Zep Authentication

[danielchalef]

Description: Add Zep API Key argument to ZepChatMessageHistory and ZepRetriever

• correct docs site links
• add zep api_key auth to constructors

ZepChatMessageHistory: @hwchase17,
ZepRetriever: @rlancemartin, @eyurtsev

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

1 Ignored Deployment

Name	Status	Preview	Comments	Updated (UTC)
langchain	⬜️ Ignored (Inspect)			Jun 30, 2023 9:17pm

[rlancemartin]

Thanks! Maybe you can also re-run the notebook to confirm all is working following the update.

[danielchalef]

Thanks! Maybe you can also [re-run the notebook]

@rlancemartin Good idea. Done!

[danielchalef]

@rlancemartin bump on this. Would love to get this merged as we have users interested in using authentication. Thanks!

[rlancemartin]

@danielchalef looks good! Can you rebase or merge master, resolve conflicts in pyproject.toml and then run poetry lock --no-update to re-generate poetry.lock? I can help if needed.

[danielchalef]

@rlancemartin Done. Waiting on CI tests to pass.

[rlancemartin]

Thanks! Hmm, still looks like a poetry.lock conflict?

[danielchalef]

Thanks! Hmm, still looks like a poetry.lock conflict?

Weird. Looks like this happened over night. There wasn't a conflict yesterday. I've tried rebasing again but am up to date. Will delete and recreate the file.

[rlancemartin]

Thanks! Hmm, still looks like a poetry.lock conflict?

Weird. Looks like this happened over night. There wasn't a conflict yesterday. I've tried rebasing again but am up to date. Will delete and recreate the file.

Thanks, ya sorry about that. LMK if issues. I can also try on my end.

Usually just:

checkout master poetry.lock
poetry lock --no-update

Should do it.

[danielchalef]

@rlancemartin done. no merge issues. tests are currently running.

[danielchalef]

Ugh. Fixing this. Two minutes

[rlancemartin]

Ha, ya. A lot of trouble w/ the pyproject.toml :P ... looks like all other tests are passing.

[danielchalef]

Yes. I keep running poetry lock --no-update and it doesn't produce a different output for the pytproject file I've checked in. The poetry.lock file has a different hash to what is in master. And that's the Only difference.

Would you mind taking it from here?

[rlancemartin]

Yes. I keep running poetry lock --no-update and it doesn't produce a different output for the pytproject file I've checked in. The poetry.lock file has a different hash to what is in master. And that's the Only difference.

Would you mind taking it from here?

Looks like I can't push to your branch. But I checked out your branch and did this:

1/ sync for your fork w/ master
2/ on master, update with git pull
3/ on your branch (zep-jwt-auth): git checkout master poetry.lock
4/ then poetry lock --no-update

This should resolve it.

Maybe your master is a bit out of date?

[rlancemartin]

1/ sync for your fork w/ master 2/ on master, update with git pull 3/ on your branch (zep-jwt-auth): git checkout master poetry.lock 4/ then poetry lock --no-update

I see the same conflict on another PR after doing the above.

Just pinged @dev2049 on it.

[rlancemartin]

1/ sync for your fork w/ master 2/ on master, update with git pull 3/ on your branch (zep-jwt-auth): git checkout master poetry.lock 4/ then poetry lock --no-update

I see the same conflict on another PR after doing the above.

Just pinged @dev2049 on it.

It was resolved by ensuring local master is up-to-date, as mentioned above.

Can you confirm that and then do the steps above?

In parallel I am getting #6772 ready to merge (to ensure no other issues).

I'd prefer to get this in first.

[danielchalef]

Doing so now. I'm a little confused why rebasing against upstream/master wasn't resolving this.

[rlancemartin]

Doing so now. I'm a little confused why rebasing against upstream/master wasn't resolving this.

It should. #6772 is now fixed, too.

I think it's possible that your local master was out of date.

Looks like the conflict is resolved!

Let's get this in once tests pass.

Sorry for the trouble.

[danielchalef]

No worries. I think that was the case. Thanks