Add support for "XXX"^^foo:typename Value Literals

This requires searching for the ':' from the right because it may appear in dateTime strings. Do the same for the '#' to keep it symmetric.
ad-freiburg · Dec 21, 2017 · d99c50d · d99c50d
1 parent a9184f3
commit d99c50d
Show file tree

Hide file tree

Showing 2 changed files with 51 additions and 7 deletions.
diff --git a/src/util/Conversions.h b/src/util/Conversions.h
@@ -70,17 +70,38 @@ inline bool isXsdValue(const string val);
 
 // _____________________________________________________________________________
 string convertValueLiteralToIndexWord(const string& orig) {
+  /*
+   * Value literals can have one of two forms
+   * 0) "123"^^<http://www.w3.org/2001/XMLSchema#integer>
+   * 1) "123"^^xsd:integer
+   *
+   * TODO: This ignores the URI such that xsd:integer == foo:integer ==
+   * <http://baz#integer>
+   */
   assert(orig.size() > 0);
   assert(orig[0] == '\"');
-  assert(orig[orig.size() - 1] == '>');
-  size_t posOfSecondQuote = orig.find('\"', 1);
-  size_t posOfHashTag = orig.find('#');
+  string value;
+  string type;
+  size_t posOfSecondQuote = orig.rfind('\"');
   assert(posOfSecondQuote != string::npos);
-  assert(posOfHashTag != string::npos);
+  // -1 for the quote and since substr takes a length
+  // not an end position
+  value = orig.substr(1, posOfSecondQuote - 1);
+  if (orig[orig.size() - 1] == '>') {
+    size_t posOfHashTag = orig.rfind('#');
+    assert(posOfHashTag != string::npos);
+
+    // +2 for '>' and Hashtag
+    type = orig.substr(posOfHashTag + 1,
+                       orig.size() - (posOfHashTag + 2));
+  } else {
+    size_t posOfDoubleDot = orig.rfind(':');
+    assert(posOfDoubleDot != string::npos);
 
-  string value = orig.substr(1, posOfSecondQuote - 1);
-  string type = orig.substr(posOfHashTag + 1,
-                            orig.size() - (posOfHashTag + 2));
+    // +1 for double dot
+    type = orig.substr(posOfDoubleDot + 1,
+                       orig.size() - (posOfDoubleDot + 1));
+  }
 
   if (type == "dateTime" || type == "gYear" || type == "gYearMonth"
       || type == "date") {

diff --git a/test/ConversionsTest.cpp b/test/ConversionsTest.cpp
@@ -128,12 +128,35 @@ TEST(ConversionsTest, convertDateToIndexWord) {
 TEST(ConversionsTest, endToEndDate) {
   string in = "\"1990-01-01T00:00:00\"^^<http://www.w3.org/2001/XMLSchema#dateTime>";
   string in2 = "\"1990-01-01\"^^<http://www.w3.org/2001/XMLSchema#date>";
+  string in3 = "\"1990-01-01\"^^xsd:date";
+  string in4 = "\"1990-01-01T00:00:00\"^^xsd:dateTime";
   string indexW = ":v:date:0000000000000001990-01-01T00:00:00";
   ASSERT_EQ(indexW, convertValueLiteralToIndexWord(in));
   ASSERT_EQ(indexW, convertValueLiteralToIndexWord(in2));
+  ASSERT_EQ(indexW, convertValueLiteralToIndexWord(in3));
+  ASSERT_EQ(indexW, convertValueLiteralToIndexWord(in4));
   ASSERT_EQ(in, convertIndexWordToValueLiteral(indexW));
 }
 
+TEST(ConversionsTest, shortFormEquivalence) {
+  string in = "\"1230.7\"^^<http://www.w3.org/2001/XMLSchema#float>";
+  string in_short = "\"1230.7\"^^xsd:float";
+  string nin = "\"-1230.7\"^^<http://www.w3.org/2001/XMLSchema#float>";
+  string nin_short = "\"-1230.7\"^^xsd:float";
+  string in2 = "\"1000\"^^<http://www.w3.org/2001/XMLSchema#int>";
+  string in2_short = "\"1000\"^^xsd:int";
+  string nin2 = "\"-1000\"^^<http://www.w3.org/2001/XMLSchema#int>";
+  string nin2_short = "\"-1000\"^^xsd:int";
+  ASSERT_EQ(convertValueLiteralToIndexWord(in),
+            convertValueLiteralToIndexWord(in_short));
+  ASSERT_EQ(convertValueLiteralToIndexWord(nin),
+            convertValueLiteralToIndexWord(nin_short));
+  ASSERT_EQ(convertValueLiteralToIndexWord(in2),
+            convertValueLiteralToIndexWord(in2_short));
+  ASSERT_EQ(convertValueLiteralToIndexWord(nin2),
+            convertValueLiteralToIndexWord(nin2_short));
+}
+
 TEST(ConversionsTest, endToEndNumbers) {
   string in = "\"1000\"^^<http://www.w3.org/2001/XMLSchema#int>";
   string nin = "\"-1000\"^^<http://www.w3.org/2001/XMLSchema#int>";