marking2/danube_test4.scala
author Christian Urban <christian dot urban at kcl dot ac dot uk>
Mon, 05 Aug 2019 20:14:06 +0100
changeset 269 3ef2542207c4
parent 259 77c3bd7a0670
child 284 fc20e5f83f0e
permissions -rw-r--r--
updated
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
259
77c3bd7a0670 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
     1
77c3bd7a0670 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
     2
// first test 
77c3bd7a0670 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
     3
77c3bd7a0670 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
     4
77c3bd7a0670 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
     5
def urban_groupById(ratings: List[(String, String)]) = 
77c3bd7a0670 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
     6
  ratings.groupBy(_._1).mapValues(_.map(_._2)) 
77c3bd7a0670 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
     7
77c3bd7a0670 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
     8
def urban_get_csv_file(name: String) : List[String] = {
77c3bd7a0670 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
     9
  import io.Source
77c3bd7a0670 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    10
  import scala.util._
77c3bd7a0670 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    11
  val csv = Source.fromFile(name)
77c3bd7a0670 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    12
  csv.mkString.split("\n").toList.drop(1)
77c3bd7a0670 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    13
}
77c3bd7a0670 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    14
77c3bd7a0670 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    15
def urban_process_ratings(lines: List[String]) : List[(String, String)] = {
77c3bd7a0670 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    16
  for (cols <- lines.map(_.split(",").toList); 
77c3bd7a0670 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    17
       if (cols(2).toFloat >= 4)) yield (cols(0), cols(1))  
77c3bd7a0670 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    18
}
77c3bd7a0670 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    19
77c3bd7a0670 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    20
77c3bd7a0670 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    21
val urban_good_ratings = process_ratings(urban_get_csv_file("ratings.csv"))
77c3bd7a0670 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    22
77c3bd7a0670 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    23
val urban_ratings_map = urban_groupById(urban_good_ratings)
77c3bd7a0670 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    24
77c3bd7a0670 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    25
assert(favourites(urban_ratings_map, "912").length  == 80)
77c3bd7a0670 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    26
assert(favourites(urban_ratings_map, "858").length  == 158)
77c3bd7a0670 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    27
assert(favourites(urban_ratings_map, "260").length  == 201)
77c3bd7a0670 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    28