Paths ordering

[carlsonngo]

Generated JSON paths is in reverse order compared to ones defined in routes file

[kailuowang]

@carlsonngo thanks for reporting, I am curious how important this order is to you?

[carlsonngo]

Feeding the data to swagger-ui will display the paths in the wrong order. This might be confusing for people if they want to preserve the order in the generated api docs.

[kailuowang]

@carlsonngo thanks for the context.

[slouc]

+1 For me it's not even reverse order, but a completely random one. Using 0.2.0.

[bilongc]

+1 for me, it is displayed as random order. Using play 2.5.4, play-swagger 0.5.0

[tculshaw]

So how do we order, using play 2.5.x, play-swagger 0.5.3? Looks random to me.

[kailuowang]

@tculshaw we haven't implement the feature the preserve the ordering in the routes file. Any PR will be more than welcome :)

[tculshaw]

If any one is interested I have a partial solution, hence this isn't a formal pull request.

Replacing the path function in SwaggerSpecGenerator, I order the main unique paths as per the ROUTE file, but NOT the order where multiple methods (GET,PUT,POST,DELETE) apply to the same path. It's a lot better than it used to be, and will suffice for my needs, but still not quite 100%!

private def paths(routes: Seq[Route], prefix: String, tag: Option[Tag]): JsObject = {
    JsObject {
      val endPointEntries = routes.flatMap(route => endPointEntry(route, prefix, tag))
      val zgbp = endPointEntries.zipWithIndex.groupBy(_._1._1)
      import collection.mutable.LinkedHashMap
      val lhm = LinkedHashMap(zgbp.toSeq sortBy (_._2.head._2): _*)
      val gbp2 = lhm mapValues (_ map (_._1)) toSeq
      gbp2.toSeq.map( x => (x._1,x._2.map(_._2).reduce(_ deepMerge _)))
    }
 }

I banged my head against the wall to get this going, so hope this is helpful to somebody!

[kailuowang]

Thanks very much @tculshaw, it's not quite 100% but it's still a step forward. I will appreciate it if you want to submit a PR (just need a simple test for the ordering).

[tculshaw]

This can now be closed? Thanks for accepting my change #152

[kailuowang]

@tculshaw it's still a partial solution right? ("NOT the order where multiple methods (GET,PUT,POST,DELETE) apply to the same path")

[tculshaw]

Yeah sorry! I wasn't going to get totally caught up in the method orders where the path is shared. It did my head in just doing this one;) Probably(?) most users who want the ordering will be happy to have the solution provided as is. I will look at if I get chance - if you think it's worth doing? Tony Culshaw MA Cantab Director Mentation Limited - Custom Software Development

…

On 7 April 2017 at 11:40, Kai(luo) Wang ***@***.***> wrote: @tculshaw <https://github.com/tculshaw> it's still a partial solution right? ("NOT the order where multiple methods (GET,PUT,POST,DELETE) apply to the same path") — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#34 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAianUJP454DiNMUeXkdxSJMZED_fvnoks5rtXfcgaJpZM4GxjIG> .

[kailuowang]

I imagine it would be nice for some users, but I don't particularly have that need in my projects. Thanks for offering! And thanks again for your contribution!

[tculshaw]

No worries:) Tony Culshaw MA Cantab Director Mentation Limited - Custom Software Development

…

On 7 April 2017 at 14:26, Kai(luo) Wang ***@***.***> wrote: I imagine it would be nice for some users, but I don't particularly have that need in my projects. Thanks for offering! And thanks again for your contribution! — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#34 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAianVIj1ZQGroZgMcdCD9Cz2McpjZzQks5rtZ6-gaJpZM4GxjIG> .