NationStates API (nationdata/regiondata)

[Kethania]

The API doc has been updated with instructions on how to answer issues via API. This is the first "Private Command" where you can actually make your nation do stuff via the API, not just gather information about it.

Wonderful!

The wording here makes me assume there are more commands to come, is that correct? If so, will there be the same system as with shards for joining several commands into one request?
It appears to be impossible to request a nation shard and execute a command in a single call, is that the intended behaviour?

I would also like to report a nasty encoding bug in the API. Specifically, I get this mess:

Code: Select all

<NATION id="saidual"><MOTTO>Oi oi its the &#x97;unicode&#x97;police&#x97;</MOTTO></NATION>

The EM DASH character is replaced by END OF GUARDED AREA. This problem only affects the API, but I have seen it also show up when dealing with dispatch shards, so I am assuming it's not local.
The issue appears to be that somewhere on the server Windows-1252 encoded text is wrongly decoded as ISO-8859-1, then passed to the clients as an NCR. While it is technically possible to write a function to unscramble this mess, I'd rather prefer not having to do that :p

[[violet]]

The wording here makes me assume there are more commands to come, is that correct?

Yes. It's a "squeaky wheel" kind of situation, where I'll probably implement what gets asked for the most persistently, since otherwise I can't tell what script authors are most keen to use.

If so, will there be the same system as with shards for joining several commands into one request?

Probably not. Chaining commands makes sense when you're grabbing information, because (a) the heaviest part of that for us is usually the initial nation load, so once we've done that, we can supply whatever data you want; and (b) there are no issues with state change. That's not the case when you're issuing commands that actually change your nation. There's also potential complication based around which commands get executed in what order.

It appears to be impossible to request a nation shard and execute a command in a single call, is that the intended behaviour?

Correct. You can't do both at the same time.

[[violet]]

I would also like to report a nasty encoding bug in the API. Specifically, I get this mess:

Code: Select all

<NATION id="saidual"><MOTTO>Oi oi its the &#x97;unicode&#x97;police&#x97;</MOTTO></NATION>

The EM DASH character is replaced by END OF GUARDED AREA. This problem only affects the API, but I have seen it also show up when dealing with dispatch shards, so I am assuming it's not local.

Unfortunately I'm not sure this is fixable, because our character encoding is a disaster area. The site began in the lovely ASCII era of 2002 and has grown a few hacks to support extended character sets since then, but it doesn't have proper Unicode support. I've looked a few times at adding it, but we have such a mishmash of different encodings spread across millions of files and data sets, I'm not sure I'll ever get there.

[Eluvatar]

Would it be possible that NS API incorporates a new public shard to query a nation's id number?
I'm referring to the number that is currently used to resolve WA census ties in the ranking system.

It would certainly be possible. As [v] and I don't see a problem with doing it, it'll probably be done soon.

It's been done.

[Drachmaland]

Would it be possible that NS API incorporates a new public shard to query a nation's id number?
I'm referring to the number that is currently used to resolve WA census ties in the ranking system.

It's been done.

Code: Select all

q=dbid

Thanks a lot, Eluvatar!

[Agadar]

Is there something of a changelog of the NS API available to us? As maintainer of one of the client libraries for the NS API, I find myself having to scour this thread for all the changes.

[Caelapes]

It's been done.

Is there any way of getting DBIDs for ex-nations?

[Eluvatar]

It's been done.

Is there any way of getting DBIDs for ex-nations?

Not at this time. Why do you want it?

Is there something of a changelog of the NS API available to us? As maintainer of one of the client libraries for the NS API, I find myself having to scour this thread for all the changes.

Not at this time. If someone were to make a record of API changes so far, however, I'm surmise that could be kept up going forward.

[Caelapes]

Is there any way of getting DBIDs for ex-nations?

Not at this time. Why do you want it?

Was just curious if it was possible.

Are there DBIDs for regions? If so, would it be possible to get a shard for them?

[Eluvatar]

Not at this time. Why do you want it?

Was just curious if it was possible.

Are there DBIDs for regions? If so, would it be possible to get a shard for them?

Drachmaland had a specific ask for dbid:

Would it be possible that NS API incorporates a new public shard to query a nation's id number?
I'm referring to the number that is currently used to resolve WA census ties in the ranking system.

What would the DBID of regions (or ex-nations that have them) tell you that you want to know?

[Caelapes]

What would the DBID of regions (or ex-nations that have them) tell you that you want to know?

Well, with all the secrecy around them, it's a fishing expedition to see what we can figure out from them.

[Eluvatar]

What would the DBID of regions (or ex-nations that have them) tell you that you want to know?

Well, with all the secrecy around them, it's a fishing expedition to see what we can figure out from them.

Helping you fish is not as exciting to me as other projects.

(Fifteen miles! In the snow! Uphill! Both ways! Off my lawn!)

[Drachmaland]

What would the DBID of regions (or ex-nations that have them) tell you that you want to know?

Isn't the regions' DBID the factor that resolves WA census ties in the ranking of regions?

And as far as ex-nations go, I believe the API can't query for them at all, can it?

[Caelapes]

What would the DBID of regions (or ex-nations that have them) tell you that you want to know?

Isn't the regions' DBID the factor that resolves WA census ties in the ranking of regions?

Hey! There we go. That's precisely why I want it.

And as far as ex-nations go, I believe the API can't query for them at all, can it?

Correct. The API returns a 404 error code for nations that do not currently exist.

[Eluvatar]

Isn't the regions' DBID the factor that resolves WA census ties in the ranking of regions?

Hey! There we go. That's precisely why I want it.

Of course.

This will probably be added fairly soon.

[Greater Tern]

[v] or Elu -- for the issues API, could you guys wrap the <DESC> field in a CDATA? Stately's deserializer freaks out when it runs into characters like '%' or 'é'. Wrapping it in CDATA should let it deserialize the data properly.

EDIT: Same thing for the <HEADLINE> tags as well.

[Auralia]

[violet], would it be possible to exempt OPTIONS requests from the API rate limit? From a JS in-browser POV, it's difficult to know whether a request will trigger an additional OPTIONS request by the browser because of CORS -- this makes it difficult to do rate limiting.

[Kethania]

Would it be possible for the macros (such as @@NAME@@, @@DEMONYM@@, etc.) to be expanded in the HEADLINE elements from the issue API command, the way they are in DESC from the same command, as well as OPTIONs and TEXT from the issues private shard?

Also, the OK/ERROR system seems rather redundant. Can't you simply use HTTP codes?

[Eluvatar]

[violet], would it be possible to exempt OPTIONS requests from the API rate limit? From a JS in-browser POV, it's difficult to know whether a request will trigger an additional OPTIONS request by the browser because of CORS -- this makes it difficult to do rate limiting.

Unfortunately currently the API will give the full response in reply to an OPTIONS request, so exempting OPTIONS requests would let people get around the API rate limit entirely. Changing this will be considered.

Would it be possible for the macros (such as @@NAME@@, @@DEMONYM@@, etc.) to be expanded in the HEADLINE elements from the issue API command, the way they are in DESC from the same command, as well as OPTIONs and TEXT from the issues private shard?

Also, the OK/ERROR system seems rather redundant. Can't you simply use HTTP codes?

What ok/error system do you mean? Do you mean for TG queueing?

[Greater Tern]

What ok/error system do you mean? Do you mean for TG queueing?

I think they're referring to the new issues API, which either returns an <OK> tag if the request was successful (along with the rest of the response), or an <ERROR> tag if the request failed. It might make more sense to just return the issue response by itself and HTTP 200 if it was successful, or return some HTTP 400 error code if it failed for some reason.

Also bumping my request to add the CDATA tag to the issues response. ^^

[[violet]]

What ok/error system do you mean? Do you mean for TG queueing?

I think they're referring to the new issues API, which either returns an <OK> tag if the request was successful (along with the rest of the response), or an <ERROR> tag if the request failed. It might make more sense to just return the issue response by itself and HTTP 200 if it was successful, or return some HTTP 400 error code if it failed for some reason.

The API is just dumping out what our internal system already provides, which is simpler than building a separate signaling system. I think that's fine.

[Greater Tern]

The API is just dumping out what our internal system already provides, which is simpler than building a separate signaling system. I think that's fine.

Fair enough. Would you be able to modify it though to wrap the content in <DESC> and <HEADLINE> inside a CDATA tag?

[Caelapes]

Since moderator-deleted nations are identifiable through the Boneyard (and thus it is not a secret that any nation has been moderator-deleted, just the reasoning) would it be possible to get some kind of indication in the nation's CTE happening (X has ceased to exist) that it was a moderator deletion?

Perhaps, only when the deletion is moderator-enacted, have an additional "TYPE" declaration as such:

Code: Select all

<EVENT id="156530714">
   <TIMESTAMP>1497896108</TIMESTAMP>
   <TEXT>
      <![CDATA[ @@bitler101@@ ceased to exist in %%the_pacific%%. ]]>
   </TEXT>
   <TYPE>DEAT</TYPE>
</EVENT>

or have TYPE declared for all, with 0 for regular lack-of-activity CTEs and 1 for moderator DEATs?

[Eluvatar]

I think that you can probably tell whether or not it was time for that nation to be updating, no?

[Caelapes]

I think that you can probably tell whether or not it was time for that nation to be updating, no?

Not always. If the nation CTEs well after update, it's obvious, but sometimes mods DEAT nations during update and currently there is no way to tell if it's from activity or moderators without scraping the Boneyard.