How to Define a Custom Analyzer and Run an Atlas Search Diacritic-Insensitive Query

In Atlas, go to the Clusters page for your project.

1. If it is not already displayed, select the organization that contains your desired project from the Organizations menu in the navigation bar.

2. If it is not already displayed, select your desired project from the Projects menu in the navigation bar.

3. If the Clusters page is not already displayed, click Database in the sidebar.

Go to the Atlas Search page for your cluster.

You can go the Atlas Search page from the sidebar, the Data Explorer, or your cluster details page.

Go to the Search Tester.

Click the Query button to the right of the index to query.

View and edit the query syntax.

Click Edit Query to view a default query syntax sample in JSON format.

Run an Atlas Search diacritic-insensitive query.

This query uses the $search stage to query the collection using the compound operator. The compound operator uses the following clauses:

• must clause to search for movie titles that begin with the term alle using the wildcard operator

• should clause to specify preference for the Drama genre using the text operator

Copy and paste the following query into the Query Editor, and then click the Search button in the Query Editor.

1
[
2
  {
3
    "$search" : {
4
      "index": "diacritic-insensitive-tutorial",
5
      "compound" : {
6
        "must": [{
7
            "wildcard" : {
8
              "query" : "alle*",
9
              "path": "title",
10
              "allowAnalyzedField": true
11
        }
12
        }],
13
        "should": [{
14
          "text": {
15
            "query" : "Drama",
16
            "path" : "genres"
17
          }
18
        }]
19
      }
20
    }
21
  }
22
]

SCORE: 1.2084882259368896  _id:  "573a13a1f29313caabd07bb6"
  plot: "A group of hip retro teenage outsiders become involved in an interscho…"
  genres:
    0: "Drama"
    1: "Family"
    2: "Sport"
  runtime: 103
  title: "Alley Cats Strike"
SCORE: 1.179288625717163  _id:  "573a13b1f29313caabd382a2"
  plot: "Famous pianist Zetterstrèm returns home to his native Denmark, to give…"
  genres:
    0: "Drama"
    1: "Romance"
    2: "Sci-Fi"
  runtime: 88
  title: "Allegro"
SCORE: 1  _id:  "573a1397f29313caabce5f15"
  plot: "An enthusiastic filmmaker thinks he's come up with a totally original …"
  genres:
    0: "Animation"
    1: "Comedy"
    2: "Fantasy"
  runtime: 75
  title: "Allegro non troppo"
SCORE: 1  _id:  "573a13d1f29313caabd8f84b"
  plot: "The eleven year old cycling talent Freddy is the son of a butcher in a…"
  genres:
    0: "Comedy"
  runtime: 100
  title: "Allez, Eddy!"

Expand your query results.

The Search Tester might not display all the fields in the documents it returns. To view all the fields, including the field that you specify in the query path, expand the document in the results.

The first document in the result includes diacritics in the title field because the diacriticFolder custom analyzer we used on the title field applied character folding on its values. Atlas Search returns documents with titles that begin with the query term alle because we used the keyword tokenizer, which tokenizes entire strings (or phrases) as a single token.

Alternatively, you can specify the standard tokenizer instead of the keyword tokenizer in the custom analyzer used on the title field. For the standard tokenizer, the Atlas Search results would contain documents with titles that begin or appear anywhere at the beginning of the word for the query term alle such as "Desde allè". To test this, edit your index definition to replace the keyword tokenizer on line 17 with standard tokenizer, save the index definition, and run the sample query.

Connect to your cluster in mongosh.

Open mongosh in a terminal window and connect to your cluster. For detailed instructions on connecting, see Connect via mongosh.

Use the sample_mflix database.

Run the following command at mongosh prompt:

use sample_mflix

Run an Atlas Search diacritic-insensitive query.

This query uses the $search stage to query the collection using the compound operator. The compound operator uses the following clauses:

• must clause to search for movie titles that begin with the term alle using the wildcard operator

• should clause to specify preference for the Drama genre using the text operator

The query uses the $project stage to:

• Exclude all fields except title and genres

• Add a field named score

1
db.movies.aggregate([
2
  {
3
    "$search" : {
4
      "index": "diacritic-insensitive-tutorial",
5
      "compound" : {
6
        "must": [{
7
            "wildcard" : {
8
              "query" : "alle*",
9
              "path": "title",
10
              "allowAnalyzedField": true
11
        }
12
        }],
13
        "should": [{
14
          "text": {
15
            "query" : "Drama",
16
            "path" : "genres"
17
          }
18
        }]
19
      }
20
    }
21
  },
22
  {
23
    "$project" : {
24
      "_id" : 0,
25
      "title" : 1,
26
      "genres" : 1,
27
      "score" : { "$meta": "searchScore" }
28
    }
29
  }
30
])

{
  genres: [ 'Drama', 'Family', 'Sport' ],
  title: 'Alley Cats Strike',
  score: 1.2084882259368896
},
{
  genres: [ 'Drama', 'Romance', 'Sci-Fi' ],
  title: 'Allegro',
  score: 1.179288625717163
},
{
  genres: [ 'Animation', 'Comedy', 'Fantasy' ],
  title: 'Allegro non troppo',
  score: 1
},
{
  genres: [ 'Comedy' ],
  title: 'Allez, Eddy!',
  score: 1
}

The first document in the result includes diacritics in the title field because the diacriticFolder custom analyzer we used on the title field applied character folding on its values. Atlas Search returns documents with titles that begin with the query term alle because we used the keyword tokenizer, which tokenizes entire strings (or phrases) as a single token.

Alternatively, you can specify the standard tokenizer instead of the keyword tokenizer in the custom analyzer used on the title field. For the standard tokenizer, the Atlas Search results would contain documents with titles that begin or appear anywhere at the beginning of the word for the query term alle such as "Desde allè". To test this, edit your index definition to replace the keyword tokenizer on line 17 with standard tokenizer, save the index definition, and run the sample query.

Connect to your cluster in MongoDB Compass.

Open MongoDB Compass and connect to your cluster. For detailed instructions on connecting, see Connect via Compass.

Use the movies collection in the sample_mflix database.

On the Database screen, click the sample_mflix database, then click the movies collection.

Run an Atlas Search diacritic-insensitive query.

This query uses the following compound operator clauses to query the collection:

• must clause to search for movie titles that begin with the term alle using the wildcard operator

• should clause to specify preference for the Drama genre using the text operator

The query uses the $project stage to:

• Exclude all fields except title and genres

• Add a field named score

To run this query in MongoDB Compass:

1. Click the Aggregations tab.

2. Click Select..., then configure each of the following pipeline stages by selecting the stage from the dropdown and adding the query for that stage. Click Add Stage to add additional stages.

{
  "index": "diacritic-insensitive-tutorial",
  "compound": {
    "must": [{
      "wildcard": {
        "path": "title",
        "query": "alle*",
        "allowAnalyzedField": true
      }
    }],
    "should": [{
      "text": {
        "query": "Drama",
        "path": "genres"
      }
    }]
  }
}

{
  "_id": 0,
  "title": 1,
  "genres": 1,
  "score": {
    "$meta": "searchScore"
  }
}

If you enabled Auto Preview, MongoDB Compass displays the following documents next to the $project pipeline stage:

{
  genres: [ 'Drama', 'Family', 'Sport' ],
  title: 'Alley Cats Strike',
  score: 1.2084882259368896
},
{
  genres: [ 'Drama', 'Romance', 'Sci-Fi' ],
  title: 'Allegro',
  score: 1.179288625717163
},
{
  genres: [ 'Animation', 'Comedy', 'Fantasy' ],
  title: 'Allegro non troppo',
  score: 1
},
{
  genres: [ 'Comedy' ],
  title: 'Allez, Eddy!',
  score: 1
}

The first document in the result includes diacritics in the title field because the diacriticFolder custom analyzer we used on the title field applied character folding on its values. Atlas Search returns documents with titles that begin with the query term alle because we used the keyword tokenizer, which tokenizes entire strings (or phrases) as a single token.

Alternatively, you can specify the standard tokenizer instead of the keyword tokenizer in the custom analyzer used on the title field. For the standard tokenizer, the Atlas Search results would contain documents with titles that begin or appear anywhere at the beginning of the word for the query term alle such as "Desde allè". To test this, edit your index definition to replace the keyword tokenizer on line 17 with standard tokenizer, save the index definition, and run the sample query.

Set up and initialize the .NET/C# project for the query.

1. Create a new directory called diacritic-insensitive-example and initialize your project with the dotnet new command.

   mkdir diacritic-insensitive-example
   cd diacritic-insensitive-example
   dotnet new console

2. Add the .NET/C# Driver to your project as a dependency.

   dotnet add package MongoDB.Driver

Create the query in the Program.cs file.

1. Replace the contents of the Program.cs file with the following code.

   The code example performs the following tasks:

   • Imports mongodb packages and dependencies.

   • Establishes a connection to your Atlas cluster.

   • Uses the following compound operator clauses to query the collection:

     • must clause to search for movie titles that begin with the term alle using the wildcard operator

     • should clause to specify preference for the Drama genre using the text operator

     The query uses the $project stage to:

     • Exclude all fields except title and genres

     • Add a field named score

   • Iterates over the cursor to print the documents that match the query.

   1
   using MongoDB.Bson;
   2
   using MongoDB.Bson.Serialization.Attributes;
   3
   using MongoDB.Bson.Serialization.Conventions;
   4
   using MongoDB.Driver;
   5
   using MongoDB.Driver.Search;
   6
   7
   public class DiacriticInsensitiveExample
   8
   {
   9
       private const string MongoConnectionString = "<connection-string>";
   10
   11
       public static void Main(string[] args)
   12
       {
   13
           // allow automapping of the camelCase database fields to our MovieDocument
   14
           var camelCaseConvention = new ConventionPack { new CamelCaseElementNameConvention() };
   15
           ConventionRegistry.Register("CamelCase", camelCaseConvention, type => true);
   16
   17
           // connect to your Atlas cluster
   18
           var mongoClient = new MongoClient(MongoConnectionString);
   19
           var mflixDatabase = mongoClient.GetDatabase("sample_mflix");
   20
           var moviesCollection = mflixDatabase.GetCollection<MovieDocument>("movies");
   21
   22
           // define and run pipeline
   23
           var results = moviesCollection.Aggregate()
   24
               .Search(Builders<MovieDocument>.Search.Compound()
   25
                   .Must(Builders<MovieDocument>.Search.Wildcard(movie => movie.Title, "alle*", true))
   26
                   .Should(Builders<MovieDocument>.Search.Text(movie => movie.Genres, "Drama")),
   27
                indexName: "diacritic-insensitive-tutorial")
   28
               .Project<MovieDocument>(Builders<MovieDocument>.Projection
   29
                   .Include(movie => movie.Title)
   30
                   .Include(movie => movie.Genres)
   31
                   .Exclude(movie => movie.Id)
   32
                   .MetaSearchScore(movie => movie.Score))
   33
               .ToList();
   34
   35
           // print results
   36
           foreach (var movie in results)
   37
           {
   38
               Console.WriteLine(movie.ToJson());
   39
           }
   40
       }
   41
   }
   42
   43
   [BsonIgnoreExtraElements]
   44
   public class MovieDocument
   45
   {
   46
       [BsonIgnoreIfDefault]
   47
       public ObjectId Id { get; set; }
   48
       public string [] Genres { get; set; }
   49
       public string Title { get; set; }
   50
       public double Score { get; set; }
   51
   }

2. Before you run the sample, replace <connection-string> with your Atlas connection string. Ensure that your connection string includes your database user's credentials. To learn more, see Connect via Drivers.

Compile and run the Program.cs file.

dotnet run diacritic-insensitive-example.csproj

{ "genres" : ["Drama", "Family", "Sport"], "title" : "Alley Cats Strike", "score" : 1.2084882259368896 }
{ "genres" : ["Drama", "Romance", "Sci-Fi"], "title" : "Allegro", "score" : 1.1792886257171631 }
{ "genres" : ["Animation", "Comedy", "Fantasy"], "title" : "Allegro non troppo", "score" : 1.0 }
{ "genres" : ["Comedy"], "title" : "Allez, Eddy!", "score" : 1.0 }

The first document in the result includes diacritics in the title field because the diacriticFolder custom analyzer we used on the title field applied character folding on its values. Atlas Search returns documents with titles that begin with the query term alle because we used the keyword tokenizer, which tokenizes entire strings (or phrases) as a single token.

Alternatively, you can specify the standard tokenizer instead of the keyword tokenizer in the custom analyzer used on the title field. For the standard tokenizer, the Atlas Search results would contain documents with titles that begin or appear anywhere at the beginning of the word for the query term alle such as "Desde allè". To test this, edit your index definition to replace the keyword tokenizer on line 17 with standard tokenizer, save the index definition, and run the sample query.

Run an Atlas Search diacritic-insensitive query.

1. Create a file named diacritic-insensitive.go.

2. Copy and paste the following code into the diacritic-insensitive.go file.

   The code example performs the following tasks:

   • Imports mongodb packages and dependencies.

   • Establishes a connection to your Atlas cluster.

   • Uses the following compound operator clauses to query the collection:

     • must clause to search for movie titles that begin with the term alle using the wildcard operator

     • should clause to specify preference for the Drama genre using the text operator

     The query uses the $project stage to:

     • Exclude all fields except title and genres

     • Add a field named score

   • Iterates over the cursor to print the documents that match the query.

   1
   package main
   2
   3
   import (
   4
   	"context"
   5
   	"fmt"
   6
   7
   	"go.mongodb.org/mongo-driver/bson"
   8
   	"go.mongodb.org/mongo-driver/mongo"
   9
   	"go.mongodb.org/mongo-driver/mongo/options"
   10
   )
   11
   12
   func main() {
   13
   	// connect to your Atlas cluster
   14
   	client, err := mongo.Connect(context.TODO(), options.Client().ApplyURI("<connection-string>"))
   15
   	if err != nil {
   16
   		panic(err)
   17
   	}
   18
   	defer client.Disconnect(context.TODO())
   19
   20
   	// set namespace
   21
   	collection := client.Database("sample_mflix").Collection("movies")
   22
   23
   	// define pipeline stages
   24
   	searchStage := bson.D{{"$search", bson.M{
   25
   		"index": "diacritic-insensitive-tutorial",
   26
   		"compound": bson.M{
   27
   			"must": bson.M{
   28
   				"wildcard": bson.M{
   29
   					"path":               "title",
   30
   					"query":              "alle*",
   31
   					"allowAnalyzedField": true,
   32
   				},
   33
   			},
   34
   			"should": bson.D{
   35
   				{"text", bson.M{
   36
   					"path":  "genres",
   37
   					"query": "Drama"}}},
   38
   		},
   39
   	}}}
   40
   	projectStage := bson.D{{"$project", bson.D{{"title", 1}, {"genres", 1}, {"_id", 0}, {"score", bson.D{{"$meta", "searchScore"}}}}}}
   41
   42
   	// run pipeline
   43
   	cursor, err := collection.Aggregate(context.TODO(), mongo.Pipeline{searchStage, projectStage})
   44
   	if err != nil {
   45
   		panic(err)
   46
   	}
   47
   48
   	// print results
   49
   	var results []bson.D
   50
   	if err = cursor.All(context.TODO(), &results); err != nil {
   51
   		panic(err)
   52
   	}
   53
   	for _, result := range results {
   54
   		fmt.Println(result)
   55
   	}
   56
   }

3. Before you run the sample, replace <connection-string> with your Atlas connection string. Ensure that your connection string includes your database user's credentials. To learn more, see Connect via Drivers.

4. Run the following command to query your collection:

   go run diacritic-insensitive.go

   HIDE OUTPUT

   [{genres [Drama Family Sport]} {title Alley Cats Strike} {score 1.2084882259368896}]
   [{genres [Drama Romance Sci-Fi]} {title Allegro} {score 1.179288625717163}]
   [{genres [Animation Comedy Fantasy]} {title Allegro non troppo} {score 1}]
   [{genres [Comedy]} {title Allez, Eddy!} {score 1}]

The first document in the result includes diacritics in the title field because the diacriticFolder custom analyzer we used on the title field applied character folding on its values. Atlas Search returns documents with titles that begin with the query term alle because we used the keyword tokenizer, which tokenizes entire strings (or phrases) as a single token.

Alternatively, you can specify the standard tokenizer instead of the keyword tokenizer in the custom analyzer used on the title field. For the standard tokenizer, the Atlas Search results would contain documents with titles that begin or appear anywhere at the beginning of the word for the query term alle such as "Desde allè". To test this, edit your index definition to replace the keyword tokenizer on line 17 with standard tokenizer, save the index definition, and run the sample query.

Ensure that your CLASSPATH contains the following libraries.

Run an Atlas Search diacritic-insensitive query.

1. Create a file named DiacriticInsensitive.java.

2. Copy and paste the following code into the DiacriticInsensitive.java file.

   The code example performs the following tasks:

   • Imports mongodb packages and dependencies.

   • Establishes a connection to your Atlas cluster.

   • Uses the following compound operator clauses to query the collection:

     • must clause to search for movie titles that begin with the term alle using the wildcard operator

     • should clause to specify preference for the Drama genre using the text operator

     The query uses the $project stage to:

     • Exclude all fields except title and genres

     • Add a field named score

   • Iterates over the cursor to print the documents that match the query.

   1
   import static com.mongodb.client.model.Aggregates.project;
   2
   import static com.mongodb.client.model.Projections.*;
   3
   import com.mongodb.client.MongoClient;
   4
   import com.mongodb.client.MongoClients;
   5
   import com.mongodb.client.MongoCollection;
   6
   import com.mongodb.client.MongoDatabase;
   7
   import org.bson.Document;
   8
   import java.util.Arrays;
   9
   import java.util.List;
   10
   11
   public class DiacriticInsensitive {
   12
       public static void main(String[] args) {
   13
           // define clauses
   14
           List<Document> mustClauses =
   15
               List.of( new Document("wildcard", 
   16
                   new Document("path", "title")
   17
                   .append("query", "alle*")
   18
                   .append("allowAnalyzedField", true)));
   19
           List<Document> shouldClauses =
   20
               List.of( new Document("text",
   21
                   new Document("query", "Drama")
   22
                   .append("path", "genres")));
   23
           // define pipeline
   24
           Document agg = new Document( "$search",
   25
               new Document("index", "diacritic-insensitive-tutorial")
   26
               .append("compound",
   27
                   new Document("must", mustClauses)
   28
                   .append("should", shouldClauses)));
   29
   30
           // connect to your Atlas cluster
   31
           String uri = "<connection-string>";
   32
   33
           try (MongoClient mongoClient = MongoClients.create(uri)) {            
   34
               // set namespace
   35
               MongoDatabase database = mongoClient.getDatabase("sample_mflix");
   36
               MongoCollection<Document> collection = database.getCollection("movies");
   37
               
   38
               // run pipeline and print results
   39
               collection.aggregate(Arrays.asList(agg,
   40
                   project(fields(
   41
                       excludeId(), 
   42
                       include("title"), 
   43
                       include("genres"), 
   44
                       computed("score", new Document("$meta", "searchScore"))))))
   45
                   .forEach(doc -> System.out.println(doc.toJson()));
   46
           }
   47
       }
   48
   }

   Note

   To run the sample code in your Maven environment, add the following code above the import statements in your file.

   package com.mongodb.drivers;

3. Before you run the sample, replace <connection-string> with your Atlas connection string. Ensure that your connection string includes your database user's credentials. To learn more, see Connect via Drivers.

4. Compile and run the DiacriticInsensitive.java file.

   javac DiacriticInsensitive.java
   java DiacriticInsensitive

   HIDE OUTPUT

   {"genres": ["Drama", "Family", "Sport"], "title": "Alley Cats Strike", "score": 1.2084882259368896}
   {"genres": ["Drama", "Romance", "Sci-Fi"], "title": "Allegro", "score": 1.179288625717163}
   {"genres": ["Animation", "Comedy", "Fantasy"], "title": "Allegro non troppo", "score": 1.0}
   {"genres": ["Comedy"], "title": "Allez, Eddy!", "score": 1.0}

The first document in the result includes diacritics in the title field because the diacriticFolder custom analyzer we used on the title field applied character folding on its values. Atlas Search returns documents with titles that begin with the query term alle because we used the keyword tokenizer, which tokenizes entire strings (or phrases) as a single token.

Alternatively, you can specify the standard tokenizer instead of the keyword tokenizer in the custom analyzer used on the title field. For the standard tokenizer, the Atlas Search results would contain documents with titles that begin or appear anywhere at the beginning of the word for the query term alle such as "Desde allè". To test this, edit your index definition to replace the keyword tokenizer on line 17 with standard tokenizer, save the index definition, and run the sample query.

Ensure that you add the following dependency to your project.

Run an Atlas Search diacritic-insensitive query.

1. Create a file named DiacriticInsensitive.kt.

2. Copy and paste the following code into the DiacriticInsensitive.kt file.

   The code example performs the following tasks:

   • Imports mongodb packages and dependencies.

   • Establishes a connection to your Atlas cluster.

   • Uses the following compound operator clauses to query the collection:

     • must clause to search for movie titles that begin with the term alle using the wildcard operator

     • should clause to specify preference for the Drama genre using the text operator

     The query uses the $project stage to:

     • Exclude all fields except title and genres

     • Add a field named score

   • Prints the documents that match the query from the AggregateFlow instance.

   1
   import com.mongodb.client.model.Aggregates.project
   2
   import com.mongodb.client.model.Projections.*
   3
   import com.mongodb.kotlin.client.coroutine.MongoClient
   4
   import kotlinx.coroutines.runBlocking
   5
   import org.bson.Document
   6
   7
   fun main() {
   8
       // connect to your Atlas cluster
   9
       val uri = "<connection-string>"
   10
       val mongoClient = MongoClient.create(uri)
   11
   12
       // set namespace
   13
       val database = mongoClient.getDatabase("sample_mflix")
   14
       val collection = database.getCollection<Document>("movies")
   15
   16
       runBlocking {
   17
           // define clauses
   18
           val mustClauses = listOf(
   19
               Document(
   20
                   "wildcard",
   21
                   Document("path", "title")
   22
                       .append("query", "alle*")
   23
                       .append("allowAnalyzedField", true)
   24
               )
   25
           )
   26
   27
           val shouldClauses = listOf(
   28
               Document(
   29
                   "text",
   30
                   Document("query", "Drama")
   31
                       .append("path", "genres")
   32
               )
   33
           )
   34
   35
           // define pipeline
   36
           val agg = Document( "\$search",
   37
               Document("index", "diacritic-insensitive-tutorial")
   38
                   .append("compound", Document("must", mustClauses)
   39
                       .append("should", shouldClauses)
   40
                   )
   41
           )
   42
   43
           // run pipeline and print results
   44
           val resultsFlow = collection.aggregate<Document>(
   45
               listOf(
   46
                   agg,
   47
                   project(fields(
   48
                       excludeId(),
   49
                       include("title", "genres"),
   50
                       computed("score", Document("\$meta", "searchScore"))))
   51
               )
   52
           )
   53
           resultsFlow.collect { println(it) }
   54
       }
   55
   56
       mongoClient.close()
   57
   }

3. Before you run the sample, replace <connection-string> with your Atlas connection string. Ensure that your connection string includes your database user's credentials. To learn more, see Connect via Drivers.

4. Run the DiacriticInsensitive.kt file.

   When you run the DiacriticInsensitive.kt program in your IDE, it prints the following documents:

   Document{{genres=[Drama, Family, Sport], title=Alley Cats Strike, score=1.2084882259368896}}
   Document{{genres=[Drama, Romance, Sci-Fi], title=Allegro, score=1.179288625717163}}
   Document{{genres=[Animation, Comedy, Fantasy], title=Allegro non troppo, score=1.0}}
   Document{{genres=[Comedy], title=Allez, Eddy!, score=1.0}}

The first document in the result includes diacritics in the title field because the diacriticFolder custom analyzer we used on the title field applied character folding on its values. Atlas Search returns documents with titles that begin with the query term alle because we used the keyword tokenizer, which tokenizes entire strings (or phrases) as a single token.

Alternatively, you can specify the standard tokenizer instead of the keyword tokenizer in the custom analyzer used on the title field. For the standard tokenizer, the Atlas Search results would contain documents with titles that begin or appear anywhere at the beginning of the word for the query term alle such as "Desde allè". To test this, edit your index definition to replace the keyword tokenizer on line 17 with standard tokenizer, save the index definition, and run the sample query.

Run an Atlas Search diacritic-insensitive query.

1. Create a file named diacritic-insensitive.js.

2. Copy and paste the following code into the diacritic-insensitive.js file.

   The code example performs the following tasks:

   • Imports mongodb, MongoDB's Node.js driver.

   • Creates an instance of the MongoClient class to establish a connection to your Atlas cluster.

   • Uses the following compound operator clauses to query the collection:

     • must clause to search for movie titles that begin with the term alle using the wildcard operator

     • should clause to specify preference for the Drama genre using the text operator

     The query uses the $project stage to:

     • Exclude all fields except title and genres

     • Add a field named score

   • Iterates over the cursor to print the documents that match the query.

   1
   const { MongoClient } = require("mongodb");
   2
   3
   // Replace the uri string with your MongoDB deployment's connection string.
   4
   const uri =
   5
     "<connection-string>";
   6
   7
   const client = new MongoClient(uri);
   8
   9
   async function run() {
   10
     try {
   11
       await client.connect();
   12
   13
       // set namespace
   14
       const database = client.db("sample_mflix");
   15
       const coll = database.collection("movies");
   16
   17
       // define pipeline
   18
       const agg = [{
   19
           '$search': {
   20
             'index': 'diacritic-insensitive-tutorial',
   21
             'compound': {
   22
                   'must': [{
   23
                       'wildcard': {
   24
                           'query': "alle*",
   25
                           'path': "title",
   26
                           'allowAnalyzedField': true
   27
                       }
   28
                   }],
   29
                   'should': [{'text': {'query': 'Drama', 'path': 'genres'}}]
   30
               }}},
   31
           { '$project': { '_id': 0, 'title': 1 , 'genres': 1, 'score': {'$meta': 'searchScore'}}}];
   32
              
   33
       // run pipeline
   34
       const result = await coll.aggregate(agg);
   35
   36
       // print results
   37
       await result.forEach((doc) => console.log(doc));
   38
       
   39
     } finally {
   40
       await client.close();
   41
     }
   42
   }
   43
   run().catch(console.dir);

3. Before you run the sample, replace <connection-string> with your Atlas connection string. Ensure that your connection string includes your database user's credentials. To learn more, see Connect via Drivers.

4. Run the following command to query your collection:

   node diacritic-insensitive.js

   HIDE OUTPUT

   {
     genres: [ 'Drama', 'Family', 'Sport' ],
     title: 'Alley Cats Strike',
     score: 1.2084882259368896
   }
   {
     genres: [ 'Drama', 'Romance', 'Sci-Fi' ],
     title: 'Allegro',
     score: 1.179288625717163
   }
   {
     genres: [ 'Animation', 'Comedy', 'Fantasy' ],
     title: 'Allegro non troppo',
     score: 1
   }
   {
     genres: [ 'Comedy' ],
     title: 'Allez, Eddy!',
     score: 1
   }

The first document in the result includes diacritics in the title field because the diacriticFolder custom analyzer we used on the title field applied character folding on its values. Atlas Search returns documents with titles that begin with the query term alle because we used the keyword tokenizer, which tokenizes entire strings (or phrases) as a single token.

Alternatively, you can specify the standard tokenizer instead of the keyword tokenizer in the custom analyzer used on the title field. For the standard tokenizer, the Atlas Search results would contain documents with titles that begin or appear anywhere at the beginning of the word for the query term alle such as "Desde allè". To test this, edit your index definition to replace the keyword tokenizer on line 17 with standard tokenizer, save the index definition, and run the sample query.

Run an Atlas Search diacritic-insensitive query.

1. Create a file named diacritic-insensitive.py.

2. Copy and paste the following code into the diacritic-insensitive.py file.

   The following code example:

   • Imports pymongo, MongoDB's Python driver, and the dns module, which is required to connect pymongo to Atlas using a DNS seed list connection string.

   • Creates an instance of the MongoClient class to establish a connection to your Atlas cluster.

   • Uses the following compound operator clauses to query the collection:

     • must clause to search for movie titles that begin with the term alle using the wildcard operator

     • should clause to specify preference for the Drama genre using the text operator

     The query uses the $project stage to:

     • Exclude all fields except title and genres

     • Add a field named score

   • Iterates over the cursor to print the documents that match the query.

   1
   import pymongo
   2
   3
   # connect to your Atlas cluster
   4
   client = pymongo.MongoClient('<connection-string>')
   5
   6
   # define pipeline
   7
   pipeline = [
   8
     {'$search': {
   9
         'index': 'diacritic-insensitive-tutorial',
   10
         'compound': {
   11
           'must': [{'wildcard': {'path': 'title', 'query': 'alle*', 'allowAnalyzedField': True}}],
   12
           'should': [{'text': {'query': 'Drama', 'path': 'genres'}}]}}},
   13
     {'$project': {'_id': 0, 'title': 1, 'genres': 1, 'score': {'$meta': 'searchScore'}}}
   14
   ]
   15
   16
   # run pipeline
   17
   result = client['sample_mflix']['movies'].aggregate(pipeline)
   18
   19
   # print results
   20
   for i in result:
   21
       print(i)

3. Before you run the sample, replace <connection-string> with your Atlas connection string. Ensure that your connection string includes your database user's credentials. To learn more, see Connect via Drivers.

4. Run the following command to query your collection:

   python diacritic-insensitive.py

   HIDE OUTPUT

   {'genres': ['Drama', 'Family', 'Sport'], 'title': 'Alley Cats Strike', 'score': 1.2084882259368896}
   {'genres': ['Drama', 'Romance', 'Sci-Fi'], 'title': 'Allegro', 'score': 1.179288625717163}
   {'genres': ['Animation', 'Comedy', 'Fantasy'], 'title': 'Allegro non troppo', 'score': 1.0}
   {'genres': ['Comedy'], 'title': 'Allez, Eddy!', 'score': 1.0}

The first document in the result includes diacritics in the title field because the diacriticFolder custom analyzer we used on the title field applied character folding on its values. Atlas Search returns documents with titles that begin with the query term alle because we used the keyword tokenizer, which tokenizes entire strings (or phrases) as a single token.

Alternatively, you can specify the standard tokenizer instead of the keyword tokenizer in the custom analyzer used on the title field. For the standard tokenizer, the Atlas Search results would contain documents with titles that begin or appear anywhere at the beginning of the word for the query term alle such as "Desde allè". To test this, edit your index definition to replace the keyword tokenizer on line 17 with standard tokenizer, save the index definition, and run the sample query.