Update roadmap with missing edge cases and KT message types

And set an optional property in AudioAttachment to be safe

ROADMAP.md

@@ -13,7 +13,12 @@
     * [ ] Replies
       * [x] In DMs
       * [ ] In multi-user rooms
-    * [x] Mentions
+      * [ ] To relay users
+    * [ ] Mentions
+      * [x] In DMs
+      * [ ] In multi-user rooms
+      * [x] To relay users
+    * [ ] Polls
   * [x] Message redactions<sup>[1]</sup>
   * [ ] Message reactions
   * [x] Read receipts
@@ -36,8 +41,11 @@
       * [x] Videos
       * [x] Images
       * [ ] Locations
+      * [ ] Links
     * [x] Replies
     * [x] Mentions
+    * [ ] Polls
+    * [ ] Posts
   * [x] Message deletion/hiding
   * [ ] Message reactions
   * [x] Message history

matrix_appservice_kakaotalk/example-config.yaml

@@ -246,10 +246,10 @@ bridge:
             m.text: '<b>$sender_displayname</b>: $message'
             m.notice: '<b>$sender_displayname<b>: $message'
             m.emote: '* <b>$sender_displayname<b> $message'
-            m.file: '<b>$sender_displayname</b> sent a file'
-            m.image: '<b>$sender_displayname</b> sent an image'
-            m.audio: '<b>$sender_displayname</b> sent an audio file'
-            m.video: '<b>$sender_displayname</b> sent a video'
+            m.file: 'File from <b>$sender_displayname</b>: $message'
+            m.image: 'Image from <b>$sender_displayname</b>: $message'
+            m.audio: 'Audio from <b>$sender_displayname</b>: $message'
+            m.video: 'Video from <b>$sender_displayname</b>: $message'
             m.location: '<b>$sender_displayname</b> sent a location'
 
 rpc:

matrix_appservice_kakaotalk/formatter/from_kakaotalk.py

@@ -24,60 +24,8 @@ from ..kt.types.chat.attachment.mention import MentionStruct
 
 from .. import puppet as pu, user as u
 
-_START = r"^|\s"
-_END = r"$|\s"
-_TEXT_NO_SURROUNDING_SPACE = r"(?:[^\s].*?[^\s])|[^\s]"
-COMMON_REGEX = re.compile(rf"({_START})([_~*])({_TEXT_NO_SURROUNDING_SPACE})\2({_END})")
-INLINE_CODE_REGEX = re.compile(rf"({_START})(`)(.+?)`({_END})")
-MENTION_REGEX = re.compile(r"@([0-9]{1,15})\u2063(.+?)\u2063")
 
-tags = {"_": "em", "*": "strong", "~": "del", "`": "code"}
-
-
-def _handle_match(html: str, match: Match, nested: bool) -> tuple[str, int]:
-    start, end = match.start(), match.end()
-    prefix, sigil, text, suffix = match.groups()
-    if nested:
-        text = _convert_formatting(text)
-    tag = tags[sigil]
-    # We don't want to include the whitespace suffix length, as that could be used as the
-    # whitespace prefix right after this formatting block.
-    pos = start + len(prefix) + (2 * len(tag) + 5) + len(text)
-    html = f"{html[:start]}{prefix}<{tag}>{text}</{tag}>{suffix}{html[end:]}"
-    return html, pos
-
-
-def _convert_formatting(html: str) -> str:
-    pos = 0
-    while pos < len(html):
-        i_match = INLINE_CODE_REGEX.search(html, pos)
-        c_match = COMMON_REGEX.search(html, pos)
-        if i_match and c_match:
-            match = min(i_match, c_match, key=lambda match: match.start())
-        else:
-            match = i_match or c_match
-
-        if match:
-            html, pos = _handle_match(html, match, nested=match != i_match)
-        else:
-            break
-    return html
-
-
-def _handle_blockquote(output: list[str], blockquote: bool, line: str) -> tuple[bool, str]:
-    if not blockquote and line.startswith("&gt; "):
-        line = line[len("&gt; ") :]
-        output.append("<blockquote>")
-        blockquote = True
-    elif blockquote:
-        if line.startswith("&gt;"):
-            line = line[len("&gt;") :]
-            if line.startswith(" "):
-                line = line[1:]
-        else:
-            output.append("</blockquote>")
-            blockquote = False
-    return blockquote, line
+MENTION_REGEX = re.compile(r"@(\d+)\u2063(.+?)\u2063")
 
 
 async def kakaotalk_to_matrix(msg: str | None, mentions: list[MentionStruct] | None) -> TextMessageEventContent:

matrix_appservice_kakaotalk/formatter/from_matrix.py

@@ -19,6 +19,7 @@ from typing import NamedTuple
 
 from mautrix.appservice import IntentAPI
 from mautrix.types import Format, MessageEventContent, RelationType, RoomID, UserID
+from mautrix.util import utf16_surrogate
 from mautrix.util.formatter import (
     EntityString,
     EntityType,
@@ -46,6 +47,7 @@ class SendParams(NamedTuple):
 
 class KakaoTalkFormatString(EntityString[SimpleEntity, EntityType], MarkdownString):
     def format(self, entity_type: EntityType, **kwargs) -> KakaoTalkFormatString:
+        prefix = suffix = ""
         if entity_type == EntityType.USER_MENTION:
             self.entities.append(
                 SimpleEntity(
@@ -55,7 +57,32 @@ class KakaoTalkFormatString(EntityString[SimpleEntity, EntityType], MarkdownStri
                     extra_info={"user_id": kwargs["user_id"]},
                 )
             )
-            self.text = f"@{self.text}"
+            return self
+        elif entity_type == EntityType.BOLD:
+            prefix = suffix = "*"
+        elif entity_type == EntityType.ITALIC:
+            prefix = suffix = "_"
+        elif entity_type == EntityType.STRIKETHROUGH:
+            prefix = suffix = "~"
+        elif entity_type == EntityType.URL:
+            if kwargs["url"] != self.text:
+                suffix = f" ({kwargs['url']})"
+        elif entity_type == EntityType.PREFORMATTED:
+            prefix = f"```{kwargs['language']}\n"
+            suffix = "\n```"
+        elif entity_type == EntityType.INLINE_CODE:
+            prefix = suffix = "`"
+        elif entity_type == EntityType.BLOCKQUOTE:
+            children = self.trim().split("\n")
+            children = [child.prepend("> ") for child in children]
+            return self.join(children, "\n")
+        elif entity_type == EntityType.HEADER:
+            prefix = "#" * kwargs["size"] + " "
+        else:
+            return self
+
+        self._offset_entities(len(prefix))
+        self.text = f"{prefix}{self.text}{suffix}"
         return self
 
 
@@ -126,8 +153,8 @@ async def matrix_to_kakaotalk(
     else:
         reply_to = None
     if content.get("format", None) == Format.HTML and content["formatted_body"] and content.msgtype.is_text:
-        parsed = await ToKakaoTalkParser().parse(content["formatted_body"])
-        text = parsed.text
+        parsed = await ToKakaoTalkParser().parse(utf16_surrogate.add(content["formatted_body"]))
+        text = utf16_surrogate.remove(parsed.text)
         mentions_by_user: dict[Long, MentionStruct] = {}
         # Make sure to not create remote mentions for any remote user not in the room
         if parsed.entities:

matrix_appservice_kakaotalk/kt/client/client.py

@@ -240,13 +240,17 @@ class Client:
         self.user.oauth_credential = oauth_info.credential
         await self.user.save()
 
-    async def connect(self) -> LoginResult:
+    async def connect(self) -> LoginResult | None:
         """
         Start a new talk session by providing a token obtained from a prior login.
         Receive a snapshot of account state in response.
         """
+        try:
             login_result = await self._api_user_request_result(LoginResult, "connect")
             assert self.user.ktid == login_result.userId, f"User ID mismatch: expected {self.user.ktid}, got {login_result.userId}"
+        except SerializerError:
+            self.log.exception("Unable to deserialize login result, but connecting anyways")
+            login_result = None
         # TODO Skip if handlers are already listening. But this is idempotent and thus probably safe
         self._start_listen()
         return login_result
@@ -322,9 +326,9 @@ class Client:
         text: str,
         reply_to: ReplyAttachment | None,
         mentions: list[MentionStruct] | None,
-    ) -> Chatlog:
+    ) -> Long:
         return await self._api_user_request_result(
-            Chatlog,
+            Long,
             "send_chat",
             channel_props=channel_props.serialize(),
             text=text,
@@ -342,9 +346,9 @@ class Client:
         width: int | None = None,
         height: int | None = None,
         ext: str | None = None,
-    ) -> Chatlog:
+    ) -> Long:
         return await self._api_user_request_result(
-            Chatlog,
+            Long,
             "send_media",
             channel_props=channel_props.serialize(),
             type=media_type,

matrix_appservice_kakaotalk/kt/types/channel/channel_type.py

@@ -30,6 +30,10 @@ class KnownChannelType(str, Enum):
     def is_direct(cls, value: Union["KnownChannelType", str]) -> bool:
         return value in [cls.DirectChat, cls.MemoChat]
 
+    @classmethod
+    def is_open(cls, value: Union["KnownChannelType", str]) -> bool:
+        return value in [cls.OM, cls.OD]
+
 
 ChannelType = Union[KnownChannelType, str] # Substitute for ChannelType = "name1" | ... | "nameN" | str
 

matrix_appservice_kakaotalk/kt/types/chat/attachment/emoticon.py

@@ -25,7 +25,7 @@ class EmoticonAttachment(Attachment):
     path: str
     name: str
     type: str
-    alt: str
+    alt: Optional[str] = None # NOTE Made optional
     s: Optional[int] = None
     sound: Optional[str] = None
     width: Optional[int] = None

matrix_appservice_kakaotalk/kt/types/chat/attachment/media.py

@@ -78,10 +78,11 @@ class FileAttachment(MediaKeyAttachment):
 
 
 @dataclass
-class AudioAttachment(MediaKeyAttachment):
-    # NOTE Changed superclass from Attachment
+class AudioAttachment(Attachment):
+    url: str
     d: int
     expire: Optional[int] = None
+    s: Optional[int] = None # NOTE Optional for inbound
 
 
 @dataclass

matrix_appservice_kakaotalk/matrix.py

@@ -157,6 +157,7 @@ class MatrixHandler(BaseMatrixHandler):
         if not user.is_connected:
             return
         message = await DBMessage.get_by_mxid(event_id, portal.mxid)
+        if message:
             await user.client.mark_read(portal.channel_props, message.ktid)
 
     async def handle_ephemeral_event(

matrix_appservice_kakaotalk/portal.py

@@ -40,6 +40,8 @@ from mautrix.types import (
     EventType,
     FileInfo,
     ImageInfo,
+    JoinRule,
+    JoinRulesStateEventContent,
     LocationMessageEventContent,
     MediaInfo,
     MediaMessageEventContent,
@@ -198,6 +200,7 @@ class Portal(DBPortal, BasePortal):
             KnownChatType.VIDEO: cls._handle_kakaotalk_video,
             KnownChatType.AUDIO: cls._handle_kakaotalk_audio,
             #KnownChatType.FILE: cls._handle_kakaotalk_file,
+            16385: cls._handle_kakaotalk_deleted,
         }
 
     # region DB conversion
@@ -629,6 +632,17 @@ class Portal(DBPortal, BasePortal):
                 "content": self.bridge_info,
             },
         ]
+        if KnownChannelType.is_open(info.channel_info.type):
+            initial_state.extend((
+                {
+                    "type": str(EventType.ROOM_JOIN_RULES),
+                    "content": JoinRulesStateEventContent(join_rule=JoinRule.PUBLIC).serialize(),
+                },
+                {
+                    "type": "m.room.guest_access",
+                    "content": {"guest_access": "forbidden"},
+                },
+            ))
         invites = []
         if self.config["bridge.encryption.default"] and self.matrix.e2ee:
             self.encrypted = True
@@ -777,6 +791,11 @@ class Portal(DBPortal, BasePortal):
         elif not sender.is_connected:
             raise Exception("not connected to KakaoTalk chats")
         elif is_relay:
+            if not message.msgtype.is_text:
+                intro_message = TextMessageEventContent(msgtype=MessageType.TEXT, body=message.body)
+                await self.apply_relay_message_format(orig_sender, intro_message)
+                await self._send_chat(sender, intro_message)
+            else:
                 await self.apply_relay_message_format(orig_sender, message)
         if message.msgtype == MessageType.TEXT or message.msgtype == MessageType.NOTICE:
             await self._handle_matrix_text(event_id, sender, message)
@@ -787,6 +806,21 @@ class Portal(DBPortal, BasePortal):
         else:
             raise NotImplementedError(f"Unsupported message type {message.msgtype}")
 
+    async def _send_chat(
+        self, sender: u.User, message: TextMessageEventContent, event_id: EventID | None = None
+    ) -> Long:
+        converted = await matrix_to_kakaotalk(message, self.mxid, self.log, self.main_intent)
+        try:
+            return await sender.client.send_chat(
+                self.channel_props,
+                text=converted.text,
+                reply_to=converted.reply_to,
+                mentions=converted.mentions,
+            )
+        except CommandException as e:
+            self.log.debug(f"Error handling Matrix message {event_id if event_id else '<extra>'}: {e!s}")
+            raise
+
     async def _make_dbm(self, event_id: EventID, ktid: Long | None = None) -> DBMessage:
         dbm = DBMessage(
             mxid=event_id,
@@ -803,19 +837,9 @@ class Portal(DBPortal, BasePortal):
     async def _handle_matrix_text(
         self, event_id: EventID, sender: u.User, message: TextMessageEventContent
     ) -> None:
-        converted = await matrix_to_kakaotalk(message, self.mxid, self.log, self.main_intent)
-        try:
-            chatlog = await sender.client.send_chat(
-                self.channel_props,
-                text=converted.text,
-                reply_to=converted.reply_to,
-                mentions=converted.mentions,
-            )
-        except CommandException as e:
-            self.log.debug(f"Error handling Matrix message {event_id}: {e!s}")
-            raise
-        await self._make_dbm(event_id, chatlog.logId)
-        self.log.debug(f"Handled Matrix message {event_id} -> {chatlog.logId}")
+        log_id = await self._send_chat(sender, message, event_id)
+        await self._make_dbm(event_id, log_id)
+        self.log.debug(f"Handled Matrix message {event_id} -> {log_id}")
         sender.send_remote_checkpoint(
             MessageSendCheckpointStatus.SUCCESS,
             event_id,
@@ -843,7 +867,7 @@ class Portal(DBPortal, BasePortal):
             width = message.info.width
             height = message.info.height
         try:
-            chatlog = await sender.client.send_media(
+            log_id = await sender.client.send_media(
                 self.channel_props,
                 TO_MSGTYPE_MAP[message.msgtype],
                 data,
@@ -855,8 +879,8 @@ class Portal(DBPortal, BasePortal):
         except CommandException as e:
             self.log.debug(f"Error uploading media for Matrix message {event_id}: {e!s}")
             raise
-        await self._make_dbm(event_id, chatlog.logId)
-        self.log.debug(f"Handled Matrix message {event_id} -> {chatlog.logId}")
+        await self._make_dbm(event_id, log_id)
+        self.log.debug(f"Handled Matrix message {event_id} -> {log_id}")
         sender.send_remote_checkpoint(
             MessageSendCheckpointStatus.SUCCESS,
             event_id,
@@ -1086,7 +1110,7 @@ class Portal(DBPortal, BasePortal):
             type_str = str(chat_type)
         self.log.warning("No handler for chat type \"%s\" (%s)",
             type_str,
-            f"text = {chat_text}" if chat_text is not None else "no text",
+            f"text = \"{chat_text}\"" if chat_text is not None else "no text",
         )
         if chat_text:
             events = await self._handle_kakaotalk_text(
@@ -1115,6 +1139,14 @@ class Portal(DBPortal, BasePortal):
         self.log.info("Got feed message at %s: %s", timestamp, chat_text or "none")
         return []
 
+    async def _handle_kakaotalk_deleted(
+        self,
+        timestamp: int,
+        **_
+    ) -> list[EventID]:
+        self.log.info(f"Got deleted (?) message at {timestamp}")
+        return []
+
     async def _handle_kakaotalk_text(
         self,
         intent: IntentAPI,

matrix_appservice_kakaotalk/rpc/rpc.py

@@ -144,9 +144,9 @@ class RPCClient:
         self._reader = r
         self._writer = w
         self._read_task = self.loop.create_task(self._try_read_loop())
+        await self._raw_request("register", peer_id=self.config["appservice.address"])
         self._is_connected.set()
         self._is_disconnected.clear()
-        await self.request("register", peer_id=self.config["appservice.address"])
 
     async def disconnect(self) -> None:
         async with self._connection_lock:
@@ -258,9 +258,11 @@ class RPCClient:
         try:
             await self._read_loop()
         except asyncio.CancelledError:
-            pass
+            return
         except:
             self.log.exception("Fatal error in read loop")
+        self.log.debug("Reader disconnected")
+        self._on_disconnect()
 
     async def _read_loop(self) -> None:
         while self._reader is not None and not self._reader.at_eof():
@@ -275,9 +277,6 @@ class RPCClient:
                 except asyncio.LimitOverrunError as e:
                     self.log.warning(f"Buffer overrun: {e}")
                     line += await self._reader.read(self._reader._limit)
-                except ConnectionResetError:
-                    if self._reader is not None:
-                        raise
                 except asyncio.CancelledError:
                     raise
             if not line:
@@ -293,8 +292,6 @@ class RPCClient:
                 raise
             except:
                 self.log.exception("Failed to handle incoming request %s", line_str)
-        self.log.debug("Reader disconnected")
-        self._on_disconnect()
 
     async def _raw_request(self, command: str, is_secret: bool = False, **data: JSON) -> asyncio.Future[JSON]:
         req_id = self._next_req_id

matrix_appservice_kakaotalk/user.py

@@ -412,6 +412,7 @@ class User(DBUser, BaseUser):
         try:
             login_result = await self.client.connect()
             await self.on_connect()
+            if login_result:
                 await self._sync_channels(login_result, sync_count)
             return True
         except AuthenticationRequired as e:

node/src/client.js

@@ -634,12 +634,13 @@ export default class PeerClient {
 	 */
 	sendChat = async (req) => {
 		const talkChannel = await this.#getUserChannel(req.mxid, req.channel_props)
-
-		return await talkChannel.sendChat({
+		const res = await talkChannel.sendChat({
 			text: req.text,
 			type: !!req.reply_to ? KnownChatType.REPLY : KnownChatType.TEXT,
 			attachment: !req.mentions ? req.reply_to : {...req.reply_to, mentions: req.mentions},
 		})
+		if (res.success) res.result = res.result.logId
+		return res
 	}
 
 	/**
@@ -655,14 +656,15 @@ export default class PeerClient {
 	 */
 	sendMedia = async (req) => {
 		const talkChannel = await this.#getUserChannel(req.mxid, req.channel_props)
-
-		return await talkChannel.sendMedia(req.type, {
+		const res = await talkChannel.sendMedia(req.type, {
 			data: Uint8Array.from(req.data),
 			name: req.name,
 			width: req.width,
 			height: req.height,
 			ext: req.ext,
 		})
+		if (res.success) res.result = res.result.logId
+		return res
 	}
 
 	/**
@@ -770,6 +772,7 @@ export default class PeerClient {
 				send_chat: this.sendChat,
 				send_media: this.sendMedia,
 				delete_chat: this.deleteChat,
+				mark_read: this.markRead,
 			}[req.command] || this.handleUnknownCommand
 		}
 		const resp = { id: req.id }