Dart/Flutter package for using Elastic App Search through a simple API returning easy to handle objects

This package is a ready-to-use API for Elastic App Search.

⚠️ You need an active deployment on Elastic to use this service.

This package does not intend to learn you how Elastic App Search is working, it is just intended to help you make queries and manipulate results easily, assuming that you already know Elastic App Search principles.

Here is a simple example of how you can query your Elastic engine:

final service = ElasticAppSearch(
    endPoint: "https://localhost:5601", 
    searchKey: "search-soaewu2ye6uc45dr8mcd54v8",
);

ElasticResponse response = await service
  .engine("national-parks-demo")
  .query("parks")
  .filter("states", isEqualTo: "California")
  .filter("world_heritage_site", isEqualTo: true)
  .resultField("title")
  .resultField("description", snippetSize: 140)
  .page(1, size: 50)
  .get();

for (ElasticResult result in response.results) {
  final data = result.data;
  if (data != null) {
    print("${data["title"]}: ${data["description"]}");
  }
}

This example will query the parks containing parks in any field of the documents. The results will be filtering the parks based in California, with the flag world_heritage_site set to true. Only the title and a snippet of 140 chars of the description of the park will be returned by the query. We request the first page, limited to 50 documents.

Table of contents

• ElasticAppSearch
• ElasticEngine
• ElasticQuery
  • Precision
  • Search filters
  • Search fields
  • Result fields
• ElasticResponse
  • ElasticResponseMeta
    • ElasticResponseMetaPage
  • ElasticResult
    • ElasticResultMeta
    • ElasticResultSnippet

ElasticAppSearch

Before any operation, you need to create an instance of ElasticAppSearch.

Both values are available in your deployment, when going to App Search then Credentials. Warning: use the search-key, not the private-key!

final service = ElasticAppSearch(
    endPoint: "https://localhost:5601", 
    searchKey: "search-soaewu2ye6uc45dr8mcd54v8",
);

ElasticEngine

All the queries must be sent to an engine. To make instantiating an engine easy, we created this syntax:

final engine = service.engine("engine_name");

ElasticQuery

The only required parameter to instanciate a query is a string which is the word you are looking for through your documents.

🔍 See https://www.elastic.co/guide/en/app-search/current/search.html

final query = engine.query("query");

Precision

This setting is available through the query modifier .precision

The value of the precision parameter must be an integer between 1 and 11, inclusive. The range of values represents a sliding scale that manages the inherent tradeoff between precision and recall. Lower values favor recall, while higher values favor precision.

🔍 See https://www.elastic.co/guide/en/app-search/current/search-api-precision.html

final query = query.precision(5);

Search filters

This setting is available through the query modifier .filter

This feature intends to filter documents that contain a specific field value. It's only available on text, number, and date fields.

🔍 See https://www.elastic.co/guide/en/app-search/current/filters.html

Important: As for now, this object only handles "all" filters, which means that all the filters added to the query will be handled as a "AND" query. The other filters available, "or" and "not", will be added in a future release of the package.

Warning: You cannot use isEqualTo and whereIn on the same field at the same time, otherwise it will raise an exception.

final query = query.filter("field", isEqualTo: "value");

final query = query.filter("field", isEqualTo: true);

final query = query.filter("field", whereIn: ["value1", "value2"]);

Search fields

This setting is available through the query modifier .searchField

It will restrict a query to search only specific fields. Restricting fields will result in faster queries, especially for schemas with many text fields Only available within text fields.

Weight is given between 10 (most relevant) to 1 (least relevant).

🔍 See https://www.elastic.co/guide/en/app-search/current/search-fields-weights.html

final query = query
  .searchField("field1", weight: 8)
  .searchField("field2", weight: 3);

Result fields

This setting is available through the query modifier .resultField

The fields which appear in search results and how their values are rendered.

Raw is an exact representation of the value within a field. Snippet is a representation of the value within a field, where query matches are returned in a specific field and other parts are splitted, in order to user [RichText] to display the results and highlight the query matches.

🔍 See https://www.elastic.co/guide/en/app-search/current/result-fields-highlights.html

final query = query
  .resultField("field1", rawSize: 80)
  .resultField("field2", snippetSize: 80)
  .resultField("field3", rawSize: 80, snippetSize: 80, fallback: true);

ElasticResponse

The response object contains two parts, the meta in a ElasticResponseMeta object and the results in a list of ElasticResult objects.

🔍 See https://www.elastic.co/guide/en/app-search/current/search.html#search-api-response-body

ElasticResponseMeta

An object containing information about the results, especially the pagination details in page.

ElasticResponseMetaPage

Object delimiting the results meta data.

ElasticResult

An object presenting a result to the query.

The data param is a map of the fields requested with the .searchResult modfifier. If this modifier was omitted, all the fields of the document are returned.

The snippets is a map of the snippets returned by the query, if requested with the .searchResult modfifier. By default, Elastic returns the snippets as HTML snippets. This package manipulates the result to return some stuff that can be handled by Flutter: an ElasticResultSnippet object.

ElasticResultMeta

An object containing information about a given result.

ElasticResultSnippet

An object contaning the snippet of the result. If you don't want to display your search results with highlights on the matching keyword, just ignore textParts and highlights and use only fullText.

textParts and highlights can be used to build a RichText widget in Flutter, please look at the example of the package to have a concrete example.

Third party packages

This app uses some external librairies:

• Freezed to generate immutable models
• Dio for http requests
• html_unescape to strip html tags from results

Credits

This package was originally created for my personnal needs but feel free to use it, it does not covers all the features available in Elastic App Search, but I will try to cover all the features over time.

I am not related to Elastic in any way, I am just a developer who needed to use this API and created this library to do so.

If you have questions, feel free to ask on Twitter.

LICENSE: MIT