postal code opendata data sources. / published by https://github.com/dacr/code-examples-manager #095ffa72-b4d9-4f3d-85b2-b3e69a302ac4/e7bbc4649a8dfe6129e20c9be9a626fba0200f77

opendata-postalcodes-2.sc

// summary : postal code opendata data sources.
// keywords : scala, opendata, data-analysis, requests, postal-codes, @testable
// publish : gist
// authors : David Crosson
// license : Apache NON-AI License Version 2.0 (https://raw.githubusercontent.com/non-ai-licenses/non-ai-licenses/main/NON-AI-APACHE2)
// id : 095ffa72-b4d9-4f3d-85b2-b3e69a302ac4
// created-on : 2020-10-10T16:21:18Z
// managed-by : https://github.com/dacr/code-examples-manager
// run-with : scala-cli $file

// ---------------------
//> using scala  "3.2.1"
//> using dep "com.lihaoyi::requests:0.8.0"
// ---------------------

case class GPS(latitude:Double, longitude:Double)

case class PostalCode(
  townCode: String,
  townName: String,
  postalCode: String,
  secondaryTownName: Option[String],
  deliveryLabel: String,
  gps: Option[GPS]
) {
  val countyCode = townCode.take(if (townCode.startsWith("97")) 3 else 2)
}

def stringToGPS(input:String):Option[GPS] = {
  input.split(",").map(_.trim) match {
    case Array(latitude, longitude) =>
      for {
        lat <- latitude.toDoubleOption
        lon <- longitude.toDoubleOption
      } yield GPS(lat,lon)
    case _ => None
  }
}

def stringToPostalCode(input:String):Option[PostalCode] = {
  input.trim.split(";") match {
    case Array(townCode, townName, postalCode, "", deliveryLabel, position) =>
      Option(PostalCode(townCode, townName, postalCode, None, deliveryLabel, stringToGPS(position)))
    case Array(townCode, townName, postalCode, secondaryTownName, deliveryLabel, position) =>
      Option(PostalCode(townCode, townName, postalCode, Some(secondaryTownName), deliveryLabel, stringToGPS(position)))
    case Array(townCode, townName, postalCode, "", deliveryLabel) =>
      Option(PostalCode(townCode, townName, postalCode, None, deliveryLabel, None))
    case Array(townCode, townName, postalCode, secondaryTownName, deliveryLabel) =>
      Option(PostalCode(townCode, townName, postalCode, Some(secondaryTownName), deliveryLabel, None))
    case data =>
      println("Unmanaged input : "+data.mkString(";"))
      None
  }
}

// code postaux https://www.data.gouv.fr/fr/datasets/base-officielle-des-codes-postaux/
val response = requests.get("https://www.data.gouv.fr/fr/datasets/r/554590ab-ae62-40ac-8353-ee75162c05ee")
val postalCodes = {
  response
    .lines()
    .drop(1) // first line == the CSV labels
    .flatMap(stringToPostalCode)
}

val townByCounty = postalCodes.toList.groupMap(_.countyCode)(_.townName)
val longestTownName = postalCodes.maxByOption(_.townName.count(_.isLetter))
val shortestTownName = postalCodes.minByOption(_.townName.count(_.isLetter))
val countyWithMostTowns = townByCounty.maxByOption { case (countyCode, towns) => towns.size }.map{ case (countyCode, towns) => countyCode -> towns.size}

postalCodes
  .sortBy(_.postalCode)
  .foreach(println)

println(
  s"""postalCodeCount : ${postalCodes.size}
     |longestTownName : ${longestTownName}
     |shortestTownName : ${shortestTownName}
     |countyWithMostTowns : ${countyWithMostTowns}
     |""".stripMargin
)