feat: add support for DataStax HCD vector store #3728
Conversation
dosubot
bot
added
size:L
github-actions
bot
added
enhancement
|
This pull request is automatically being deployed by Amplify Hosting (learn more). |
bradfordcp
force-pushed
the
feature/datastax-hcd
branch
from
41ae08d
to
1503b0c
Compare
| class HCDVectorStoreComponent(LCVectorStoreComponent): | ||
| display_name: str = "Hyper-Converged Database" | ||
| description: str = "Implementation of Vector Store using Hyper-Converged Database (HCD) with search capabilities" | ||
| documentation: str = "https://python.langchain.com/docs/integrations/vectorstores/astradb" |
There was a problem hiding this comment.
Is there any specific HCD documentation we want to link to?
There was a problem hiding this comment.
Good point, we can link to our install docs. It's worth noting that we are using AstraDB langchain integration so that URL is valid despite it talking to HCD.
|
|
||
|
|
||
| class HCDVectorStoreComponent(LCVectorStoreComponent): | ||
| display_name: str = "Hyper-Converged Database" |
There was a problem hiding this comment.
Can you clarify how we'll use HCD re: naming? Do we want it to be so strongly associated with Datastax that we don't need to specify "Datastax HCD"?
There was a problem hiding this comment.
I asked about this internally and the decision was to keep it as Hyper-Converged Database. Thanks for checking!
| name="api_endpoint", | ||
| display_name="HCD API Endpoint", | ||
| info="API endpoint URL for the HCD service.", | ||
| value="HCD_API_ENDPOINT", |
There was a problem hiding this comment.
Does langchain-astradb support interpreting this env variable as the endpoint (instead of ASTRA_DB_API_ENDPOINT?). I think this naming makes sense either way though, and should be supported in langchain-astradb if not.
There was a problem hiding this comment.
Unfortunately no, langchain-astradb only looks for ASTRA_DB_API_ENDPOINT. There is not associated HCD_API_ENDPOINT, instead we set an environment parameter during client initialization environment=Environment.HCD.
| ), | ||
| HandleInput( | ||
| name="embedding", | ||
| display_name="Embedding or Astra Vectorize", |
There was a problem hiding this comment.
Is vectorize supported in HCD?
There was a problem hiding this comment.
I believe the code paths are present, but I am unclear on how it is configured at this time.
Added support for DataStax Hyper-Converged Database (HCD) backed vector store.