fix: Introduced new index_document() to fix chunking related issue

[chandrasekharan-zipstack]

What

• Deprecated old function get_vector_store_index_from_storage_context() to index and created a new one index_document() in vector_db.py
• Moved attributes within new function to simplify logic at caller side
• Renamed deprecated SimpleNodeParser to the intended SentenceSplitter
• Bumped SDK to 0.48.0

Why

• The llama index function VectorStoreIndex.from_documents() expects an attribute transformations which was not set correctly. As a result the user configured values for chunk size and chunk overlap was never set and used

Notes on Testing

• Attempted indexing locally

Screenshots

1. With 500 and 50 chunk size / overlap
   
2. With 2048 and 256 chunk size / overlap
   

Checklist

I have read and understood the Contribution Guidelines.

[gaya3-zipstack]

Looks good