Improved title case conversion, add tests

qgis · May 16, 2018 · 3949802 · DelazJ · Aug 16, 2018 · 3949802
1 parent 825a697
commit 3949802
Show file tree

Hide file tree

Showing 2 changed files with 45 additions and 7 deletions.
diff --git a/src/core/qgsstringutils.cpp b/src/core/qgsstringutils.cpp
@@ -18,6 +18,7 @@
 #include <QRegExp>
 #include <QStringList>
 #include <QTextBoundaryFinder>
+#include <QRegularExpression>
 
 QString QgsStringUtils::capitalize( const QString &string, QgsStringUtils::Capitalization capitalization )
 {
@@ -61,29 +62,37 @@ QString QgsStringUtils::capitalize( const QString &string, QgsStringUtils::Capit
       // yes, this is MASSIVELY simplifying the problem!!
 
       static QStringList smallWords;
+      static QStringList newPhraseSeparators;
+      static QRegularExpression splitWords;
       if ( smallWords.empty() )
       {
-        smallWords = QObject::tr( "a|an|and|as|at|but|by|en|for|if|in|nor|of|on|or|per|the|to|vs.|vs|via" ).split( '|' );
+        smallWords = QObject::tr( "a|an|and|as|at|but|by|en|for|if|in|nor|of|on|or|per|s|the|to|vs.|vs|via" ).split( '|' );
+        newPhraseSeparators = QObject::tr( ".|:" ).split( '|' );
+        splitWords = QRegularExpression( QStringLiteral( "\\b" ) );
       }
 
-      const QStringList parts = string.split( ' ' );
+      const QStringList parts = string.split( splitWords, QString::SkipEmptyParts );
       QString result;
       bool firstWord = true;
+      int i = 0;
+      int lastWord = parts.count() - 1;
       for ( const QString &word : qgis::as_const( parts ) )
       {
-        if ( word.isEmpty() )
+        if ( newPhraseSeparators.contains( word.trimmed() ) )
         {
-          result += ' ';
+          firstWord = true;
+          result += word;
         }
-        else if ( firstWord || !smallWords.contains( word ) )
+        else if ( firstWord || ( i == lastWord ) || !smallWords.contains( word ) )
         {
-          result += ' ' + word.at( 0 ).toUpper() + word.mid( 1 );
+          result += word.at( 0 ).toUpper() + word.mid( 1 );
           firstWord = false;
         }
         else
         {
-          result += ' ' + word;
+          result += word;
         }
+        i++;
       }
       return result;
     }

diff --git a/tests/src/core/testqgsstringutils.cpp b/tests/src/core/testqgsstringutils.cpp
@@ -34,6 +34,8 @@ class TestQgsStringUtils : public QObject
     void hammingDistance();
     void soundex();
     void insertLinks();
+    void titleCase_data();
+    void titleCase();
 
 };
 
@@ -154,6 +156,33 @@ void TestQgsStringUtils::insertLinks()
   QVERIFY( !found );
 }
 
+void TestQgsStringUtils::titleCase_data()
+{
+  QTest::addColumn<QString>( "input" );
+  QTest::addColumn<QString>( "expected" );
+
+  // invalid strings
+  QTest::newRow( "empty string" ) << "" << "";
+  QTest::newRow( "single character" ) << "a" << "A";
+  QTest::newRow( "string 1" ) << "follow step-by-step instructions" << "Follow Step-by-Step Instructions";
+  QTest::newRow( "string 2" ) << "this sub-phrase is nice" << "This Sub-Phrase Is Nice";
+  QTest::newRow( "" ) << "catchy title: a subtitle" << "Catchy Title: A Subtitle";
+  QTest::newRow( "string 3" ) << "all words capitalized" << "All Words Capitalized";
+  QTest::newRow( "string 4" ) << "small words are for by and of lowercase" << "Small Words Are for by and of Lowercase";
+  QTest::newRow( "string 5" ) << "a small word starts" << "A Small Word Starts";
+  QTest::newRow( "last word" ) << "a small word it ends on" << "A Small Word It Ends On";
+  QTest::newRow( "last word2" ) << "Ends with small word of" << "Ends With Small Word Of";
+  QTest::newRow( "string 6" ) << "Merge VRT(s)" << "Merge VRT(s)";
+  QTest::newRow( "string 6" ) << "multiple sentences. more than one." << "Multiple Sentences. More Than One.";
+}
+
+void TestQgsStringUtils::titleCase()
+{
+  QFETCH( QString, input );
+  QFETCH( QString, expected );
+  QCOMPARE( QgsStringUtils::capitalize( input, QgsStringUtils::TitleCase ), expected );
+}
+
 
 QGSTEST_MAIN( TestQgsStringUtils )
 #include "testqgsstringutils.moc"