Python: add diskann index type

[SelinaTang21]

Motivation and Context

Add a new vector index type diskann to Azure Cosmos DB Mongo vCore vector store.
Issue #9676

Why is this change required? This change enables MongoDB vCore to support the new diskann vector index as a preview feature for vector search. The diskann index, as described in the DiskANN paper, offers an additional option for fast nearest neighbor searches.

Description

Contribution Checklist

• The code builds clean without any errors or warnings
• The PR follows the SK Contribution Guidelines and the pre-submission formatting script raises no violations
• All unit tests pass, and I have added new tests where possible
• I didn't break anyone 😄

[markwallace-microsoft]

Python Unit Test Overview

Tests	Skipped	Failures	Errors	Time
2869	4 💤	0 ❌	0 🔥	1m 15s ⏱️

[eavanvalkenburg]

Hi @SelinaTang21 thanks for this, however we are no longer making updates to the existing memory stores, we are moving to the new model with a VectorStore and VectorStoreRecordCollection, this has not been done yet for MongoDB version of Cosmos, the NoSQL version was just added to main this week, we would love contributions to the new model and we already have DISK_ANN in there (see data/const.py), so for that one this can be adopted directly. We still need to create samples showing how to go from the old to the new stores but in essence that is quite easy since with the new model you can just create your own data model, including one that matches the old one, off course that does depend on the new implementation of the Mongo DB store and collection! Implementing the new one shouldn't be too difficult, it takes only a couple of functions to implement but it is currently not at the top of our list, so have a look!