WhatsApp "Someone" nicknaming

[DummyThatMatters]

Is your feature request related to a problem? Please describe.
WhatsApp tend to relay messages with "<Someone>" nickname in some cases (randomly) even if actual WhatsApp nickname presented. So it is completely impossible to distinguish authors of the posts on the other side of the bridge.

Describe the solution you'd like
To fix WhatsApp "<Someone>" nicknames with actual nicknames and add the option to use phone numbers as <NICK> source (if actual nickname is not set).

Describe alternatives you've considered
To generate unique random pseudonym nickname for each "anonymous" user instead of static "<Someone>" nickname.

Additional context
Group chat, WhatsApp - Mattermost bridge, ~50 users, some users have specified nickname, some of them are not.

[JEK58]

Why did you close this? Did you find a solution?

[joko2020]

Dear Sir,
I need help to integrate Whatsapp to Nextcloud Talk, I have Installed matterbridge on Nextcloud 20.0.7 with matterbridge Running version 1.22.0... but I cannot find whatsapp channel in the talk matterbridge setting. after googling seems to manually edit the matterbridge.toml config file.. but i' cannot find that file.. where is the file path of matterbridge.toml
thanks

[42wim]

@joko2020 you should look at the nextcloud forums, I don't know where they put the matterbridge.toml

[joko2020]

@joko2020 you should look at the nextcloud forums, I don't know where they put the matterbridge.toml

Thanks for your reply.. unfortunately , I cannot get the right inormation Also

[micw]

Solution for this particular issue is IMO to allow to show the phone number instead of "someone". While that's not wanted in public groups, it's absolutely wanted in most private group bridging.
I'll create another issue for the fact that jid->account nick name resolution seems to be broken.

[micw]

When reading the code og go-whatsapp, I found:

matterbridge/vendor/github.com/Rhymen/go-whatsapp/store.go

Line 48 in 4fcad8e

jid := strings.Replace(contactNode.Attributes["jid"], "@c.us", "@s.whatsapp.net", 1)

May it be that the key in the contatc list is different from the actual sender jid, causing the lookup failures?

[micw]

It seems that if I restart matterbridge, after someone has sent a message to a WA group, this one's nick is resolved properly.

[micw]

There was already a related fix in #796. But it seems that this is insufficient. I have tried to analyze when someone is printed and when the nick name.
It seems that if a user posts a second message after he posts a message, the nick is resolved properly. After restart, the nick name cache seems to be re-filled but only from the recent messages. So if one posts now again and he had posted recently, the nick is resolved at the first message.

My assumption: When a message arrives and the nick is not known, it is retrieved somehow. But this happens after the nick for this first message is already resolved to someone. But it should happen before this.

Could anyone with knowledge about the code verify (and hopefully fix) this?

[micw]

@nathanaelhoun I ping you in this ticket since your PR #1160 seems to be related to this issue.

[nathanaelhoun]

Hi @micw, I tried to fix this issue as you noted.

It seems that if a user posts a second message after he posts a message, the nick is resolved properly. After restart, the nick name cache seems to be re-filled but only from the recent messages. So if one posts now again and he had posted recently, the nick is resolved at the first message.

I arrived to the same conclusion and it's why I made #1160 to reload the user list if their is no nickname. Because this changes did not solve the issue (at least, not everytime), I believe your assumption may be right, from what I remember from the last time I checked the code:

My assumption: When a message arrives and the nick is not known, it is retrieved somehow. But this happens after the nick for this first message is already resolved to someone. But it should happen before this

[micw]

@nathanaelhoun
I'm reviewing

matterbridge/bridge/whatsapp/helpers.go

Lines 75 to 111 in 78f3820

	func (b *Bwhatsapp) getSenderName(senderJid string) string {
	if sender, exists := b.users[senderJid]; exists {
	if sender.Name != "" {
	return sender.Name
	}
	// if user is not in phone contacts
	// it is the most obvious scenario unless you sync your phone contacts with some remote updated source
	// users can change it in their WhatsApp settings -> profile -> click on Avatar
	if sender.Notify != "" {
	return sender.Notify
	}
	if sender.Short != "" {
	return sender.Short
	}
	}
	// try to reload this contact
	_, err := b.conn.Contacts()
	if err != nil {
	b.Log.Errorf("error on update of contacts: %v", err)
	}
	if contact, exists := b.conn.Store.Contacts[senderJid]; exists {
	// Add it to the user map
	b.users[senderJid] = contact
	if contact.Name != "" {
	return contact.Name
	}
	// if user is not in phone contacts
	// same as above
	return contact.Notify
	}
	return ""
	}

To avoid code duplication (which leads to diverge in code as already happened, compare L77-L89 with L102-107) I think the check should be earlier, something like:

sender, exists := b.users[senderJid]
if  !exists {
  // reload, add, set exists=true
}
if  exists {
  // code from L77-L89
}

Besides that, Contacts() looks to me like is jkust sends out the request. The response will probably asynchronous. So message processing should somehow block until the contacts are available. This might be quite complicated because another message might arrive meanwhile which also triggers a contacts reload.

[ghost]

Is there a solution or workaround to this?