Merge pull request #143 from drivergroup/improve-phone-number-parsing

SCALA-20 Add ability to parse and store chinese phone numbers (and other countries)

3 files changed, 106 insertions, 28 deletions

diff --git a/build.sbt b/build.sbt
index 9f878f1..88e4582 100644
--- a/build.sbt
+++ b/build.sbt
@@ -7,22 +7,23 @@ lazy val core = (project in file("."))
   .driverLibrary("core")
   .settings(lintingSettings ++ formatSettings)
   .settings(libraryDependencies ++= Seq(
-    "xyz.driver"                   %% "tracing"                        % "0.0.2",
-    "com.typesafe.akka"            %% "akka-http-core"                 % akkaHttpV,
-    "com.typesafe.akka"            %% "akka-http-spray-json"           % akkaHttpV,
-    "com.typesafe.akka"            %% "akka-http-testkit"              % akkaHttpV,
-    "com.pauldijou"                %% "jwt-core"                       % "0.14.0",
-    "org.scalatest"                %% "scalatest"                      % "3.0.2" % "test",
-    "org.scalacheck"               %% "scalacheck"                     % "1.13.4" % "test",
-    "org.scalaz"                   %% "scalaz-core"                    % "7.2.19",
-    "org.mockito"                  % "mockito-core"                    % "1.9.5" % "test",
-    "com.github.swagger-akka-http" %% "swagger-akka-http"              % "0.11.2",
-    "com.amazonaws"                % "aws-java-sdk-s3"                 % "1.11.26",
-    "com.google.cloud"             % "google-cloud-pubsub"             % "0.25.0-beta",
-    "com.google.cloud"             % "google-cloud-storage"            % "1.7.0",
-    "com.typesafe.slick"           %% "slick"                          % "3.2.1",
-    "com.typesafe"                 % "config"                          % "1.3.1",
-    "com.typesafe.scala-logging"   %% "scala-logging"                  % "3.5.0",
-    "eu.timepit"                   %% "refined"                        % "0.8.4",
-    "ch.qos.logback"               % "logback-classic"                 % "1.1.11"
+    "xyz.driver"                    %% "tracing"                        % "0.0.2",
+    "com.typesafe.akka"             %% "akka-http-core"                 % akkaHttpV,
+    "com.typesafe.akka"             %% "akka-http-spray-json"           % akkaHttpV,
+    "com.typesafe.akka"             %% "akka-http-testkit"              % akkaHttpV,
+    "com.pauldijou"                 %% "jwt-core"                       % "0.14.0",
+    "org.scalatest"                 %% "scalatest"                      % "3.0.2" % "test",
+    "org.scalacheck"                %% "scalacheck"                     % "1.13.4" % "test",
+    "org.scalaz"                    %% "scalaz-core"                    % "7.2.19",
+    "com.github.swagger-akka-http"  %% "swagger-akka-http"              % "0.11.2",
+    "com.typesafe.scala-logging"    %% "scala-logging"                  % "3.5.0",
+    "eu.timepit"                    %% "refined"                        % "0.8.4",
+    "com.typesafe.slick"            %% "slick"                          % "3.2.1",
+    "org.mockito"                   %  "mockito-core"                   % "1.9.5" % "test",
+    "com.amazonaws"                 %  "aws-java-sdk-s3"                % "1.11.26",
+    "com.google.cloud"              %  "google-cloud-pubsub"            % "0.25.0-beta",
+    "com.google.cloud"              %  "google-cloud-storage"           % "1.7.0",
+    "com.typesafe"                  %  "config"                         % "1.3.1",
+    "ch.qos.logback"                %  "logback-classic"                % "1.1.11",
+    "com.googlecode.libphonenumber" %  "libphonenumber"                 % "8.9.2"
   ))
diff --git a/src/main/scala/xyz/driver/core/domain.scala b/src/main/scala/xyz/driver/core/domain.scala
index 48943a7..7731345 100644
--- a/src/main/scala/xyz/driver/core/domain.scala
+++ b/src/main/scala/xyz/driver/core/domain.scala
@@ -1,13 +1,14 @@
 package xyz.driver.core
 
+import com.google.i18n.phonenumbers.PhoneNumberUtil
 import scalaz.Equal
-import scalaz.syntax.equal._
 import scalaz.std.string._
+import scalaz.syntax.equal._
 
 object domain {
 
   final case class Email(username: String, domain: String) {
-    override def toString = username + "@" + domain
+    override def toString: String = username + "@" + domain
   }
 
   object Email {
@@ -27,16 +28,13 @@ object domain {
   }
 
   object PhoneNumber {
-    def parse(phoneNumberString: String): Option[PhoneNumber] = {
-      val onlyDigits = phoneNumberString.replaceAll("[^\\d.]", "")
 
-      if (onlyDigits.length < 10) None
-      else {
-        val tenDigitNumber = onlyDigits.takeRight(10)
-        val countryCode    = Option(onlyDigits.dropRight(10)).filter(_.nonEmpty).getOrElse("1")
+    private val phoneUtil = PhoneNumberUtil.getInstance()
 
-        Some(PhoneNumber(countryCode, tenDigitNumber))
-      }
+    def parse(phoneNumber: String): Option[PhoneNumber] = {
+      val phone = phoneUtil.parseAndKeepRawInput(phoneNumber, "US")
+      if (!phoneUtil.isValidNumber(phone)) None
+      else Some(PhoneNumber(phone.getCountryCode.toString, phone.getNationalNumber.toString))
     }
   }
 }
diff --git a/src/test/scala/xyz/driver/core/PhoneNumberTest.scala b/src/test/scala/xyz/driver/core/PhoneNumberTest.scala
new file mode 100644
index 0000000..384c7be
--- /dev/null
+++ b/src/test/scala/xyz/driver/core/PhoneNumberTest.scala
@@ -0,0 +1,79 @@
+package xyz.driver.core
+
+import org.scalatest.{FlatSpec, Matchers}
+import xyz.driver.core.domain.PhoneNumber
+
+class PhoneNumberTest extends FlatSpec with Matchers {
+
+  "PhoneNumber.parse" should "recognize US numbers in international format, ignoring non-digits" in {
+    // format: off
+    val numbers = List(
+      "+18005252225",
+      "+1 800 525 2225",
+      "+1 (800) 525-2225",
+      "+1.800.525.2225")
+    // format: on
+
+    val parsed = numbers.flatMap(PhoneNumber.parse)
+
+    parsed should have size numbers.size
+    parsed should contain only PhoneNumber("1", "8005252225")
+  }
+
+  it should "recognize US numbers without the plus sign" in {
+    PhoneNumber.parse("18005252225") shouldBe Some(PhoneNumber("1", "8005252225"))
+  }
+
+  it should "recognize US numbers without country code" in {
+    // format: off
+    val numbers = List(
+      "8005252225",
+      "800 525 2225",
+      "(800) 525-2225",
+      "800.525.2225")
+    // format: on
+
+    val parsed = numbers.flatMap(PhoneNumber.parse)
+
+    parsed should have size numbers.size
+    parsed should contain only PhoneNumber("1", "8005252225")
+  }
+
+  it should "recognize CN numbers in international format" in {
+    PhoneNumber.parse("+868005252225") shouldBe Some(PhoneNumber("86", "8005252225"))
+    PhoneNumber.parse("+86 134 52 52 2256") shouldBe Some(PhoneNumber("86", "13452522256"))
+  }
+
+  it should "return None on numbers that are shorter than the minimum number of digits for the country (i.e. US - 10, AR - 11)" in {
+    withClue("US and CN numbers are 10 digits - 9 digit (and shorter) numbers should not fit") {
+      // format: off
+      val numbers = List(
+        "+1 800 525-222",
+        "+1 800 525-2",
+        "+86 800 525-222",
+        "+86 800 525-2")
+      // format: on
+
+      numbers.flatMap(PhoneNumber.parse) shouldBe empty
+    }
+
+    withClue("Argentinian numbers are 11 digits (when prefixed with 0) - 10 digit numbers shouldn't fit") {
+      // format: off
+      val numbers = List(
+        "+54 011 525-22256",
+        "+54 011 525-2225",
+        "+54 011 525-222")
+      // format: on
+
+      numbers.flatMap(PhoneNumber.parse) should contain theSameElementsAs List(PhoneNumber("54", "1152522256"))
+    }
+  }
+
+  it should "return None on numbers that are longer than the maximum number of digits for the country (i.e. DK - 8, CN - 11)" in {
+    val numbers = List("+45 27 45 25 22", "+45 135 525 223", "+86 134 525 22256", "+86 135 525 22256 7")
+
+    numbers.flatMap(PhoneNumber.parse) should contain theSameElementsAs
+      List(PhoneNumber("45", "27452522"), PhoneNumber("86", "13452522256"))
+  }
+
+}