orchardcore-custom-indexing-azureai

Orchard Core Custom Indexing for Azure AI Search

Index arbitrary records into Azure AI Search

You are an Orchard Core expert. Generate code that indexes custom records (not content items) into an Azure AI Search index profile, optionally with vectors. Model it on CrestApps.OrchardCore.AI.Memory (records, settings, indexing service) and CrestApps.OrchardCore.AI.Memory.AzureAI (provider registration and mappings).

When to use it

• Records you store yourself (CRM rows, generated summaries, user-scoped facts, imported data) that need full-text, filtered, or vector search.
• For Orchard content items use the built-in content index (see orchardcore-azure-ai-search). For RAG over an index you already have, an AI data source is usually enough (see orchardcore-ai-data-sources).

Architecture

The shared module (records, neutral document, master-index setting, indexing service, lifecycle handler, re-sync handler) is identical to the Elasticsearch variant; build it as shown in orchardcore-custom-indexing-elasticsearch. This skill covers the Azure AI Search provider module and what differs.

1. Register the provider

public sealed class Startup : StartupBase
{
    internal readonly IStringLocalizer S;

    public Startup(IStringLocalizer<Startup> stringLocalizer)
    {
        S = stringLocalizer;
    }

    public override void ConfigureServices(IServiceCollection services)
    {
        services.AddIndexProfileHandler<CustomerInsightAzureAISearchIndexProfileHandler>();
        services.AddScoped<IDocumentIndexHandler, CustomerInsightAzureAISearchDocumentIndexHandler>();

        services.AddAzureAISearchIndexingSource(CustomerInsightsConstants.IndexingTaskType, o =>
        {
            o.DisplayName = S["Customer insights (Azure AI Search)"];
            o.Description = S["Create an Azure AI Search index for customer insight records."];
        });
    }
}