რით შეგვიძლია დახმარება?
Didn't find what you were looking for?
New post32 comments
-
Hi Liftoff,
Yes, it would be okay to explain how these API requests are used for educational purposes, as long as nothing is being done with malicious intent.
However, one thing to note is that KA's public API is being removed, so these endpoints will eventually stop working and your tutorial will become outdated in the near future.
I hope this helps! Please let me know if you have any additional questions!
-
The API endpoints are all being converted to a single GraphQL endpoint, which will not be accessible from within the CS projects (its blocked by CORS). There's also no longer a public API or documentation for the endpoints.
In short, any existing endpoints that your projects use/teach will eventually be removed entirely and will no longer function.
-
SpencerKuan There's no definitive time-frame as to when specific endpoints will be removed, and it's more likely that the remaining REST endpoints will be gradually transferred to the new API. However, I believe all of the public endpoints will be fully removed within about a year.
-
Well, this is very disappointing. There are hundreds of projects that exist that use KA's API and rely on the API to function. Especially commonly used programs such as project searches, user searches, and project flag checkers all use the API. At least I know that I am using Willard's project search and Programming King's flag check ALL THE TIME. His project search is the number one most useful thing for finding out if something is plagiarism. It would be terrible to not be able to find the original of a plagiarised project anymore. Not only that, but I think learning how to work with API requests is a very important thing to learn for any person wanting to do web development.
And just to clarify, the old API endpoints are being completely removed and gone over the period of a year. It's not going to happen all at once, and the old API is going to be completely gone and not just archived.
Is there any way that KA's developers can let us make GET requests to the one GraphQL endpoint just like we can now with the current API? Like somehow disable CORS or something.
-
Unfortunately, the KACP community is probably not major concern of the KA development team. Access within KA programs has never been completely secure anyways, I believe. Disabling CORS is definitely not going to happen, or it could be a major security risk. I suspect most large tools can still be used off of KA through a proxy.
-
Yes Leaf
-
As of now, all user profile requests will no longer work.
̶Q̶u̶i̶n̶n̶ ̶(̶a̶n̶d̶ ̶m̶o̶s̶t̶l̶y̶ ̶Q̶u̶i̶n̶n̶)̶,̶ ̶S̶h̶i̶p̶m̶e̶n̶t̶,̶ ̶a̶n̶d̶ ̶I̶ ̶f̶i̶g̶u̶r̶e̶d̶ ̶o̶u̶t̶ ̶h̶o̶w̶ ̶t̶o̶ ̶u̶s̶e̶ ̶t̶h̶e̶ ̶n̶e̶w̶ ̶A̶P̶I̶.̶ ̶H̶o̶w̶e̶v̶e̶r̶,̶ ̶i̶t̶ ̶c̶a̶n̶n̶o̶t̶ ̶b̶e̶ ̶d̶o̶n̶e̶ ̶f̶r̶o̶m̶ ̶t̶h̶e̶ ̶f̶r̶o̶n̶t̶ ̶e̶n̶d̶.̶ ̶Y̶o̶u̶ ̶m̶u̶s̶t̶ ̶s̶e̶t̶ ̶u̶p̶ ̶a̶ ̶N̶o̶d̶e̶.̶j̶s̶ ̶s̶e̶r̶v̶e̶r̶ ̶a̶n̶d̶ ̶u̶s̶e̶ ̶n̶o̶d̶e̶-̶f̶e̶t̶c̶h̶ ̶(̶s̶p̶e̶c̶i̶f̶i̶c̶a̶l̶l̶y̶ ̶v̶e̶r̶s̶i̶o̶n̶ ̶2̶)̶ ̶t̶o̶ ̶m̶a̶k̶e̶ ̶t̶h̶e̶ ̶G̶r̶a̶p̶h̶Q̶L̶ ̶r̶e̶q̶u̶e̶s̶t̶s̶ ̶t̶o̶ ̶K̶h̶a̶n̶ ̶A̶c̶a̶d̶e̶m̶y̶'̶s̶ ̶s̶e̶r̶v̶e̶r̶s̶.̶
(Edit: pretty much all of that is not 100% correct, so just scroll down and read Quinn's post to get 100% correct information)I also have mention that I am disappointed that the new API is MUCH MUCH slower than the old API. The time it takes to make a request and receive a response using the old API is approx 200 milliseconds, but with the new API it takes approx 1200 milliseconds. I was hoping that the new API would at least be slightly faster.
-
There are two KA API tutorials that I know of.
Firstly, Quinn Hardbrook's (which is where I learned how to use the API). I think it's great and I am constantly going back the that project because it has a good documentation.
https://www.khanacademy.org/computer-programming/sal-ka-an-api-labs/5414181885788160Secondly, Jett Burn's
https://www.khanacademy.org/computer-programming/i/6272867024306176Both those tutorials teach you how to use the old API. There are no tutorials yet on how to use the new API. But you say you want to fetch info about a program which is still done via the old API so it's fine. There may be tutorials on how to use the new API in the future. Or instead of learning how to use the new API, someone may make a proxy that you can make requests to and it will use the API for you and return you the data it got.
-
I'm very disappointed. I just learned how to use API not to long ago. So does this mean https://willard.fun/ will no longer work?
Since Shipment, Quinn, and yourself have figured out how to use the new API, can you guys make a tutorial please?
-
I believe credit was given out a little in-proportionately. I looked up the new endpoint and got it working while Vexcess, Shipment, and I were discussing it. They weren’t too familiar with Node Fetch but got it working soon. Willard, Squishy, and a few others were also using it within a few days. We didn’t tell them, so it can be assumed they worked it out on their own. It’s not a hard thing to do if you’re familiar with the process.
A few corrections to what Vexcess said:
1. The new API can be used outside of NodeJS. The reason for accessing it backend is because you don’t have to worry about CORS that way. You can also use the new API in frontend application such as bookmarklets and extensions (content scripts only), since that won’t violate CORS either.
2. You can use version 3 of node-fetch if you prefer modules.
3. You don’t really need to set up a server. You can run it locally when you need and end the process.
The new API relies on GraphQL. That means you provide a format of the data to return when you make the request. Since GET requests can’t contain a body, the requests will be POST. Currently, the requests are very slow. It could take several seconds to load one.
The entire point of GraphQL is to allow the client to only fetch what data they need. Currently, requests do not work unless they have the exact right format. It’s a little bit silly, if you ask me. I’d hope Khan Academy’s dev team only has it set up like this for now, but I could be wrong.
I won’t be creating a tutorial. Since the code cannot be run in KA programs, it doesn’t make sense to put it on KA. If you really want to use it, it’s as simple as going to Network in your dev tools, finding the request, and copying as a Node Fetch, or your preferred method. Remove the excess and set the cookies to "fkey=0" so you don’t give out your KA account info.
Edit: in addition, my API tutorial will be deprecated soon, due to the change. I won’t be updating for the reasons stated above. I encourage you to learn elsewhere about REST APIs. You should look into MDN’s documentation on fetch.
-
This is super disappointing and a huge backwards step for the KA programming community. There are so many use cases that rely on the ability to use the public API. I know quite a few of my programs will die as a result, and almost all the community driven competitions use the API for managing participants, etc.
Also, as mentioned, it provided a good initial introduction to learning how an API works, albeit at a basic level, which will no longer exist.
@EvanLewis - do you know if anything will likely be made available within the KA environment that will give us the same functionality (retrieve user/program info, etc.) or are we just not that important :(
-
Hey I have a question. My website:
https://wintertheprogrammer.github.io/
uses KA API in both the about part and KA programs. In the KA programs tab I use API to fetch all the data about my programs on KA. Will all those programs go away or when I make a new program (which gets more votes than the the ones currently on the list), it won't be added to the list?
-
I agree about the whole API thing. My website uses it to retrieve project information, and one day it just decided not to load. Is there a place to ask to get it back? If there is, let me know, or else I have to either figure out the new API or connect my projects with nothing but a simple underwhelming link of text 😞
-
Nice, so you're destroying your own site? This morning I couldn't get into my account - saying it had trouble loading my profile info. Makes almost no sense why you're doing this - people don't like it, and it might even make the site slower—maybe.
Like Gray Wolf said, huge step backwards. We should be moving forwards, so why are we doing the opposite?
-
Lots of people are getting that error. It's caused by the Khan Academy Extension making a get notifs request every 750 milliseconds I think.
But overall Khan Academy is for sure devolving same way humans are. They say that we started off as tiny organisms and then evolved into monkeys and then humans over billions of years, but now humans are wanting to go backwards identify as animals again.
-
Don’t literally all websites on the internet use some kind of API?
There are some comments that there is still an API, its just been changed so that other people can’t use it.
With the way programs and comments work, Khan needs an API to function right?
If there was an API, how would I use it? (and what would i use it for)
- Boo
P.S. I am only now learning what an API is, so I might be totally wrong
-
Hello everyone, I have now reviewed most of the discussion around the Khan API and no one ever really addresses why the old API is being depreciated (apart from common security concerns) and if the is or will be a replacement (using GraphQL)? Could some give a clear answer on what objectives KA is perusing for its developer community? For a non-profit producing a public good not having an API (and worse not having a strategy) is really not acceptable IMHOP. Can some one step up and make a clear statement with a clear rationale?
-
Tim Delhaes Good question, I would love to know the answer as well.
-
Feel free to chime in bool frightening noodles
https://twitter.com/tim_delhaes/status/1538189988965367811
Any comments maybe from Quinn Hardbrook or Evan Lewis ?
You guys seem to be more on the inside :) -
Communication across channels and tools :)
From discussion in Github"I disagree with that. It would not be acceptable if KA was a government entity or some state-sponsored enterprise, or state-sanctioned monopoly. Since KA is neither, it is IMO free to do whatever it believes best serves to further its goals and those of its sponsors, using the limited resources it has available. Having an API is a great enabling facility and it can be very helpful (I used to use that API myself in the past for my project), but we don't have a right to demand such API, or to shame KA into providing it to us."
-
How the current API works (surface-level): There is a set of data offered by Khan Academy that is available through the front-end. We can make JSONP requests with callbacks to use this data in the front-end. JavaScript is front-end.
The old API is being removed. This does not mean there is no more API, it just means that they are adding a new API which is not accessible through the front-end anymore. Since there are no servers user-accessible, and therefore no back-end languages available in the KA environment, there is no way to access the new API within Khan Academy.
The keys are slowly being removed to opt for a more modern API, and there is not much we can do about it. Recently, we have seen small changes and small bugs that follow, as the developers carefully make the move to the new API. You can read more about the way the new API works at https://graphql.org
While a lot of people are disappointed with the new changes, it is not really the fault of the developers nor the responsibility to heed our complaints. It is a free website, after all. Besides, I am sure somebody will find a way to access it through the front-end at some point. We are all coders here.
კომენტარის დასატოვებლად გთხოვთ გაიაროთ ავტორიზაცია.