Updating indexes in Pincone to reflect an updated source

NElliott · January 30, 2024, 3:48pm

I have a chatbot hosted on AWS that makes use of PineCone and relies on the Langchain library.
I have many PDFs in an AWS S3 bucket which the vector store is based on.
I see an upsert facility for PineCone but that seems like it’s used to redefine the specifics of an index rather than reflecting that the content that relates to an index has changed.
If a PDF is updated, how can I update the vector store to reflect the new version of the document.
Appreciate any comments / guidance or pointing me to relevant articles.
Thanks,
Neil

Topic		Replies	Views
RAG - Can't upsert embeddings - Prepare the Embeddings and Upsert to Pinecone Building Applications with Vector Databases week-1	1	146	March 3, 2024
Pinecone - Lesson 1 - Semantic Search: Fails to Create Embeddings and Upsert to Pinecone Building Applications with Vector Databases week-1	1	146	March 3, 2024
Langchain - How to get the "Source Documents" information as output? AI Discussions ai-discussions , langchain	3	169	July 16, 2023
Lesson 3 - Recommender Systems - can't embed vectors into Pinecone Building Applications with Vector Databases	0	141	March 3, 2024
Help needed with vector database AI Discussions ai-discussions	2	119	February 8, 2024

Updating indexes in Pincone to reflect an updated source

Related topics