testing2/danube_test6.scala
author Christian Urban <urbanc@in.tum.de>
Tue, 19 Nov 2019 06:38:20 +0000
changeset 322 856aaa591840
parent 284 fc20e5f83f0e
permissions -rw-r--r--
updated
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
284
fc20e5f83f0e updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
     1
fc20e5f83f0e updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
     2
import CW7b._
fc20e5f83f0e updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
     3
fc20e5f83f0e updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
     4
// first test 
fc20e5f83f0e updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
     5
fc20e5f83f0e updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
     6
fc20e5f83f0e updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
     7
def urban_groupById(ratings: List[(String, String)]) = 
fc20e5f83f0e updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
     8
  ratings.groupBy(_._1).view.mapValues(_.map(_._2)).toMap 
fc20e5f83f0e updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
     9
fc20e5f83f0e updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    10
def urban_get_csv_file(name: String) : List[String] = {
fc20e5f83f0e updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    11
  import io.Source
fc20e5f83f0e updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    12
  import scala.util._
fc20e5f83f0e updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    13
  val csv = Source.fromFile(name)
fc20e5f83f0e updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    14
  csv.mkString.split("\n").toList.drop(1)
fc20e5f83f0e updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    15
}
fc20e5f83f0e updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    16
fc20e5f83f0e updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    17
def urban_process_ratings(lines: List[String]) : List[(String, String)] = {
fc20e5f83f0e updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    18
  for (cols <- lines.map(_.split(",").toList); 
fc20e5f83f0e updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    19
       if (cols(2).toFloat >= 4)) yield (cols(0), cols(1))  
fc20e5f83f0e updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    20
}
fc20e5f83f0e updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    21
fc20e5f83f0e updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    22
def urban_process_movies(lines: List[String]) : List[(String, String)] = {
fc20e5f83f0e updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    23
  for (cols <- lines.map(_.split(",").toList)) yield (cols(0), cols(1))  
fc20e5f83f0e updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    24
}
fc20e5f83f0e updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    25
fc20e5f83f0e updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    26
fc20e5f83f0e updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    27
val urban_good_ratings = process_ratings(urban_get_csv_file("ratings.csv"))
fc20e5f83f0e updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    28
val urban_movie_names = process_movies(urban_get_csv_file("movies.csv")).toMap
fc20e5f83f0e updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    29
fc20e5f83f0e updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    30
val urban_ratings_map = urban_groupById(urban_good_ratings)
fc20e5f83f0e updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    31
fc20e5f83f0e updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    32
assert((for (n <- List("1", "2", "3", "4", "5")) yield {
fc20e5f83f0e updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    33
  recommendations(urban_ratings_map, urban_movie_names, n).length
fc20e5f83f0e updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    34
}) == List(2, 2, 2, 0, 2))
fc20e5f83f0e updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    35