fix: preserve falsy metadata values in Azure AI Search vector store

[voidborne-d]

Problem

Falsy metadata values (0, "", [], False) are silently dropped when building the Azure AI Search index document.

The current truthiness check on line 899:

if metadata_value:

skips any value that evaluates to False in a boolean context, which includes valid values like integer 0, empty strings, and empty lists.

Fix

Replace the truthiness check with an explicit None check:

if metadata_value is not None:

This ensures only truly absent metadata keys are skipped, while all valid falsy values are preserved in the index.

Reproducing

1. Create an Azure AI Search index with an integer metadata field
2. Add a node with that metadata field set to 0
3. Retrieve the node — the value comes back as None instead of 0

Fixes #21385