Open davidADSP opened 3 months ago
When using the Vertex AI RAG service here:
https://github.com/googleapis/python-aiplatform/blob/b5c3cdd737acd695301c9a564d8f91371288f9f1/vertexai/preview/rag/rag_data.py
, it says that you can use this with Google Drive or Google Cloud Storage links
import vertexai vertexai.init(project="my-project") # Google Drive example paths = ["https://drive.google.com/file/123", "https://drive.google.com/file/456"] # Google Cloud Storage example paths = ["gs://my_bucket/my_files_dir", ...] response = vertexai.preview.rag.import_files( corpus_name="projects/my-project/locations/us-central1/ragCorpora/my-corpus-1", paths=paths, chunk_size=512, chunk_overlap=100, )
I have found that this doesn't work with Google Shared Drive links, even if the service account is added as per the documentation.
Is this intentional? If so, it would be useful to add this caveat to the documentation here:
https://cloud.google.com/vertex-ai/generative-ai/docs/llamaindex-on-vertexai#supported-doc-types
Could you share the error message? You may strip of any sensitive info. Thanks!
When using the Vertex AI RAG service here:
https://github.com/googleapis/python-aiplatform/blob/b5c3cdd737acd695301c9a564d8f91371288f9f1/vertexai/preview/rag/rag_data.py
, it says that you can use this with Google Drive or Google Cloud Storage links
I have found that this doesn't work with Google Shared Drive links, even if the service account is added as per the documentation.
Is this intentional? If so, it would be useful to add this caveat to the documentation here:
https://cloud.google.com/vertex-ai/generative-ai/docs/llamaindex-on-vertexai#supported-doc-types