Search and Retrieval

Overview

The Search flow creates and maintains a searchable index over the structured data stored in Azure SQL. Its primary purpose is to make it easy for operators to search across refugee documents and family members without querying the database directly.

This flow is currently triggered manually, though it may be automated in a future iteration.

Steps

1. Trigger — Manual

The indexing process is started manually. When triggered, it reads the current state of Azure SQL and synchronizes it with the Azure AI Search index.

2. Index — Azure AI Search

The data from Azure SQL is used to build a shadow clone — an indexed copy of the records optimized for search. Azure AI Search handles full-text search, fuzzy matching, and filtering across all indexed fields.

3. Search — Document & Member Lookup

Once the index is up to date, users can search through:

• Refugee family members
• Document records (Master Cards, Index Cards, Red Cross Cards)

The search is designed to surface relevant results quickly, even with partial or approximate input.

Flow Summary

Manual Trigger
      ↓
Read structured data from Azure SQL
      ↓
Build / refresh index in Azure AI Search
      ↓
Index available for document & member search