Feature/multiple feishu bot support

[MarkHoch]

Pull request overview

Adds support for running multiple Feishu (and other) channel instances in a single DeerFlow deployment by introducing list-based channel configuration and per-instance naming, alongside a Feishu WebSocket initialization rewrite to isolate event loops and avoid uvloop conflicts.

Changes:

• Extend ChannelService config parsing to support list-format multi-instance channels with explicit unique instance names and a stored _base_type.
• Update ChannelManager logic to better handle multi-instance channel names when resolving streaming capability and inbound file readers.
• Refactor Feishu channel startup to import/initialize lark_oapi inside a dedicated thread with its own asyncio loop and a global init lock.

Reviewed changes

Copilot reviewed 4 out of 5 changed files in this pull request and generated 5 comments.

Show a summary per file

File	Description
backend/app/channels/feishu.py	Reworks Feishu WS startup to isolate event loops and serialize SDK initialization across instances.
backend/app/channels/service.py	Adds list-format multi-instance channel config support and base-type resolution for channel startup/status.
backend/app/channels/manager.py	Adjusts streaming + inbound file reader resolution for multi-instance channel names.
backend/app/channels/wecom.py	Changes default working_message behavior for WeCom streaming replies.
.gitignore	Adds an ignore entry for a tracked source file (likely unintended).

Comments suppressed due to low confidence (1)

backend/app/channels/feishu.py:98

• start() no longer checks for lark-oapi availability before marking the channel as running. If lark-oapi isn’t installed, the WS thread will error, but the channel will remain subscribed and appear running, and no friendly install hint is logged. Consider a non-import check (e.g., importlib.util.find_spec("lark_oapi")) in start() and/or setting _running back to False when _run_ws fails during initialization.

    async def start(self) -> None:
        if self._running:
            return

        app_id = self.config.get("app_id", "")
        app_secret = self.config.get("app_secret", "")
        domain = self.config.get("domain", "https://open.feishu.cn")

        if not app_id or not app_secret:
            logger.error("Feishu channel requires app_id and app_secret")
            return

        logger.info("[Feishu] using domain: %s", domain)
        self._main_loop = asyncio.get_event_loop()

        self._running = True
        self.bus.subscribe_outbound(self._on_outbound)

        # Both ws.Client construction and start() must happen in a dedicated
        # thread with its own event loop.  lark-oapi caches the running loop
        # at construction time and later calls loop.run_until_complete(),
        # which conflicts with an already-running uvloop.
        self._thread = threading.Thread(
            target=self._run_ws,
            args=(app_id, app_secret, domain),
            daemon=True,
        )
        self._thread.start()
        logger.info("Feishu channel started")

[Copilot]

Pull request overview

Adds support for running multiple Feishu (and other) channel instances in a single DeerFlow deployment by introducing list-based channel configuration and per-instance naming, alongside a Feishu WebSocket initialization rewrite to isolate event loops and avoid uvloop conflicts.

Changes:

• Extend ChannelService config parsing to support list-format multi-instance channels with explicit unique instance names and a stored _base_type.
• Update ChannelManager logic to better handle multi-instance channel names when resolving streaming capability and inbound file readers.
• Refactor Feishu channel startup to import/initialize lark_oapi inside a dedicated thread with its own asyncio loop and a global init lock.

Reviewed changes

Copilot reviewed 4 out of 5 changed files in this pull request and generated 5 comments.

Show a summary per file

File	Description
backend/app/channels/feishu.py	Reworks Feishu WS startup to isolate event loops and serialize SDK initialization across instances.
backend/app/channels/service.py	Adds list-format multi-instance channel config support and base-type resolution for channel startup/status.
backend/app/channels/manager.py	Adjusts streaming + inbound file reader resolution for multi-instance channel names.
backend/app/channels/wecom.py	Changes default working_message behavior for WeCom streaming replies.
.gitignore	Adds an ignore entry for a tracked source file (likely unintended).

Comments suppressed due to low confidence (1)

backend/app/channels/feishu.py:98

• start() no longer checks for lark-oapi availability before marking the channel as running. If lark-oapi isn’t installed, the WS thread will error, but the channel will remain subscribed and appear running, and no friendly install hint is logged. Consider a non-import check (e.g., importlib.util.find_spec("lark_oapi")) in start() and/or setting _running back to False when _run_ws fails during initialization.

    async def start(self) -> None:
        if self._running:
            return

        app_id = self.config.get("app_id", "")
        app_secret = self.config.get("app_secret", "")
        domain = self.config.get("domain", "https://open.feishu.cn")

        if not app_id or not app_secret:
            logger.error("Feishu channel requires app_id and app_secret")
            return

        logger.info("[Feishu] using domain: %s", domain)
        self._main_loop = asyncio.get_event_loop()

        self._running = True
        self.bus.subscribe_outbound(self._on_outbound)

        # Both ws.Client construction and start() must happen in a dedicated
        # thread with its own event loop.  lark-oapi caches the running loop
        # at construction time and later calls loop.run_until_complete(),
        # which conflicts with an already-running uvloop.
        self._thread = threading.Thread(
            target=self._run_ws,
            args=(app_id, app_secret, domain),
            daemon=True,
        )
        self._thread.start()
        logger.info("Feishu channel started")

[MarkHoch]

@copilot apply changes based on the comments in this thread

[Copilot]

Pull request overview

Copilot reviewed 4 out of 4 changed files in this pull request and generated 3 comments.

Comments suppressed due to low confidence (1)

backend/app/channels/feishu.py:100

• start() marks the channel as running and subscribes to outbound messages before the WS thread has imported/initialized lark_oapi and built _api_client. If the SDK isn’t installed or initialization fails, the channel will still appear running, and outbound messages will be dropped (send() just logs a warning when _api_client is None). Consider a fast availability check before starting the thread (e.g. importlib.util.find_spec), and/or waiting for _api_client_ready with a short timeout and rolling back _running + outbound subscription when initialization fails (surfacing _api_client_init_error in logs/status).

        logger.info("[Feishu] using domain: %s", domain)
        self._main_loop = asyncio.get_event_loop()

        self._running = True
        self.bus.subscribe_outbound(self._on_outbound)

        # Both ws.Client construction and start() must happen in a dedicated
        # thread with its own event loop.  lark-oapi caches the running loop
        # at construction time and later calls loop.run_until_complete(),
        # which conflicts with an already-running uvloop.
        self._thread = threading.Thread(
            target=self._run_ws,
            args=(app_id, app_secret, domain),
            daemon=True,
        )
        self._thread.start()
        logger.info("Feishu channel started")

[MarkHoch]

multiple feishu bot support