Zep cloud setup #1760
Zep cloud setup #1760
Conversation
| @@ -22,6 +22,7 @@ | |||
| "@datastax/astra-db-ts": "^0.1.2", | |||
| "@dqbd/tiktoken": "^1.0.7", | |||
| "@elastic/elasticsearch": "^8.9.0", | |||
| "@getzep/zep-cloud": "npm:@getzep/zep-js@next", | |||
There was a problem hiding this comment.
Note: We are aliasing a different version of our ts sdk (tagged with next) as we currently only have support for Zep Cloud API in Release Candidate of our ts sdk
| @@ -0,0 +1,181 @@ | |||
| import { IMessage, INode, INodeData, INodeParams, MemoryMethods, MessageType } from '../../../src/Interface' | |||
There was a problem hiding this comment.
instead of creating a new node, what do u think if we just add a new options field to the ZepMemory that allow users to select OpenSource or Cloud, then according to selection, use the respective library for init
There was a problem hiding this comment.
I think this would be doable, but I have some reservations.
Our interfaces are not the same for open source and cloud sdk. For memory, for example, we don't offer the ability to specify memoryType on Open Source version, whereas this is one of the new things we have introduced with Zep Cloud. The same applies to vector stores, in cloud fro example we don't accept embeddings and embedding dimensions any more.
There was a problem hiding this comment.
So we would have to have a lot of additional if blocks in the nodes.
There was a problem hiding this comment.
Oh okay i thought the difference is only the initialization. In that case, make sense to have separate nodes
There was a problem hiding this comment.
Cool! Marked it as ready for review then!
|
@paul-paliychuk some linting issues if you can do |
|
@HenryHengZJ My bad, just ran the linter-fixer |
|
|
||
| constructor() { | ||
| this.label = 'Zep Collection - Cloud' | ||
| this.name = 'Zep Collection - Cloud' |
There was a problem hiding this comment.
the name must be in camelcase zepCloud, otherwise the upsert button will not be shown:
No description provided.