Skip to content

Latest commit

 

History

History
84 lines (66 loc) · 3.98 KB

README.md

File metadata and controls

84 lines (66 loc) · 3.98 KB
Raw

Canvas API Wrapper in C#

A wrapper function for HTTP calls in Canvas that provides these key features:

  • Adaptive Semaphore: This uses a counting semaphore that has the max number of concurrent calls set adaptively while the program is running.
  • Rate Limiting: Canvas limits the number of concurrent API calls that a single user can make. This checks to see if you went over your rate limit then waits, slows down the rate of API calls, and retries the API call that exceeded the rate limit.
  • Automatic Pagination Handling: If there are multiple pages in a GET API call then this will get all of the pages. You don't have to add any pagination parameters, it will get everything for you! (currently doesn't work with any parameters)
  • Limited Retries: If an individual api call returns 403 forbidden more than 10 times (customizable via Wrapper.Http.APIRetryLimit) then the call will fail. If that happens you likely have too many api calls happening all at once and should add an "await" or two between some calls (this probably won't happen)
  • Other exceptions: If the status code is anything other than 200 or 403, then it will fail and display the API call that failed, along with the response.
  • POCOs: Some POCOs have been added based on the API reference, more are planned in the future. In the meantime you'll have to work with the JSON string response (for unsupported objects).
  • CRUD: Currently only read (GET) is supported.

Important!

Make sure your canvas API token is set in the command line!

The domain and "/api/v1/" are taken care of for you, you only need to add the endpoint, or choose the sub-class that matches.

The best way to use this system is to make similar sized API calls together, and then await. If you want to make 1000 calls that cost 0.5 points, and then 300 calls that cost 12.0 points, you should make all of one cost first, await all of those, and then make all the rest.

Working with POCOs

using CanvasAPIWrapper;
Wrapper Canvas = new Wrapper();

CoursesObject MyCourse = await Canvas.Courses.Show("61116");
CoursesObject MyCourseWithParameters = await Canvas.Courses.Show("61116", "?include[]=term&include[]=syllabus_body");

Console.WriteLine(MyCourseWithParameters.SyllabusBody);

Custom Get (Working with JSON directly)

using CanvasAPIWrapper;
Wrapper Canvas = new Wrapper();

string CourseFeatures = await Canvas.Http.Get("/courses/61116/features/enabled");

Console.WriteLine(CourseFeatures);

// then you can parse the json to your own POCO
// or you can just work with the string

Running concurrent API calls

using CanvasAPIWrapper;
Wrapper Canvas = new Wrapper();

var courses = new List<Task<CourseObject>>();
courses.Add(Canvas.Courses.Show("40958"));
courses.Add(Canvas.Courses.Show("40960"));
courses.Add(Canvas.Courses.Show("40962"));
courses.Add(Canvas.Courses.Show("40964"));
courses.Add(Canvas.Courses.Show("40966"));

Task<CourseObject[]> t = Task.WhenAll(courses.ToArray());
t.Wait();

foreach (CourseObject item in t.Result)
{
    Console.Write(item.Id + " ");
    Console.WriteLine(item.Name);
}

Supported API Calls

  • Http.Get("custom_api_call")
  • Courses
  • BlueprintSubscriptions.SubscriptionsIndex = courses/ + id + /blueprint_subscriptions

We are following the canvas api documentation for our naming convention: https://canvas.instructure.com/doc/api/index.html

Improvement Ideas

  • Should we add Polly.Cache? (reduce the number of calls to the server)
  • We may be able to replace my concurrent API call limiter with Polly.Bulkhead
  • Redo branch created
    • Project moved to private server