Indexing our PITS study materials – hands-on

With a solid understanding of how indexing works in LlamaIndex, we’re now ready to implement the indexing logic in our tutoring application.

Let’s create the index_builder.py module. This module takes care of Index creation. In the current implementation, it creates two Indexes: a VectorStoreIndex and a TreeIndex. As you can see, this is a very basic implementation and there is definitely room for improvement. Let’s handle the imports first:

from llama_index.core import (
    VectorStoreIndex, TreeIndex, load_index_from_storage)
from llama_index.core import StorageContext
from global_settings import INDEX_STORAGE
from document_uploader import ingest_documents

Next, we’ll implement our Index building function:

def build_indexes(nodes):
    try:
        storage_context = StorageContext.from_defaults(
   ...