Skip to main content

Posts tagged with 'csharp'

This is a repost that originally appeared on the Couchbase Blog: Geospatial Search with ASP.NET Core, Aurelia, and Google Maps.

Geospatial search is now fully supported in Couchbase Server 5.5. Check out the Couchbase Server 5.5 announcement, and download the developer build for free right now.

In this post, I’m going to demonstrate the geospatial search capabilities of Couchbase Full Text Search by creating a web-based UI that performs searches. Whenever I think of geospatial searches, I think about Yelp, which is great at helping me find restaurants in a specific area.

So I’m going to have a little fun and create a very bare-bones version of Yelp, but only for hotels.

If you want to follow along, the full source code is available on Github.

Getting set up

Here are the steps I took to create a new project before I started writing code.

  1. At the command line: dotnet new aurelia. This assumes that you have .NET Core installed. Note that Geospatial Search is not a .NET-only feature: you can use it with the other Couchbase SDKs like Node.js, Java, etc. It also assumes that you’ve installed a SPA template for Aurelia. You can also go with Angular or React if you’d like, but I really like Aurelia, and I think you should give it a chance.

  2. The above command will create a shell of an ASP.NET Core project. In this blog post, I’m not going to use Razor. I’m just using ASP.NET as a backend for REST API endpoints.

  3. npm install aurelia-google-maps. You don’t have to use this, but the aurelia-google-maps plugin will make it easy for me to interact with Google Maps in my app.

  4. I opened this project in Visual Studio 2017. I added Couchbase.Extensions.DependencyInjection with NuGet. You don’t have to use this extension but it makes things easier.

  5. I installed Couchbase Server 5.5, including the Full Text Search service. I setup the travel-sample bucket. I created a user "matt" with full access to that bucket.

Create a Geospatial Index

Before building the ASP.NET backend, we need to create a geospatial index in Couchbase Server. Once you log in, click "Search" on the menu (it’s under "Workbench"). Click "Add Index" to get started.

Create Geospatial index

I named my index "mygeoindex". I selected travel-sample as the bucket to index.

In "Type Mappings", I uncheck the default. I add a new type mapping with a type name of "hotel". Every hotel document in "travel-sample" has a type with a value of "hotel". Check the "only index specified fields" box.

I’m going to add two child fields. One is "geo", which contains the geospatial coordinates inside a hotel document. Make sure to select "geopoint" as the type. The other is "name", which will be the name of the hotel. I choose to "store" each of these: it will make the index larger, but I can avoid a secondary lookup if I store the information in the index.

Important Note: There is a bug (NCBC-1651) in the current release of the .NET SDK that will cause an error if you try to read from a geopoint field. In the code samples, I’ve created a workaround: I don’t actually get the geo & name fields from the search index. I instead use the document key returned by search to make a secondary "get" call and get the full document. Keep in mind this is still a technique you may want to consider if you want to keep the size of your index down. This bug has already been fixed and will be in a future release. Such is the peril of being on the cutting edge!

That’s all there is to it. Click "Create Index". Watch the "indexing progress" on the next screen until it gets to 100% (it should not take very long, assuming you remembered to uncheck "default").

ASP.NET Core REST Endpoints

Next, let’s move over to ASP.NET. I’ll create two endpoints. One endpoint will demonstrate the bounding box search method, and the other will demonstrate the distance search method.

I’ll need a Couchbase bucket object to execute the queries. Follow the examples in my blog post about dependency injection or check out the source code on Github if you’ve never done this before.

Bounding Box

A "bounding box" search means that you define a box on a map, and you want to search for points of interest that are inside of that box. You only need two points to define a box: the top right corner coordinates and the bottom left corner coordinates. (Coordinates are latitude and longitude).

public class BoxSearch
{
    public double LatitudeTopLeft { get; set; }
    public double LongitudeTopLeft { get; set; }
    public double LatitudeBottomRight { get; set; }
    public double LongitudeBottomRight { get; set; }
}

To create a bounding box geospatial query, use the GeoBoundingBoxQuery class available in the .NET SDK. I’ll do this inside of a POST method with the above BoxSearch class as a parameter.

        [Route("api/Box")]
        [HttpPost]
        public IActionResult Box([FromBody] BoxSearch box)
        {
            var query = new GeoBoundingBoxQuery();
            query.TopLeft(box.LongitudeTopLeft, box.LatitudeTopLeft);
            query.BottomRight(box.LongitudeBottomRight, box.LatitudeBottomRight);
            var searchParams = new SearchParams()
                // .Fields("geo", "name") // omitting because of bug NCBC-1651
                .Limit(10)
                .Timeout(TimeSpan.FromMilliseconds(10000));
            var searchQuery = new SearchQuery
            {
                Query = query,
                Index = "mygeoindex",
                SearchParams = searchParams
            };
            var results = _bucket.Query(searchQuery);

// ... snip ...

All I need to return from this endpoint is a list of the results: each hotel’s coordinates and the hotel’s name & location. I created a GeoSearchResult class for this.

public class GeoSearchResult
{
    public double Latitude { get; set; }
    public double Longitude { get; set; }
    public InfoWindow InfoWindow { get; set; }
}

public class InfoWindow
{
    public string Content { get; set; }
}

I’ve constructed this class to match the Google Maps plugin that I’ll be using later.

Finally, I’ll use this class to return some results from the endpoint.

// ... snip ...

            var list = new List<GeoSearchResult>();
            foreach (var hit in results.Hits)
            {
                // *** this part shouldn't be necessary
                // the geo and name should come with the search results
                // but there's an SDK bug NCBC-1651
                var doc = _bucket.Get<dynamic>(hit.Id).Value;
                // ****************
                list.Add(new GeoSearchResult
                {
                    Latitude = doc.geo.lat,
                    Longitude = doc.geo.lon,
                    InfoWindow = new InfoWindow
                    {
                        Content = doc.name + "<br />" +
                            doc.city + ", " +
                            doc.state + " " +
                            doc.country
                    }
                });
            }
            return Ok(list);
        }

A "distance" search is another way to perform geospatial queries. This time, instead of a box, it will be more like a circle. You supply a single coordinate, and a distance. The distance will be the radius from that point.

public class PointSearch
{
    public double Latitude { get; set; }
    public double Longitude { get; set; }
    public int Distance { get; set; }
    // miles is being assumed as the unit
    public string DistanceWithUnits => Distance + "mi";
}

I’m defaulting it to miles, but certainly you can use kilometers instead, or present the option in the UI.

The endpoint will be very similar to the bounding box endpoint, except that it uses GeoDistanceQuery.

[Route("api/Point")]
[HttpPost]
public IActionResult Point([FromBody] PointSearch point)
{
    var query = new GeoDistanceQuery();
    query.Latitude(point.Latitude);
    query.Longitude(point.Longitude);
    query.Distance(point.DistanceWithUnits);
    var searchParams = new SearchParams()
        // .Fields("geo", "name") // omitting because of bug NCBC-1651
        .Limit(10)
        .Timeout(TimeSpan.FromMilliseconds(10000));
    var searchQuery = new SearchQuery
    {
        Query = query,
        Index = "mygeoindex",
        SearchParams = searchParams
    };
    var results = _bucket.Query(searchQuery);

    var list = new List<GeoSearchResult>();
    foreach (var hit in results.Hits)
    {
        // *** this part shouldn't be necessary
        // the geo and name should come with the search results
        // but there's an SDK bug NCBC-1651
        var doc = _bucket.Get<dynamic>(hit.Id).Value;
        // ****************
        list.Add(new GeoSearchResult
        {
            Latitude = doc.geo.lat,
            Longitude = doc.geo.lon,
            InfoWindow = new InfoWindow
            {
                Content = doc.name + "<br />" +
                          doc.city + ", " +
                          doc.state + " " +
                          doc.country
            }
        });
    }
    return Ok(list);
}

At this point, you can start testing these endpoint with Postman or Fiddler if you’d like. But it will be so much nice to see this on a map.

Auerlia and Google Maps

In Aurelia, I’ve created two components: geosearchbox and geosearchpoint.

Auerlia components

Each of them will have a Google Maps component that the user can interact with. These maps will be centered on San Francisco, because that’s where a lot of the hotels in "travel-sample" are located.

Bounding Box search component

The google-map` component has a map-click.delegate that will will fire whenever the users clicks on the map. In geosearchbox.html:

<google-map
    if.bind="markers"
    map-click.delegate="clickMap($event)"
    latitude="37.780986253433895"
    longitude="-122.45291600632277"
    zoom="12"
    markers.bind="markers">
</google-map>

markers is simply an array containing coordinates of search results that should appear on the map. Initially it will be empty.

When the user first clicks the map, this will set the first coordinate (top left) in the form. In geosearchbox.ts:

public clickMap(event : any) {
    var latLng = event.detail.latLng,
        lat = latLng.lat(),
        lng = latLng.lng();

    // only update top left if it hasn't been set yet
    // or if bottom right is already set
    if (!this.longitudeTopLeft || this.longitudeBottomRight) {
        this.longitudeTopLeft = lng;
        this.latitudeTopLeft = lat;
        this.longitudeBottomRight = null;
        this.latitudeBottomRight = null;
    } else {
        this.longitudeBottomRight = lng;
        this.latitudeBottomRight = lat;
    }
}

Then, click another spot on the map. This will set the second coordinate (bottom right).

My implementation is very bare bones. No fancy graphics and no validation of the second coordinate being to the bottom right of the first. The fields on a form will simply be populated with the latitude and longitude. In geosearchbox.html:

<p>
    Bounding box search:
    <br />
    Latitude (top left):
        <input type="text" value="${ latitudeTopLeft }" />
    Longitude (top left):
        <input type="text" value="${ longitudeTopLeft }" />
    <br />
    Latitude (bottom right):
        <input type="text" value="${ latitudeBottomRight }" />
    Longitude (bottom right):
        <input type="text" value="${ longitudeBottomRight }" />
    <br />
    <input
        if.bind="latitudeTopLeft && latitudeBottomRight"
        click.trigger="searchClick()"
        type="button"
        name="search"
        value="Search" />
</p>

Once you’ve selected two coordinates, a search button will appear. Click that to post these coordinates to the endpoint created earlier, and it will trigger the searchClick() method as seen in geosearchbox.ts:

public searchClick() {
    let boxSearch = {
        latitudeTopLeft: this.latitudeTopLeft,
        longitudeTopLeft: this.longitudeTopLeft,
        latitudeBottomRight: this.latitudeBottomRight,
        longitudeBottomRight: this.longitudeBottomRight
    };

    console.log("POSTing to api/Box: " + JSON.stringify(boxSearch));

    this.http.fetch('api/Box', { method: "POST", body: json(boxSearch) })
        .then(result => result.json() as Promise<any[]>)
        .then(data => {
            this.markers = data;
        });
}

When Aurelia, Google Maps, ASP.NET Core, and Couchbase all work together, it looks like this:

Geospatial bounding box

Distance Search

Implementing the "distance" geostatial query will be similar to the bounding box UI. This time, you only need to click a single point on the map. But, you will need to type in a distance (in miles).

The google-map component will look identical. The clickMap function is different:

public clickMap(event: any) {
    var latLng = event.detail.latLng,
        lat = latLng.lat(),
        lng = latLng.lng();

    this.longitude = lng;
    this.latitude = lat;
}

Specify a distance (in miles), and then click 'search' to make a POST request to the endpoint we wrote earlier.

geosearchbox.html:

    <p>
        Distance search:
        <br />
        Latitude: <input type="text" value="${ latitude }" />
        Longitude: <input type="text" value="${ longitude }" />
        <br />
        Distance (miles): <input type="text" value="${ distance }" />
        <br />
        <input if.bind="latitude" click.trigger="searchClick()" type="button" name="search" value="Search" />
    </p>

geosearchbox.ts:

    public searchClick() {
        let pointSearch = {
            latitude: this.latitude,
            longitude: this.longitude,
            distance: this.distance
        };

        console.log("POSTing to api/Point: " + JSON.stringify(pointSearch));

        this.http.fetch('api/Point', { method: "POST", body: json(pointSearch) })
            .then(result => result.json() as Promise<any[]>)
            .then(data => {
                this.markers = data;
            });
    }
}

Below is a clip of the search in motion. Note how the results change as I move the coordinate around.

Geospatial distance point search query

Summary

With Couchbase’s built-in geospatial indexing and search feature, all the math and the searching is delegated to the Couchbase Data Platform. So you can focus on building a killer UI (better than mine anyway) and rock-solid business logic.

Be sure to check out the documentation for a complete overview of the geospatial capabilities of Couchbase.

If you need help or have questions, please check out the Couchbase Server forums, and if you have any questions about the Couchbase .NET SDK, check out the .NET SDK forums.

If you’d like to get in touch with me, please leave a comment or find me on Twitter @mgroves.

Swashbuckle is a handy library to easily bring Swagger support to your ASP.NET Core (or ASP.NET) application. It is especially handy when developing an HTTP based API. It creates a form of interactive documentation based on the OpenAPI Specification.

Before diving into Swashbuckle: Merry Christmas! This blog is being posted on December 25th, 2017. It’s the final post of the very first C# Advent Calendar. Please check out the other 24 posts in the series! This event has gone so well, that I’m already planning on doing it again in 2018. Thank you, again, to everyone who participated (whether you are a writer or you’ve just been following along).

The full source code used in this example is available on Github.

ASP.NET Core HTTP API

I’m going to assume some level of familiarity with ASP.NET Core and creating a REST API. Here’s an example of a GET and a POST. These endpoints are reading/writing from a JSON text file (in a way that is probably not thread-safe and definitely not efficient, but it’s fine for this example).

public class ValuesController : Controller
{
    [HttpGet]
    [Route("api/teams")]
    public IActionResult GetTeams()
    {
        var jsonFile = System.IO.File.ReadAllText("jsonFile.json");
        var teams = JsonConvert.DeserializeObject<List<Team>>(jsonFile);
        return Ok(teams);
    }

    [HttpPost]
    [Route("api/team")]
    public IActionResult PostTeam([FromBody]Team team)
    {
        var jsonFile = System.IO.File.ReadAllText("jsonFile.json");
        var teams = JsonConvert.DeserializeObject<List<Team>>(jsonFile);
        teams.Add(team);
        System.IO.File.WriteAllText("jsonFile.json",JsonConvert.SerializeObject(teams));
        return Ok(team);
    }

    // etc...

To try out the GET endpoint, the simplest thing I can do is open a browser and view the results. But to try out the POST endpoint, I need something else. I could install Postman or Fiddler (and you should). Here’s how that would look.

Postman

Postman is great for interacting with endpoints, but Postman alone doesn’t really tell us anything about the endpoint or the system as a whole. This is where Swagger comes in.

Swagger

Swagger is a standard way to provide specifications for endpoints. Usually, that specification is automatically generated and then used to generate an interactive UI.

We could write the Swagger spec out by hand, but fortunately ASP.NET Core provides enough information to generate a spec for us. Look at the PostTeam action above. Just from reading that we know:

  • It expects a POST

  • The URL for it is /api/team

  • There’s a Team class that we can look at to see what kind of body is expected

From that, we could construct a Swagger spec like the following (I used JSON, you can also use YAML).

{
	"swagger": "2.0",
	"info": { "version": "v1", "title": "Sports API" },
	"basePath": "/",
	"paths": {
		"/api/team": {
			"post": {
				"consumes": ["application/json"],
				"parameters": [{
					"name": "team",
					"in": "body",
					"required": false,
					"schema": { "$ref": "#/definitions/Team" }
				}]
			}
		}
	},
	"definitions": {
		"Team": {
			"type": "object",
			"properties": {
				"name": { "type": "string" },
				"stadiumName": { "type": "string" },
				"sport": { "type": "string" }
			}
		}
	}
}

But why on earth would you want to type that out? Let’s bring in a .NET library to do the job. Install Swashbuckle.AspNetCore with NuGet (there’s a different package if you want to do this with ASP.NET).

You’ll need to add a few things to Startup.cs:

In the ConfigureServices method:

services.AddSwaggerGen(c =>
{
    c.SwaggerDoc("v1", new Info { Title = "Sports API", Version = "v1"});
});

In the Configure method:

app.UseSwagger();

Aside: With ASP.NET, NuGet actually does all this setup work for you.

Once you’ve done this, you can open a URL like http://localhost:9119/swagger/v1/swagger.json and see the generated JSON spec.

Swagger spec in JSON

Swagger UI with Swashbuckle

That spec is nice, but it would be even nicer if we could use the spec to generate a UI.

Back in the Configure method, add this:

app.UseSwaggerUI(c =>
{
    c.SwaggerEndpoint("/swagger/v1/swagger.json", "Sports API v1");
});

Now, open your site and go to /swagger:

Swagger UI generated by Swashbuckle

Some cool things to notice:

  • Expand/collapse by clicking the URL of an endpoint (note that you must use Route attributes for Swashbuckle to work with ASP.NET Core).

  • "Try it out!" buttons. You can execute GET/POST right from the browser

  • The "parameter" of the POST method. Not only can you paste in some content, but you get an example value that acts like a template (just click it).

Giving some swagger to your Swagger

Swagger and Swashbuckle have done a lot with just a little bit. It can do even more if we add a little more information in the code.

  • Response: The ProducesResponseType attribute will let Swagger know what the response will look like (this is especially useful if you are using IActionResult and/or an endpoint could return different types in different situations).

  • Comments: If you are using XML comments, you can have these included with the Swagger output.

services.AddSwaggerGen(c =>
{
    c.SwaggerDoc("v1", new Info { Title = "Sports API", Version = "v1" });
    var filePath = Path.Combine(PlatformServices.Default.Application.ApplicationBasePath, "swashbuckle-example.xml");
    c.IncludeXmlComments(filePath);
});

(Also make sure you XML Documentation output for your project enabled)

Here’s an example of a GetTeams method with both XML comments and ProducesResponseType:

/// <summary>
/// Gets all the teams stored in the file
/// </summary>
/// <remarks>Baseball is the best sport</remarks>
/// <response code="200">List returned succesfully</response>
/// <response code="500">Something went wrong</response>
[HttpGet]
[Route("api/teams2")]
[ProducesResponseType(typeof(Team), 200)]
public IActionResult GetTeams2()
{
    var jsonFile = System.IO.File.ReadAllText("jsonFile.json");
    var teams = JsonConvert.DeserializeObject<List<Team>>(jsonFile);
    return Ok(teams);
}
  • Customize your info: there’s more to the Info class than just Title and Version. You can specify a license, contact, etc.

services.AddSwaggerGen(c =>
{
    c.SwaggerDoc("v1", new Info
    {
        Title = "Sports API",
        Version = "v1",
        Description = "An API to list and add sports teams",
        TermsOfService = "This is just an example, not for production!",
        Contact = new Contact
        {
            Name = "Matthew Groves",
            Url = "https://crosscuttingconcerns.com"
        },
        License = new License
        {
            Name = "Apache 2.0",
            Url = "http://www.apache.org/licenses/LICENSE-2.0.html"
        }
    });
    var filePath = Path.Combine(PlatformServices.Default.Application.ApplicationBasePath, "swashbuckle-example.xml");
    c.IncludeXmlComments(filePath);
});

Here’s a screenshot of the UI that has all three of the above enhancements: response type, XML comments, and more customized info.

Swagger with more swagger

Summary

Working on HTTP-based APIs? Bring Swashbuckle and Swagger into your life!

More resources:

Thanks again for reading the 2017 C# Advent!

This is a repost that originally appeared on the Couchbase Blog: ASP.NET Core with Couchbase: Getting Started.

ASP.NET Core is the newest development platform for Microsoft developers. If you are looking for information about plain old ASP.NET, check out ASP.NET with Couchbase: Getting Started.

ASP.NET Core Tools to Get Started

The following video will take you from having no code to having an HTTP REST API that uses Couchbase Server, built with ASP.NET Core.

These tools are used in the video:

Getting Started Video

In the video, I touch quickly on Scan Consistency. For more details on that, check out the Scan Consistency documentation or read a blog post that I wrote introducing AtPlus, which also covers the other types of Scan Consistency.

Summary

This video gives you the absolute minimum to get started with Couchbase by walking you through a simple CRUD application.

If you have any questions, please leave a comment. Or, you can always ask me questions on Twitter @mgroves.

UPDATE: The calendar is full. You can sign up to be an alternate, in case someone drops out or fails to deliver. I can't give you a date though, so you'd essentially have to have a post ready to go as soon as December 1st. And please, don't let the lack of an advent date keep you from writing that C# blog post! Finally, due to the tremendous response, I will double up the slots next year (from 25 to 50), assuming this advent goes well :)

I heard about the F# Advent Calendar, a traditional that's been carried on since 2010 (2014 in English). I think this is a great idea, and there needs to be one for C# too!

(I asked Sergey Tihon for permission!)

So, I need you to write a C# blog post!

Here are the rules:

  1. Reserve a slot on Twitter (with hash tag #csadvent) or leave a comment on this post. You do not have to announce your topic until the day you reserve.
  2. Prepare a blog post (in English).
  3. Add a link in your blog post that links back to here, so that your readers may find the entire advent.
  4. Publish your blog post on the specified date. Your post must be related to C# in some way, but otherwise the content is completely up to you. I've posted a few ideas below to get your creativity flowing.
  5. Post the link to your post on Twitter with hashtags #csharp and #csadvent

Below are all the slots, and who has claimed each date. I chose to go with 25 total slots (yes I know some advent calendars are 24).

I will do my best to keep this up to date. The slots will be first come first serve. I have already claimed December 25th for myself, since I assume that will be the least desirable date. But I'm happy to swap if you really really want that day.

DateClaimed byBlog Post
Dec 1, 2017 Hilary Weaver-Robb Combining Integration and UI Automation in C#
Dec 2, 2017 Jamie Rees Open Source, 2 years retrospective
Dec 3, 2017 Lucas Lansky Analyzing unit-ness of white-box tests using OpenCover
Dec 4, 2017 Andrei Ignat AOP with Roslyn–part 3–custom code at beginning of each method
Dec 5, 2017 Baskar rao Dsn Quick Actions in Visual Studio 2017
Dec 6, 2017 Bill Sempf Coding for an encrypted service
Dec 7, 2017 Brant Burnett Who says C# interfaces can't have implementations?
Dec 8, 2017 Lee Englestone Creating Alexa Skills with Web-API and hosting on Microsoft Azure
Dec 9, 2017 James Hickey Deck the Halls With Strategy Pattern Implementations in C#: Basic to Advanced
Dec 10, 2017 Carl Layton Using C# dynamic Keyword To Replace Data Transfer Objects
Dec 11, 2017 James Curran An Exercise in Refactoring - Specification Pattern
Dec 12, 2017 Andrew Lock Creating a .NET Standard Roslyn Analyzer in Visual Studio 2017
Dec 13, 2017 Jonathan Danylko 5 More C# Extension Methods for the Stocking! (plus a bonus method for enums)
Dec 14, 2017 Gérald Barré (aka Meziantou) Interpolated strings: advanced usages
Dec 15, 2017 Baskar rao Dsn Using BenchMarkDotNet for Performance BenchMarking
Dec 16, 2017 Chris Bohatka Breaking Down Dependency Injection and Inversion of Control in C# with SimpleInjector
Dec 17, 2017 David Pine C# Special Edition
Dec 18, 2017 Barret Blake Using C# and Azure Cognitive Services Text Analytics to Identify and Relate Text Documents
Dec 19, 2017 Ed Charbeneau 5 .NET Standard Features You Shouldn’t Miss
Dec 20, 2017 Kevin Miller Things you Should do with Strings While You’re Coworkers are on Holiday and No One is Checking the Production Code Branch
Dec 21, 2017 Angus Kinsey What makes C# great?
Dec 22, 2017 Jim Wilcox Unloading the UI Thread in C# on Windows 10 + UWP
Dec 23, 2017 Duane Newman All I want for Christmas is a C# Build System
Dec 24, 2017 Calvin Allen Live Unit Testing in Visual Studio 2017
Dec 25, 2017 Matthew D. Groves Swashbuckle and Swagger with ASP.NET Core

Alternates:

  • Leave a comment or tweet with #csadvent to be put on this list!

Some ideas:

  1. Introduction to [your favorite NuGet package]
  2. X C# Tricks and Tips
  3. JSON (De)serialization
  4. Streams
  5. The newer C# 6 and 7 features
  6. Async/await

Thanks to everyone who is participating! I hope this goes well and it can become an annual tradition.

This is a repost that originally appeared on the Couchbase Blog: Azure Functions and Lazy Initialization with Couchbase Server.

Azure Functions are still new to me, and I’m learning as I’m going. I blogged about my foray into Azure Functions with Couchbase over a month ago. Right after I posted that, I got some helpful feedback about the way I was instantiating a Couchbase cluster (and bucket).

I had (wrongly) assumed that there was no way to save state between Azure Function calls. This is why I created a GetCluster() method that was called each time the function ran. But, initializing a Couchbase Cluster object is an expensive operation. The less often you instantiate it, the better.

You can follow along with the updated source code for this blog post on Github.

Static state

I had a hard time finding documentation on whether I could use a static object for reuse between function calls. I suppose I should have experimented, like fellow Microsoft MVP Mark Heath did. Instead, I posed the question to StackOverflow.

In short: yes. A Cluster, instantiated and saved to a static member, can is reusable between function calls. According to Mark’s post above, there’s no guarantee how long this value will survive. But that’s an expected trade-off that you make when going "serverless".

Lazy initializing within Azure Functions

Simply using a static member would work, but it’s not thread-safe. There are a few ways to tackle that issue, but an easy way that’s built right into the .NET framework is to use Lazy Initialization with Lazy<T>.

Lazy Initialization in Azure Functions

First, I removed the GetBucket and GetCluster methods. Next, I created a Lazy<IBucket> property to replace them.

private static readonly Lazy<IBucket> Bucket = new Lazy<IBucket>(() =>
{
    var uri = ConfigurationManager.AppSettings["couchbaseUri"];
    var bucketName = ConfigurationManager.AppSettings["couchbaseBucketName"];
    var bucketPassword = ConfigurationManager.AppSettings["couchbaseBucketPassword"];
    var cluster = new Cluster(new ClientConfiguration
    {
        Servers = new List<Uri> { new Uri(uri) }
    });
    return cluster.OpenBucket(bucketName, bucketPassword);
});

I just made a single property for a bucket, since that’s all I need for this example. But if you need to use the cluster, you can easily make that its own Lazy property. (Once you have a cluster, getting a bucket is a relatively cheap operation).

Using a Lazy property

When you instantiate a Lazy<T> object, you supply it with an initialization lambda. That lambda won’t execute until the Value property is actually called for the first time.

var lazyObject = new Lazy<string>(() =>
{
    // this code won't be called until 'lazyObject.Value' is referenced
    // for the first time
    return "I'm lazy!";
});

For instance, notice the Value between Bucket and GetAsync in the updated version of my Azure Functions:

var doc = await Bucket.Value.GetAsync<MyDocument>(id);

If that’s the first time Value is used, the cluster will be initialized. Otherwise, it will use the already initialized cluster (try experimenting with a Guid instead of a Bucket).

Summary

State can be saved between Azure Function calls by using a static member. Make sure that it’s thread-safe (by using Lazy<T> or something like it). Don’t make any assumptions about how long that object will be around.

Anything else I missed? Are you using Azure Functions with Couchbase? I would love to hear from you. Please leave a comment below or ping me on Twitter @mgroves.

Matthew D. Groves

About the Author

Matthew D. Groves lives in Central Ohio. He works remotely, loves to code, and is a Microsoft MVP.

Latest Comments

Twitter