UmbracoAzureSearch allows you to completely replace Examine from your Umbraco website and replace it with Azure Search.

The first thing required to get your UmbracoAzureSearch service up and running will be to create an Azure search service.

Steps to setting up an Azure Search instance

Inside the ~/config/AzureSearch.config you will want to populate the SearchServiceName, the SearchServiceAdminApiKey (which you can find in your Azure Search instance, under Settings > Keys), and lastly the IndexName of which you have named your search index.

{
    "SearchServiceName" : "my-umbraco-search",
    "SearchServiceAdminApiKey" : "XXXXXXXXXXXXXXXXXXXXXX",
    "IndexName": "MyUmbracoIndex"
}

The AzureSearchClient provides a fluent interface to allow you to build and execute your Azure Search queries against your newly defined index.

public class SearchController : RenderMvcController 
{
    private readonly IAzureSearchClient _search;
    
    public SearchController()
    {
        this._search = new AzureSearchClient(HttpContext.Current.Server.MapPath("/"));
    }
}

public ISearchResult GetContent(string searchTerm)
{
    return this._search.Content().Term(searchTerm).Results();
}

Media will also return an IAzureSearchClient so will allow you to use any of the following fluent methods for filtering.

public ISearchResult GetMedia()
{
    return this_.search.Media().Results();
}

Filter by a single DocumentType

public ISearchResult GetContent(string searchTerm)
{
    return this._search.Content().Term(searchTerm).DocumentType("newsArticle").Results();
}

Filter by multiple DocumentTypes

public ISearchResult GetContent(string searchTerm)
{
    return this._search.Content().Term(searchTerm).DocumentTypes(new { "newsArticle", "contentPage" }).Results();
}

Order by a property alias

public ISearchResult GetContent(string searchTerm)
{
    return this._search.Content().Term(searchTerm).OrderBy("publicationDate").Results();
}

Paginate results

public ISearchResult GetContent(string searchTerm, int page)
{
    return this._search.Content().Term(searchTerm).PageSize(10).Results(page);
}

Define the base node from which all of your search queries originate.

public ISearchResult GetContent(string searchTerm)
{
    return this._search.Content().Term(searchTerm).SearchRoot(1).Results();
}

Filtering allows overloads for string,int and bool to filter on a property value. Properties must include the IsFilterable: True flag within the index configuration.

Config:

{
    "Name":"published",
    "Type":"bool",
    "IsFilterable":"True"
},

Code:

public ISearchResult GetContent(string searchTerm)
{
    return this._search.Content().Term(searchTerm).Filter("published", true).Results();
}

Facets are configured from within the index configuration by assigning the IsFacetable: True flag on a property.

Config:

{
    "Name":"year",
    "Type":"string",
    "IsFacetable":"True"
},

Code:

public ISearchResult GetFacets()
{
    return this._search.Content().Facet("year").Results();
}

In order to search on a custom property, you must ensure that the IsSearchable: True flag is set on the property. This will allow us to use the Contains() method.

Config:

{
    "Name":"title",
    "Type":"string",
    "IsSearchable":"True"
},

Code:

public ISearchResult GetContent(string searchTerm)
{
    return this._search.Content().Contains("title", searchTerm).Results();
}

In order to enable autocomplete functionality in Azure Search, you must first add a suggester to your index. The standard suggester will allow you to chose which of your indexed fields the autocomplete search will be based on.

"Suggesters": [
    {
        "Name": "sg",
        "SearchFields": [
            "Name",
            "title"
        ]
    }
]

After this is defined and your index is reconstructed, we can start autocomplete based on search terms.

public IEnumerable<string> Autocomplete(string searchTerm)
{
    IList<SuggestResult> result = this._searchService.Suggest(searchTerm, 10);
    return result.Select(x => x.Text).Distinct().ToList();
}

You can use Azure Search with external data in a similar way to http://sleslie.me/2016/indexing-external-data-using-examineumbraco/.

Add separate config files for each index, that use the following pattern AzureSearch.{{simpledata}}.config, where simpledata is replaced with the name of the index.

In the config file itself add the extra property, DataService and point this to the Class and Assembly name of your data service. There's an example in the Umbraco project of this repo.

As with examine your data service will implement an interface, IAzureSearchSimpleDataService, using the GettBatchData method to fill instances of AzureSearchSimpleDataSet.

Because Azure Search works better with batches the interface also defines the method GetAllIds(), this is used to return a list of primary key id's for batching in the proevios method call.

Once the data service is added into the project and the config, log into Umbraco. In the Moroyama Azure Search config in there should be additional tabs for each added index. From here you can manage each separately.

There is an example of using the index within your application in the Moriyama.AzureSearch.Umbraco.ExampleSimpleData. It should all look very similar to working with Examine.

For the search code, again the implementation is very similar to Examine. Apart from the filter and the search fields, these are added via the SearchParameters rather than an in text query. The actual query body works in a very similar way and should make moving over from Examine very easy.