MongoFlix - interactive demo for MongoDB Atlas Search, MongoDB Realm, GraphQL and so much more.
Also available on Realm as a static site!
You can of course clone the repo and run the project locally npm install && npm run build.
Alternatively, you can open the project in your browser without any installation required on your machine.
Open the project live on StackBlitz:
Duplicate the file .env.local.example-add-app-id-here and name it: .env.local.
You will need to change the <APP_ID> value to the app id of your MongoDB Realm app, which will be created at a later step.
You have also to update the NEXT_PUBLIC_REALM_BASE_URL value if you have a different base URL for your MongoDB Realm app.
This value will depend on the deployment region of your MongoDB Realm app.
All Steps
To follow along with the demo, you will need to create a MongoDB Atlas cluster and load the sample data set into your cluster. Please create an account on MongoDB Atlas and follow the instructions. If it is the first time you use Atlas you will need to create an organization and a project. After you complete the account setup, you will see the Atlas UI. If you don not have any cluster click the Build a Database button.
In the following dialog select Shared and click Create. The following screen will provide an interface to configure the cluster.
If you choose a region other than Frankfurt you will need to update the endpoint in the app in .env.local to match the region.
Here are the settings for the cluster:
- Cloud Provider & Region:
AWS, Frankfurt (eu-central-1) - Cluster Tier:
MO Sandbox (Shared RAM, 512 MB Storage) - Cluster Name:
Cluster0
After your cluster was deployed in the region of your choice, you will need to load the sample data set into your cluster. Click the three dots menu in the top headline of the cluster card. Click Load Sample Dataset. Click the Load Sample Dataset button in the overlay to start the process. (It should take about 5-10 minutes. ☕️ 🍵)
Click the Cluster name to open it. In your cluster on Atlas click the Search tab. Click the Create Search Index button to create an index.
- Select the JSON editor and click Next.
- In the Database and Collection sidebar select
sample_mflixand selectmovies. - For the name leave it as
defaultand paste the following JSON. - Click Next.
- After reviewing it, click Create Search Index.
{
"mappings": {
"dynamic": true,
"fields": {
"title": [
{
"dynamic": true,
"type": "document"
},
{
"type": "autocomplete"
}
]
}
}
}The index creation should take less tha a minute.
Lets test it, to verify that it works.
Still in the Search tab, click the Query button besides the newly created index.
Enter the following querry to find all movies containing the text time in any text values.
{ "$search": { "text": "time travel" } }In the Atlas UI click the Realm tab at the top. If you are using Realm for the first time, you will see a dialog with additonal instructions. You can safely select Build your own App it and click the Next. The information should be populated automatically. Make sure to use the same name for simplicity.
In the following dialog, setup the name of the Realm App, connect it to your newly created cluster and select a local (single region) deployment model. It should be preferable to use the region closest to your cluster region.
- Name: Application-0
- Cluster: Cluster0
- Deployment Model: Local
To create the app click Create Realm Application.
Hint: Now with the app created you can update the .env.local file to include the App ID value from your Realm app.
On the left side bar of the Atlas UI, within Data Access, click Authentication. As you see Realm provides many authentication methods, we will use Anonymous for this demo. Click on the Edit button and set the checkbox to ON for this authentication method.
On the left side bar of the Atlas UI, within Data Access, click Rules. Rules provide you many ways to limit and configure data access per collection and user role, deep down to the document level. For this demo we will allow all users to only read all documents in the movies colelction. Realm provides templates for many scenarios and we will use the Users can only read all data template.
On the left side bar of the Atlas UI, within Data Access, click Schema. Schema defines the data structures and types for documents in each collection in the databases. Select the movies collection within the sample_mflix database. Click the generate schema button. Select just the movies collection, leave the samling size as default and click the Generate Schema button. This will also generate all the neccessary types and queries for a GraphQL schema. Which can be used immediately to access the data through the GraphQL endpoint managed by Realm.
Click the Review Draft & Deploy button at the top of the page and Deploy your changes.
Hint: Now with the schema generated you can update the .env.local file to include the following base URL from your Realm app.
Lets test how GraphQL actually works. In the GraphQL tab, within the GraphQL editor paste in the following code snippet to test the generated scheme.
query {
movie(query: { title: "The Godfather" }) {
_id
title
metacritic
num_mflix_comments
fullplot
}
}Now with the correct rules and schema in place we can start creating functions for the app. For the first feature we will create a function that will return a list of movies that match the search term by the title. It will use our dynamic index created in the previous step with the autocomplete functionality. This enables us to provide autocomplete and fuzzy search for movie tiltes in the search bar of the frontend app.
On the left side bar of the Atlas UI, within Build, click Functions. Functions provide a way to execute serverside logic on Realm integrating data from the connected cluster. With the Aggregation Framework at your disposal you can create very powerful aggregations, even without a driver.
Click the Create New Function button and enter autocompleteTitle as the name for the function.
Now click the Function Editor tab.
Paste the following code into the Function Editor:
exports = async (title) => {
const collection = context.services.get("mongodb-atlas").db("sample_mflix").collection("movies");
return await collection
.aggregate([
{
$search: {
autocomplete: {
path: "title",
query: title,
fuzzy: { maxEdits: 1 },
},
},
},
{
$project: {
title: 1,
},
},
{
$limit: 10,
},
])
.toArray();
};Click the Save Draft button to save the function.
We want to use the autocomplete function in our GraphQL schema. To do this we need to create a custom resolver. Custom resolvers allow us to define custom queries and mutations for our GraphQL schema, backed by Functions created on Realm.
On the left side bar of the Atlas UI, within Build, click GraphQL. Click the Custom Resolvers tab and click the Add a Custom Resolver button. For the GraphQL Field Name enter autocompleteTitle, for the Parent Type select Query and for the Function Name select the newly created function autocompleteTitle.
The input type defines the data type of what will be send to the GraphQL API as input for this resolver. The return type defines the data type of what will be returned by the API. We will send a string as input and expect a list of movie objects as output.
- Input Type:
Scalar Type,String
- Return Type:
Existing Type (List),[Movie]
Click the Save Draft button to save the custom resolver.
Click the Review Draft & Deploy button at the top of the page and Deploy your changes.
Now with the first feature setup take the time to test the app, enter some movie titles into the search bar and see the autocomplete results.
Now with the autocomplete function in place we can create a new function for highlights and scoring. This function will return a list of movies that match the search term by the title, the genres selected and the country where a certain movie was produced. Additionally, it will return highlights and search scores for the results. The highlights contain the exact substring within the title and the plot strings, containing the matched search term. This will allow us to highlight the found search terms within the frontend UI.
Similar to the previous function we will create a new function for highlights and scoring.
On the left side bar of the Atlas UI, within Build, click Functions.
Click the Create New Function button and enter filteredMovies as the name for the function.
Now click the Function Editor tab.
Paste the following code into the Function Editor:
exports = async (input) => {
const collection = context.services.get("mongodb-atlas").db("sample_mflix").collection("movies");
const { term, genres, countries } = input;
const searchShould = [];
const searchMust = [];
if (term.length > 0) {
const termStage = {
autocomplete: {
path: "title",
query: term,
fuzzy: { maxEdits: 1.0 },
score: {
boost: {
path: "imdb.rating",
undefined: 1,
},
},
},
};
searchMust.push(termStage);
const plotStage = {
text: {
query: term,
path: "plot",
},
};
searchShould.push(plotStage);
}
if (genres.length > 0) {
const genresStage = {
text: {
query: genres,
path: "genres",
},
};
searchMust.push(genresStage);
}
if (countries.length > 0) {
const countryStage = {
text: {
query: countries,
path: "countries",
},
};
searchMust.push(countryStage);
}
const searchQuery = [
{
$search: {
compound: {
should: searchShould,
must: searchMust,
},
highlight: { path: ["title", "genres", "countries", "plot"] },
},
},
{
$project: {
_id: 1,
title: 1,
poster: 1,
cast: 1,
directors: 1,
plot: 1,
fullplot: 1,
year: 1,
genres: 1,
countries: 1,
imdb: 1,
score: { $meta: "searchScore" },
highlights: { $meta: "searchHighlights" },
},
},
{ $limit: 20 },
];
return await collection.aggregate(searchQuery).toArray();
};On the left side bar of the Atlas UI, within Build, click GraphQL. Click the Custom Resolvers tab and click the Add a Custom Resolver button. For the GraphQL Field Name enter filteredMovies, for the Parent Type select Query and for the Function Name select the newly created function filteredMovies.
We will send a string as input and expect a list of custom movie objects, containing the scores and highlights for each movie as output.
- Input Type:
Custom Type
{
"type": "object",
"title": "FilteredMoviesInput",
"properties": {
"term": {
"bsonType": "string"
},
"genres": {
"bsonType": "array",
"items": {
"bsonType": "string"
}
},
"countries": {
"bsonType": "array",
"items": {
"bsonType": "string"
}
}
}
}- Return Type:
Custom Type
{
"items": {
"bsonType": "object",
"properties": {
"_id": {
"bsonType": "objectId"
},
"cast": {
"bsonType": "array",
"items": {
"bsonType": "string"
}
},
"countries": {
"bsonType": "array",
"items": {
"bsonType": "string"
}
},
"directors": {
"bsonType": "array",
"items": {
"bsonType": "string"
}
},
"fullplot": {
"bsonType": "string"
},
"genres": {
"bsonType": "array",
"items": {
"bsonType": "string"
}
},
"highlights": {
"bsonType": "array",
"items": {
"bsonType": "object",
"properties": {
"path": {
"bsonType": "string"
},
"score": {
"bsonType": "double"
},
"texts": {
"bsonType": "array",
"items": {
"bsonType": "object",
"properties": {
"type": {
"bsonType": "string"
},
"value": {
"bsonType": "string"
}
}
}
}
}
}
},
"imdb": {
"bsonType": "object",
"properties": {
"id": {
"bsonType": "int"
},
"rating": {
"bsonType": "double"
},
"votes": {
"bsonType": "int"
}
}
},
"plot": {
"bsonType": "string"
},
"poster": {
"bsonType": "string"
},
"score": {
"bsonType": "double"
},
"title": {
"bsonType": "string"
},
"year": {
"bsonType": "int"
}
}
},
"title": "FilteredMovies",
"type": "array"
}Click the Save Draft button to save the custom resolver.
Click the Review Draft & Deploy button at the top of the page and Deploy your changes.
Now with the highlights feature setup take the time to test the app, enter some movie titles into the search bar scroll in the results list and verify that the fuzzy matched search term is highlighted within the movie title and the short plot when there is a match.
Facets open many powerful use cases for grouping your search results. The following feature shows how to run an Atlas Search query to get results grouped by values for genres of each movie in the movies collection, including the count for each of those groups.
In your cluster on Atlas in the Search tab, create a new index with the name facets and the following JSON for the movies collection.
{
"mappings": {
"dynamic": false,
"fields": {
"genres": [
{
"dynamic": true,
"type": "document"
},
{
"type": "stringFacet"
}
],
"year": [
{
"dynamic": true,
"type": "document"
},
{
"representation": "int64",
"type": "number"
}
]
}
}
}Now with the index created, in the Atlas UI click the Realm tab. Click Application-0 in the UI. On the left side bar of the Atlas UI, within Build, click Functions.
Click the Create New Function button and enter facetsGenres as the name for the function.
Now click the Function Editor tab.
Paste the following code into the Function Editor:
exports = async (arg) => {
const collection = context.services.get("mongodb-atlas").db("sample_mflix").collection("movies");
return await collection
.aggregate([
{
$searchMeta: {
index: "facets",
facet: {
operator: {
range: {
path: "year",
gte: 1900,
},
},
facets: {
genresFacet: {
type: "string",
path: "genres",
},
},
},
},
},
])
.toArray();
};On the left side bar of the Atlas UI, within Build, click GraphQL. Click the Custom Resolvers tab and click the Add a Custom Resolver button. For the GraphQL Field Name enter facetsGenres, for the Parent Type select Query and for the Function Name select the newly created function facetsGenres.
We won't send input to this query and expect a list of custom objects representing the facets for each genre, containing the number of movies for each genre.
- Input Type:
None
- Return Type:
Custom Type
{
"title": "GenresMeta",
"type": "array",
"items": {
"bsonType": "object",
"properties": {
"count": {
"bsonType": "double"
},
"facet": {
"bsonType": "object",
"properties": {
"genresFacet": {
"bsonType": "object",
"properties": {
"buckets": {
"bsonType": "array",
"items": {
"bsonType": "object",
"properties": {
"_id": {
"bsonType": "string"
},
"count": {
"bsonType": "double"
}
}
}
}
}
}
}
}
}
}
}Click the Save Draft button to save the custom resolver.
Click the Review Draft & Deploy button at the top of the page and Deploy your changes.
Now with the facets setup test the app and open the dropdown for Genres. Notice that there is now a number besides each genre representing the total number of movies for that genre.
MongoDB Realm Hosting allows you to host, manage, and serve your application's static media and document files. You can use Hosting to store individual pieces of content or to upload and serve your entire client application.
Our frontend app contains all the necessary calls to the GraphQL API on Realm. We can export the whole frontend app as a static site and host it on MongoDB Realm.
For this you need to execute the follwing code in the root folfder of the project. Make sure that you have the dependencies installed with.
npm installand then build and export the site with an npm script using nextjs.
npm run build && npm run exportThis will create a folder called out in the root folder of the project.
On the MongoDB Atlas UI on the Realm tab. On the left side bar of the Atlas UI, within Manage, click Hosting. Click the Enable Hosting button. Drag and drop the contents of the folder out into the Hosting tab to upload all files.
Click the Review Draft & Deploy button at the top of the page and Deploy your changes.
Click now the Settings tab copy the Realm Domain paste it in a browser of your choice and press enter to view the site. 🎉