example-output.txt

[info] Running Example 
EXTRACTED USERS:
(0,{"id_str":"0","name":"Petra Sievinen","testing":"petra","followers_count":1,"foobar":[1,2,3]})
(2,{"id_str":"2","name":"Retweeter"})
(1,{"testing":"jepjep","id_str":"1","name":"Pyppe"})
(5,{"id_str":"5","name":"Quoter","testing":"aino","followers_count":333,"lang":"en"})
WHOLE TWEET WITH LITE USERS:
{
  "created_at" : "Wed Jun 10 20:01:01 +0000 2015",
  "id_str" : "608725574924808192",
  "text" : "Testing",
  "truncated" : false,
  "retweet_count" : 0,
  "favorite_count" : 0,
  "entities" : {
    "hashtags" : [ {
      "text" : "lovedungeon",
      "indices" : [ 57, 70 ]
    } ],
    "trends" : [ ],
    "urls" : [ {
      "url" : "https://t.co/abc",
      "expanded_url" : "https://twitter.com/diipadaapa/status/1234",
      "display_url" : "twitter.com/diipadaapa/sta…",
      "indices" : [ 72, 95 ]
    } ],
    "user_mentions" : [ {
      "screen_name" : "foo",
      "name" : "Foo bar",
      "id" : 102025211,
      "id_str" : "102025211",
      "indices" : [ 0, 13 ]
    } ],
    "symbols" : [ ]
  },
  "favorited" : false,
  "retweeted" : false,
  "filter_level" : "low",
  "lang" : "fi",
  "timestamp_ms" : "1433966461293",
  "user" : {
    "id_str" : "0",
    "name" : "Petra Sievinen"
  },
  "retweeted_status" : {
    "hip" : "hei",
    "user" : {
      "id_str" : "2",
      "name" : "Retweeter"
    },
    "quoted_status" : {
      "some" : "property",
      "user" : {
        "id_str" : "1",
        "name" : "Pyppe"
      }
    }
  },
  "quoted_status" : {
    "id_str" : "4",
    "geo" : null,
    "favorited" : false,
    "retweeted" : false,
    "possibly_sensitive" : false,
    "filter_level" : "low",
    "lang" : "fi",
    "user" : {
      "id_str" : "5",
      "name" : "Quoter"
    }
  }
}

example-tweet.json

{
  "created_at": "Wed Jun 10 20:01:01 +0000 2015",
  "id_str": "608725574924808192",
  "text": "Testing",
  "truncated": false,
  "retweeted_status": {
    "hip": "hei",
    "user": {
      "id_str": "2",
      "name": "Retweeter"
    },
    "quoted_status": {
      "some": "property",
      "user": {
        "testing": "jepjep",
        "id_str": "1",
        "name": "Pyppe"
      }
    }
  },
  "quoted_status": {
    "id_str": "4",
    "user": {
      "id_str": "5",
      "name": "Quoter",
      "testing": "aino",
      "followers_count": 333,
      "lang": "en"
    },
    "geo": null,
    "favorited": false,
    "retweeted": false,
    "possibly_sensitive": false,
    "filter_level": "low",
    "lang": "fi"
  },
  "retweet_count": 0,
  "favorite_count": 0,
  "entities": {
    "hashtags": [
      {
        "text": "lovedungeon",
        "indices": [
          57,
          70
        ]
      }
    ],
    "trends": [],
    "urls": [
      {
        "url": "https://t.co/abc",
        "expanded_url": "https://twitter.com/diipadaapa/status/1234",
        "display_url": "twitter.com/diipadaapa/sta…",
        "indices": [
          72,
          95
        ]
      }
    ],
    "user_mentions": [
      {
        "screen_name": "foo",
        "name": "Foo bar",
        "id": 102025211,
        "id_str": "102025211",
        "indices": [
          0,
          13
        ]
      }
    ],
    "symbols": []
  },
  "favorited": false,
  "retweeted": false,
  "filter_level": "low",
  "lang": "fi",
  "timestamp_ms": "1433966461293",
  "user": {
    "id_str": "0",
    "name": "Petra Sievinen",
    "testing": "petra",
    "followers_count": 1,
    "foobar": [1,2,3]
  }
}

Example.scala

object Example extends App {

  /**
   * Extract users from following paths (from given tweet-JSON):
   *   __ \ "user"
   *   __ \ "retweeted_status" \ "user"
   *   __ \ "retweeted_status" \ "quoted_status" \ "user"
   *   __ \ "quoted_status" \ "user"
   * and replace those where only user-fields `name` and `id_str` remain
   *
   * @return a tuple of:
   *         1. tweet-JSON with "lite" users (only fields `id_str` and `name`)
   *         2. Map of full user-JSONs (key id_str)
   */
  def extractUsers(js: JsObject, userFieldFilter: (String => Boolean)) = Try {
    @tailrec
    def extract(tweet: JsObject,
                users: Map[String, JsObject],
                pathsToCheck: List[JsPath]): (JsObject, Map[String, JsObject]) = {
      pathsToCheck match {
        case Nil => (tweet, users)
        case path :: tail =>
          path.asSingleJson(tweet) match {
            case user: JsObject =>
              val id = (user \ "id_str").as[String]
              val liteUser = JsObject(user.fields.filter { kv => userFieldFilter(kv._1)})
              val withLiteUser =
                tweet.transform(path.json.prune).get.
                  deepMerge(Json.obj().transform(path.json.put(liteUser)).get)
              extract(withLiteUser, users + (id -> user), tail)
            case _ =>
              extract(tweet, users, tail)
          }
      }
    }
    extract(js, Map.empty, List(
      (__ \ "user"),
      (__ \ "retweeted_status" \ "user"),
      (__ \ "retweeted_status" \ "quoted_status" \ "user"),
      (__ \ "quoted_status" \ "user")
    ))
  }

  val (tweetWithLiteUsers, users) = extractUsers(exampleTweet, Set("id_str", "name")).get
  println("EXTRACTED USERS:")
  users foreach println
  println("WHOLE TWEET WITH LITE USERS:")
  println(Json.prettyPrint(tweetWithLiteUsers))

}