Upgrading RAG Pipeline

We are at a point where we need to upgrade our RAG Pipeline and I am uptaking Managing ImportsUnderstanding imports - Pinecone Docs , a couple of quick questions.

  1. Does this assume conversion of any form of file to parquet , pdf, text,csv,json → parquet
  2. What are mechanisms to handle incremental doc updates ? a new file is dropped in a bucket
  3. Is the chunking/embedding handled as a config at the index level ?
    As I am writing an agentic workflow + airflow to continuously process docs to the vector database , this is a crucial step to continuously update our namespaces.