Skip to content

mdread/buhtig

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

28 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

buhtig

Scala client for Github rest API. The main goal is to be damn easy, and to adapt to changes on the API. In fact it is just a utility to help you build requests against github in a fluid way.

why?

The idea is to have a basic utility to interact with Github REST API, that does not require you to study a new API. If you already know github REST API you already know how to use this library, and because it is no more than a utility to help you build requests, if the API changes you don't need to wait for a new version to be released.

Installation

sbt

libraryDependencies += "net.caoticode.buhtig" %% "buhtig" % "0.3.1"

The library is cross compiled for scala 2.10 / 2.11

sources

:> git clone https://github.com/mdread/buhtig.git
:> cd buhtig
:> sbt package

The generated jar file can be found under target/scala_{version} (don't forget to add Dispatch and json4s as dependency).

To execute tests create a file named github.token under src/test/resources with a valid github api token.

Usage

Getting a client

To start using the library we need an instance of the client, and for that we first need to instantiate Buhtig class with a valid github token.

val token = "..." // your personal API token
val buhtig = new Buhtig(token)

Now we can choose to get a synchronous or an asynchronous client, with the synchronous one the execution of the program is going to stop on every request waiting for the response from github servers, while with the asynchronous one the execution is not going to stop and the requests are going to be processed in background by Scala Futures.

To get the synchronous client do:

val client = buhtig.syncClient

For the asynchronous one:

val client = buhtig.asyncClient

The other method you are going to use from Buhtig class is close, which closes the underlying Dispatch client. For this just call buhtig.close() when all your operations are done.

Constructing URLs

The client we get before is actually an immutable class which holds a fragment of the URL we want to compose, and has some handy methods to build it and execute the request.

Concatenating URL fragments

The easiest way is to use method chaining, where every method call make a fragment of the URL we want to compose, let's say we want to get information about this repository, from github API we need to make a GET request to the URL https://api.github.com/repos/mdread/buhtig, which we can do as follow:

val repo = client.repos.mdread.buhtig.get[JSON]

Leaving out for the moment the last get[JSON] call, the URL https://api.github.com/repos/mdread/buhtig gets constructed with client.repos.mdread.buhtig, actually any URL can be constructed this way. Also you can choose to build part of the URL as parameter of a method call, as in the following example:

val gistID = "7655832"
val gist = client.gists(gistID).get[JSON]

Here the URL that gets constructed is https://api.github.com/gists/7655832, where 7655832 is the ID of a Gist. Can be useful to build requests this way when a part of the URL can come from an external source and can not be hard-coded.

Another way of composing URL fragments is to use the / method of the client, this way the first example can be rewritten as:

val repo = (client / "repos" / "mdread" / "buhtig").get[JSON]

Query parameters

Query parameters can be added with the ? method of the client, for example to build the URL https://api.github.com/search/repositories?q=tetris&sort=stars to search for tetris repos we can do as follows:

val tetris = (client.search.repositories ? ("q" -> "tetris", "sort" -> "stars")).get[JSON]

Write operations

So far we have only made GET calls to Github API, but some parts of it requires to use POST, PUT, PATCH or DELETE. To do so just call the appropriate method after constructing the url, and pass the body of the request as parameter, for example for creating a new Gist:

val request = client.gists POST """
  {
    "description": "the description for this gist",
    "public": true,
    "files": {
      "file1.txt": {
        "content": "String file contents"
      }
    }
  }
  """

println(request.get[String])

The body of the request can also be a JSON built with json4s, so the last example can be rewritten as follows:

val request = client.gists POST
  ("description" -> "the description for this gist") ~
  ("public" -> true) ~
  ("files" ->
    ("file1.txt" ->
      ("content" -> "String file contents")))

println(request.get[String])

Actually if we check how the function POST is defined

def POST[S](value: S)(implicit converter: S => String) = { ... }

we see it can take any parameter provided there is a function in scope which can convert it to a String, so if you use a different library for JSON parsing, or some custom object holding your data, you can always implement an S => String function and declare it implicit in scope to use the custom object directly.

For the DELETE method there are no parameters, so for example to delete a gist just do:

(client / "gists" / "113533257213d2883e74").DELETE.get[String]

Getting the result

In all previous examples we have used the method get[T] at the end of the constructed URL, this method is the one which effectively executes the request, there is also a getOpt[T] method which returns an Option[T]. Depending on the client you are using (synchronous or asynchronous) those methods will return directly the expected result, or a Future which will eventually return the expected value when is ready.

Looking at the definition of get method:

def get[T](implicit x: (Response => T)): T  = { ... }

we see it also depends on a function which converts a Response object (com.ning.http.client.Response) to the expected value, there are predefined converters for String, JSON (type alias for JValue), Response, Byte and also for getting the HEADER (type alias for com.ning.http.client.FluentCaseInsensitiveStringsMap) and a tuple of HEADER_BODY (tuple of type (HEADER, JSON) ). To convert to any custom type implement a Response => T function and declare it implicit in scope.

Examples

Here some complete examples with all needed imports and declarations

Creating a new Gist

Creating a new Gist and printing the id

import net.caoticode.buhtig.Buhtig
import net.caoticode.buhtig.Converters._
import org.json4s.native.JsonMethods._

val token = "..." // your personal API token
val buhtig = new Buhtig(token)
val client = buhtig.syncClient

val request = client.gists POST """
 {
   "description": "the description for this gist",
   "public": true,
   "files": {
     "file1.txt": {
       "content": "String file contents"
     }
   }
 }
 """

val gist = request.get[JSON]

println(compact(render( (gist \ "id") )))

client.close()

Asynchronously listing all contributors of all repos of a user

In this example we are going to use the asynchronous client to list all contributors of all repositories of the user apache

import net.caoticode.buhtig.Buhtig
import net.caoticode.buhtig.Converters._
import org.json4s._
import org.json4s.native.JsonMethods._
import scala.concurrent.Future
import scala.concurrent.ExecutionContext.Implicits.global

val token = "..." // your personal API token
val buhtig = new Buhtig(token)
val client = buhtig.asyncClient

def listRepos(user: String) = client.users(user).repos.get[JSON]
def listContributors(user: String, repo: String) = client.repos(user, repo).stats.contributors.get[JSON]

// wait for successful resolution of the Future
for(repo <- listRepos("apache")) {

  // extract repository names with json4s api
  val names = for {
    JObject(obj) <- repo
    JField("name", JString(name)) <- obj
  } yield name

  // convert the List of futures to a future of List
  val contrsList = Future.sequence(names map {e => listContributors("apache", e)})

  // wait for completion of all futures
  contrsList.onSuccess {
    case contribtors => {

      // extract the list of contributors from the list of JSON responses
      val logins = contribtors.flatMap(cont => cont filterField {
       case JField("login", _) => true
       case _ => false
      }).map(e => compact(render(e._2))).mkString(", ")

      println(logins)

      buhtig.close()
    }
  }
}

How does it works?

Internally all HTTP requests are managed by Dispatch. The external API is build around Scala Type Dynamic functionality, which basically allows to intercept calls to missing methods, this is how you construct URLs making method invocations. Also it uses json4s to parse JSON.

About

painless Github client

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages