Features

This page provides an overview of the features and capabilities provided by the Algolia integration.

Core functionality

The Algolia integration implements the framework's Search.Service interface, providing two main search methods that work with Algolia indexes.

Search methods

Generic search: search<T>()

Performs a generic search across any type of data stored in Algolia indexes. Returns raw search results from Algolia.

Method signature:

search<T>(indexName: string, payload: Search.Model.SearchPayload): Observable<Search.Model.SearchResult<T>>

What it does:

• Executes a search query against the specified Algolia index
• Transforms the SearchPayload into Algolia query parameters
• Returns results in the framework's SearchResult<T> format
• Handles errors gracefully (returns empty results for 404 errors)

Use cases:

• Searching any custom data types stored in Algolia
• Generic content search across different index types
• When you need raw Algolia results without transformation

Article search: searchArticles()

Performs a specialized search for articles with automatic mapping to the framework's article model.

Method signature:

searchArticles(indexName: string, payload: Search.Model.SearchPayload): Observable<Articles.Model.Articles>

What it does:

• Uses search<SearchEngineArticleModel>() internally
• Automatically maps SearchEngineArticleModel results to Articles.Model.Articles
• Returns articles in the framework's standard article format

Use cases:

• Knowledge base article search
• Documentation search
• Content search that needs to match the framework's article model

Supported query features

The integration supports the following query parameters from SearchPayload:

Text search (query)

Full-text search across indexed content. The query string is passed directly to Algolia's search API.

{
    query: 'getting started';
}

Locale filtering (locale)

Filters results by locale using Algolia facet filters. The locale is converted to locale:{locale} format.

{
    locale: 'en';
}

Exact matching (exact)

Filters by exact field values using Algolia facet filters. Supports:

• Single values: { category: "guides" }
• Array values: { tags: ["javascript", "typescript"] } (creates multiple filters)

{
  exact: {
    category: "guides",
    status: "published",
    tags: ["javascript", "typescript"]
  }
}

Range filtering (range)

Filters numeric or date fields by range using Algolia numeric filters. Supports min and max values.

{
  range: {
    price: {
      min: 10,
      max: 100
    },
    publishedAt: {
      min: "2024-01-01"
    }
  }
}

Pagination (pagination)

Controls result pagination. The integration converts offset to Algolia's page parameter:

• limit: Number of results per page
• offset: Starting position (converted to page number)

Conversion formula: page = Math.floor(offset / limit) Default limit: 20 (if not specified)

{
  pagination: {
    limit: 10,
    offset: 40  // Will be converted to page 4
  }
}

Sorting (sort)

Sorts results by field. Important: Only the first sort configuration is used, and it modifies the index name.

Index name modification:

• Original: articles
• With sort: { field: "publishedAt", order: "desc" }
• Modified: articles_publishedAt_desc

This means you need separate Algolia indexes for each sort configuration.

{
    sort: [
        {
            field: 'publishedAt',
            order: 'desc',
        },
    ];
}

Automatic features

Facets

The integration automatically adds facets: ['*'] to all search queries. This enables facet filtering capabilities in Algolia, allowing you to use the exact filter on any facetable attribute.

Error handling

• Missing environment variables: Throws error during service initialization
• Missing index name: Throws error if index name is not provided
• Index not found (404): Returns empty result set ({ hits: [], total: 0 }) gracefully
• Other API errors: Logged and re-thrown for framework error handling

Logging

The integration logs at debug level:

• Original search payload
• Transformed Algolia query

Errors are logged at error level with full details.

Article data model

When using searchArticles(), the integration expects articles in Algolia to follow the SearchEngineArticleModel structure:

type SearchEngineArticleModel = {
    id: string;
    documentId: string;
    slug: string;
    locale?: string;
    createdAt: string;
    updatedAt: string;
    publishedAt: string;
    SEO: {
        title: string;
        noIndex: boolean;
        noFollow: boolean;
        description: string;
        keywords?: Array<{ keyword: string }>;
        image?: {
            url: string;
            alternativeText?: string;
            width?: number;
            height?: number;
            name: string;
        };
    };
};

Mapping to Articles.Model.Article

The searchArticles() method maps fields as follows:

Article Model Field	Source Field	Notes
id	documentId	Uses documentId as the article ID
slug	slug	Direct mapping
title	SEO.title	From SEO object
lead	SEO.description	From SEO object
createdAt	updatedAt	Uses updatedAt for both
updatedAt	updatedAt	Uses updatedAt for both
permissions	-	Initialized as empty array
tags	-	Initialized as empty array
sections	-	Initialized as empty array

Note: The mapper uses updatedAt for both createdAt and updatedAt fields, and initializes permissions, tags, and sections as empty arrays.

Limitations

The following SearchPayload fields are not implemented in the Algolia integration:

• exists - Field existence filtering (not converted to Algolia query)
• notExists - Field non-existence filtering (not converted to Algolia query)
• filter - Custom filters (not converted to Algolia query)

These fields are part of the SearchPayload model but are not processed by the buildQuery() method.

Response structure

All search methods return results in the SearchResult<T> format:

{
  hits: T[];              // Array of search results
  total: number;          // Total number of matching results
  page?: number;          // Current page (if pagination used)
  nbPages?: number;      // Total number of pages (if pagination used)
  processingTimeMS?: number; // Query processing time
}