Hello, is there a way to define the chunk size for the VectorIndex during its creation? In the past, it could be set via the ServiceContext but now VectorStoreIndex.from_documents(documents, embed_model=my_embd_model) ignores any chunk_size argument.