Feature: RDF/XML parser

CMakeLists.txt

@@ -23,6 +23,7 @@ find_package(highway REQUIRED)
 find_package(dice-hash REQUIRED)
 find_package(dice-sparse-map REQUIRED)
 find_package(dice-template-library REQUIRED)
+find_package(libxml2 REQUIRED)
 
 configure_file(${CMAKE_CURRENT_SOURCE_DIR}/cmake/version.hpp.in ${CMAKE_CURRENT_SOURCE_DIR}/src/rdf4cpp/version.hpp)
 
@@ -149,6 +150,7 @@ add_library(rdf4cpp
         src/rdf4cpp/IRIFactory.cpp
         src/rdf4cpp/util/Anonymizer.cpp
         private/rdf4cpp/parser/IStreamQuadIteratorSerdImpl.cpp
+        private/rdf4cpp/parser/XMLParser.cpp
         private/rdf4cpp/regex/RegexImpl.cpp
         private/rdf4cpp/regex/RegexReplacerImpl.cpp
         ${serd_source_files}
@@ -178,6 +180,7 @@ target_link_libraries(rdf4cpp
         OpenSSL::Crypto
         uni-algo::uni-algo
         highway::highway
+        LibXml2::LibXml2
         )
 
 set_target_properties(rdf4cpp PROPERTIES

conanfile.py

@@ -39,6 +39,7 @@ def requirements(self):
         self.requires("dice-hash/0.4.11", transitive_headers=True)
         self.requires("dice-sparse-map/0.2.9", transitive_headers=True)
         self.requires("dice-template-library/1.19.0", transitive_headers=True)
+        self.requires("libxml2/2.15.0")
 
         if self.options.with_test_deps:
             self.test_requires("doctest/2.4.11")

private/rdf4cpp/parser/IStreamQuadIteratorSerdImpl.hpp

@@ -14,17 +14,25 @@
 
 namespace rdf4cpp::parser {
 
-struct IStreamQuadIterator::Impl {
+struct IStreamQuadIterator::ImplSerd final : Impl {
     using flags_type = IStreamQuadIterator::flags_type;
     using state_type = IStreamQuadIterator::state_type;
     using ok_type = IStreamQuadIterator::ok_type;
     using error_type = IStreamQuadIterator::error_type;
 
 private:
-    SerdReader *reader;
+    // workaround for gcc-14 bug, erroneously warns on unsing a lambda here
+    // see https://github.com/NVIDIA/stdexec/issues/1143
+    struct SerdReaderDtorLambda {
+        void operator()(SerdReader* r) const {
+            serd_reader_end_stream(r);
+            serd_reader_free(r);
+        }
+    };
+    std::unique_ptr<SerdReader, SerdReaderDtorLambda> reader;
 
+    std::unique_ptr<state_type> state_owned = nullptr;
     state_type *state;
-    bool state_is_owned;
 
     std::deque<Quad> quad_buffer;
     std::optional<ParsingError> last_error;
@@ -33,11 +41,9 @@ struct IStreamQuadIterator::Impl {
 
     flags_type flags;
 
-private:
     static std::string_view node_into_string_view(SerdNode const *node) noexcept;
     static ParsingError::Type parsing_error_type_from_serd(SerdStatus st) noexcept;
 
-private:
     nonstd::expected<Node, SerdStatus> get_bnode(std::string &&graph_str, SerdNode const *node) noexcept;
     nonstd::expected<IRI, SerdStatus> get_iri(SerdNode const *node) noexcept;
     nonstd::expected<IRI, SerdStatus> get_prefixed_iri(SerdNode const *node) noexcept;
@@ -63,13 +69,13 @@ struct IStreamQuadIterator::Impl {
     }
 
 public:
-    Impl(void *stream,
+    ImplSerd(void *stream,
          ReadFunc read,
          ErrorFunc,
          flags_type flags,
          state_type *state) noexcept;
 
-    ~Impl() noexcept;
+    ~ImplSerd() override;
 
     /**
      * Tries to extract the next element from the serd backend.
@@ -81,10 +87,10 @@ struct IStreamQuadIterator::Impl {
      *      expected Quad: if there was a next element and it could be parsed
      *      unexpected ParsingError: if there was a next element but it could not be parsed
      */
-    [[nodiscard]] std::optional<nonstd::expected<ok_type, error_type>> next();
+    [[nodiscard]] std::optional<nonstd::expected<ok_type, error_type>> next() override;
 
-    [[nodiscard]] uint64_t current_line() const noexcept;
-    [[nodiscard]] uint64_t current_column() const noexcept;
+    [[nodiscard]] uint64_t current_line() const noexcept override;
+    [[nodiscard]] uint64_t current_column() const noexcept override;
 };
 
 }  // namespace rdf4cpp::parser