diff --git a/.gitignore b/.gitignore
index c58d83b..0f79af5 100644
--- a/.gitignore
+++ b/.gitignore
@@ -15,3 +15,6 @@ project/plugins/project/
 # Scala-IDE specific
 .scala_dependencies
 .worksheet
+
+# IntelliJ IDEA
+/.idea
\ No newline at end of file
diff --git a/build.sbt b/build.sbt
new file mode 100644
index 0000000..76be9ef
--- /dev/null
+++ b/build.sbt
@@ -0,0 +1,12 @@
+name := "lyrics-engine"
+
+version := "1.0"
+
+scalaVersion := "2.11.8"
+
+
+lazy val raw_data_processor = project in file("raw_data_processor")
+lazy val data_analyzer = project in file("data_analyzer")
+lazy val http_api = project in file("http_api")
+
+// TODO Setup visibility common dependencies.
\ No newline at end of file
diff --git a/data_analyzer/build.sbt b/data_analyzer/build.sbt
new file mode 100644
index 0000000..306924f
--- /dev/null
+++ b/data_analyzer/build.sbt
@@ -0,0 +1,8 @@
+name := "data_analyzer"
+
+version := "1.0"
+
+scalaVersion := "2.11.8"
+
+//libraryDependencies += "org.apache.spark" % "spark-core_2.10" % "1.6.1"
+//libraryDependencies += "org.apache.spark" % "spark-mllib_2.10" % "1.6.1"
diff --git a/http_api/build.sbt b/http_api/build.sbt
new file mode 100644
index 0000000..761a2ad
--- /dev/null
+++ b/http_api/build.sbt
@@ -0,0 +1,5 @@
+name := "http_api"
+
+version := "1.0"
+
+scalaVersion := "2.11.8"
\ No newline at end of file
diff --git a/raw_data_processor/build.sbt b/raw_data_processor/build.sbt
new file mode 100644
index 0000000..d080a34
--- /dev/null
+++ b/raw_data_processor/build.sbt
@@ -0,0 +1,26 @@
+name := "raw_data_processor"
+
+version := "1.0"
+
+scalaVersion := "2.11.8"
+
+// -------- -------- Dependencies: -------- --------
+
+// -------- Config:
+// https://github.com/typesafehub/config
+libraryDependencies += "com.typesafe" % "config" % "1.3.0"
+
+// -------- Logger:
+// http://logback.qos.ch/
+libraryDependencies += "ch.qos.logback" % "logback-classic" % "1.1.7"
+// https://github.com/typesafehub/scala-logging
+libraryDependencies += "com.typesafe.scala-logging" %% "scala-logging" % "3.4.0"
+
+// -------- Test:
+// http://www.scalatest.org/
+libraryDependencies += "org.scalatest" % "scalatest_2.11" % "2.2.6"
+
+// -------- DB:
+// mongo:
+// http://mongodb.github.io/mongo-scala-driver/1.1/
+libraryDependencies += "org.mongodb.scala" % "mongo-scala-driver_2.11" % "1.1.0"
\ No newline at end of file
diff --git a/raw_data_processor/project/build.properties b/raw_data_processor/project/build.properties
new file mode 100644
index 0000000..d638b4f
--- /dev/null
+++ b/raw_data_processor/project/build.properties
@@ -0,0 +1 @@
+sbt.version = 0.13.8
\ No newline at end of file
diff --git a/raw_data_processor/project/plugins.sbt b/raw_data_processor/project/plugins.sbt
new file mode 100644
index 0000000..14a6ca1
--- /dev/null
+++ b/raw_data_processor/project/plugins.sbt
@@ -0,0 +1 @@
+logLevel := Level.Warn
\ No newline at end of file
diff --git a/raw_data_processor/src/main/resources/application.conf b/raw_data_processor/src/main/resources/application.conf
new file mode 100644
index 0000000..1b66bcf
--- /dev/null
+++ b/raw_data_processor/src/main/resources/application.conf
@@ -0,0 +1,12 @@
+// ---- ---- Million Song Dataset data:
+// subset from: http://labrosa.ee.columbia.edu/millionsong/sites/default/files/AdditionalFiles/mxm_dataset_train.txt.zip
+msd.pathToMsd = "C:\\Temp\\mxm_dataset_train.txt"
+
+// ---- ---- DB:
+// mongoDB:
+// Nice mongo admin: https://github.com/mrvautin/adminMongo
+// Short introduction to mongoDB installation:
+// http://stackoverflow.com/questions/26585433/mongodb-failed-to-connect-to-127-0-0-127017-reason-errno10061
+mongoDb.host = localhost
+mongoDb.port = 27017
+mongoDb.dbName = liric
\ No newline at end of file
diff --git a/raw_data_processor/src/main/scala/scalalab3/lyricsengine/processor/SongProcessor.scala b/raw_data_processor/src/main/scala/scalalab3/lyricsengine/processor/SongProcessor.scala
new file mode 100644
index 0000000..cb083ab
--- /dev/null
+++ b/raw_data_processor/src/main/scala/scalalab3/lyricsengine/processor/SongProcessor.scala
@@ -0,0 +1,7 @@
+package scalalab3.lyricsengine.processor
+
+/**
+  * Performs operations above songs.
+  */
+object SongProcessor {
+}
diff --git a/raw_data_processor/src/main/scala/scalalab3/lyricsengine/processor/WordsProcessor.scala b/raw_data_processor/src/main/scala/scalalab3/lyricsengine/processor/WordsProcessor.scala
new file mode 100644
index 0000000..8e6e3c5
--- /dev/null
+++ b/raw_data_processor/src/main/scala/scalalab3/lyricsengine/processor/WordsProcessor.scala
@@ -0,0 +1,7 @@
+package scalalab3.lyricsengine.processor
+
+/**
+  * Performs operations above words.
+  */
+object WordsProcessor {
+}
diff --git a/raw_data_processor/src/main/scala/scalalab3/lyricsengine/reader/MillionSongDataSubSetReader.scala b/raw_data_processor/src/main/scala/scalalab3/lyricsengine/reader/MillionSongDataSubSetReader.scala
new file mode 100644
index 0000000..e0cdec1
--- /dev/null
+++ b/raw_data_processor/src/main/scala/scalalab3/lyricsengine/reader/MillionSongDataSubSetReader.scala
@@ -0,0 +1,49 @@
+package scalalab3.lyricsengine.reader
+
+import scala.io.Source
+
+object MillionSongDataSubSetReader {
+  def extract(path: String): MillionSongDataSubSet = {
+    val wordsSongsPair = readRawData(path)
+    new MillionSongDataSubSet(extractWords(wordsSongsPair._1), extractSongs(wordsSongsPair._2))
+  }
+
+  private def readRawData(path: String) = {
+    val lines = Source.fromFile(path).getLines().toSeq
+    // 17 - Line that contains the words.
+    val words = lines(17)
+    // 18 - First line that contains the data by songs.
+    val songs = lines.slice(18, lines.size)
+    words -> songs
+  }
+
+  private def extractWords(words: String) = words.substring(1).split(",");
+
+  private def extractSongs(songsData: Seq[String]) = songsData.map(
+    songData => {
+      val elements = songData.split(",")
+      val trackId = elements(0)
+      val mxmTrackId = elements(1)
+      val words_map = extractSongWords(elements.slice(2, elements.size))
+      new Song(trackId, mxmTrackId, words_map);
+    }).toSeq
+
+  private def extractSongWords(words: Array[String]) = words.toSeq.map(
+    pairInStringForm => {
+      val pair = pairInStringForm.split(":")
+      val wordIndex = pair(0).toInt
+      val count = pair(1).toInt
+      wordIndex -> count
+    }).toMap
+}
+
+/**
+  * That class represent subset of data from Million Song Dataset (MSD).
+  * http://labrosa.ee.columbia.edu/millionsong/sites/default/files/AdditionalFiles/mxm_dataset_train.txt.zip
+  */
+case class MillionSongDataSubSet(words: Seq[String], songs: Seq[Song])
+
+case class Song(trackId: String, mxmTrackId: String, words: Map[Int, Int])
+
+
+
diff --git a/raw_data_processor/src/main/scala/scalalab3/lyricsengine/repositories/MillionSongDataSubSetRepository.scala b/raw_data_processor/src/main/scala/scalalab3/lyricsengine/repositories/MillionSongDataSubSetRepository.scala
new file mode 100644
index 0000000..f83aadb
--- /dev/null
+++ b/raw_data_processor/src/main/scala/scalalab3/lyricsengine/repositories/MillionSongDataSubSetRepository.scala
@@ -0,0 +1,18 @@
+package scalalab3.lyricsengine.repositories
+
+
+import scalalab3.lyricsengine.system.{ProjectConfig, ProjectLogger}
+
+/**
+  * Repository for mongoDB
+  */
+// TODO: move to API?
+class MillionSongDataSubSetRepository extends ProjectConfig with ProjectLogger {
+  def saveSongs(songs: Any) = ???
+
+  def saveWords(songs: Any) = ???
+
+
+  def findAll() = ???
+}
+
diff --git a/raw_data_processor/src/main/scala/scalalab3/lyricsengine/system/ProjectConfig.scala b/raw_data_processor/src/main/scala/scalalab3/lyricsengine/system/ProjectConfig.scala
new file mode 100644
index 0000000..e6dc7bd
--- /dev/null
+++ b/raw_data_processor/src/main/scala/scalalab3/lyricsengine/system/ProjectConfig.scala
@@ -0,0 +1,24 @@
+package scalalab3.lyricsengine.system
+
+import com.typesafe.config.ConfigFactory
+
+
+trait ProjectConfig {
+  val conf = ConfigFactory.load()
+  // Config variable:
+  // ---- Million Song Dataset data:
+  val PATH_TO_MSD_SUBSET_VAR_NAME = "msd.pathToMsd"
+  // ---- MongoDB:
+  val MONGODB_HOST_VAR_NAME = "mongoDb.host"
+  val MONGODB_PORT_VAR_NAME = "mongoDb.port"
+  val MONGODB_DB_NAME_VAR_NAME = "mongoDb.dbName"
+
+
+  def getPathToData() = conf.getString(PATH_TO_MSD_SUBSET_VAR_NAME)
+
+  def getMongoDbHost() = conf.getString(MONGODB_HOST_VAR_NAME)
+
+  def getMongoDbPort() = conf.getInt(MONGODB_PORT_VAR_NAME)
+
+  def getMongoDbName() = conf.getString(MONGODB_DB_NAME_VAR_NAME)
+}
diff --git a/raw_data_processor/src/main/scala/scalalab3/lyricsengine/system/ProjectLogger.scala b/raw_data_processor/src/main/scala/scalalab3/lyricsengine/system/ProjectLogger.scala
new file mode 100644
index 0000000..dc5c72c
--- /dev/null
+++ b/raw_data_processor/src/main/scala/scalalab3/lyricsengine/system/ProjectLogger.scala
@@ -0,0 +1,8 @@
+package scalalab3.lyricsengine.system
+
+import com.typesafe.scalalogging.Logger
+import org.slf4j.LoggerFactory
+
+trait ProjectLogger {
+  val logger = Logger(LoggerFactory.getLogger("common"))
+}
diff --git a/raw_data_processor/src/test/resources/test_msd.txt b/raw_data_processor/src/test/resources/test_msd.txt
new file mode 100644
index 0000000..e04a3dd
--- /dev/null
+++ b/raw_data_processor/src/test/resources/test_msd.txt
@@ -0,0 +1,20 @@
+# TRAINING SET
+# MusiXmatch dataset, the official lyrics dataset
+# of the Million Song Dataset
+#    file created on Tue Mar 29 04:28:44 2011
+#    contact: T. Bertin-Mahieux (Columbia University)
+#             tb2332@columbia.edu
+#    also: http://labrosa.ee.columbia.edu/millionsong/musixmatch
+#          http://www.musixmatch.com
+# FORMAT:
+#     #   - comment, to ignore
+#     %   - list of top words, comma-separated
+#         - normal line, contains track_id, mxm track id,
+#           then word count for each of the top words, comma-separated
+#           word count is in sparse format -> ...,<word idx>:<cnt>,...
+#           <word idx> starts at 1 (not zero!)
+# All our work is done using UTF-8 encoding.
+# enjoy!
+%i,the,you,to,and
+TRZZZYV128F92E996D,6849828,1:10,2:6,3:20,5:2,7:30
+TRZZZYX128F92D32C6,681124,1:4,2:18,4:3,5:6,6:9
\ No newline at end of file
diff --git a/raw_data_processor/src/test/scala/MillionSongDataSubSetReaderTest.scala b/raw_data_processor/src/test/scala/MillionSongDataSubSetReaderTest.scala
new file mode 100644
index 0000000..ff1d300
--- /dev/null
+++ b/raw_data_processor/src/test/scala/MillionSongDataSubSetReaderTest.scala
@@ -0,0 +1,30 @@
+import org.scalatest.{FlatSpec, Matchers}
+
+import scalalab3.lyricsengine.reader.{MillionSongDataSubSet, MillionSongDataSubSetReader, Song}
+
+class MillionSongDataSubSetReaderTest extends FlatSpec with Matchers {
+
+  "A MillionSongDataSubSetReader" should "correct load all data" in {
+    // --- Prepare expected data:
+    val expectedData: MillionSongDataSubSet = prepareTestData();
+
+    // --- Computation actual data:
+    val path = getClass.getResource("/test_msd.txt").getPath
+    val actualData: MillionSongDataSubSet = MillionSongDataSubSetReader.extract(path)
+
+    // --- Checking data:
+    assert(expectedData.words == actualData.words)
+    assert(expectedData.songs.size == actualData.songs.size)
+    actualData.songs(0) == expectedData.songs(0)
+    actualData.songs(1) == expectedData.songs(1)
+    //Just in case.
+    assert(expectedData.songs == actualData.songs)
+  }
+
+  def prepareTestData() = {
+    val words = List("i", "the", "you", "to", "and")
+    val firstSong = new Song("TRZZZYV128F92E996D", "6849828", Map(1 -> 10, 2 -> 6, 3 -> 20, 5 -> 2, 7 -> 30))
+    val secondSong = new Song("TRZZZYX128F92D32C6", "681124", Map(1 -> 4, 2 -> 18, 4 -> 3, 5 -> 6, 6 -> 9))
+    new MillionSongDataSubSet(words, Seq(firstSong, secondSong))
+  }
+}