feat: Vertex Check Grounding API integration #186
Conversation
|
|
||
| from google.api_core import exceptions as core_exceptions # type: ignore | ||
| from google.auth.credentials import Credentials # type: ignore | ||
| from google.cloud import discoveryengine_v1alpha # type: ignore |
There was a problem hiding this comment.
it should not be on the top level, otherwise the library can't be used (since it makes an attempt to import on the module level, at __init__.py)
|
/gcbrun |
| }, | ||
| ) | ||
| for doc in documents | ||
| if doc.page_content and doc.metadata |
There was a problem hiding this comment.
why are we ignoring documents with empty metadata?
| ) | ||
|
|
||
|
|
||
| def test_integration_parse( |
There was a problem hiding this comment.
should any additional APIs be enabled on the project where tests are running?
There was a problem hiding this comment.
the only API this integration requires is google-cloud-discoveryengine, which should automatically be installed from langchain-google-community[vertexaisearch]
There was a problem hiding this comment.
I mean, should any GCP APIs be enabled / whitelisted for the GCP project where this test is running at?
There was a problem hiding this comment.
Yes, we would need to whitelist Discovery Engine API if it has not been whitelisted for this to run.
There was a problem hiding this comment.
we need to mark it as extended then
This PR is to integrate the Vertex Standalone API for Check Grounding.
The check grounding API determines how grounded a given piece of text is in a given set of reference texts. The API can generate supporting citations from the reference text to indicate where the given text is supported by the reference texts.
This is implemented as a output parser as Check Grounding is usually done once the generation step of the RAG pipeline is done.