Slack Connector issue with Conversations API and 'is_archived' about danswer

[chrisedington]

I'm testing a potential fix:

def get_channel_messages(
    client: WebClient,
    channel: dict[str, Any],
    oldest: str | None = None,
    latest: str | None = None,
) -> Generator[list[MessageType], None, None]:
    """Get all messages in a channel"""
    # join so that the bot can access messages
    if not channel["is_member"] and not channel["is_archived"]:
        client.conversations_join(
            channel=channel["id"], is_private=channel["is_private"]
        )

    for result in _make_slack_api_call(
        client.conversations_history,
        channel=channel["id"],
        oldest=oldest,
        latest=latest,
    ):
        yield cast(list[MessageType], result["messages"])

[chrisedington]

That failed, I think it needs to move up to get_channels.

Python SDK provides:

Archived channels are included by default. You can exclude them by passing exclude_archived=True to your request.

response = client.conversations_list(exclude_archived=True)

See conversations.list for more info.

[Weves]

Good catch! We hadn't tested on slack workspaces with archived channels yet. Approved / merged #208.

I will say that ideally, in the long term, we would be able to scrape archived channels as well (since they often contain some useful, relevant information), but it doesn't seem like that's possible based on my reading of the Slack API :(