.get_by_profile(participant, client)
         # TODO Consider setting no room name for non-group chats.
         #      But then the LINE bot itself may appear in the title...
         changed = await self._update_name(f"{conv.name} (LINE)")
@@ -480,7 +495,12 @@ class Portal(DBPortal, BasePortal):
             return
 
         # Store the current member list to prevent unnecessary updates
-        current_members = {participant.id for participant in participants}
+        current_members = set()
+        for participant in participants:
+            current_members.add(
+                participant.id if participant.id != None else \
+                (await p.Puppet.get_by_profile(participant)).mid)
+
         if current_members == self._last_participant_update:
             self.log.trace("Not updating participants: list matches cached list")
             return
@@ -495,7 +515,7 @@ class Portal(DBPortal, BasePortal):
         for participant in participants:
             if forbid_own_puppets and p.Puppet.is_mid_for_own_puppet(participant.id):
                 continue
-            intent = (await p.Puppet.get_by_mid(participant.id)).intent
+            intent = (await p.Puppet.get_by_sender(participant)).intent
             await intent.ensure_joined(self.mxid)
 
         print(current_members)
@@ -536,15 +556,8 @@ class Portal(DBPortal, BasePortal):
 
         self.log.debug("Got %d messages from server", len(messages))
         async with NotificationDisabler(self.mxid, source):
-            # Member joins/leaves are not shown in chat history.
-            # Best we can do is have a puppet join if its user had sent a message.
-            members_known = set(await self.main_intent.get_room_members(self.mxid)) if not self.is_direct else None
             for evt in messages:
-                puppet = await p.Puppet.get_by_mid(evt.sender.id) if not self.is_direct else None
+                await self.handle_remote_message(source, evt)
-                if puppet and evt.sender.id not in members_known:
-                    await puppet.update_info(evt.sender, source.client)
-                    members_known.add(evt.sender.id)
-                await self.handle_remote_message(source, puppet, evt)
         self.log.info("Backfilled %d messages through %s", len(messages), source.mxid)
 
     @property
@@ -680,10 +693,8 @@ class Portal(DBPortal, BasePortal):
             self.by_mxid[self.mxid] = self
             await self.backfill(source)
             if not self.is_direct:
-                # For multi-user chats, backfill before updating participants,
+                # TODO Joins and leaves are (usually) shown after all, so track them properly.
-                # to act as as a best guess of when users actually joined.
+                #      In the meantime, just check the participants list after backfilling.
-                # No way to tell when a user actually left, so just check the
-                # participants list after backfilling.
                 await self._update_participants(info.participants)
 
         return self.mxid

matrix_puppeteer_line/puppet.py

@@ -19,7 +19,7 @@ from mautrix.bridge import BasePuppet
 from mautrix.types import UserID, ContentURI
 from mautrix.util.simple_template import SimpleTemplate
 
-from .db import Puppet as DBPuppet
+from .db import Puppet as DBPuppet, Stranger
 from .config import Config
 from .rpc import Participant, Client, PathImage
 from . import user as u
@@ -141,6 +141,9 @@ class Puppet(DBPuppet, BasePuppet):
 
     @classmethod
     async def get_by_mid(cls, mid: str, create: bool = True) -> Optional['Puppet']:
+        if mid is None:
+            return None
+
         # TODO Might need to parse a real id from "_OWN"
         try:
             return cls.by_mid[mid]
@@ -160,10 +163,38 @@ class Puppet(DBPuppet, BasePuppet):
 
         return None
 
+    @classmethod
+    async def get_by_profile(cls, info: Participant, client: Optional[Client] = None) -> 'Puppet':
+        stranger = await Stranger.get_by_profile(info)
+        if not stranger:
+            stranger = await Stranger.init_available_or_new()
+
+            puppet = cls(stranger.fake_mid)
+            # NOTE An update will insert anyways, so just do it now
+            await puppet.insert()
+            await puppet.update_info(info, client)
+            puppet._add_to_cache()
+
+            # Get path from puppet in case it uses the URL as the path.
+            # But that should never happen in practice for strangers,
+            # which should only occur in rooms, where avatars have paths.
+            stranger.avatar_path = puppet.avatar_path
+            stranger.name = info.name
+            await stranger.insert()
+            # TODO Need a way to keep stranger name/avatar up to date,
+            #      lest name/avatar changes get seen as another stranger.
+            #      Also need to detect when a stranger becomes a friend.
+        return await cls.get_by_mid(stranger.fake_mid)
+
+    @classmethod
+    async def get_by_sender(cls, info: Participant, client: Optional[Client] = None) -> 'Puppet':
+        puppet = await cls.get_by_mid(info.id)
+        return puppet if puppet else await cls.get_by_profile(info, client)
+
     # TODO When supporting multiple bridge users, this should return the user whose puppet this is
     @classmethod
     def is_mid_for_own_puppet(cls, mid) -> bool:
-        return mid.startswith("_OWN_") if mid else False
+        return mid and mid.startswith("_OWN_")
 
     @classmethod
     async def get_by_custom_mxid(cls, mxid: UserID) -> Optional['u.User']:

matrix_puppeteer_line/rpc/types.py

@@ -41,9 +41,9 @@ class ChatListInfo(SerializableAttrs['ChatListInfo']):
 
 @dataclass
 class Participant(SerializableAttrs['Participant']):
-    id: str
-    avatar: Optional[PathImage]
     name: str
+    avatar: Optional[PathImage]
+    id: Optional[str] = None
 
 
 @dataclass

matrix_puppeteer_line/user.py

@@ -162,12 +162,11 @@ class User(DBUser, BaseUser):
     async def handle_message(self, evt: Message) -> None:
         self.log.trace("Received message %s", evt)
         portal = await po.Portal.get_by_chat_id(evt.chat_id, create=True)
-        puppet = await pu.Puppet.get_by_mid(evt.sender.id) if not portal.is_direct else None
         if not portal.mxid:
             await self.client.set_last_message_ids(await DBMessage.get_max_mids())
             chat_info = await self.client.get_chat(evt.chat_id)
             await portal.create_matrix_room(self, chat_info)
-        await portal.handle_remote_message(self, puppet, evt)
+        await portal.handle_remote_message(self, evt)
 
     async def handle_receipt(self, receipt: Receipt) -> None:
         self.log.trace(f"Received receipt for chat {receipt.chat_id}")

puppet/src/contentscript.js

@@ -238,10 +238,18 @@ class MautrixController {
 			sender.id = this.getUserIdFromFriendsList(sender.name)
 			// Group members aren't necessarily friends,
 			// but the participant list includes their ID.
+			// ROOMS DO NOT!! Ugh.
 			if (!sender.id) {
 				const participantsList = document.querySelector(participantsListSelector)
-				imgElement = participantsList.querySelector(`img[alt='${sender.name}'`)
+				// Groups use a participant's name as the alt text of their avatar image,
-				sender.id = imgElement.parentElement.parentElement.getAttribute("data-mid")
+				// but rooms do not...ARGH! But they both use a dedicated element for it.
+				const participantNameElement =
+					Array.from(participantsList.querySelectorAll(`.mdRGT13Ttl`))
+					.find(e => e.innerText == "EN-MY translator")
+				if (participantNameElement) {
+					imgElement = participantNameElement.previousElementSibling.firstElementChild
+					sender.id = imgElement.parentElement.parentElement.getAttribute("data-mid")
+				}
 			} else {
 				imgElement = element.querySelector(".mdRGT07Img > img")
 			}