fauna / faunadb-csharp Goto Github PK

Package FaunaDB.Client 1.1.0 is not compatible with netcoreapp1.1 (.NETCoreApp,Version=v1.1). Package FaunaDB.Client 1.1.0 supports: net45 (.NETFramework,Version=v4.5)

Please add support to dotnet core. It was used in modern platform and especially Amazon Lambda

Hey folks,

Just came across FaunaDb. While checking out this driver I noticed that it seems to create a new HttpClient every time.

A lot of the serverless platforms (FaaS ones like Lambda and Azure Functions) have limits on the number of connections that can be created inside a function and inappropriate usage HttpClient is frequently the cause of many errors.

The old way to fix this was to use a static http client but that has problems as well. The newer way is to use an HttpClientFactory which does a much better job of managing connections.

Anyways. Just wanted to bring that up. The serverless tier of FaunaDb is what caught my eye and this is a frequent gotcha for those scenarios.

Hi,

Regarding JSON serializer, I see that FaunaDB.Client is using Newtonsoft.Json !!
Check out the System.Text.Json package, I highly recommend using it, it is Fast plus it's owned and maintained by Microsoft.

Starting ASP.NET Core 3.1, System.Text.Json is the default JSON Serializer, and Newtonsoft.Json will be removed starting .NET 5.

I'm new to Fauna and having a VERY hard time understanding how to query and return a collection of objects.

In EntityFramework I would access a property on my database connection like:

dbContext.MyCollectionName

and be able to interact with the contents with LINQ syntax. I struggle to find the right combination of Query(Paginate(Match()))).To<MyObject[]>().Value to get objects read from Fauna. Can we put together a simple example document that shows Fauna 101 for each of the CRUD operations?

I have created a custom type called User, which utilises the FaunaField reflection method for specifying properties of the objects.

Please provide more information on how to decode arrays of custom types, as this id poorly described in your documentation for the C# driver.

The problem can be repreduced from the following example.

Example:

public class User
    {
        [FaunaField("id")]
        public string Id { get; set; }
        [FaunaField("name")]
        public string Name { get; set; }
        ...
    } 

I then created a mutation for creating a new document in my 'user' collection, which works fine, and returns the expected result.

Example:

var result = await _fauna.client.Query(
                    Create(
                        Collection("user"),
                        Obj("data", Encoder.Encode(upsertUser))));
                
return Decoder.Decode<User>(result.At("data"));

Now my issue is retrieving the full list of user documents in my 'users' collection. I verified that i retrieve all the right data, but once i decode the data with Decoder.Decode or <<....>>.To<User[]>() i get NULL values.

Example:

public async Task<User[]> GetUser()
        {
            var result = await _fauna.client.Query(
                Map(
                    Paginate(
                        Documents(Collection("user"))), 
                    Lambda("X", Get(Var("X")))));
            
            var userResult = result.At("data").To<User[]>().Value;

            return userResult;
        }

Raw Output:

Arr(
  ObjectV(
    ref: RefV(id = "286263727644213761", 
    collection = RefV(id = "user", collection = RefV(id = "collections"))),
    ts: LongV(1609261214770000),
    data: ObjectV(id: StringV(d9853632-7b8f-42d1-b217-eb314e58d417),
      name: StringV(Test User),
      email: StringV([email protected]),
  ObjectV(
    ref: RefV(id = "286263756128780801", 
    collection = RefV(id = "user", collection = RefV(id = "collections"))),
    ts: LongV(1609261241940000),
    data: ObjectV(id: StringV(56a06ea3-ef0b-4ded-a627-ceab63385fe2),
      name: StringV(Test User),
      email: StringV([email protected]),
    )
)

Decoded Output:

{
  "data": {
    "user": [
      {
        "id": null,
        "name": null,
        "email": null,
        "access": null
      },
      {
        "id": null,
        "name": null,
        "email": null,
        "access": null
      }
    ]
  }
}

Expected Output:

{
  "data": {
    "user": [
      {
        "id": "Sample ID",
        "name": "Sample Name",
        "email": "Sample Email",
      },
      {
        "id": "Sample ID",
        "name": "Sample Name",
        "email": "Sample Email",
      }
    ]
  }
}

After learning about faunadb API for sometimes I felt unfamiliar with the style of API

Such as

client.Query(
  Paginate(Match(Index("spells_by_element"), "fire")));

client.Query(
  Do(
    Create(
      Ref(Class("magical_creatures"), 1),
      Obj("data", Obj("name", "Hen Wen"))),
    Get(Ref(Class("magical_creatures"), 1))));

I felt that some functions that require Expr as parameter should be extension method. And so we could write it like this

client.Query(Index("spells_by_element").Match("fire").Paginate());

client.Query(Do(Class("magical_creatures").Ref(1).Create(Obj("data", Obj("name", "Hen Wen")))
     ,Class("magical_creatures").Ref(1).Get()));

/// or

client.Query(Class("magical_creatures").Ref(1).Create(Obj("data", Obj("name", "Hen Wen")))
    .Then(Class("magical_creatures").Ref(1).Get()));

Could we have a sub namespace for this style api? (FaunaDB.LINQ maybe?)

StringV is sealed class, so the only way to create a derived class is to inherit from ScalarValue<string>. The internal Decoder works one way only: ScalarValue<string> Derived Class to StringV but not the other way around.

Digging deeper into internal decoder implementation, the root cause is the usage of

typeof(StringV).IsAssignableFrom(typeof(DerivedClass))
typeof(DerivedClass).IsAssignableFrom(typeof(StringV))

which in both ways evaluates to false and throws Invalid Cast Exception.

This can be fixed in one of two ways:

1. Making StringV not sealed
2. Replacing IsAssignableFrom check with string or ScalarValue<string> check.

Save instance of User without filling out one of FirstName, LastName or Login makes Decoder.Decode throw exception with message: Missing required property: PropertyName

public class User 
{
        public string FirstName { get; set; }
        public string LastName { get; set; }
        public Login Login { get; set; }
    }

    public class Login
    {
        public LoginTypeEnum LoginType { get; set; }

        public string Username { get; set; }
        public string PasswordHash { get; set; } 
        public string FacebookToken { get; set; }
    }

It would be nice if Decoder handles properties with null values simply by setting these to null when calling Decode.

would be nice if you could have a Uri be encoded as a string into faunadb. this can be accomplished similar to the [FaunaDate] and [FaunaTime] attributes. You can have a [FaunaString] that encodes the object by calling the .ToString() and the decoder calls a constructor with a single parameter, string.

Same as fauna/faunadb-ruby#135

User's feedback:

C# driver: Noticing that enums are getting encoded as int's. Would it be possible to configure the Encoder to use the string value for an enum instead?
This makes it much easier when quering using the Query Console.

In addition to DateTime. class DateTimeOffset is more useful for working with time in C#. Please add support to it like DateTime and please make implicit conversion for these classes

Consider the following structure

public class Person
{
    public string Name { get; set; }
    public Dictionary<string, object> Metadata { get; set; }
}

The encoder doesn't seem to have an issue and serializes the dictionary correctly; here's the encoded value ToString().

ObjectV(name: StringV(Japanfy),metadata: ObjectV(comm.status: StringV(connected),comm.satellites: LongV(8)))

However, the decoder doesn't behave well with this and doesn't decode the values of the dictionary. It just creates an empty object for each.

{"Name":"Japanfy","Metadata":{"comm.status":{},"comm.satellites":{}}}

Because C# has the limitation of 100 nanosecond precision on the DateTime object there is no way to add nanosecond time to fauna. Sometimes a user may want to store a nanosecond precision Time via a Unix epoch timestamp. Fauna has the Epoch function that can accomplish this. The encoder could take a long and run it through that function after detecting a [FaunaEpoch(EpochPrecision.Nanosecond)] attribute. Then the decoder would convert fauna time into a ts long value.

My Intellisense is complaining "No overload for method 'Create' takes two arguments" with the following snippet:

serverClient.Query(
  Create(
    Collection("posts"),
    Obj("data", Obj("title", "What I had for breakfast .."))));

I'm following the docs exactly. I'm not sure where this function comes from. Also, if you navigate to 'Create' in the official csharp docs, 'CreateRole' is instead the example code snippet displayed':

https://docs.fauna.com/fauna/current/api/fql/functions/create?lang=csharp

...what? 😑

There is a good documentation about this process on nuget page
https://docs.nuget.org/create/creating-and-publishing-a-package

If a non 2XX error code is received, then resultRequest.ResponseContent is deserialized by Newtonsoft.Json.

However, the ResponseContent received is sometimes not JSON. One user received the following error; it's possible that the ResponseContent is html rather than a JSON string.

Error :Newtonsoft.Json.JsonReaderException: Unexpected character encountered while parsing value: <. Path '', line 0, position 0.

It may be helpful to handle this error more gracefully and provide the user with a better message.

Intellisense is not available for any objects in this package, and that makes discoverability of the features and interactions with the Fauna client very difficult for new developers

I would like to know whether there would ever be a faunadb extension for EFcore since it's one of the main ways that people manipulate databases in .Net

The follow links seems to be a good starting point:
https://docs.nuget.org/create/enforced-package-conventions
http://blogs.msmvps.com/punitganshani/2015/06/21/5-steps-to-targeting-multiple-net-frameworks/

This library isn't working on Blazor WASM because of this 'x-query-timeout' header usage.

All HTTPClient calls get exceptions like:
Access to fetch at 'https://db.fauna.com/' from origin 'https://.....' has been blocked by CORS policy: Request header field x-query-timeout is not allowed by Access-Control-Allow-Headers in preflight response.

Until this issue is solved on the FaunaDB cloud side you can use the following workaround:
Remove the 'x-query-timeout' header in the request.

		public class MessageHandler : DelegatingHandler
		{
			public MessageHandler(HttpClientHandler httpClientHandler) :
				base(httpClientHandler) 
			{  }

			protected override Task<HttpResponseMessage> SendAsync(HttpRequestMessage request, CancellationToken cancellationToken)
			{
				// Dont use header 'X-Query-Timeout' for CORS policy
				request.Headers.Remove("X-Query-Timeout");
				return base.SendAsync(request, cancellationToken);
			}
		}

		public async void ContactFaunaDb()
		{
			var httpClient = new HttpClient(new MessageHandler(new HttpClientHandler()));
			var faunaDb = new FaunaClient(endpoint: Endpoint, secret: Secret, httpClient: httpClient);

			var rawIndexValues = await faunaDb.Query(Paginate(Match(Index("some_index"))));
			/// etc....
		}

This is a feature request. JSON convention says field names should start lowercase, while C# Properties start with an uppercase letter. Newtonsoft.Json converts the casing when serializing/deserializing, the FaunaDB driver should do so as well. Right now I either have to lowercase my model properties, or uppercase JSON fields, due to the lack of conversion (neither of which is a satisfactory solution).