Skip to content

An example API wrapper for Twitter using Hammock and T4

License

Notifications You must be signed in to change notification settings

jacobavenkraft/tweetsharp

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

87 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Tweetsharp - Short, sweet, social.

PM> Install-Package TweetSharp

Introduction

TweetSharp is a Twitter API library that greatly simplifies the task of adding Twitter to your desktop, web, and mobile applications. You can build simple widgets, or complex application suites using TweetSharp. The second version, a rewrite, was designed to be lighter, faster, and more intuitive than the original. You write fewer lines of code, make fewer decisions, and get better results. Visual Studio T4 templates are employed to automatically generate new API methods from a simple text-based DSL.

Caveat Tractor / Public Service Announcement

This project is open source software, not a commercial product, so please curb your enthusiasm around expecting support or feature requests. After two and a half years of life and two years of active development, this project is no longer supported freely. I am happy to accept any reasonable pull requests, and the code is written with Visual Studio T4 templates, making it ridiculously easy to extend for any API methods that Twitter introduces in the future.

Due to popular request, I will offer commercial support services, depending on the scope and time requirements of your project (spare time only, nothing mission critical, etc.); you may inquire about support requests by contacting me through GitHub. For the majority of cases, though, you should be able to get on fine by forking, adding your desired API changes through the DSL, and issuing a pull request.

API wrappers, in general, are a maintenance nightmare for both the consumer and the maintainer (this particular project tries very hard to be maintainable through code generation, DSLs, and a flexible data serialization scheme, but nothing is perfect), and are a malnourished response to the fact that most APIs are not truly RESTful; if they were, we could consume them at the same level of abstraction as a browser, rather than constantly creating and maintaining bespoke mappings across service boundaries. If you would rather maintain your own micro-libraries at the service level using a general purpose HTTP API client, I suggest you use Hammock or RestSharp.

Learn the Twitter API

Make sure you visit (http://dev.twitter.com) to get acquainted with the Twitter API. Most of the time, confusion around the methods in this library are a result of not understanding Twitter's requirements, or the OAuth authentication workflow.

Hello, Twitter

using TweetSharp;

TwitterService service = new TwitterService();
IEnumerable<TwitterStatus> tweets = service.ListTweetsOnPublicTimeline();
foreach (var tweet in tweets)
{
    Console.WriteLine("{0} says '{1}'", tweet.User.ScreenName, tweet.Text);
}

OAuth Authentication

The first step to accessing the Twitter API is to create an application at (http://dev.twitter.com). When that process is complete, your application is issued a Consumer Key and Consumer Secret. These tokens are responsible for identifying your application when it is in use by your customers. Once you have these values, you can create a new service and pass them in.

Authenticating a client application (i.e. desktop)

using TweetSharp;

// Pass your credentials to the service
TwitterService service = new TwitterService("consumerKey", "consumerSecret");

// Step 1 - Retrieve an OAuth Request Token
OAuthRequestToken requestToken = service.GetRequestToken();

// Step 2 - Redirect to the OAuth Authorization URL
Uri uri = service.GetAuthorizationUri(requestToken);
Process.Start(uri.ToString());

// Step 3 - Exchange the Request Token for an Access Token
string verifier = "123456"; // <-- This is input into your application by your user
OAuthAccessToken access = service.GetAccessToken(requestToken, verifier);

// Step 4 - User authenticates using the Access Token
service.AuthenticateWith(access.Token, access.TokenSecret);
IEnumerable<TwitterStatus> mentions = service.ListTweetsMentioningMe();

Authenticating a browser application

using TweetSharp;

public ActionResult Authorize()
{
    // Step 1 - Retrieve an OAuth Request Token
    TwitterService service = new TwitterService("consumerKey", "consumerSecret");
    
    // This is the registered callback URL
    OAuthRequestToken requestToken = service.GetRequestToken("http://localhost:9090/AuthorizeCallback"); 
    
    // Step 2 - Redirect to the OAuth Authorization URL
    Uri uri = service.GetAuthorizationUri(requestToken);
    return new RedirectResult(uri.ToString(), false /*permanent*/);
}

    // This URL is registered as the application's callback at http://dev.twitter.com
    public ActionResult AuthorizeCallback(string oauth_token, string oauth_verifier)
    {
        var requestToken = new OAuthRequestToken {Token = oauth_token};
        
        // Step 3 - Exchange the Request Token for an Access Token
        TwitterService service = new TwitterService(_consumerKey, _consumerSecret);
        OAuthAccessToken accessToken = service.GetAccessToken(requestToken, oauth_verifier);

        // Step 4 - User authenticates using the Access Token
        service.AuthenticateWith(accessToken.Token, accessToken.TokenSecret);
        TwitterUser user = service.VerifyCredentials();
        ViewModel.Message = string.Format("Your username is {0}", user.ScreenName);
        return View();
}

xAuth Authentication

If you are building a mobile application and want to benefit from a seamless authentication experience with no additional steps for the user, you need to enroll your application in Twitter's xAuth support. You must complete this step in order for xAuth to function correctly.

using TweetSharp;

// OAuth Access Token Exchange
TwitterService service = new TwitterService("consumerKey", "consumerSecret");
OAuthAccessToken access = service.GetAccessTokenWithXAuth("username", "password");

OAuth Delegation with Echo

Twitter provides OAuth Echo support, which allows you to use other services like TwitPic by delegating the user's existing credentials. TweetSharp uses Hammock both internally and as a tool for you to make delegated requests. This example shows how you would use TweetSharp and Hammock together to post an image on TwitPic using OAuth.

using TweetSharp;
using Hammock;

TwitterService service = new TwitterService(consumerKey, consumerSecret);
service.AuthenticateWith("accessToken", "accessTokenSecret");

// Prepare an OAuth Echo request to TwitPic
RestRequest request = service.PrepareEchoRequest(); 
request.Path = "uploadAndPost.xml";
request.AddFile("media", "failwhale", "failwhale.jpg", "image/jpeg");
request.AddField("key", "apiKey"); // <-- Sign up with TwitPic to get an API key
request.AddField("message", "Failwhale!");

// Post photo to TwitPic with Hammock
RestClient client = new RestClient { Authority = "http://api.twitpic.com/", VersionPath = "2"};
RestResponse response = client.Request(request);

Discovering API Methods

TweetSharp uses a consistent method naming convention to help you locate the method you're looking for. In general, methods that return multiple results begin with List, while methods that return a single result begin with Get. Most methods, like the Twitter API, have additional parameters for obtaining pages of results rather than the default count. Keep in mind that paging methods have limits that you can confirm at (http://dev.twitter.com). Here's a sample of some of the most common Twitter API methods:

using TweetSharp;

TwitterStatus GetTweet(long id);
TwitterStatus SendTweet(string text);

IEnumerable<TwitterStatus> ListTweetsOnPublicTimeline();
IEnumerable<TwitterStatus> ListTweetsOnHomeTimeline();
IEnumerable<TwitterStatus> ListTweetsOnHomeTimeline(int count);
IEnumerable<TwitterStatus> ListTweetsOnHomeTimeline(int page, int count);

IEnumerable<TwitterStatus> ListTweetsOnHomeTimelineBefore(long maxId);
IEnumerable<TwitterStatus> ListTweetsOnHomeTimelineBefore(long maxId, int count)
IEnumerable<TwitterStatus> ListTweetsOnHomeTimelineBefore(long maxId, int page, int count);

IEnumerable<TwitterStatus> ListTweetsOnHomeTimelineSince(long sinceId);
IEnumerable<TwitterStatus> ListTweetsOnHomeTimelineSince(long sinceId, int count);
IEnumerable<TwitterStatus> ListTweetsOnHomeTimelineSince(long sinceId, int page, int count);

Dealing with Twitter API Rate Limiting

Twitter limits the frequency of all API calls in a variety of ways, to help ensure the service is not abused. This means that your applications will have to account for possible rate limit shortages at the user and IP address level. Client applications that are meant for public consumption usually use the rate limit profile of users that are logging in to their application. Server-side integration applications usually use their own, white-listed account's rate limit so that they can process large jobs without exhausting their allowance. You can find out more about rate limiting at (http://dev.twitter.com/pages/rate_limiting_faq) and (http://dev.twitter.com/pages/rate_limiting).

TweetSharp provides two ways to access rate limiting data. You can either make an explicit API call to retrieve it, or inspect the TwitterResponse's RateLimitStatus property, if it's available. The latter option conserves HTTP traffic, as the information is embedded in the HTTP Response itself.

using TweetSharp;

TwitterService service = new TwitterService("consumerKey", "consumerSecret");
service.AuthenticateWith("accessToken", "accessTokenSecret");

// Option 1 - Retrieve from the API
TwitterRateLimitStatus rate = service.GetRateLimitStatus();
Console.WriteLine("You have used " + rate.RemainingHits + " out of your " + rate.HourlyLimit);

// Option 2 - Retrieve from the response
IEnumerable<TwitterStatus> mentions = service.GetMentions();
TwitterRateLimitStatus rate = service.Response.RateLimitStatus;
Console.WriteLine("You have used " + rate.RemainingHits + " out of your " + rate.HourlyLimit);

Asynchronous Methods

TweetSharp supports executing methods asynchronously, and provides two styles of operation. The first is a delegate style, where you pass an Action into the named method after any optional parameters. This Action provides you with both the expected response that you would get if you called the method sequentially, as well as the response info you would have accessed on TwitterService's Response property.

Asynchronous operation (delegate style)

using TweetSharp;

TwitterService service = new TwitterService();
IAsyncResult result = service.ListTweetsOnPublicTimeline(
    (tweets, response) =>
        {
            if(response.StatusCode == HttpStatusCode.OK)
            {
                foreach (var tweet in tweets)
                {
                    Console.WriteLine("{0} said '{1}'", tweet.User.ScreenName, tweet.Text);
                }
            }
        });

In addition to delegate-based asynchronous methods, TweetSharp lets you simplify asynchronous operations with the familiar .NET Begin/End pattern. This style of operation involves calling the same methods as the synchronous style, but prefixing the method with Begin. Similarly, to retrieve the results you were looking for, you call the appropriate method beginning with End, with the option to provide a timeout value.

Asynchronous operation (begin/end style)

using TweetSharp;

var service = new TwitterService();
IAsyncResult result = service.BeginListTweetsOnPublicTimeline();
IEnumerable<TwitterStatus> tweets = service.EndListTweetsOnPublicTimeline(result);

foreach (var tweet in tweets)
{
    Console.WriteLine("{0} said '{1}'", tweet.User.ScreenName, tweet.Text);
}

Using Windows Phone 7

TweetSharp is designed with Windows Phone 7 in mind. Each sequential method on TwitterService also has an asynchronous equivalent for Windows Phone 7. Rather than expect a response, each method asks for a delegation Action to perform, which provides the expected result, as well as a wrapper class to help you handle unexpected results in your application.

using TweetSharp;

Dispatcher dispatcher = Deployment.Current.Dispatcher;
TwitterService service = new TwitterService("consumerKey", "consumerSecret");
service.AuthenticateWith("accessToken", "accessTokenSecret");

// Example: Getting Mentions
service.ListTweetsMentioningMe((statuses, response) =>
{
    if(response.StatusCode == HttpStatusCode.OK)
    {
        foreach (var status in statuses)
        {
            TwitterStatus tweet = status;
            dispatcher.BeginInvoke(() => tweets.Items.Add(tweet));
        }
    }
    else
    {
        throw new Exception(response.StatusCode.ToString());
    }
});

// Example: Posting a Tweet
service.SendTweet("Tweeting with #tweetsharp for #wp7", (tweet, response) =>
{
    if (response.StatusCode == HttpStatusCode.OK)
    {
        dispatcher.BeginInvoke(() => tweets.Items.Add(tweet));
    }
    else
    {
        throw new Exception(response.StatusCode.ToString());
    }
});

Data Format Handling

By default, TweetSharp handles serialization and deserialization details for you, preferring JSON for its compact size, which leads to better performance. If you want to switch TweetSharp's internal serialization mechanism to XML, you can do this by changing the TwitterService's Format enum property to TwitterServiceFormat.Xml. You may want to do this if your application requires lower level control of the content returned by Twitter. If you change the format, the RawSource property on each model object will contain the format you selected. On Windows Phone 7, JSON is the only supported format.

using TweetSharp;

TwitterService service = new TwitterService("consumerKey", "consumerSecret");
service.Format = TwitterServiceFormat.Xml;

If you go one step further and decide you don't trust our serializer, you can change TwitterService's Serializer and Deserializer properties, setting them to Hammock-compatible interfaces, and TwitterService will then defer to your custom serializer in all requests.

using TweetSharp;
using Hammock.Serialization;

MyAwesomeSerializer serializer = new MyAwesomeSerializer();
TwitterService service = new TwitterService("consumerKey", "consumerSecret");
service.Serializer = serializer;
service.Deserializer = serializer;

Handling Errors

There are four ways of handling errors at the Twitter API level.

  • You can use the TwitterResponse object to inspect details about the request and act accordingly. This object is available sequentially using TwitterService's Response property, which means the Response property will update after each API call is complete. If you're using Windows Phone 7, you get the TwitterResponse object passed into each Action delegate, so you know the response you're accessing belongs to the request that's returning through the callback.

  • TweetSharp uses a relaxed JSON parsing strategy to mitigate exceptions when API objects change. This means that you won't receive a null response if your API call returns an error instead of the expected object; you'll get the object with default initialization. This means, for major objects like TwitterUser and TwitterStatus, that if the Id property is 0, something went wrong.

  • You can use TwitterService's Deserialize<T>(ITwitterModel model) method to attempt to cast any result you get back from TwitterService into a TwitterError object. This will help you capture any details you get back from the Twitter API explicitly. You can check the ErrorMessage property to ensure the JSON parser returned a valid error message hash.

  • If you're not confident with the deserialization of your object, you can use the RawSource property that exists on all Twitter model objects to inspect the actual JSON response that was returned by Twitter, specific to that object. This means if you returned a collection of tweets, each tweet's RawSource will contain the JSON for that specific tweet. This is helpful if you want to perform custom parsing or tracing of the raw data.

Error Handling Examples (sequential service calls)

using TweetSharp;

TwitterService service = new TwitterService(_consumerKey, _consumerSecret)();

// Missing authentication; this call will fail
IEnumerable<TwitterStatus> mentions = service.ListTweetsMentioningMe(); 

// Strategy 1 - Look for bad requests by inspecting the response for important info
if(service.Response.StatusCode == HttpStatusCode.OK) // <-- Should be 401 - Unauthorized
{
    // Strategy 2 - If you get back an 200 - OK response, you might have received an error, not the objects you wanted;
    // if you were trying to get a collection, any errors are added to it, so look for only one collection item
    if(mentions.Count() == 1) // <-- result
    {
        // Serialization failures will result in "bare" objects; if this tweet's Id is zero, 
        // you know it either failed to deserialize properly, or, in the case of requesting collections, 
        // it contains the error returned from Twitter in the first and only result.

        TwitterStatus mention = mentions.First();
        if(mention.Id == 0)
        {
            // This was not a successful deserialization...
        }

        // Strategy 3 - All model objects store their raw content in RawSource, even if serialization fails; 
        // This means you can check to see if the content is a real error vs. a failed deserialization.

        TwitterError error = service.Deserialize<TwitterError>(mentions.First());
        if(!string.IsNullOrEmpty(error.ErrorMessage))
        {
            // You now know you have a real error from Twitter, and can handle it
        }
    }
}
else
{
    // Likely this is an error; we don't have to go fishing for it
    TwitterError error = service.Deserialize<TwitterError>(service.Response.Response);
    if(!string.IsNullOrEmpty(error.ErrorMessage))
    {
        // You now know you have a real error from Twitter, and can handle it
    }
}

Using Twitter Entities

TweetSharp supports Twitter's entities feature. This feature provides additional metadata for locating mentions, links, and hashtags embedded in tweet text. TweetSharp goes a step further, emulating the entities support for classes that Twitter currently doesn't support, that implement ITweetable; namely TwitterStatus, TwitterDirectMessage and TwitterSearchStatus. All three of these classes contain an Entities property. This allows you to find these elements in UI columns that contain multiple tweet types, via the ITweetable interface. To retrieve all of the entities in an ITweetable ordered by where they appear in the text itself, you can call the Coalesce method on the relevant TwitterEntities instance.

Client Development Features

TweetSharp is a bit more than an API wrapper, it also provides support for client application development beyond the data. In this section, TweetSharp's features specific to developing client applications on the .NET Framework are highlighted.

  • All model objects are [Serializable], implement IPropertyNotifyChanged, use virtual signatures, and implement DataContract / DataMember where supported. This means all of our model objects are persistable in frameworks like NHibernate, observable in WPF, and serializable for over-the-wire communication in WCF.

  • ITweetable and ITweeter: TwitterStatus, TwitterDirectMessage, and TwitterSearchStatus all implement ITweetable, which is an interface to make it easier to blend tweets from different sources into the same UI display. ITweeter helps encapsulate common user display properties, and TweetSharp even provides emulation of the entity metadata for non-timeline ITweetables; this means that you can access mentions, hashtags, and links embedded in the text of any ITweetable, in the order they appear.

These are the UI interfaces:

public interface ITweetable
{
    long Id { get; }
    string Text { get; }
    ITweeter Author { get; }
    DateTime CreatedDate { get; }
    TwitterEntities Entities { get; }
}

public interface ITweeter
{
    string ScreenName { get; }
    string ProfileImageUrl { get; }
}

About

An example API wrapper for Twitter using Hammock and T4

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published