Public Stats API

Nope (I have a process that runs a few times a day and samples all of the data and looks for new stuff).

 

I must admit that I am green to everything coding. That said, I went by the elasticsearch website and took a look at the client options and wanted to try and put something together in Javascript as per this document. However, I'm getting a 403 Forbidden error just trying to check to see if the cluster exists. Is the shroudoftheavatar elasticsearch setup so differently that the standard documentation no longer applies? Or any other insight here on something obvious I'm missing?

 

IIRC, Shroud's elasticsearch only responds to GET type requests. So you can't use POST/HEAD/etc. Hence the 403 error.

 

Awesome! Thanks, I now have something that works to play with.

 

@Chris & @Undone - any chance of enabling POST, since body on GET requests is incredibly bad form (I have no idea why ES even still allows that), and a lot of systems (and several proper elastic clients reject any such attempt)

 

GET, POST, PUT, DELETE etc all have almost identical behaviour. Like 99,9% identical. Difference comes from contract, that API developer has used correct VERB to annotate certain action. GET should and is used as a end point, that only gets data from server. It does NOT alter the data, it does not change the state of the server. GET with same parameters, should always return same response, unless state has changed.
POST, is something that is used to send data to server, new data.

GET and POST are just verbs. Http works almost identical no matter what you use. Its all about the contract. Internet and browsers are built on that contract, that get gets, post posts, everything else, is bad.

 

I'm completely lost here, are you trying to agree with me? Or are you trying to explain how the internet works?

If it's the latter and you are going to talk technical, please review sections 4.3 and 9.3 of the Hypertext Transfer Protocol -- HTTP/1.1

 

Just realized how weird the ES api is. It indeed allows and consumes GET bodies for the cases where client does not have POST or some other reasons.
Thats my bad.

However, if using GET, i think using the source param is the way to go?

 

The source parameter in the above examples provides the query as message body on the GET requests. If you do "clean" GET requests with everything in the URL (query string) you cannot structure your query as a JSON document.
You can do a lot with just query string parameters, it just doesn't facilitate a very clean interface when doing complex queries.

While the node.js api does allow changing the method on the search endpoint to GET by using the source parameter it is essentially a hack, and while that works for the search api, is does not work with the scroll api. So if for some reason you should wish to scroll through a larger data set, you cannot do this without writing a framework to do the requests yourself - again a hack of sorts.

Hence my proposal to enable POST on the endpoint so we don't have to "break" the protocol or write our own libraries to utilize the ES api.

 

@Undone / @Chris, Are any changes to the API planned for R45?

 

is thre a delay in the merchSale data? I fully admit I'm a hack at this api thing but I was trying to see the lottery transactions and I'm not seeing any for the last 3 days and I gotta believe there were tons of them 2 days ago(1 day before the draw). I removed lottery fields for Ss and Giggles and it seems I can't see any merch data for the last 3 days? Is that on purpose?(or am I just stupid, certainly possible)

Gets data(I think the last 4 days):
http://54.224.243.98:9200/_all/_sea...tionEvent:ItemGained_Merchant&size=500&pretty

No data(I think the last 3 days):
http://54.224.243.98:9200/_all/_sea...tionEvent:ItemGained_Merchant&size=500&pretty

 

****! It's the APIcolypse!

 

Was just about to bug you about the same thing with pvp

 

@Undone is it down again?

 

Looks like it's operating fine to me, SotaSTATS is still chugging away.