Own classes for both Literal and Iri

CMakeLists.txt

@@ -414,9 +414,6 @@ add_executable(ServerMain src/ServerMain.cpp)
 qlever_target_link_libraries (ServerMain engine ${CMAKE_THREAD_LIBS_INIT} Boost::program_options)
 target_precompile_headers(ServerMain REUSE_FROM engine)
 
-add_executable(TurtleParserMain src/TurtleParserMain.cpp)
-qlever_target_link_libraries(TurtleParserMain parser ${CMAKE_THREAD_LIBS_INIT})
-
 add_executable(VocabularyMergerMain src/VocabularyMergerMain.cpp)
 qlever_target_link_libraries(VocabularyMergerMain index ${CMAKE_THREAD_LIBS_INIT})
 

src/engine/ExportQueryExecutionTrees.cpp

@@ -206,15 +206,19 @@ ExportQueryExecutionTrees::idToStringAndType(const Index& index, Id id,
       std::optional<string> entity =
           index.idToOptionalString(id.getVocabIndex());
       AD_CONTRACT_CHECK(entity.has_value());
+      // TODO<joka921> make this more efficient AND more correct
+      auto litOrIri =
+          ad_utility::triple_component::LiteralOrIri::fromStringRepresentation(
+              entity.value());
       if constexpr (onlyReturnLiterals) {
-        if (!entity.value().starts_with('"')) {
+        if (!litOrIri.isLiteral()) {
           return std::nullopt;
         }
       }
       if constexpr (removeQuotesAndAngleBrackets) {
-        entity = RdfEscaping::normalizedContentFromLiteralOrIri(
-            std::move(entity.value()));
+        entity = asStringViewUnsafe(litOrIri.getContent());
       }
+      // TODO<joka921> handle the exporting of literals more correctly.
       return std::pair{escapeFunction(std::move(entity.value())), nullptr};
     }
     case LocalVocabIndex: {

src/engine/sparqlExpressions/LiteralExpression.h

@@ -5,6 +5,7 @@
 #pragma once
 
 #include "engine/sparqlExpressions/SparqlExpression.h"
+#include "util/TypeTraits.h"
 
 namespace sparqlExpression {
 namespace detail {
@@ -38,24 +39,27 @@ class LiteralExpression : public SparqlExpression {
 
   // Evaluating just returns the constant/literal value.
   ExpressionResult evaluate(EvaluationContext* context) const override {
-    // Common code for the `Literal` and `std::string` case.
-    auto getIdOrString = [this,
-                          &context](const std::string& s) -> ExpressionResult {
+    // Common code for the `Literal` and `Iri` case.
+    auto getIdOrString =
+        [this,
+         &context](const ad_utility::SameAsAny<TripleComponent::Literal,
+                                               TripleComponent::Iri> auto& s)
+        -> ExpressionResult {
       if (auto ptr = cachedResult_.load(std::memory_order_relaxed)) {
         return *ptr;
       }
-      Id id;
-      bool idWasFound = context->_qec.getIndex().getId(s, &id);
-      IdOrString result = idWasFound ? IdOrString{id} : IdOrString{s};
+      auto id = context->_qec.getIndex().getId(s);
+      IdOrString result =
+          id.has_value() ? IdOrString{id.value()}
+                         : IdOrString{std::string{s.toStringRepresentation()}};
       auto ptrForCache = std::make_unique<IdOrString>(result);
       ptrForCache.reset(std::atomic_exchange_explicit(
           &cachedResult_, ptrForCache.release(), std::memory_order_relaxed));
       context->cancellationHandle_->throwIfCancelled();
       return result;
     };
-    if constexpr (std::is_same_v<TripleComponent::Literal, T>) {
-      return getIdOrString(_value.rawContent());
-    } else if constexpr (std::is_same_v<string, T>) {
+    if constexpr (ad_utility::SameAsAny<T, TripleComponent::Literal,
+                                        TripleComponent::Iri>) {
       return getIdOrString(_value);
     } else if constexpr (std::is_same_v<Variable, T>) {
       return evaluateIfVariable(context, _value);
@@ -99,7 +103,9 @@ class LiteralExpression : public SparqlExpression {
     } else if constexpr (std::is_same_v<T, ValueId>) {
       return absl::StrCat("#valueId ", _value.getBits(), "#");
     } else if constexpr (std::is_same_v<T, TripleComponent::Literal>) {
-      return absl::StrCat("#literal: ", _value.rawContent());
+      return absl::StrCat("#literal: ", _value.toStringRepresentation());
+    } else if constexpr (std::is_same_v<T, TripleComponent::Iri>) {
+      return absl::StrCat("#iri: ", _value.toStringRepresentation());
     } else if constexpr (std::is_same_v<T, VectorWithMemoryLimit<ValueId>>) {
       // We should never cache this, as objects of this type of expression are
       // used exactly *once* in the HashMap optimization of the GROUP BY
@@ -175,7 +181,7 @@ class LiteralExpression : public SparqlExpression {
 
 ///  The actual instantiations and aliases of LiteralExpressions.
 using VariableExpression = detail::LiteralExpression<::Variable>;
-using IriExpression = detail::LiteralExpression<string>;
+using IriExpression = detail::LiteralExpression<TripleComponent::Iri>;
 using StringLiteralExpression =
     detail::LiteralExpression<TripleComponent::Literal>;
 using IdExpression = detail::LiteralExpression<ValueId>;

src/engine/sparqlExpressions/RegexExpression.cpp

@@ -67,15 +67,6 @@ std::optional<std::string> getPrefixRegex(std::string regex) {
   return regex;
 }
 
-// Assert that `input` starts and ends with double quotes `"` and remove those
-// quotes.
-std::string removeQuotes(std::string_view input) {
-  AD_CORRECTNESS_CHECK(input.size() >= 2 && input.starts_with('"') &&
-                       input.ends_with('"'));
-  input.remove_prefix(1);
-  input.remove_suffix(1);
-  return std::string{input};
-}
 }  // namespace sparqlExpression::detail
 
 namespace sparqlExpression {
@@ -93,16 +84,15 @@ RegexExpression::RegexExpression(
         "REGEX expressions are currently supported only on variables.");
   }
   std::string regexString;
-  std::string originalRegexString;
   if (auto regexPtr =
           dynamic_cast<const StringLiteralExpression*>(regex.get())) {
-    originalRegexString = regexPtr->value().normalizedLiteralContent().get();
-    if (!regexPtr->value().datatypeOrLangtag().empty()) {
+    const auto& regexLiteral = regexPtr->value();
+    regexString = asStringViewUnsafe(regexLiteral.getContent());
+    if (regexLiteral.hasDatatype() || regexLiteral.hasLanguageTag()) {
       throw std::runtime_error(
           "The second argument to the REGEX function (which contains the "
           "regular expression) must not contain a language tag or a datatype");
     }
-    regexString = detail::removeQuotes(originalRegexString);
   } else {
     throw std::runtime_error(
         "The second argument to the REGEX function must be a "
@@ -111,15 +101,14 @@ RegexExpression::RegexExpression(
   if (optionalFlags.has_value()) {
     if (auto flagsPtr = dynamic_cast<const StringLiteralExpression*>(
             optionalFlags.value().get())) {
-      std::string_view originalFlags =
-          flagsPtr->value().normalizedLiteralContent().get();
-      if (!flagsPtr->value().datatypeOrLangtag().empty()) {
+      const auto& flagsLiteral = flagsPtr->value();
+      std::string_view flags = asStringViewUnsafe(flagsLiteral.getContent());
+      if (flagsLiteral.hasDatatype() || flagsLiteral.hasLanguageTag()) {
         throw std::runtime_error(
             "The third argument to the REGEX function (which contains optional "
             "flags to configure the evaluation) must not contain a language "
             "tag or a datatype");
       }
-      auto flags = detail::removeQuotes(originalFlags);
       auto firstInvalidFlag = flags.find_first_not_of("imsu");
       if (firstInvalidFlag != std::string::npos) {
         throw std::runtime_error{absl::StrCat(
@@ -148,8 +137,8 @@ RegexExpression::RegexExpression(
     const auto& r = std::get<RE2>(regex_);
     if (r.error_code() != RE2::NoError) {
       throw std::runtime_error{absl::StrCat(
-          "The regex ", originalRegexString,
-          " is not supported by QLever (which uses Google's RE2 library). "
+          "The regex \"", regexString,
+          "\" is not supported by QLever (which uses Google's RE2 library). "
           "Error from RE2 is: ",
           r.error())};
     }

src/engine/sparqlExpressions/RelationalExpressions.cpp

@@ -299,7 +299,7 @@ RelationalExpression<Comp>::getLanguageFilterExpression() const {
     // TODO<joka921> Is this even allowed by the grammar?
     return LangFilterData{
         varPtr->variable(),
-        std::string{langPtr->value().normalizedLiteralContent().get()}};
+        std::string{asStringViewUnsafe(langPtr->value().getContent())}};
   };
 
   const auto& child1 = children_[0];

src/global/Constants.h

@@ -64,9 +64,9 @@ static const std::string HAS_PREDICATE_PREDICATE =
     makeInternalIri("has-predicate");
 static const std::string HAS_PATTERN_PREDICATE = makeInternalIri("has-pattern");
 static constexpr std::pair<std::string_view, std::string_view> GEOF_PREFIX = {
-    "geof:", "<http://www.opengis.net/def/function/geosparql/"};
+    "geof:", "http://www.opengis.net/def/function/geosparql/"};
 static constexpr std::pair<std::string_view, std::string_view> MATH_PREFIX = {
-    "math:", "<http://www.w3.org/2005/xpath-functions/math#"};
+    "math:", "http://www.w3.org/2005/xpath-functions/math#"};
 
 static const std::string INTERNAL_VARIABLE_PREFIX =
     "?_QLever_internal_variable_";