AzureAISearchService

Kind: Service

Source: atloria-monorepo/apps/api/src/web-search/services/azure-ai-search.service.ts

Azure AI Search Service Queries the indexed competitor documentation

AzureAISearchService is a NestJS backend service that queries the Azure AI Search index containing competitor documentation. It provides search, category-filtered search, document lookup by URL, and health-check capabilities for web-search consumers.

Methods

Where it refuses work

• AzureAISearchService stops the work with an early return when !this.searchClient, in 2 places.

When something fails

• AzureAISearchService handles failure in 3 places: it turns it into a return value in all 3.

Diagram

mermaid
sequenceDiagram
    participant Client
    participant Consumer as Controller/Service Consumer
    participant Search as AzureAISearchService
    participant Azure as Azure AI Search Index

    Client->>Consumer: Request competitor documentation
    Consumer->>Search: search(query)
    Search->>Azure: Execute indexed document query
    Azure-->>Search: Matching search results
    Search-->>Consumer: CompetitorSearchResult
    Consumer-->>Client: Search response

    Note over Consumer,Search: Other supported operations:<br/>searchByCategory(), getDocumentByUrl(), healthCheck()

Usage

ts
import { Injectable } from '@nestjs/common';
import { AzureAISearchService } from './azure-ai-search.service';

@Injectable()
export class CompetitorResearchService {
  constructor(
    private readonly azureAISearchService: AzureAISearchService,
  ) {}

  async findAuthenticationDocs() {
    const results = await this.azureAISearchService.search('authentication');

    return results;
  }

  async findDocument(url: string) {
    const document = await this.azureAISearchService.getDocumentByUrl(url);

    if (!document) {
      return null;
    }

    return document;
  }

  async isSearchAvailable(): Promise<boolean> {
    return this.azureAISearchService.healthCheck();
  }
}

AI Coding Instructions

• Inject AzureAISearchService through NestJS dependency injection; do not instantiate it directly.
• Use search() for general competitor-documentation queries and searchByCategory() when the caller already has a category constraint.
• Handle getDocumentByUrl() returning null when the indexed URL cannot be found.
• Use healthCheck() for readiness or dependency monitoring rather than issuing a full search request.
• Treat returned search results as indexed data that may be stale or incomplete; validate URLs and metadata before using them in downstream workflows.