Merge pull request #87 from valydia/feature_enforce_column_case

Feature enforce column case
digital-preservation · Nov 11, 2015 · 88b0978 · 88b0978
2 parents e95a5c5 + 98bdcb8
commit 88b0978
Show file tree

Hide file tree

Showing 4 changed files with 386 additions and 2 deletions.
diff --git a/csv-validator-core/src/main/scala/uk/gov/nationalarchives/csv/validator/schema/Rule.scala b/csv-validator-core/src/main/scala/uk/gov/nationalarchives/csv/validator/schema/Rule.scala
@@ -339,6 +339,10 @@ case class Uuid4Rule() extends PatternRule("uuid4", Uuid4Regex)
 
 case class PositiveIntegerRule() extends PatternRule("positiveInteger", PositiveIntegerRegex)
 
+case class UpperCaseRule() extends PatternRule("upperCase", UpperCaseRegex)
+
+case class LowerCaseRule() extends PatternRule("lowerCase", LowerCaseRegex)
+
 case class IdenticalRule() extends Rule("identical") {
 
   var lastValue: Option[String] = None

diff --git a/...dator-core/src/main/scala/uk/gov/nationalarchives/csv/validator/schema/SchemaParser.scala b/...dator-core/src/main/scala/uk/gov/nationalarchives/csv/validator/schema/SchemaParser.scala
@@ -263,7 +263,7 @@ trait SchemaParser extends RegexParsers
    *                        XsdDateTimeExpr | XsdDateExpr | XsdTimeExpr |
    *                        UkDateExpr | DateExpr | PartialUkDateExpr | PartialDateExpr |
    *                        uuid4Expr |
-   *                        PositiveIntegerExpr)
+   *                        PositiveIntegerExpr | UpperCaseExpr | LowerCaseExpr)
    */
   //TODO need to implement and add DateExpr, PartialDateExpr
   lazy val singleExpr: PackratParser[Rule] = "SingleExpr" ::=
@@ -275,7 +275,7 @@ trait SchemaParser extends RegexParsers
       xsdDateTimeExpr | xsdDateExpr | xsdTimeExpr |
       ukDateExpr | partialUkDateExpr |
       uuid4Expr |
-      positiveIntegerExpr) ^^ {
+      positiveIntegerExpr | upperCaseExpr | lowerCaseExpr) ^^ {
     case explicitContext ~ rule =>
       rule.explicitColumn = explicitContext
       rule
@@ -475,6 +475,12 @@ trait SchemaParser extends RegexParsers
    */
   lazy val positiveIntegerExpr: PackratParser[PositiveIntegerRule] = "PositiveIntegerExpr" ::= "positiveInteger" ^^^ PositiveIntegerRule()
 
+
+  lazy val upperCaseExpr: PackratParser[UpperCaseRule] = "UpperCaseExpr" ::= "upperCase" ^^^ UpperCaseRule()
+
+
+  lazy val lowerCaseExpr: PackratParser[LowerCaseRule] = "LowerCaseExpr" ::= "lowerCase" ^^^ LowerCaseRule()
+
   /**
    * [59] StringProvider ::= ColumnRef | StringLiteral
    */
@@ -501,6 +507,7 @@ trait SchemaParser extends RegexParsers
       FileExistsRule(pathSubstitutions, enforceCaseSensitivePathChecks, s)
   }
 
+
   /**
    * [62] ChecksumExpr ::= "checksum(" FileExpr "," StringLiteral ")" /* first arg is file expr, second arg is algorithm to use for checksum */
    */

diff --git a/csv-validator-core/src/main/scala/uk/gov/nationalarchives/csv/validator/schema/package.scala b/csv-validator-core/src/main/scala/uk/gov/nationalarchives/csv/validator/schema/package.scala
@@ -22,6 +22,8 @@ package object schema {
 
   val UkDateRegex = """(((0[1-9]|(1|2)[0-9]|3[0-1])\/(0(1|3|5|7|8)|1(0|2)))|((0[1-9]|(1|2)[0-9]|30)\/(0(4|6|9)|11))|((0[1-9]|(1|2)[0-9])\/02))\/[0-9]{4}"""
   val PositiveIntegerRegex = "[0-9]+"
+  val UpperCaseRegex = "^[\\p{Lu}\\p{N}\\p{P}\\s]*$" //"^[\\p{Lu}\\p{N}\\p{P}\\s]+$" if we ant to dismiss empty string
+  val LowerCaseRegex = "^[\\p{Ll}\\p{N}\\p{P}\\s]*$"
   val UkDateFormat = "dd/MM/YYYY"
   val PartUkDateRegex = """(([0\?][1-9\?])|([1-2\?][0-9\?])|([3\?][0-1\?])|\*)\/(January|February|March|April|May|June|July|August|September|October|November|December|\?|\*)\/([0-9\?]{4}|\*)"""