orchardcore-custom-indexing-azureai

Orchard Core Custom Indexing for Azure AI Search - Prompt Templates

Create a custom Azure AI Search index for arbitrary data

You are an Orchard Core expert. Generate code and configuration for indexing arbitrary records into Azure AI Search using Orchard Core index profiles, document handlers, and provider-specific mappings.

When to use this skill

Use this skill when Orchard content-item indexing is not enough and you need an index for custom records such as:

• generated AI artifacts
• domain records stored in a custom catalog
• user-scoped memory or preference records
• imported external business data
• custom vector-search sources

Architecture to follow

Use the CrestApps AI Memory modules as the reference architecture:

• CrestApps.OrchardCore.AI.Memory contains the shared record/indexing logic and the master index setting.
• CrestApps.OrchardCore.AI.Memory.AzureAI registers the Azure AI Search indexing source plus provider-specific handlers.
• CrestApps.OrchardCore.AI.Memory.Elasticsearch proves the same neutral-document approach can target another provider with different mappings.

Master index pattern

For arbitrary data, create a single logical index profile type for that record family and let the tenant choose the active master index profile by name.

Then:

1. persist the source record in your own store
2. build a neutral index document model from it
3. use IDocumentIndexHandler to map that neutral document into Azure AI Search fields
4. write through the keyed IDocumentIndexManager
5. trigger add/update/delete from the record lifecycle, not a single UI path

Key Orchard Core pieces

• IIndexProfileStore
• IndexProfileHandlerBase
• IDocumentIndexHandler
• keyed IDocumentIndexManager
• services.AddAzureAISearchIndexingSource(type, options => ...)

Recommended implementation steps

1. Define the custom index profile type

public static class CustomerInsightsConstants
{
    public const string IndexingTaskType = "CustomerInsights";
}