Does email/IMAP support threading? If not, will it ever?

I’m not able to get the email integration working at the moment, and I don’t find any mention of this in the docs (might want to add this there once answered :grinning_face_with_smiling_eyes:). So I thought I’d ask here:

Does the email integration support “threading”? (grouping of messages on a single topic/chain of replies) If not, does this mean that e.g. every new reply to an email thread will create a new Entity? This seems rather… non-ideal. For me I think I’d actually prefer if it worked like the Discourse integration currently does, and simply updated the Rich Text field with the full contents of the email thread whenever a new message came in.

If threading is not supported, is there any suggested approach to make it more manageable to deal with email in Fibery?

I’m hoping that this will be solved based on the discussion in “Conversations” for integrations / Email integration and the fact that it is being considered during the Blocks planning and implementation process.

1 Like

Hi, @Oshyan

Thanks for asking. We are going to add threading support after blocks implementation in Fibery. Unfortunately no workarounds are coming to my mind as alternative to handle it right now.

Thanks,
Oleg

2 Likes

@Oleg,
Should we still expect to have thread/conversation support for GMail/IMAP integrations in Fibery?

Hi, @Matt_Blais

That is in our plans. Unfortunately we can’t tell exactly when it will be implemented.

Thanks,
Oleg

just bumping this if there is any development :slight_smile:

someone posted that they had programmed a solution with stripping the RE: and then grouping by name or something, but i cannot find it. also wondering if that is something that breaks horribly when sync updates all fields again?

i dont even know if this is something that exists or how to even name this … but a sorting that is begin to sort from the last character (right to left) would be a “workaround” … (which sparks the question if you have support for languages that are right to left …)

I don’t know if it got posted ever, but I have previously experimented with stripping the RE: FW: etc. from the message subject using regex.
It is potentially quite complicated though, because different email clients and different languages give prefixes with a variety of formats.

However, if you get such a field working, there’s no reason to think it will break with a sync - you can add fields to an integration database without affecting the fields that are being synced.

If all you wanted to do was sort messages so that ones with the same subject (minus the Re: bits) then reversing the string would be a solution, but it wouldn’t really create a thread. For that, you would ideally need to use an auto-relation that linked all emails to others with the same fundamental subject.

We don’t have any specific localisation support. Of course, users can choose to name their databases however they like (with some minor restrictions).

for this you would work within a formula field or write an automation with a script and then add that self-relation looking for same name of field, correct?

would that be something that an AI could solve as well? or would that be a bit too much in regards of cost

I was wondering if google is tagging these msgs somehow as replies in their API to get a read out of that … ( i guess there is more to this threads connections than just the Subject field, or inst it?)

If it were me, I would use a formula field to extract the fundamental subject from the name field, and then I would use auto-relation to link messages together based on them having the same fundamental subject and the same sender.

It might be possible, but AI is currently rather unreliable/untrustworthy, so I wouldn’t be confident that it had done its job correctly.

The other, simpler, solution would be to adjust the Email integration to include the email headers as a synced field, or at least those headers that are used for thread ID.

3 Likes

is that something a user can do or does fibery need to do that?

this is the formula I ended up using: Trim(ReplaceRegex(Lower(Name), "^.*(?:re:|aw:|fwd:)", ""))

it changes everything to lower case and then looks for the last occurence of either re:, aw: or fwd: and gets rid of that. (chatgpt helped with the regex)

then i made a relation many-to-many with the auto match on the formula field.
ultimately adding a feed-view of the related msgs to the view and sorting by date.

awesome!

… now how to strip the “quoted” parts of the msgs … (it never stops hahaha … )

Fibery would need to add that.

my integration of the “strip” field (that gets rid of all the re) works great … until the same title is there from two different mail-chains … any way of adding a filter that is matching if an email is either from or to the same person …?

the feed-view can not be filtered by formula … so i dont know how to do this on a dynamic end …

Well, assuming that for every message you’re either the sender or the recipient, you could create a formula field which gets the other field.
I.e. if the Message is To dukevannori then return the From email, and vice versa.
Then you can add matching on this formula field to your auto-linking.
However, the To field can actually theoretically contain multiple people, so it’s not infallible. You can just pick the first listed name, but there’s a small chance that you’ll miss a match.

i always make the mistake that I think there is some shortcut :wink: but it’s just adding some more fields to make the validations. easy! :thinking::upside_down_face:

i guess the chance of both the banner missing and the title being the same is quite low. i take it. until you make some internal magic with the thread stuff :wink:

thanks! have a great weekend