override mode

* and_condition_impl::combine_equal - Remove UB container modification.

The container is modified while iterating it.
Switched to indexed based iteration to avoid UB.

* Update condition.cpp

Sorry missed this line.

CMakeLists.txt

@@ -27,11 +27,12 @@ cmake_minimum_required(VERSION 3.23)
 # set the project name
 project(
 	libcifpp
-	VERSION 7.0.4
+	VERSION 8.0.0
 	LANGUAGES CXX)
 
 list(PREPEND CMAKE_MODULE_PATH "${CMAKE_CURRENT_SOURCE_DIR}/cmake")
 
+include(FindAtomic)
 include(CheckFunctionExists)
 include(CheckIncludeFiles)
 include(CheckLibraryExists)
@@ -40,8 +41,12 @@ include(CheckCXXSourceCompiles)
 include(GenerateExportHeader)
 include(CTest)
 include(FetchContent)
+include(ExternalProject)
 
-message(STATUS "DESTDIR is '${DESTDIR}'")
+# FindBoost, take care of it now.
+if(CMAKE_VERSION VERSION_GREATER_EQUAL 3.30)
+	cmake_policy(SET CMP0167 NEW)
+endif()
 
 # When building with ninja-multiconfig, build both debug and release by default
 if(CMAKE_GENERATOR STREQUAL "Ninja Multi-Config")
@@ -188,10 +193,6 @@ if(GXX_LIBSTDCPP)
 	endif()
 endif()
 
-set(CMAKE_THREAD_PREFER_PTHREAD)
-set(THREADS_PREFER_PTHREAD_FLAG)
-find_package(Threads)
-
 if(MSVC)
 	# Avoid linking the shared library of zlib Search ZLIB_ROOT first if it is
 	# set.
@@ -221,6 +222,7 @@ if(MSVC)
 endif()
 
 find_package(ZLIB QUIET)
+find_package(Threads)
 
 if(NOT ZLIB_FOUND)
 	message(FATAL_ERROR "The zlib development files were not found you this system, please install them and try again (hint: on debian/ubuntu use apt-get install zlib1g-dev)")
@@ -235,21 +237,18 @@ if(Eigen3_FOUND AND TARGET Eigen3::Eigen)
 	get_target_property(EIGEN_INCLUDE_DIR Eigen3::Eigen
 		INTERFACE_INCLUDE_DIRECTORIES)
 else()
-	# Create a private copy of eigen3 and populate it only, no need to build
-	FetchContent_Declare(
-		my-eigen3
+	# Use ExternalProject since FetchContent always tries to install the result...
+	ExternalProject_Add(my-eigen3
 		GIT_REPOSITORY https://gitlab.com/libeigen/eigen.git
-		GIT_TAG 3.4.0)
-
-	FetchContent_GetProperties(my-eigen3)
-
-	if(NOT my-eigen3_POPULATED)
-		FetchContent_Populate(my-eigen3)
-	endif()
-
-	set(EIGEN_INCLUDE_DIR ${my-eigen3_SOURCE_DIR})
+		GIT_TAG 3.4.0
+		INSTALL_COMMAND "")
+	
+	ExternalProject_Get_Property(my-eigen3 SOURCE_DIR)
+	set(EIGEN_INCLUDE_DIR ${SOURCE_DIR})
 endif()
 
+message(STATUS "Eigen include dir is ${EIGEN_INCLUDE_DIR}")
+
 # Create a revision file, containing the current git version info
 include(VersionString)
 write_version_header(${CMAKE_CURRENT_SOURCE_DIR}/src/ LIB_NAME "LibCIFPP")
@@ -260,6 +259,8 @@ if(CIFPP_RECREATE_SYMOP_DATA)
 	add_executable(symop-map-generator
 		"${CMAKE_CURRENT_SOURCE_DIR}/src/symop-map-generator.cpp")
 
+	target_compile_features(symop-map-generator PUBLIC cxx_std_20)
+
 	add_custom_command(
 		OUTPUT ${CMAKE_CURRENT_SOURCE_DIR}/src/symop_table_data.hpp
 		COMMAND
@@ -331,9 +332,13 @@ set(project_headers
 	include/cif++/validate.hpp
 )
 
-add_library(cifpp STATIC)
+add_library(cifpp)
 add_library(cifpp::cifpp ALIAS cifpp)
 
+if(TARGET my-eigen3)
+	add_dependencies(cifpp my-eigen3)
+endif()
+
 target_sources(cifpp
 	PRIVATE ${project_sources}
 	${CMAKE_CURRENT_SOURCE_DIR}/src/symop_table_data.hpp
@@ -371,7 +376,7 @@ target_include_directories(
 	"$<INSTALL_INTERFACE:${CMAKE_INSTALL_INCLUDEDIR}>"
 	PRIVATE "${BOOST_REGEX_INCLUDE_DIR}" "${EIGEN_INCLUDE_DIR}")
 
-target_link_libraries(cifpp PUBLIC Threads::Threads ZLIB::ZLIB)
+target_link_libraries(cifpp PUBLIC Threads::Threads ZLIB::ZLIB $<$<TARGET_EXISTS:std::atomic>:std::atomic>)
 
 if(CMAKE_CXX_COMPILER_ID STREQUAL "AppleClang")
 	target_link_options(cifpp PRIVATE -undefined dynamic_lookup)
@@ -442,6 +447,11 @@ endif()
 
 if(CIFPP_DATA_DIR)
 	target_compile_definitions(cifpp PUBLIC DATA_DIR="${CIFPP_DATA_DIR}")
+	set_target_properties(cifpp PROPERTIES CIFPP_DATA_DIR ${CIFPP_DATA_DIR})
+endif()
+
+if(NOT PROJECT_IS_TOP_LEVEL)
+	set(CIFPP_SHARE_DIR ${CIFPP_DATA_DIR} PARENT_SCOPE)
 endif()
 
 if(UNIX AND NOT BUILD_FOR_CCP4)
@@ -464,77 +474,74 @@ else()
 	unset(CIFPP_CACHE_DIR)
 endif()
 
-# Avoid full installation in case we are not the top level target
-if(PROJECT_IS_TOP_LEVEL OR BUILD_FOR_CCP4)
-	# Install rules
-	install(TARGETS cifpp
-		EXPORT cifpp
-		FILE_SET cifpp_headers DESTINATION ${CMAKE_INSTALL_INCLUDEDIR})
-
-	if(MSVC AND BUILD_SHARED_LIBS)
-		install(
-			FILES $<TARGET_PDB_FILE:cifpp>
-			DESTINATION ${CMAKE_INSTALL_LIBDIR}
-			OPTIONAL)
-	endif()
-
-	# Clean up old config files (with old names)
-	file(GLOB OLD_CONFIG_FILES
-		${CMAKE_INSTALL_FULL_LIBDIR}/cmake/cifpp/cifppConfig*.cmake
-		${CMAKE_INSTALL_FULL_LIBDIR}/cmake/cifpp/cifppTargets*.cmake)
-
-	if(OLD_CONFIG_FILES)
-		message(
-			STATUS "Installation will remove old config files: ${OLD_CONFIG_FILES}")
-		install(CODE "file(REMOVE ${OLD_CONFIG_FILES})")
-	endif()
-
-	install(EXPORT cifpp
-		NAMESPACE cifpp::
-		FILE "cifpp-targets.cmake"
-		DESTINATION lib/cmake/cifpp)
-
-	if(CIFPP_DATA_DIR AND CIFPP_DOWNLOAD_CCD)
-		install(
-			FILES ${CMAKE_CURRENT_SOURCE_DIR}/rsrc/mmcif_ddl.dic
-			${CMAKE_CURRENT_SOURCE_DIR}/rsrc/mmcif_pdbx.dic
-			${CMAKE_CURRENT_SOURCE_DIR}/rsrc/mmcif_ma.dic ${COMPONENTS_CIF}
-			DESTINATION ${CMAKE_INSTALL_DATADIR}/libcifpp)
-	endif()
-
-	set(CONFIG_TEMPLATE_FILE ${CMAKE_CURRENT_SOURCE_DIR}/cmake/cifpp-config.cmake.in)
-
-	configure_package_config_file(
-		${CONFIG_TEMPLATE_FILE} ${CMAKE_CURRENT_BINARY_DIR}/cifpp/cifpp-config.cmake
-		INSTALL_DESTINATION lib/cmake/cifpp
-		PATH_VARS CIFPP_DATA_DIR)
+# Install rules
+install(TARGETS cifpp
+	EXPORT cifpp
+	FILE_SET cifpp_headers DESTINATION ${CMAKE_INSTALL_INCLUDEDIR})
 
+if(MSVC AND BUILD_SHARED_LIBS)
 	install(
-		FILES "${CMAKE_CURRENT_BINARY_DIR}/cifpp/cifpp-config.cmake"
-		"${CMAKE_CURRENT_BINARY_DIR}/cifpp/cifpp-config-version.cmake"
-		DESTINATION lib/cmake/cifpp)
-
-	set_target_properties(
-		cifpp
-		PROPERTIES VERSION ${PROJECT_VERSION}
-		SOVERSION "${PROJECT_VERSION_MAJOR}.${PROJECT_VERSION_MINOR}"
-		INTERFACE_cifpp_MAJOR_VERSION ${PROJECT_VERSION_MAJOR})
-
-	set_property(
-		TARGET cifpp
-		APPEND
-		PROPERTY COMPATIBLE_INTERFACE_STRING cifpp_MAJOR_VERSION)
-
-	write_basic_package_version_file(
-		"${CMAKE_CURRENT_BINARY_DIR}/cifpp/cifpp-config-version.cmake"
-		VERSION ${PROJECT_VERSION}
-		COMPATIBILITY AnyNewerVersion)
-else()
-	# Set this variable so that consumers can find the files in rsrc
-	set(CIFPP_SHARE_DIR ${CMAKE_CURRENT_SOURCE_DIR}/rsrc PARENT_SCOPE)
+		FILES $<TARGET_PDB_FILE:cifpp>
+		DESTINATION ${CMAKE_INSTALL_LIBDIR}
+		OPTIONAL)
 endif()
 
-if(BUILD_TESTING)
+# Clean up old config files (with old names)
+file(GLOB OLD_CONFIG_FILES
+	${CMAKE_INSTALL_FULL_LIBDIR}/cmake/cifpp/cifppConfig*.cmake
+	${CMAKE_INSTALL_FULL_LIBDIR}/cmake/cifpp/cifppTargets*.cmake)
+
+if(OLD_CONFIG_FILES)
+	message(
+		STATUS "Installation will remove old config files: ${OLD_CONFIG_FILES}")
+	install(CODE "file(REMOVE ${OLD_CONFIG_FILES})")
+endif()
+
+install(EXPORT cifpp
+	NAMESPACE cifpp::
+	FILE "cifpp-targets.cmake"
+	DESTINATION lib/cmake/cifpp)
+
+install(
+	FILES ${CMAKE_CURRENT_SOURCE_DIR}/rsrc/mmcif_ddl.dic
+	${CMAKE_CURRENT_SOURCE_DIR}/rsrc/mmcif_pdbx.dic
+	${CMAKE_CURRENT_SOURCE_DIR}/rsrc/mmcif_ma.dic
+	DESTINATION ${CMAKE_INSTALL_DATADIR}/libcifpp)
+
+if(CIFPP_DATA_DIR AND CIFPP_DOWNLOAD_CCD)
+	install(FILES ${COMPONENTS_CIF}
+		DESTINATION ${CMAKE_INSTALL_DATADIR}/libcifpp)
+endif()
+
+set(CONFIG_TEMPLATE_FILE ${CMAKE_CURRENT_SOURCE_DIR}/cmake/cifpp-config.cmake.in)
+
+configure_package_config_file(
+	${CONFIG_TEMPLATE_FILE} ${CMAKE_CURRENT_BINARY_DIR}/cifpp/cifpp-config.cmake
+	INSTALL_DESTINATION lib/cmake/cifpp
+	PATH_VARS CIFPP_DATA_DIR)
+
+install(
+	FILES "${CMAKE_CURRENT_BINARY_DIR}/cifpp/cifpp-config.cmake"
+	"${CMAKE_CURRENT_BINARY_DIR}/cifpp/cifpp-config-version.cmake"
+	DESTINATION lib/cmake/cifpp)
+
+set_target_properties(
+	cifpp
+	PROPERTIES VERSION ${PROJECT_VERSION}
+	SOVERSION "${PROJECT_VERSION_MAJOR}.${PROJECT_VERSION_MINOR}"
+	INTERFACE_cifpp_MAJOR_VERSION ${PROJECT_VERSION_MAJOR})
+
+set_property(
+	TARGET cifpp
+	APPEND
+	PROPERTY COMPATIBLE_INTERFACE_STRING cifpp_MAJOR_VERSION)
+
+write_basic_package_version_file(
+	"${CMAKE_CURRENT_BINARY_DIR}/cifpp/cifpp-config-version.cmake"
+	VERSION ${PROJECT_VERSION}
+	COMPATIBILITY AnyNewerVersion)
+
+if(BUILD_TESTING AND PROJECT_IS_TOP_LEVEL)
 	add_subdirectory(test)
 endif()
 

README.md

@@ -52,7 +52,7 @@ int main(int argc, char *argv[])
 
     if (file.empty())
     {
-        std::cerr << "Empty file" << std::endl;
+        std::cerr << "Empty file\n";
         exit(1);
     }
 
@@ -66,8 +66,8 @@ int main(int argc, char *argv[])
     auto n = atom_site.count(cif::key("label_atom_id") == "OXT");
 
     std::cout << "File contains " << atom_site.size() << " atoms of which "
-              << n << (n == 1 ? " is" : " are") << " OXT" << std::endl
-              << "residues with an OXT are:" << std::endl;
+              << n << (n == 1 ? " is" : " are") << " OXT\n"
+              << "residues with an OXT are:\n";
 
     // Loop over all atoms with atom-id "OXT" and print out some info.
     // That info is extracted using structured binding in C++
@@ -76,7 +76,7 @@ int main(int argc, char *argv[])
                 cif::key("label_atom_id") == "OXT",
                 "label_asym_id", "label_comp_id", "label_seq_id"))
     {
-        std::cout << asym << ' ' << comp << ' ' << seqnr << std::endl;
+        std::cout << asym << ' ' << comp << ' ' << seqnr << '\n';
     }
 
     return 0;

changelog

@@ -1,3 +1,35 @@
+Version 8.0.0
+- A dictionary is for a datablock and a file can have
+  datablocks with differing dictionaries.
+
+Version 7.0.10
+- Deal with missing _entity.type in reconstructing mmCIF files
+- Replace code creating quaternions from rotation matrices
+  that might sometimes give incorrect results. Or at least,
+  the test code failed on this particular kind of code. Sometimes.
+
+Version 7.0.9
+- Using cif::file::load_dictionary it is now possible to
+  load a dictionary along with its extensions in one go.
+  E.g. file.load_dictionary("mmcif_pdbx;dssp-extension")
+- Fix in compound factory to avoid errors with lower case
+  compound id's
+- Fix sac_parser's index to be case insensitive
+
+Version 7.0.8
+- Fix PDB Remark 3 parser
+- Added three way comparison for point
+
+Version 7.0.7
+- Set CIFPP_DATA_DIR on target cifpp for use in projects that include
+  libcifpp directly
+
+Version 7.0.6
+- Fix linking to std::atomic
+
+Version 7.0.5
+- Fix case where category index was not updated for updated value
+
 Version 7.0.4
 - Do not install headers and library in case we're not the top project
 

cmake/FindAtomic.cmake

@@ -0,0 +1,63 @@
+# Simple check to see if we need a library for std::atomic
+
+if(TARGET std::atomic)
+	return()
+endif()
+
+cmake_minimum_required(VERSION 3.10)
+
+include(CMakePushCheckState)
+include(CheckIncludeFileCXX)
+include(CheckCXXSourceRuns)
+
+cmake_push_check_state()
+
+check_include_file_cxx("atomic" _CXX_ATOMIC_HAVE_HEADER)
+mark_as_advanced(_CXX_ATOMIC_HAVE_HEADER)
+
+set(code [[
+#include <atomic>
+int main(int argc, char** argv) {
+  std::atomic<long long> s;
+  ++s;
+  return 0;
+}
+]])
+
+check_cxx_source_runs("${code}" _CXX_ATOMIC_BUILTIN)
+
+if(_CXX_ATOMIC_BUILTIN)
+	set(_found 1)
+else()
+  list(APPEND CMAKE_REQUIRED_LIBRARIES atomic)
+  list(APPEND FOLLY_LINK_LIBRARIES atomic)
+
+  check_cxx_source_runs("${code}" _CXX_ATOMIC_LIB_NEEDED)
+  if (NOT _CXX_ATOMIC_LIB_NEEDED)
+    message(FATAL_ERROR "unable to link C++ std::atomic code: you may need \
+      to install GNU libatomic")
+  else()
+	set(_found 1)
+  endif()
+endif()
+
+if(_found)
+	add_library(std::atomic INTERFACE IMPORTED)
+	set_property(TARGET std::atomic APPEND PROPERTY INTERFACE_COMPILE_FEATURES cxx_std_14)
+
+	if(_CXX_ATOMIC_BUILTIN)
+		# Nothing to add...
+	elseif(_CXX_ATOMIC_LIB_NEEDED)
+		set_target_properties(std::atomic PROPERTIES IMPORTED_LIBNAME atomic)
+		set(STDCPPATOMIC_LIBRARY atomic)
+	endif()
+endif()
+
+cmake_pop_check_state()
+
+set(Atomic_FOUND ${_found} CACHE BOOL "TRUE if we can run a program using std::atomic" FORCE)
+mark_as_advanced(Atomic_FOUND)
+
+if(Atomic_FIND_REQUIRED AND NOT Atomic_FOUND)
+    message(FATAL_ERROR "Cannot run simple program using std::atomic")
+endif()

examples/example.cpp

@@ -13,7 +13,7 @@ int main(int argc, char *argv[])
 		exit(1);
 	}
 
-	cif::file file = cif::pdb::read(argv[1]);
+	cif::file file(argv[1]);
 
 	if (file.empty())
 	{

include/cif++/category.hpp

@@ -178,7 +178,7 @@ class category
 	/// @brief Set the validator for this category to @a v
 	/// @param v The category_validator to assign. A nullptr value is allowed.
 	/// @param db The enclosing @ref datablock
-	void set_validator(const validator *v, datablock &db);
+	void set_validator(const validator_base *v, datablock &db);
 
 	/// @brief Update the links in this category
 	/// @param db The enclosing @ref datablock
@@ -186,7 +186,7 @@ class category
 
 	/// @brief Return the global @ref validator for the data
 	/// @return The @ref validator or nullptr if not assigned
-	const validator *get_validator() const { return m_validator; }
+	const validator_base *get_validator() const { return m_validator; }
 
 	/// @brief Return the category validator for this category
 	/// @return The @ref category_validator or nullptr if not assigned
@@ -1285,7 +1285,7 @@ class category
 
 	std::string m_name;
 	std::vector<item_entry> m_items;
-	const validator *m_validator = nullptr;
+	const validator_base *m_validator = nullptr;
 	const category_validator *m_cat_validator = nullptr;
 	std::vector<link> m_parent_links, m_child_links;
 	bool m_cascade = true;

include/cif++/compound.hpp

@@ -138,7 +138,7 @@ struct compound_bond
 /// This information is derived from the CDD by default.
 ///
 /// To create compounds, you use the factory method. You can add your own
-/// compound definitions by calling the addExtraComponents function and
+/// compound definitions by calling the push_dictionary function and
 /// pass it a valid CCD formatted file.
 
 class compound
@@ -180,8 +180,7 @@ class compound
 	friend class local_compound_factory_impl;
 
 	compound(cif::datablock &db);
-	compound(cif::datablock &db, int);
-
+	
 	std::string m_id;
 	std::string m_name;
 	std::string m_type;

include/cif++/condition.hpp

@@ -1284,6 +1284,19 @@ condition operator==(const key &key, const std::optional<T> &v)
 		return condition(new detail::key_is_empty_condition_impl(key.m_item_name));
 }
 
+/**
+ * @brief Create a condition to search any item for a value @a v if @a v contains a value
+ * compare to null if not.
+ */
+template <typename T>
+condition operator!=(const key &key, const std::optional<T> &v)
+{
+	if (v.has_value())
+		return condition(new detail::not_condition_impl(condition(new detail::key_equals_condition_impl({ key.m_item_name, *v }))));
+	else
+		return condition(new detail::not_condition_impl(condition(new detail::key_is_empty_condition_impl(key.m_item_name))));
+}
+
 /**
  * @brief Operator to create a boolean opposite of the condition in @a rhs
  */

include/cif++/datablock.hpp

@@ -98,19 +98,32 @@ class datablock : public std::list<category>
 		m_name = name;
 	}
 
+	/**
+	 * @brief Attempt to load the dictionary specified in audit_conform category
+	 * 
+	 */
+	void load_dictionary();
+
+	/**
+	 * @brief Load the dictionary named @a dict_name
+	 * 
+	 * @param dict_name 
+	 */
+	void load_dictionary(std::string_view dict_name);
+
 	/**
 	 * @brief Set the validator object to @a v
 	 * 
 	 * @param v The new validator object, may be null
 	 */
-	void set_validator(const validator *v);
+	void set_validator(const validator_base *v);
 
 	/**
 	 * @brief Get the validator object
 	 * 
 	 * @return const validator* The validator or nullptr if there is none
 	 */
-	const validator *get_validator() const;
+	const validator_base *get_validator() const;
 
 	/**
 	 * @brief Validates the content of this datablock and all its content
@@ -231,7 +244,7 @@ class datablock : public std::list<category>
 
   private:
 	std::string m_name;
-	const validator *m_validator = nullptr;
+	const validator_base *m_validator = nullptr;
 };
 
 } // namespace cif

include/cif++/dictionary_parser.hpp

@@ -42,9 +42,4 @@ namespace cif
  */
 validator parse_dictionary(std::string_view name, std::istream &is);
 
-/**
- * @brief Extend the definitions in validator @a v with the contents of stream @a is
- */
-void extend_dictionary(validator &v, std::istream &is);
-
 } // namespace cif

include/cif++/file.hpp

@@ -118,19 +118,6 @@ class file : public std::list<datablock>
 
 	/** @endcond */
 
-	/**
-	 * @brief Set the validator object to @a v
-	 */
-	void set_validator(const validator *v);
-
-	/**
-	 * @brief Get the validator object
-	 */
-	const validator *get_validator() const
-	{
-		return m_validator;
-	}
-
 	/**
 	 * @brief Validate the content and return true if everything was valid.
 	 * 
@@ -165,21 +152,6 @@ class file : public std::list<datablock>
 	 */
 	bool validate_links() const;
 
-	/**
-	 * @brief Attempt to load a dictionary (validator) based on
-	 * the contents of the *audit_conform* category, if available.
-	 */
-	void load_dictionary();
-
-
-	/**
-	 * @brief Attempt to load the named dictionary @a name and
-	 * create a validator based on it.
-	 * 
-	 * @param name The name of the dictionary to load
-	 */
-	void load_dictionary(std::string_view name);
-
 	/**
 	 * @brief Return true if a datablock with the name @a name is part of this file
 	 */
@@ -232,6 +204,18 @@ class file : public std::list<datablock>
 	/** Load the data from @a is */
 	void load(std::istream &is);
 
+	/** Load the data from the file specified by @a p using validator @a v */
+	void load(const std::filesystem::path &p, const validator_base &v);
+
+	/** Load the data from @a is using validator @a v */
+	void load(std::istream &is, const validator_base &v);
+
+	/** Load the data from the file specified by @a p using a validator constructed from dictionary @a dict */
+	void load(const std::filesystem::path &p, std::string_view dict);
+
+	/** Load the data from @a is using a validator constructed from dictionary @a dict */
+	void load(std::istream &is, std::string_view dict);
+
 	/** Save the data to the file specified by @a p */
 	void save(const std::filesystem::path &p) const;
 
@@ -246,9 +230,6 @@ class file : public std::list<datablock>
 		f.save(os);
 		return os;
 	}
-
-  private:
-	const validator *m_validator = nullptr;
 };
 
 } // namespace cif

include/cif++/item.hpp

@@ -378,7 +378,7 @@ struct item_handle
 	template <typename T>
 	item_handle &operator=(T &&value)
 	{
-		assign_value(item{ "", std::move(value) }.value());
+		assign_value(item{ "", std::forward<T>(value) }.value());
 		return *this;
 	}
 
@@ -608,6 +608,8 @@ struct item_handle::item_value_as<T, std::enable_if_t<std::is_arithmetic_v<T> an
 				}
 				result = 1;
 			}
+			else if (std::abs(v - value) <= std::numeric_limits<value_type>::epsilon())
+				result = 0;
 			else if (v < value)
 				result = -1;
 			else if (v > value)

include/cif++/model.hpp

@@ -350,7 +350,12 @@ class atom
 	std::string get_pdb_ins_code() const { return get_property("pdbx_PDB_ins_code"); } ///< Return the pdb_ins_code property
 
 	/// Return true if this atom is an alternate
-	bool is_alternate() const { return not get_label_alt_id().empty(); }
+	bool is_alternate() const
+	{
+		if (auto alt_id = get_label_alt_id(); alt_id.empty() or alt_id == ".")
+			return false;
+		return true;
+	}
 
 	/// Convenience method to return a string that might be ID in PDB space
 	std::string pdb_id() const
@@ -550,6 +555,9 @@ class residue
 	/// \brief Return true if this residue has alternate atoms
 	bool has_alternate_atoms() const;
 
+	/// \brief Return true if this residue has alternate atoms for the atom \a atomID
+	bool has_alternate_atoms_for(const std::string &atomID) const;
+
 	/// \brief Return the list of unique alt ID's present in this residue
 	std::set<std::string> get_alternate_ids() const;
 
@@ -572,6 +580,10 @@ class residue
 								   m_auth_seq_id == rhs.m_auth_seq_id);
 	}
 
+	/// @brief Create a new atom and add it to the list
+	/// @return newly created atom
+	virtual atom create_new_atom(atom_type inType, const std::string &inAtomID, point inLocation);
+
   protected:
 	/** @cond */
 	residue() {}
@@ -672,6 +684,8 @@ class monomer : public residue
 		return m_polymer == rhs.m_polymer and m_index == rhs.m_index;
 	}
 
+	atom create_new_atom(atom_type inType, const std::string &inAtomID, point inLocation) override;
+
   private:
 	const polymer *m_polymer;
 	std::size_t m_index;
@@ -1091,6 +1105,9 @@ class structure
 	/// \brief emplace the moved atom @a atom
 	atom &emplace_atom(atom &&atom);
 
+	/// \brief Reorder atom_site atoms based on 'natural' ordering
+	void reorder_atoms();
+
   private:
 	friend polymer;
 	friend residue;

include/cif++/parser.hpp

@@ -39,6 +39,8 @@
 namespace cif
 {
 
+class validator_base;
+
 // --------------------------------------------------------------------
 
 /** Exception that is thrown when the mmCIF file contains a parsing error */
@@ -73,7 +75,15 @@ class sac_parser
 {
   public:
 	/** @cond */
-	using datablock_index = std::map<std::string, std::size_t>;
+	struct iless_op
+	{
+		bool operator()(std::string_view a, std::string_view b) const
+		{
+			return icompare(a, b) < 0;
+		}
+	};
+
+	using datablock_index = std::map<std::string, std::size_t, iless_op>;
 
 	virtual ~sac_parser() = default;
 	/** @endcond */
@@ -299,6 +309,14 @@ class sac_parser
 class parser : public sac_parser
 {
   public:
+	/// \brief constructor, generates data into @a file from @a is using validator @a v
+	parser(std::istream &is, file &file, const validator_base *v)
+		: sac_parser(is)
+		, m_file(file)
+		, m_validator(v)
+	{
+	}
+
 	/// \brief constructor, generates data into @a file from @a is
 	parser(std::istream &is, file &file)
 		: sac_parser(is)
@@ -319,6 +337,7 @@ class parser : public sac_parser
 	file &m_file;
 	datablock *m_datablock = nullptr;
 	category *m_category = nullptr;
+	const validator_base *m_validator = nullptr;
 	row_handle m_row;
 
 	/** @endcond */

include/cif++/point.hpp

@@ -662,12 +662,23 @@ struct point_type
 		return std::make_tuple(std::ref(m_x), std::ref(m_y), std::ref(m_z));
 	}
 
-	/// \brief Compare with @a rhs
+#if defined(__cpp_impl_three_way_comparison)
+	/// \brief a default spaceship operator
+	constexpr auto operator<=>(const point_type &rhs) const = default;
+#else
+	/// \brief a default equals operator
 	constexpr bool operator==(const point_type &rhs) const
 	{
 		return m_x == rhs.m_x and m_y == rhs.m_y and m_z == rhs.m_z;
 	}
 
+	/// \brief a default not-equals operator
+	constexpr bool operator!=(const point_type &rhs) const
+	{
+		return not operator==(rhs);
+	}
+#endif
+
 	// consider point as a vector... perhaps I should rename point?
 
 	/// \brief looking at the point as if it is a vector, return the squared length

include/cif++/text.hpp

@@ -378,7 +378,7 @@ std::from_chars_result from_chars(const char *first, const char *last, FloatType
 	} state = IntegerSign;
 	int sign = 1;
 	unsigned long long vi = 0;
-	long double f = 1;
+	int fl = 0, tz = 0;
 	int exponent_sign = 1;
 	int exponent = 0;
 	bool done = false;
@@ -427,7 +427,14 @@ std::from_chars_result from_chars(const char *first, const char *last, FloatType
 				if (ch >= '0' and ch <= '9')
 				{
 					vi = 10 * vi + (ch - '0');
-					f /= 10;
+
+					if (ch == '0')
+						tz += 1;
+					else
+					{
+						fl += tz + 1;
+						tz = 0;
+					}
 				}
 				else if (ch == 'e' or ch == 'E')
 					state = ExponentSign;
@@ -469,7 +476,10 @@ std::from_chars_result from_chars(const char *first, const char *last, FloatType
 
 	if (not (bool)result.ec)
 	{
-		long double v = f * vi * sign;
+		while (tz-- > 0)
+			vi /= 10;
+
+		long double v = std::pow(10, -fl) * vi * sign;
 		if (exponent != 0)
 			v *= std::pow(10, exponent * exponent_sign);
 

include/cif++/validate.hpp

@@ -32,6 +32,7 @@
 #include <filesystem>
 #include <list>
 #include <mutex>
+#include <optional>
 #include <system_error>
 #include <utility>
 
@@ -48,6 +49,7 @@
 namespace cif
 {
 
+class category;
 struct category_validator;
 
 // --------------------------------------------------------------------
@@ -384,57 +386,32 @@ struct link_validator
 
 // --------------------------------------------------------------------
 
-/**
- * @brief The validator class combines all the link, category and item validator classes
- *
- */
-class validator
+class validator_base
 {
   public:
-	/**
-	 * @brief Construct a new validator object
-	 *
-	 * @param name The name of the underlying dictionary
-	 */
-	validator(std::string_view name)
-		: m_name(name)
-	{
-	}
-
 	/// @brief destructor
-	~validator() = default;
+	virtual ~validator_base() = default;
 
-	validator(const validator &rhs) = delete;
-	validator &operator=(const validator &rhs) = delete;
+	validator_base(const validator_base &rhs) = delete;
+	validator_base &operator=(const validator_base &rhs) = delete;
 
 	/// @brief move constructor
-	validator(validator &&rhs) = default;
+	validator_base(validator_base &&rhs) = default;
 
 	/// @brief move assignment operator
-	validator &operator=(validator &&rhs) = default;
-
-	friend class dictionary_parser;
-
-	/// @brief Add type_validator @a v to the list of type validators
-	void add_type_validator(type_validator &&v);
+	validator_base &operator=(validator_base &&rhs) = default;
 
 	/// @brief Return the type validator for @a type_code, may return nullptr
-	const type_validator *get_validator_for_type(std::string_view type_code) const;
-
-	/// @brief Add category_validator @a v to the list of category validators
-	void add_category_validator(category_validator &&v);
+	virtual const type_validator *get_validator_for_type(std::string_view type_code) const = 0;
 
 	/// @brief Return the category validator for @a category, may return nullptr
-	const category_validator *get_validator_for_category(std::string_view category) const;
-
-	/// @brief Add link_validator @a v to the list of link validators
-	void add_link_validator(link_validator &&v);
+	virtual const category_validator *get_validator_for_category(std::string_view category) const = 0;
 
 	/// @brief Return the list of link validators for which the parent is @a category
-	std::vector<const link_validator *> get_links_for_parent(std::string_view category) const;
+	virtual std::vector<const link_validator *> get_links_for_parent(std::string_view category) const = 0;
 
 	/// @brief Return the list of link validators for which the child is @a category
-	std::vector<const link_validator *> get_links_for_child(std::string_view category) const;
+	virtual std::vector<const link_validator *> get_links_for_child(std::string_view category) const = 0;
 
 	/// @brief Bottleneck function to report an error in validation
 	void report_error(validation_error err, bool fatal = true) const
@@ -456,19 +433,87 @@ class validator
 	void report_error(std::error_code ec, std::string_view category,
 		std::string_view item, bool fatal = true) const;
 
-	const std::string &name() const { return m_name; }        ///< Get the name of this validator
-	void set_name(const std::string &name) { m_name = name; } ///< Set the name of this validator
+	const std::string &name() const { return m_name; }       ///< Get the name of this validator
+	const std::string &version() const { return m_version; } ///< Get the version of this validator
+	bool is_strict() const { return m_strict; }              ///< Get the strict flag of this validator
 
-	const std::string &version() const { return m_version; }              ///< Get the version of this validator
+  protected:
+	/**
+	 * @brief Construct a new validator object
+	 *
+	 * @param name The name of the underlying dictionary
+	 */
+	validator_base(std::string_view name)
+		: m_name(name)
+	{
+	}
+
+	validator_base() = default;
+
+	std::string m_name;
+	std::string m_version;
+	bool m_strict = false;
+};
+
+/**
+ * @brief The validator class combines all the link, category and item validator classes
+ *
+ */
+class validator : public validator_base
+{
+  public:
+	/**
+	 * @brief Construct a new validator object
+	 *
+	 * @param name The name of the underlying dictionary
+	 */
+	validator(std::string_view name)
+		: validator_base(name)
+	{
+	}
+
+	/// @brief destructor
+	~validator() = default;
+
+	validator(const validator &rhs) = delete;
+	validator &operator=(const validator &rhs) = delete;
+
+	/// @brief move constructor
+	validator(validator &&rhs) = default;
+
+	/// @brief move assignment operator
+	validator &operator=(validator &&rhs) = default;
+
+	friend class dictionary_parser;
+
+	/// @brief Add type_validator @a v to the list of type validators
+	void add_type_validator(type_validator &&v);
+
+	/// @brief Return the type validator for @a type_code, may return nullptr
+	const type_validator *get_validator_for_type(std::string_view type_code) const override;
+
+	/// @brief Add category_validator @a v to the list of category validators
+	void add_category_validator(category_validator &&v);
+
+	/// @brief Return the category validator for @a category, may return nullptr
+	const category_validator *get_validator_for_category(std::string_view category) const override;
+
+	/// @brief Add link_validator @a v to the list of link validators
+	void add_link_validator(link_validator &&v);
+
+	/// @brief Return the list of link validators for which the parent is @a category
+	std::vector<const link_validator *> get_links_for_parent(std::string_view category) const override;
+
+	/// @brief Return the list of link validators for which the child is @a category
+	std::vector<const link_validator *> get_links_for_child(std::string_view category) const override;
+
+	void set_name(const std::string &name) { m_name = name; }             ///< Set the name of this validator
 	void set_version(const std::string &version) { m_version = version; } ///< Set the version of this validator
 
   private:
 	// name is fully qualified here:
 	item_validator *get_validator_for_item(std::string_view name) const;
 
-	std::string m_name;
-	std::string m_version;
-	bool m_strict = false;
 	std::set<type_validator> m_type_validators;
 	std::set<category_validator> m_category_validators;
 	std::vector<link_validator> m_link_validators;
@@ -476,6 +521,46 @@ class validator
 
 // --------------------------------------------------------------------
 
+class extended_validator : public validator_base
+{
+  public:
+	/**
+	 * @brief Construct a new validator object
+	 *
+	 * @param name The name of the underlying dictionary
+	 * @param validators The validators this extended validator is composed off
+	 */
+	extended_validator(std::vector<const validator *> validators);
+
+	extended_validator(const extended_validator &rhs) = delete;
+	extended_validator &operator=(const extended_validator &rhs) = delete;
+
+	/// @brief move constructor
+	extended_validator(extended_validator &&rhs) = default;
+
+	/// @brief move assignment operator
+	extended_validator &operator=(extended_validator &&rhs) = default;
+
+	/// @brief Return the type validator for @a type_code, may return nullptr
+	virtual const type_validator *get_validator_for_type(std::string_view type_code) const override;
+
+	/// @brief Return the category validator for @a category, may return nullptr
+	virtual const category_validator *get_validator_for_category(std::string_view category) const override;
+
+	/// @brief Return the list of link validators for which the parent is @a category
+	virtual std::vector<const link_validator *> get_links_for_parent(std::string_view category) const override;
+
+	/// @brief Return the list of link validators for which the child is @a category
+	virtual std::vector<const link_validator *> get_links_for_child(std::string_view category) const override;
+
+  protected:
+	friend class validator_factory;
+
+	std::vector<const validator *> m_validators;
+};
+
+// --------------------------------------------------------------------
+
 /**
  * @brief Validators are globally unique objects, use the validator_factory
  * class to construct them. This class is a singleton.
@@ -488,18 +573,36 @@ class validator_factory
 	static validator_factory &instance();
 
 	/// @brief Return the validator with name @a dictionary_name
-	const validator &operator[](std::string_view dictionary_name);
+	[[deprecated("use construct_validator(const category &audit_conform) instead")]]
+	const validator_base &operator[](std::string_view dictionary_name);
+
+	/// @brief Return a validator for the data contained in an audit_conform category
+	const validator_base &construct_validator(const category &audit_conform);
+
+	/// @brief Construct a new validator with name @a name from resource data with at least version @a version if specified
+	const validator &construct_validator(std::string_view name,
+		std::optional<std::string> version);
+
+	/// @brief Construct a new validator with name @a name from the data in @a is with at least version @a version if specified
+	const validator &construct_validator(std::string_view name,
+		std::optional<std::string> version, std::istream &is);
 
 	/// @brief Construct a new validator with name @a name from the data in @a is
-	const validator &construct_validator(std::string_view name, std::istream &is);
+	const validator &construct_validator(std::string_view name, std::istream &is)
+	{
+		return construct_validator(name, {}, is);
+	}
 
   private:
 	// --------------------------------------------------------------------
 
 	validator_factory() = default;
 
+	static bool check_version(std::string_view name, std::string_view expected, std::string_view found);
+
 	std::mutex m_mutex;
 	std::list<validator> m_validators;
+	std::list<extended_validator> m_extended_validators;
 };
 
 } // namespace cif

src/category.cpp

@@ -617,7 +617,7 @@ std::set<uint16_t> category::key_item_indices() const
 
 // --------------------------------------------------------------------
 
-void category::set_validator(const validator *v, datablock &db)
+void category::set_validator(const validator_base *v, datablock &db)
 {
 	m_validator = v;
 
@@ -696,7 +696,7 @@ bool category::is_valid() const
 	bool result = true;
 
 	if (m_validator == nullptr)
-		throw std::runtime_error("no Validator specified");
+		throw std::runtime_error("no validator specified");
 
 	if (empty())
 	{
@@ -718,7 +718,7 @@ bool category::is_valid() const
 		auto iv = m_cat_validator->get_validator_for_item(col.m_name);
 		if (iv == nullptr)
 		{
-			m_validator->report_error(validation_error::unknown_item, col.m_name, m_name, false);
+			m_validator->report_error(validation_error::unknown_item, m_name, col.m_name, false);
 			result = false;
 		}
 
@@ -1321,7 +1321,7 @@ void category::update_value(const std::vector<row_handle> &rows, std::string_vie
 		std::string oldValue{ parent[item_name].text() };
 		std::string value{ value_provider(oldValue) };
 
-		parent.assign(colIx, value, false);
+		update_value(parent.get_row(), colIx, value, false, false);
 
 		for (auto &&[childCat, linked] : m_child_links)
 		{
@@ -1444,8 +1444,7 @@ void category::update_value(row *row, uint16_t item, std::string_view value, boo
 	// before updating
 
 	bool reinsert = false;
-	if (updateLinked and // an update of an Item's value
-		m_index != nullptr and key_item_indices().count(item))
+	if (m_index != nullptr and key_item_indices().count(item))
 	{
 		reinsert = m_index->find(*this, row);
 		if (reinsert)
@@ -1698,6 +1697,12 @@ void category::swap_item(uint16_t item_ix, row_handle &a, row_handle &b)
 	auto &ra = *a.m_row;
 	auto &rb = *b.m_row;
 
+	while (ra.size() <= item_ix)
+		ra.emplace_back("");
+
+	while (rb.size() <= item_ix)
+		rb.emplace_back("");
+
 	std::swap(ra.at(item_ix), rb.at(item_ix));
 }
 

src/compound.cpp

@@ -177,89 +177,6 @@ compound::compound(cif::datablock &db)
 	}
 }
 
-compound::compound(cif::datablock &db, int)
-{
-	auto &chemComp = db["chem_comp"];
-
-	if (chemComp.size() != 1)
-		throw std::runtime_error("Invalid compound file, chem_comp should contain a single row");
-
-	cif::tie(m_id, m_name) =
-		chemComp.front().get("id", "name");
-	
-	cif::trim(m_name);
-
-	m_type = "NON-POLYMER";
-
-	auto &chemCompAtom = db["chem_comp_atom"];
-	for (auto row : chemCompAtom)
-	{
-		compound_atom atom;
-		std::string type_symbol;
-		cif::tie(atom.id, type_symbol, atom.charge, atom.x, atom.y, atom.z) =
-			row.get("atom_id", "type_symbol", "charge", "x", "y", "z");
-		atom.type_symbol = atom_type_traits(type_symbol).type();
-
-		m_formal_charge += atom.charge;
-
-		m_atoms.push_back(std::move(atom));
-	}
-
-	auto &chemCompBond = db["chem_comp_bond"];
-	for (auto row : chemCompBond)
-	{
-		compound_bond bond;
-		std::string btype;
-		cif::tie(bond.atom_id[0], bond.atom_id[1], btype, bond.aromatic) = row.get("atom_id_1", "atom_id_2", "type", "aromatic");
-
-		using cif::iequals;
-
-		if (iequals(btype, "single"))
-			bond.type = bond_type::sing;
-		else if (iequals(btype, "double"))
-			bond.type = bond_type::doub;
-		else if (iequals(btype, "triple"))
-			bond.type = bond_type::trip;
-		else if (iequals(btype, "deloc") or iequals(btype, "aromat") or iequals(btype, "aromatic"))
-			bond.type = bond_type::delo;
-		else
-		{
-			if (cif::VERBOSE > 0)
-				std::cerr << "Unimplemented chem_comp_bond.type " << btype << " in " << db.name() << '\n';
-			bond.type = bond_type::sing;
-		}
-		m_bonds.push_back(std::move(bond));
-	}
-
-	// reconstruct a formula and weight
-
-	m_formula_weight = 0;
-
-	std::map<atom_type, int> f;
-	for (auto &atom : m_atoms)
-		f[atom.type_symbol] += 1;
-	
-	if (f.count(atom_type::C))
-	{
-		atom_type_traits att(atom_type::C);
-		m_formula += att.symbol() + std::to_string(f[atom_type::C]) + ' ';
-		m_formula_weight += att.weight() * f[atom_type::C];
-	}
-
-	for (const auto &[type, count] : f)
-	{
-		if (type == atom_type::C)
-			continue;
-
-		atom_type_traits att(type);
-		m_formula += att.symbol() + std::to_string(count) + ' ';
-		m_formula_weight += att.weight() * count;
-	}
-	
-	if (not m_formula.empty())
-		m_formula.pop_back();
-}
-
 compound_atom compound::get_atom_by_atom_id(const std::string &atom_id) const
 {
 	compound_atom result = {};
@@ -379,41 +296,17 @@ class compound_factory_impl : public std::enable_shared_from_this<compound_facto
 
 	compound *get(std::string id)
 	{
-		cif::to_upper(id);
-
 		std::shared_lock lock(mMutex);
 
 		compound *result = nullptr;
-
-		// walk the list, see if any of the implementations has the compound already
+		
 		for (auto impl = shared_from_this(); impl; impl = impl->m_next)
 		{
-			for (auto cmp : impl->m_compounds)
-			{
-				if (iequals(cmp->id(), id))
-				{
-					result = cmp;
-					break;
-				}
-			}
-
-			if (result)
+			result = impl->create(id);
+			if (result != nullptr)
 				break;
 		}
 
-		if (result == nullptr and m_missing.count(id) == 0)
-		{
-			for (auto impl = shared_from_this(); impl; impl = impl->m_next)
-			{
-				result = impl->create(id);
-				if (result != nullptr)
-					break;
-			}
-
-			if (result == nullptr)
-				m_missing.insert(id);
-		}
-
 		return result;
 	}
 
@@ -444,7 +337,7 @@ class compound_factory_impl : public std::enable_shared_from_this<compound_facto
 	cif::parser::datablock_index m_index;
 
 	std::vector<compound *> m_compounds;
-	std::set<std::string> m_missing;
+	cif::iset m_missing;
 	std::shared_ptr<compound_factory_impl> m_next;
 };
 
@@ -465,6 +358,14 @@ compound_factory_impl::compound_factory_impl(const fs::path &file, std::shared_p
 
 compound *compound_factory_impl::create(const std::string &id)
 {
+	// shortcut
+
+	if (m_missing.contains(id))
+		return nullptr;
+
+	if (auto i = find_if(m_compounds.begin(), m_compounds.end(), [id](compound *c) { return c->id() == id; }); i != m_compounds.end())
+		return *i;
+
 	compound *result = nullptr;
 
 	std::unique_ptr<std::istream> ccd;
@@ -496,7 +397,7 @@ compound *compound_factory_impl::create(const std::string &id)
 		m_index = parser.index_datablocks();
 
 		if (cif::VERBOSE > 1)
-			std::cout << " done" << std::endl;
+			std::cout << " done\n";
 
 		// reload the resource, perhaps this should be improved...
 		if (m_file.empty())
@@ -519,7 +420,7 @@ compound *compound_factory_impl::create(const std::string &id)
 	parser.parse_single_datablock(id, m_index);
 
 	if (cif::VERBOSE > 1)
-		std::cout << " done" << std::endl;
+		std::cout << " done\n";
 
 	if (not file.empty())
 	{
@@ -533,6 +434,9 @@ compound *compound_factory_impl::create(const std::string &id)
 		}
 	}
 
+	if (result == nullptr)
+		m_missing.insert(id);
+
 	return result;
 }
 
@@ -545,20 +449,6 @@ class local_compound_factory_impl : public compound_factory_impl
 		: compound_factory_impl(next)
 		, m_local_file(file)
 	{
-		const std::regex peptideRx("(?:[lmp]-)?peptide", std::regex::icase);
-
-		for (const auto &[id, name, threeLetterCode, group] :
-			file["comp_list"]["chem_comp"].rows<std::string, std::string, std::string, std::string>("id", "name", "three_letter_code", "group"))
-		{
-			auto &rdb = m_local_file["comp_" + id];
-			if (rdb.empty())
-			{
-				std::cerr << "Missing data in restraint file for id " + id + '\n';
-				continue;
-			}
-
-			construct_compound(rdb, id, name, threeLetterCode, group);
-		}
 	}
 
 	compound *create(const std::string &id) override;
@@ -572,11 +462,17 @@ class local_compound_factory_impl : public compound_factory_impl
 
 compound *local_compound_factory_impl::create(const std::string &id)
 {
+	if (m_missing.contains(id))
+		return nullptr;
+
+	if (auto i = find_if(m_compounds.begin(), m_compounds.end(), [id](compound *c) { return c->id() == id; }); i != m_compounds.end())
+		return *i;
+
 	compound *result = nullptr;
 
 	for (auto &db : m_local_file)
 	{
-		if (db.name() == "comp_" + id)
+		if (db.name() == id)
 		{
 			auto chem_comp = db.get("chem_comp");
 			if (not chem_comp)
@@ -598,6 +494,9 @@ compound *local_compound_factory_impl::create(const std::string &id)
 		}
 	}
 
+	if (result == nullptr)
+		m_missing.insert(id);
+
 	return result;
 }
 
@@ -610,9 +509,13 @@ compound *local_compound_factory_impl::construct_compound(const datablock &rdb,
 	int formal_charge = 0;
 	std::map<std::string,std::size_t> formula_data;
 
-	for (std::size_t ord = 1; const auto &[atom_id, type_symbol, type, charge, x, y, z] :
-		rdb["chem_comp_atom"].rows<std::string, std::string, std::string, int, float, float, float>(
-			"atom_id", "type_symbol", "type", "charge", "x", "y", "z"))
+	for (std::size_t ord = 1; const auto &[atom_id, type_symbol, type, charge, x, y, z, xi, yi, zi] :
+		rdb["chem_comp_atom"].rows<std::string, std::string, std::string, int,
+			std::optional<float>, std::optional<float>, std::optional<float>,
+			std::optional<float>, std::optional<float>, std::optional<float>>(
+			"atom_id", "type_symbol", "type", "charge",
+			"model_Cartn_x", "model_Cartn_y", "model_Cartn_z",
+			"pdbx_model_Cartn_x_ideal", "pdbx_model_Cartn_y_ideal", "pdbx_model_Cartn_z_ideal"))
 	{
 		auto atom = cif::atom_type_traits(type_symbol);
 		formula_weight += atom.weight();
@@ -624,9 +527,9 @@ compound *local_compound_factory_impl::construct_compound(const datablock &rdb,
 			{ "atom_id", atom_id },
 			{ "type_symbol", type_symbol },
 			{ "charge", charge },
-			{ "model_Cartn_x",  x, 3 },
-			{ "model_Cartn_y",  y, 3 },
-			{ "model_Cartn_z",  z, 3 },
+			{ "model_Cartn_x",  x.has_value() ? x : xi, 3 },
+			{ "model_Cartn_y",  y.has_value() ? y : yi, 3 },
+			{ "model_Cartn_z",  z.has_value() ? z : zi, 3 },
 			{ "pdbx_ordinal", ord++ }
 		});
 

src/condition.cpp

@@ -115,27 +115,33 @@ namespace detail
 		auto first = subs.front();
 		auto &fc = first->m_sub;
 
-		for (auto c : fc)
+		for (size_t fc_i = 0; fc_i < fc.size();)
 		{
-			if (not found_in_range(c, subs.begin() + 1, subs.end()))
+			auto c = fc[fc_i];
+			if (not found_in_range(c, subs.begin() + 1, subs.end())) {
+				++fc_i;
 				continue;
+			}
 
 			if (and_result == nullptr)
 				and_result = new and_condition_impl();
 
 			and_result->m_sub.push_back(c);
-			fc.erase(remove(fc.begin(), fc.end(), c), fc.end());
+			fc.erase(fc.begin() + fc_i);
 
 			for (auto sub : subs)
 			{
 				auto &ssub = sub->m_sub;
 
-				for (auto sc : ssub)
+				for (size_t ssub_i = 0; ssub_i < ssub.size();)
 				{
-					if (not sc->equals(c))
+					auto sc = ssub[ssub_i];
+					if (not sc->equals(c)) {
+						++ssub_i;
 						continue;
+					}
 					
-					ssub.erase(remove(ssub.begin(), ssub.end(), sc), ssub.end());
+					ssub.erase(ssub.begin() + ssub_i);
 					delete sc;
 					break;
 				}

src/datablock.cpp

@@ -38,7 +38,18 @@ datablock::datablock(const datablock &db)
 		cat.update_links(*this);
 }
 
-void datablock::set_validator(const validator *v)
+void datablock::load_dictionary()
+{
+	if (auto *audit_conform = get("audit_conform"); audit_conform and not audit_conform->empty())
+		set_validator(&validator_factory::instance().construct_validator(*audit_conform));
+}
+
+void datablock::load_dictionary(std::string_view name)
+{
+	set_validator(&validator_factory::instance()[name]);
+}
+
+void datablock::set_validator(const validator_base *v)
 {
 	m_validator = v;
 
@@ -49,11 +60,12 @@ void datablock::set_validator(const validator *v)
 	}
 	catch (const std::exception &)
 	{
+		m_validator = nullptr;
 		throw_with_nested(std::runtime_error("Error while setting validator in datablock " + m_name));
 	}
 }
 
-const validator *datablock::get_validator() const
+const validator_base *datablock::get_validator() const
 {
 	return m_validator;
 }
@@ -61,7 +73,7 @@ const validator *datablock::get_validator() const
 bool datablock::is_valid() const
 {
 	if (m_validator == nullptr)
-		throw std::runtime_error("Validator not specified");
+		throw std::runtime_error("Validator not specified for datablock data_" + name());
 
 	bool result = true;
 	for (auto &cat : *this)
@@ -73,12 +85,12 @@ bool datablock::is_valid() const
 bool datablock::is_valid()
 {
 	if (m_validator == nullptr)
-		throw std::runtime_error("Validator not specified");
+		throw std::runtime_error("Validator not specified for datablock data_" + name());
 
 	bool result = true;
 	for (auto &cat : *this)
 		result = cat.is_valid() and result;
-	
+
 	// Add or remove the audit_conform block here.
 	if (result)
 	{
@@ -99,7 +111,9 @@ bool datablock::is_valid()
 		}
 	}
 	else
-		erase(std::find_if(begin(), end(), [](category &cat) { return cat.name() == "audit_conform"; }), end());
+		erase(std::find_if(begin(), end(), [](category &cat)
+				  { return cat.name() == "audit_conform"; }),
+			end());
 
 	return result;
 }
@@ -173,7 +187,7 @@ std::tuple<datablock::iterator, bool> datablock::emplace(std::string_view name)
 
 	if (is_new)
 	{
-		i = insert(end(), {name});
+		i = insert(end(), { name });
 		i->set_validator(m_validator, *this);
 	}
 
@@ -234,7 +248,7 @@ namespace
 		return std::get<2>(*i);
 	}
 
-	void calculate_cat_order(cat_order_t &cat_order, iter_t i, const validator &validator)
+	void calculate_cat_order(cat_order_t &cat_order, iter_t i, const validator_base &validator)
 	{
 		if (i == cat_order.end() or get_count(i) >= 0)
 			return;
@@ -307,7 +321,7 @@ void datablock::write(std::ostream &os) const
 	else
 	{
 		// mmcif support, sort of. First write the 'entry' Category
-		// and if it exists, _AND_ we have a Validator, write out the
+		// and if it exists, _AND_ we have a validator, write out the
 		// audit_conform record.
 
 		if (auto entry = get("entry"); entry != nullptr)

src/dictionary_parser.cpp

@@ -488,11 +488,4 @@ validator parse_dictionary(std::string_view name, std::istream &is)
 	return result;
 }
 
-void extend_dictionary(validator &v, std::istream &is)
-{
-	file f;
-	dictionary_parser p(v, is, f);
-	p.load_dictionary();
-}
-
 } // namespace cif

src/file.cpp

@@ -30,19 +30,8 @@
 namespace cif
 {
 
-// --------------------------------------------------------------------
-void file::set_validator(const validator *v)
-{
-	m_validator = v;
-	for (auto &db : *this)
-		db.set_validator(v);
-}
-
 bool file::is_valid() const
 {
-	if (m_validator == nullptr)
-		std::runtime_error("No validator loaded explicitly, cannot continue");
-
 	bool result = true;
 	for (auto &d : *this)
 		result = d.is_valid() and result;
@@ -55,14 +44,6 @@ bool file::is_valid() const
 
 bool file::is_valid()
 {
-	if (m_validator == nullptr)
-	{
-		if (VERBOSE > 0)
-			std::cerr << "No dictionary loaded explicitly, loading default\n";
-
-		load_dictionary();
-	}
-
 	bool result = not empty();
 
 	for (auto &d : *this)
@@ -76,56 +57,54 @@ bool file::is_valid()
 
 bool file::validate_links() const
 {
-	if (m_validator == nullptr)
-		std::runtime_error("No validator loaded explicitly, cannot continue");
-	
 	bool result = true;
 
 	for (auto &db : *this)
 		result = db.validate_links() and result;
-	
+
 	return result;
 }
 
-void file::load_dictionary()
-{
-	if (not empty())
-	{
-		auto *audit_conform = front().get("audit_conform");
-		if (audit_conform and not audit_conform->empty())
-		{
-			std::string name = audit_conform->front().get<std::string>("dict_name");
+// void file::load_dictionary()
+// {
+// 	if (not empty())
+// 	{
+// 		auto *audit_conform = front().get("audit_conform");
+// 		if (audit_conform and not audit_conform->empty())
+// 		{
+// 			std::string name = audit_conform->front().get<std::string>("dict_name");
 
-			if (name == "mmcif_pdbx_v50")
-				name = "mmcif_pdbx.dic";	// we had a bug here in libcifpp... 
+// 			if (name == "mmcif_pdbx_v50")
+// 				name = "mmcif_pdbx.dic"; // we had a bug here in libcifpp...
 
-			if (not name.empty())
-			{
-				try
-				{
-					load_dictionary(name);
-				}
-				catch (const std::exception &ex)
-				{
-					if (VERBOSE)
-						std::cerr << "Failed to load dictionary " << std::quoted(name) << ": " << ex.what() << '\n';
-				}
-			}
-		}
-	}
+// 			if (not name.empty())
+// 			{
+// 				try
+// 				{
+// 					load_dictionary(name);
+// 				}
+// 				catch (const std::exception &ex)
+// 				{
+// 					if (VERBOSE)
+// 						std::cerr << "Failed to load dictionary " << std::quoted(name) << ": " << ex.what() << '\n';
+// 				}
+// 			}
+// 		}
+// 	}
 
-	// if (not m_validator)
-	// 	load_dictionary("mmcif_pdbx.dic");	// TODO: maybe incorrect? Perhaps improve?
-}
+// 	// if (not m_validator)
+// 	// 	load_dictionary("mmcif_pdbx.dic");	// TODO: maybe incorrect? Perhaps improve?
+// }
 
-void file::load_dictionary(std::string_view name)
-{
-	set_validator(&validator_factory::instance()[name]);
-}
+// void file::load_dictionary(std::string_view name)
+// {
+// 	set_validator(&validator_factory::instance()[name]);
+// }
 
 bool file::contains(std::string_view name) const
 {
-	return std::find_if(begin(), end(), [name](const datablock &db) { return iequals(db.name(), name); }) != end();
+	return std::find_if(begin(), end(), [name](const datablock &db)
+			   { return iequals(db.name(), name); }) != end();
 }
 
 datablock &file::operator[](std::string_view name)
@@ -165,10 +144,7 @@ std::tuple<file::iterator, bool> file::emplace(std::string_view name)
 	}
 
 	if (is_new)
-	{
 		i = insert(end(), { name });
-		i->set_validator(m_validator);
-	}
 
 	assert(i != end());
 	return std::make_tuple(i, is_new);
@@ -190,18 +166,44 @@ void file::load(const std::filesystem::path &p)
 	}
 }
 
-void file::load(std::istream &is)
+void file::load(const std::filesystem::path &p, std::string_view dict)
 {
-	auto saved = m_validator;
-	set_validator(nullptr);
+	load(p, validator_factory::instance().operator[](dict));
+}
 
+void file::load(std::istream &is, std::string_view dict)
+{
+	load(is, validator_factory::instance().operator[](dict));
+}
+
+void file::load(const std::filesystem::path &p, const validator_base &v)
+{
+	gzio::ifstream in(p);
+	if (not in.is_open())
+		throw std::runtime_error("Could not open file '" + p.string() + '\'');
+
+	try
+	{
+		load(in, v);
+	}
+	catch (const std::exception &)
+	{
+		throw_with_nested(std::runtime_error("Error reading file '" + p.string() + '\''));
+	}
+}
+
+void file::load(std::istream &is, const validator_base &v)
+{
 	parser p(is, *this);
 	p.parse_file();
+	for (auto &db : *this)
+		db.set_validator(&v);
+}
 
-	if (saved != nullptr)
-		set_validator(saved);
-	else
-		load_dictionary();
+void file::load(std::istream &is)
+{
+	parser p(is, *this);
+	p.parse_file();
 }
 
 void file::save(const std::filesystem::path &p) const

src/model.cpp

@@ -313,7 +313,7 @@ residue::residue(structure &structure, const std::vector<atom> &atoms)
 {
 	if (atoms.empty())
 		throw std::runtime_error("Empty list of atoms");
-	
+
 	auto &a = atoms.front();
 
 	m_compound_id = a.get_label_comp_id();
@@ -352,9 +352,42 @@ EntityType residue::entity_type() const
 
 void residue::add_atom(atom &atom)
 {
+	// update atom since it is now part of this residue
 	m_atoms.push_back(atom);
 }
 
+atom residue::create_new_atom(atom_type inType, const std::string &inAtomID, point inLocation)
+{
+	auto &db = m_structure->get_datablock();
+	auto &atom_site = db["atom_site"];
+
+	auto ai = atom_site.emplace({
+		{ "group_PDB", "HETATM" },
+		{ "id", atom_site.get_unique_id("") },
+		{ "type_symbol", atom_type_traits(inType).symbol() },
+		{ "label_entity_id", get_entity_id() },
+		{ "label_atom_id", inAtomID },
+		{ "label_asym_id", m_asym_id },
+		{ "label_alt_id", "." },
+		{ "label_comp_id", m_compound_id },
+		{ "label_seq_id", m_seq_id },
+		{ "auth_asym_id", m_auth_asym_id },
+		{ "auth_atom_id", inAtomID },
+		{ "auth_comp_id", m_compound_id },
+		{ "auth_seq_id", m_auth_seq_id },
+		{ "occupancy", 1.0f, 2 },
+		{ "pdbx_PDB_model_num", m_structure->get_model_nr() },
+	});
+
+	atom a(db, *ai);
+
+	m_atoms.push_back(a);
+
+	a.set_location(inLocation);
+
+	return a;
+}
+
 std::vector<atom> residue::unique_atoms() const
 {
 	std::vector<atom> result;
@@ -455,6 +488,12 @@ bool residue::has_alternate_atoms() const
 			   { return atom.is_alternate(); }) != m_atoms.end();
 }
 
+bool residue::has_alternate_atoms_for(const std::string &atomID) const
+{
+	return std::find_if(m_atoms.begin(), m_atoms.end(), [atomID](const atom &atom)
+			   { return atom.get_label_atom_id() == atomID and atom.is_alternate(); }) != m_atoms.end();
+}
+
 std::set<std::string> residue::get_atom_ids() const
 {
 	std::set<std::string> ids;
@@ -673,25 +712,26 @@ float monomer::omega() const
 }
 
 const std::map<std::string, std::vector<std::string>> kChiAtomsMap = {
-	{"ASP", {"CG", "OD1"}},
-	{"ASN", {"CG", "OD1"}},
-	{"ARG", {"CG", "CD", "NE", "CZ"}},
-	{"HIS", {"CG", "ND1"}},
-	{"GLN", {"CG", "CD", "OE1"}},
-	{"GLU", {"CG", "CD", "OE1"}},
-	{"SER", {"OG"}},
-	{"THR", {"OG1"}},
-	{"LYS", {"CG", "CD", "CE", "NZ"}},
-	{"TYR", {"CG", "CD1"}},
-	{"PHE", {"CG", "CD1"}},
-	{"LEU", {"CG", "CD1"}},
-	{"TRP", {"CG", "CD1"}},
-	{"CYS", {"SG"}},
-	{"ILE", {"CG1", "CD1"}},
-	{"MET", {"CG", "SD", "CE"}},
-	{"MSE", {"CG", "SE", "CE"}},
-	{"PRO", {"CG", "CD"}},
-	{"VAL", {"CG1"}}};
+	{ "ASP", { "CG", "OD1" } },
+	{ "ASN", { "CG", "OD1" } },
+	{ "ARG", { "CG", "CD", "NE", "CZ" } },
+	{ "HIS", { "CG", "ND1" } },
+	{ "GLN", { "CG", "CD", "OE1" } },
+	{ "GLU", { "CG", "CD", "OE1" } },
+	{ "SER", { "OG" } },
+	{ "THR", { "OG1" } },
+	{ "LYS", { "CG", "CD", "CE", "NZ" } },
+	{ "TYR", { "CG", "CD1" } },
+	{ "PHE", { "CG", "CD1" } },
+	{ "LEU", { "CG", "CD1" } },
+	{ "TRP", { "CG", "CD1" } },
+	{ "CYS", { "SG" } },
+	{ "ILE", { "CG1", "CD1" } },
+	{ "MET", { "CG", "SD", "CE" } },
+	{ "MSE", { "CG", "SE", "CE" } },
+	{ "PRO", { "CG", "CD" } },
+	{ "VAL", { "CG1" } }
+};
 
 std::size_t monomer::nr_of_chis() const
 {
@@ -713,7 +753,7 @@ float monomer::chi(std::size_t nr) const
 		auto i = kChiAtomsMap.find(m_compound_id);
 		if (i != kChiAtomsMap.end() and nr < i->second.size())
 		{
-			std::vector<std::string> atoms{"N", "CA", "CB"};
+			std::vector<std::string> atoms{ "N", "CA", "CB" };
 
 			atoms.insert(atoms.end(), i->second.begin(), i->second.end());
 
@@ -839,7 +879,8 @@ bool monomer::are_bonded(const monomer &a, const monomer &b, float errorMargin)
 			a.get_atom_by_atom_id("CA").get_location(),
 			a.get_atom_by_atom_id("C").get_location(),
 			b.get_atom_by_atom_id("N").get_location(),
-			b.get_atom_by_atom_id("CA").get_location()};
+			b.get_atom_by_atom_id("CA").get_location()
+		};
 
 		auto distanceCACA = distance(atoms[0], atoms[3]);
 		double omega = dihedral_angle(atoms[0], atoms[1], atoms[2], atoms[3]);
@@ -880,6 +921,15 @@ bool monomer::is_cis(const monomer &a, const monomer &b)
 	return std::abs(omega(a, b)) < 30.0f;
 }
 
+atom monomer::create_new_atom(atom_type inType, const std::string &inAtomID, point inLocation)
+{
+	atom a = residue::create_new_atom(inType, inAtomID, inLocation);
+
+	a.set_property("group_PDB", "ATOM");
+
+	return a;
+}
+
 // --------------------------------------------------------------------
 // polymer
 
@@ -916,7 +966,7 @@ polymer::polymer(structure &s, const std::string &entityID, const std::string &a
 		}
 		else if (VERBOSE > 0)
 		{
-			monomer m{*this, index, seqID, authSeqID, pdbInsCode, compoundID};
+			monomer m{ *this, index, seqID, authSeqID, pdbInsCode, compoundID };
 			std::cerr << "Dropping alternate residue " << m << '\n';
 		}
 	}
@@ -984,7 +1034,6 @@ sugar::sugar(sugar &&rhs)
 	: residue(std::forward<residue>(rhs))
 	, m_branch(rhs.m_branch)
 {
-
 }
 
 sugar &sugar::operator=(sugar &&rhs)
@@ -1048,19 +1097,19 @@ cif::mm::atom sugar::add_atom(row_initializer atom_info)
 
 	auto atom_id = atom_site.get_unique_id("");
 
-	atom_info.set_value({"group_PDB", "HETATM"});
-	atom_info.set_value({"id", atom_id});
-	atom_info.set_value({"label_entity_id", m_branch->get_entity_id()});
-	atom_info.set_value({"label_asym_id", m_branch->get_asym_id()});
-	atom_info.set_value({"label_comp_id", m_compound_id});
-	atom_info.set_value({"label_seq_id", "."});
-	atom_info.set_value({"label_alt_id", "."});
-	atom_info.set_value({"auth_asym_id", m_branch->get_asym_id()});
-	atom_info.set_value({"auth_comp_id", m_compound_id});
-	atom_info.set_value({"auth_seq_id", m_auth_seq_id});
-	atom_info.set_value({"occupancy", 1.0, 2});
-	atom_info.set_value({"B_iso_or_equiv", 30.0, 2});
-	atom_info.set_value({"pdbx_PDB_model_num", 1});
+	atom_info.set_value({ "group_PDB", "HETATM" });
+	atom_info.set_value({ "id", atom_id });
+	atom_info.set_value({ "label_entity_id", m_branch->get_entity_id() });
+	atom_info.set_value({ "label_asym_id", m_branch->get_asym_id() });
+	atom_info.set_value({ "label_comp_id", m_compound_id });
+	atom_info.set_value({ "label_seq_id", "." });
+	atom_info.set_value({ "label_alt_id", "." });
+	atom_info.set_value({ "auth_asym_id", m_branch->get_asym_id() });
+	atom_info.set_value({ "auth_comp_id", m_compound_id });
+	atom_info.set_value({ "auth_seq_id", m_auth_seq_id });
+	atom_info.set_value({ "occupancy", 1.0, 2 });
+	atom_info.set_value({ "B_iso_or_equiv", 30.0, 2 });
+	atom_info.set_value({ "pdbx_PDB_model_num", 1 });
 
 	auto row = atom_site.emplace(std::move(atom_info));
 	auto result = m_structure->emplace_atom(db, row);
@@ -1118,7 +1167,7 @@ void branch::link_atoms()
 		auto entity_id = front().get_entity_id();
 
 		for (const auto &[num1, num2, atom1, atom2] : branch_link.find<std::size_t, std::size_t, std::string, std::string>(
-				"entity_id"_key == entity_id, "entity_branch_list_num_1", "entity_branch_list_num_2", "atom_id_1", "atom_id_2"))
+				 "entity_id"_key == entity_id, "entity_branch_list_num_1", "entity_branch_list_num_2", "atom_id_1", "atom_id_2"))
 		{
 			// if (not iequals(atom1, "c1"))
 			// 	throw std::runtime_error("invalid pdbx_entity_branch_link");
@@ -1133,10 +1182,11 @@ void branch::link_atoms()
 
 sugar &branch::get_sugar_by_num(int nr)
 {
-	auto i = find_if(begin(), end(), [nr](const sugar &s) { return s.num() == nr; });
+	auto i = find_if(begin(), end(), [nr](const sugar &s)
+		{ return s.num() == nr; });
 	if (i == end())
 		throw std::out_of_range("Sugar with num " + std::to_string(nr) + " not found in branch " + m_asym_id);
-	
+
 	return *i;
 }
 
@@ -1157,32 +1207,29 @@ sugar &branch::construct_sugar(const std::string &compound_id)
 	auto r = chemComp.find(key("id") == compound_id);
 	if (r.empty())
 	{
-		chemComp.emplace({
-			{"id", compound_id},
-			{"name", compound->name()},
-			{"formula", compound->formula()},
-			{"formula_weight", compound->formula_weight()},
-			{"type", compound->type()}});
+		chemComp.emplace({ { "id", compound_id },
+			{ "name", compound->name() },
+			{ "formula", compound->formula() },
+			{ "formula_weight", compound->formula_weight() },
+			{ "type", compound->type() } });
 	}
 
 	sugar &result = emplace_back(*this, compound_id, m_asym_id, static_cast<int>(size() + 1));
 
-	db["pdbx_branch_scheme"].emplace({
-		{"asym_id", result.get_asym_id()},
-		{"entity_id", result.get_entity_id()},
-		{"num", result.num()},
-		{"mon_id", result.get_compound_id()},
+	db["pdbx_branch_scheme"].emplace({ { "asym_id", result.get_asym_id() },
+		{ "entity_id", result.get_entity_id() },
+		{ "num", result.num() },
+		{ "mon_id", result.get_compound_id() },
 
-		{"pdb_asym_id", result.get_asym_id()},
-		{"pdb_seq_num", result.num()},
-		{"pdb_mon_id", result.get_compound_id()},
+		{ "pdb_asym_id", result.get_asym_id() },
+		{ "pdb_seq_num", result.num() },
+		{ "pdb_mon_id", result.get_compound_id() },
 
-		{"auth_asym_id", result.get_auth_asym_id()},
-		{"auth_mon_id", result.get_compound_id()},
-		{"auth_seq_num", result.get_auth_seq_id()},
+		{ "auth_asym_id", result.get_auth_asym_id() },
+		{ "auth_mon_id", result.get_compound_id() },
+		{ "auth_seq_num", result.get_auth_seq_id() },
 
-		{"hetero", "n"}
-	});
+		{ "hetero", "n" } });
 
 	return result;
 }
@@ -1200,19 +1247,17 @@ sugar &branch::construct_sugar(const std::string &compound_id, const std::string
 	auto &pdbx_entity_branch_link = db["pdbx_entity_branch_link"];
 	auto linkID = pdbx_entity_branch_link.get_unique_id("");
 
-	db["pdbx_entity_branch_link"].emplace({
-		{ "link_id", linkID },
+	db["pdbx_entity_branch_link"].emplace({ { "link_id", linkID },
 		{ "entity_id", get_entity_id() },
-		{ "entity_branch_list_num_1", result.num() }, 
-		{ "comp_id_1", compound_id }, 
+		{ "entity_branch_list_num_1", result.num() },
+		{ "comp_id_1", compound_id },
 		{ "atom_id_1", atom_id },
-		{ "leaving_atom_id_1", "O1" }, 	/// TODO: Need to fix this!
-		{ "entity_branch_list_num_2", linked.num() }, 
-		{ "comp_id_2", linked.get_compound_id() }, 
-		{ "atom_id_2", linked_atom_id }, 
-		{ "leaving_atom_id_2", "." }, 
-		{ "value_order", "sing" }
-	});
+		{ "leaving_atom_id_1", "O1" }, /// TODO: Need to fix this!
+		{ "entity_branch_list_num_2", linked.num() },
+		{ "comp_id_2", linked.get_compound_id() },
+		{ "atom_id_2", linked_atom_id },
+		{ "leaving_atom_id_2", "." },
+		{ "value_order", "sing" } });
 
 	return result;
 }
@@ -1261,6 +1306,9 @@ structure::structure(datablock &db, std::size_t modelNr, StructureOpenOptions op
 	: m_db(db)
 	, m_model_nr(modelNr)
 {
+	if (db.get_validator() == nullptr)
+		db.load_dictionary();
+
 	auto &atomCat = db["atom_site"];
 
 	load_atoms_for_model(options);
@@ -1321,7 +1369,7 @@ void structure::load_data()
 {
 	auto &polySeqScheme = m_db["pdbx_poly_seq_scheme"];
 
-	for (const auto &[asym_id, auth_asym_id, entityID] : polySeqScheme.rows<std::string,std::string,std::string>("asym_id", "pdb_strand_id", "entity_id"))
+	for (const auto &[asym_id, auth_asym_id, entityID] : polySeqScheme.rows<std::string, std::string, std::string>("asym_id", "pdb_strand_id", "entity_id"))
 	{
 		if (m_polymers.empty() or m_polymers.back().get_asym_id() != asym_id or m_polymers.back().get_entity_id() != entityID)
 			m_polymers.emplace_back(*this, entityID, asym_id, auth_asym_id);
@@ -1329,7 +1377,7 @@ void structure::load_data()
 
 	auto &branchScheme = m_db["pdbx_branch_scheme"];
 
-	for (const auto &[asym_id, entity_id] : branchScheme.rows<std::string,std::string>("asym_id", "entity_id"))
+	for (const auto &[asym_id, entity_id] : branchScheme.rows<std::string, std::string>("asym_id", "entity_id"))
 	{
 		if (m_branches.empty() or m_branches.back().get_asym_id() != asym_id)
 			m_branches.emplace_back(*this, asym_id, entity_id);
@@ -1337,8 +1385,8 @@ void structure::load_data()
 
 	auto &nonPolyScheme = m_db["pdbx_nonpoly_scheme"];
 
-	for (const auto&[asym_id, monID, pdbStrandID, pdbSeqNum, pdbInsCode] :
-			nonPolyScheme.rows<std::string,std::string,std::string,std::string,std::string>("asym_id", "mon_id", "pdb_strand_id", "pdb_seq_num", "pdb_ins_code"))
+	for (const auto &[asym_id, monID, pdbStrandID, pdbSeqNum, pdbInsCode] :
+		nonPolyScheme.rows<std::string, std::string, std::string, std::string, std::string>("asym_id", "mon_id", "pdb_strand_id", "pdb_seq_num", "pdb_ins_code"))
 		m_non_polymers.emplace_back(*this, monID, asym_id, 0, pdbStrandID, pdbSeqNum, pdbInsCode);
 
 	// place atoms in residues
@@ -1349,18 +1397,18 @@ void structure::load_data()
 	for (auto &poly : m_polymers)
 	{
 		for (auto &res : poly)
-			resMap[{res.get_asym_id(), res.get_seq_id(), res.get_auth_seq_id()}] = &res;
+			resMap[{ res.get_asym_id(), res.get_seq_id(), res.get_auth_seq_id() }] = &res;
 	}
 
 	for (auto &res : m_non_polymers)
-		resMap[{res.get_asym_id(), res.get_seq_id(), res.get_auth_seq_id()}] = &res;
+		resMap[{ res.get_asym_id(), res.get_seq_id(), res.get_auth_seq_id() }] = &res;
 
 	std::set<std::string> sugars;
 	for (auto &branch : m_branches)
 	{
 		for (auto &sugar : branch)
 		{
-			resMap[{sugar.get_asym_id(), sugar.get_seq_id(), sugar.get_auth_seq_id()}] = &sugar;
+			resMap[{ sugar.get_asym_id(), sugar.get_seq_id(), sugar.get_auth_seq_id() }] = &sugar;
 			sugars.insert(sugar.get_compound_id());
 		}
 	}
@@ -1392,7 +1440,9 @@ void structure::load_data()
 	}
 
 	// what the ...
-	m_branches.erase(std::remove_if(m_branches.begin(), m_branches.end(), [](const branch &b) { return b.empty(); }), m_branches.end());
+	m_branches.erase(std::remove_if(m_branches.begin(), m_branches.end(), [](const branch &b)
+						 { return b.empty(); }),
+		m_branches.end());
 
 	for (auto &branch : m_branches)
 		branch.link_atoms();
@@ -1638,7 +1688,7 @@ residue &structure::get_residue(const std::string &asym_id, int seqID, const std
 
 	if (seqID != 0)
 		desc += "/" + std::to_string(seqID);
-	
+
 	if (not authSeqID.empty())
 		desc += "-" + authSeqID;
 
@@ -1684,7 +1734,7 @@ residue &structure::get_residue(const std::string &asym_id, const std::string &c
 
 	if (seqID != 0)
 		desc += "/" + std::to_string(seqID);
-	
+
 	if (not authSeqID.empty())
 		desc += "-" + authSeqID;
 
@@ -1714,12 +1764,11 @@ std::string structure::insert_compound(const std::string &compoundID, bool is_en
 	auto r = chemComp.find(key("id") == compoundID);
 	if (r.empty())
 	{
-		chemComp.emplace({
-			{"id", compoundID},
-			{"name", compound->name()},
-			{"formula", compound->formula()},
-			{"formula_weight", compound->formula_weight()},
-			{"type", compound->type()}});
+		chemComp.emplace({ { "id", compoundID },
+			{ "name", compound->name() },
+			{ "formula", compound->formula() },
+			{ "formula_weight", compound->formula_weight() },
+			{ "type", compound->type() } });
 	}
 
 	std::string entity_id;
@@ -1735,16 +1784,14 @@ std::string structure::insert_compound(const std::string &compoundID, bool is_en
 			auto &entity = m_db["entity"];
 			entity_id = entity.get_unique_id("");
 
-			entity.emplace({
-				{"id", entity_id},
-				{"type", "non-polymer"},
-				{"pdbx_description", compound->name()},
-				{"formula_weight", compound->formula_weight()}});
+			entity.emplace({ { "id", entity_id },
+				{ "type", "non-polymer" },
+				{ "pdbx_description", compound->name() },
+				{ "formula_weight", compound->formula_weight() } });
 
-			pdbxEntityNonpoly.emplace({
-				{"entity_id", entity_id},
-				{"name", compound->name()},
-				{"comp_id", compoundID}});
+			pdbxEntityNonpoly.emplace({ { "entity_id", entity_id },
+				{ "name", compound->name() },
+				{ "comp_id", compoundID } });
 		}
 	}
 
@@ -1773,7 +1820,7 @@ atom &structure::emplace_atom(atom &&atom)
 			R = i - 1;
 	}
 
-	if (R == -1)	// msvc... 
+	if (R == -1) // msvc...
 		m_atom_index.insert(m_atom_index.begin(), m_atoms.size());
 	else
 		m_atom_index.insert(m_atom_index.begin() + R + 1, m_atoms.size());
@@ -1829,19 +1876,15 @@ void structure::remove_atom(atom &a, bool removeFromResidue)
 		for (std::string prefix : { "ptnr1_", "ptnr2_", "pdbx_ptnr3_" })
 		{
 			if (a.get_label_seq_id() == 0)
-				cond = std::move(cond) or (
-					cif::key(prefix + "label_asym_id") == a.get_label_asym_id() and
-					cif::key(prefix + "label_seq_id") == null and
-					cif::key(prefix + "auth_seq_id") == a.get_auth_seq_id() and
-					cif::key(prefix + "label_atom_id") == a.get_label_atom_id()
-				);
+				cond = std::move(cond) or (cif::key(prefix + "label_asym_id") == a.get_label_asym_id() and
+											  cif::key(prefix + "label_seq_id") == null and
+											  cif::key(prefix + "auth_seq_id") == a.get_auth_seq_id() and
+											  cif::key(prefix + "label_atom_id") == a.get_label_atom_id());
 			else
-				cond = std::move(cond) or (
-					cif::key(prefix + "label_asym_id") == a.get_label_asym_id() and
-					cif::key(prefix + "label_seq_id") == a.get_label_seq_id() and
-					cif::key(prefix + "auth_seq_id") == a.get_auth_seq_id() and
-					cif::key(prefix + "label_atom_id") == a.get_label_atom_id()
-				);
+				cond = std::move(cond) or (cif::key(prefix + "label_asym_id") == a.get_label_asym_id() and
+											  cif::key(prefix + "label_seq_id") == a.get_label_seq_id() and
+											  cif::key(prefix + "auth_seq_id") == a.get_auth_seq_id() and
+											  cif::key(prefix + "label_atom_id") == a.get_label_atom_id());
 		}
 
 		if (cond)
@@ -1947,15 +1990,15 @@ void structure::change_residue(residue &res, const std::string &newCompound,
 		if (entityID.empty())
 		{
 			entityID = entity.get_unique_id("");
-			entity.emplace({{"id", entityID},
-				{"type", "non-polymer"},
-				{"pdbx_description", compound->name()},
-				{"formula_weight", compound->formula_weight()}});
+			entity.emplace({ { "id", entityID },
+				{ "type", "non-polymer" },
+				{ "pdbx_description", compound->name() },
+				{ "formula_weight", compound->formula_weight() } });
 
 			auto &pdbxEntityNonpoly = m_db["pdbx_entity_nonpoly"];
-			pdbxEntityNonpoly.emplace({{"entity_id", entityID},
-				{"name", compound->name()},
-				{"comp_id", newCompound}});
+			pdbxEntityNonpoly.emplace({ { "entity_id", entityID },
+				{ "name", compound->name() },
+				{ "comp_id", newCompound } });
 		}
 
 		auto &pdbxNonPolyScheme = m_db["pdbx_nonpoly_scheme"];
@@ -1971,11 +2014,11 @@ void structure::change_residue(residue &res, const std::string &newCompound,
 		auto &chemComp = m_db["chem_comp"];
 		if (not chemComp.contains(key("id") == newCompound))
 		{
-			chemComp.emplace({{"id", newCompound},
-				{"name", compound->name()},
-				{"formula", compound->formula()},
-				{"formula_weight", compound->formula_weight()},
-				{"type", compound->type()}});
+			chemComp.emplace({ { "id", newCompound },
+				{ "name", compound->name() },
+				{ "formula", compound->formula() },
+				{ "formula_weight", compound->formula_weight() },
+				{ "type", compound->type() } });
 		}
 
 		// update the struct_asym for the new entity
@@ -2105,7 +2148,7 @@ void structure::remove_residue(residue &res)
 
 		case EntityType::Branched:
 		{
-			auto &s = dynamic_cast<sugar&>(res);
+			auto &s = dynamic_cast<sugar &>(res);
 
 			remove_sugar(s);
 
@@ -2133,7 +2176,7 @@ void structure::remove_sugar(sugar &s)
 		throw std::runtime_error("sugar not part of branch");
 	std::size_t six = si - branch.begin();
 
-	if (six == 0)	// first sugar, means the death of this branch
+	if (six == 0) // first sugar, means the death of this branch
 		remove_branch(branch);
 	else
 	{
@@ -2148,7 +2191,7 @@ void structure::remove_sugar(sugar &s)
 
 			if (dix.count(tix))
 				continue;
-			
+
 			dix.insert(tix);
 
 			for (auto &s2 : branch)
@@ -2161,7 +2204,9 @@ void structure::remove_sugar(sugar &s)
 				remove_atom(atom, false);
 		}
 
-		branch.erase(remove_if(branch.begin(), branch.end(), [dix](const sugar &s) { return dix.count(s.num()); }), branch.end());
+		branch.erase(remove_if(branch.begin(), branch.end(), [dix](const sugar &s)
+						 { return dix.count(s.num()); }),
+			branch.end());
 
 		auto entity_id = create_entity_for_branch(branch);
 
@@ -2181,23 +2226,21 @@ void structure::remove_sugar(sugar &s)
 
 		for (auto &sugar : branch)
 		{
-			pdbx_branch_scheme.emplace({
-				{"asym_id", asym_id},
-				{"entity_id", entity_id},
-				{"num", sugar.num()},
-				{"mon_id", sugar.get_compound_id()},
+			pdbx_branch_scheme.emplace({ { "asym_id", asym_id },
+				{ "entity_id", entity_id },
+				{ "num", sugar.num() },
+				{ "mon_id", sugar.get_compound_id() },
 
-				{"pdb_asym_id", asym_id},
-				{"pdb_seq_num", sugar.num()},
-				{"pdb_mon_id", sugar.get_compound_id()},
+				{ "pdb_asym_id", asym_id },
+				{ "pdb_seq_num", sugar.num() },
+				{ "pdb_mon_id", sugar.get_compound_id() },
 
 				// TODO: need fix, collect from nag_atoms?
-				{"auth_asym_id", asym_id},
-				{"auth_mon_id", sugar.get_compound_id()},
-				{"auth_seq_num", sugar.get_auth_seq_id()},
+				{ "auth_asym_id", asym_id },
+				{ "auth_mon_id", sugar.get_compound_id() },
+				{ "auth_seq_num", sugar.get_auth_seq_id() },
 
-				{"hetero", "n"}
-			});
+				{ "hetero", "n" } });
 		}
 	}
 }
@@ -2232,13 +2275,11 @@ std::string structure::create_non_poly(const std::string &entity_id, const std::
 	auto &struct_asym = m_db["struct_asym"];
 	std::string asym_id = struct_asym.get_unique_id();
 
-	struct_asym.emplace({
-		{"id", asym_id},
-		{"pdbx_blank_PDB_chainid_flag", "N"},
-		{"pdbx_modified", "N"},
-		{"entity_id", entity_id},
-		{"details", "?"}
-	});
+	struct_asym.emplace({ { "id", asym_id },
+		{ "pdbx_blank_PDB_chainid_flag", "N" },
+		{ "pdbx_modified", "N" },
+		{ "entity_id", entity_id },
+		{ "details", "?" } });
 
 	std::string comp_id = m_db["pdbx_entity_nonpoly"].find1<std::string>("entity_id"_key == entity_id, "comp_id");
 
@@ -2250,29 +2291,27 @@ std::string structure::create_non_poly(const std::string &entity_id, const std::
 	{
 		auto atom_id = atom_site.get_unique_id("");
 
-		auto row = atom_site.emplace({
-			{"group_PDB", atom.get_property("group_PDB")},
-			{"id", atom_id},
-			{"type_symbol", atom.get_property("type_symbol")},
-			{"label_atom_id", atom.get_property("label_atom_id")},
-			{"label_alt_id", atom.get_property("label_alt_id")},
-			{"label_comp_id", comp_id},
-			{"label_asym_id", asym_id},
-			{"label_entity_id", entity_id},
-			{"label_seq_id", "."},
-			{"pdbx_PDB_ins_code", ""},
-			{"Cartn_x", atom.get_property("Cartn_x")},
-			{"Cartn_y", atom.get_property("Cartn_y")},
-			{"Cartn_z", atom.get_property("Cartn_z")},
-			{"occupancy", atom.get_property("occupancy")},
-			{"B_iso_or_equiv", atom.get_property("B_iso_or_equiv")},
-			{"pdbx_formal_charge", atom.get_property("pdbx_formal_charge")},
-			{"auth_seq_id", 1},
-			{"auth_comp_id", comp_id},
-			{"auth_asym_id", asym_id},
-			{"auth_atom_id", atom.get_property("label_atom_id")},
-			{"pdbx_PDB_model_num", 1}
-		});
+		auto row = atom_site.emplace({ { "group_PDB", atom.get_property("group_PDB") },
+			{ "id", atom_id },
+			{ "type_symbol", atom.get_property("type_symbol") },
+			{ "label_atom_id", atom.get_property("label_atom_id") },
+			{ "label_alt_id", atom.get_property("label_alt_id") },
+			{ "label_comp_id", comp_id },
+			{ "label_asym_id", asym_id },
+			{ "label_entity_id", entity_id },
+			{ "label_seq_id", "." },
+			{ "pdbx_PDB_ins_code", "" },
+			{ "Cartn_x", atom.get_property("Cartn_x") },
+			{ "Cartn_y", atom.get_property("Cartn_y") },
+			{ "Cartn_z", atom.get_property("Cartn_z") },
+			{ "occupancy", atom.get_property("occupancy") },
+			{ "B_iso_or_equiv", atom.get_property("B_iso_or_equiv") },
+			{ "pdbx_formal_charge", atom.get_property("pdbx_formal_charge") },
+			{ "auth_seq_id", 1 },
+			{ "auth_comp_id", comp_id },
+			{ "auth_asym_id", asym_id },
+			{ "auth_atom_id", atom.get_property("label_atom_id") },
+			{ "pdbx_PDB_model_num", 1 } });
 
 		auto &newAtom = emplace_atom(std::make_shared<atom::atom_impl>(m_db, atom_id));
 		res.add_atom(newAtom);
@@ -2281,16 +2320,16 @@ std::string structure::create_non_poly(const std::string &entity_id, const std::
 	auto &pdbx_nonpoly_scheme = m_db["pdbx_nonpoly_scheme"];
 	std::size_t ndb_nr = pdbx_nonpoly_scheme.find("asym_id"_key == asym_id and "entity_id"_key == entity_id).size() + 1;
 	pdbx_nonpoly_scheme.emplace({
-		{"asym_id", asym_id},
-		{"entity_id", entity_id},
-		{"mon_id", comp_id},
-		{"ndb_seq_num", ndb_nr},
-		{"pdb_seq_num", res.get_auth_seq_id()},
-		{"auth_seq_num", res.get_auth_seq_id()},
-		{"pdb_mon_id", comp_id},
-		{"auth_mon_id", comp_id},
-		{"pdb_strand_id", asym_id},
-		{"pdb_ins_code", "."},
+		{ "asym_id", asym_id },
+		{ "entity_id", entity_id },
+		{ "mon_id", comp_id },
+		{ "ndb_seq_num", ndb_nr },
+		{ "pdb_seq_num", res.get_auth_seq_id() },
+		{ "auth_seq_num", res.get_auth_seq_id() },
+		{ "pdb_mon_id", comp_id },
+		{ "auth_mon_id", comp_id },
+		{ "pdb_strand_id", asym_id },
+		{ "pdb_ins_code", "." },
 	});
 
 	return asym_id;
@@ -2303,13 +2342,11 @@ std::string structure::create_non_poly(const std::string &entity_id, std::vector
 	auto &struct_asym = m_db["struct_asym"];
 	std::string asym_id = struct_asym.get_unique_id();
 
-	struct_asym.emplace({
-		{"id", asym_id},
-		{"pdbx_blank_PDB_chainid_flag", "N"},
-		{"pdbx_modified", "N"},
-		{"entity_id", entity_id},
-		{"details", "?"}
-	});
+	struct_asym.emplace({ { "id", asym_id },
+		{ "pdbx_blank_PDB_chainid_flag", "N" },
+		{ "pdbx_modified", "N" },
+		{ "entity_id", entity_id },
+		{ "details", "?" } });
 
 	std::string comp_id = m_db["pdbx_entity_nonpoly"].find1<std::string>("entity_id"_key == entity_id, "comp_id");
 
@@ -2326,14 +2363,14 @@ std::string structure::create_non_poly(const std::string &entity_id, std::vector
 		atom.set_value("auth_asym_id", asym_id);
 		atom.set_value("label_entity_id", entity_id);
 
-		atom.set_value_if_empty({"group_PDB", "HETATM"});
-		atom.set_value_if_empty({"label_comp_id", comp_id});
-		atom.set_value_if_empty({"label_seq_id", "."});
-		atom.set_value_if_empty({"auth_comp_id", comp_id});
-		atom.set_value_if_empty({"auth_seq_id", 1});
-		atom.set_value_if_empty({"pdbx_PDB_model_num", 1});
-		atom.set_value_if_empty({"label_alt_id", ""});
-		atom.set_value_if_empty({"occupancy", 1.0, 2});
+		atom.set_value_if_empty({ "group_PDB", "HETATM" });
+		atom.set_value_if_empty({ "label_comp_id", comp_id });
+		atom.set_value_if_empty({ "label_seq_id", "." });
+		atom.set_value_if_empty({ "auth_comp_id", comp_id });
+		atom.set_value_if_empty({ "auth_seq_id", 1 });
+		atom.set_value_if_empty({ "pdbx_PDB_model_num", 1 });
+		atom.set_value_if_empty({ "label_alt_id", "" });
+		atom.set_value_if_empty({ "occupancy", 1.0, 2 });
 
 		auto row = atom_site.emplace(atom.begin(), atom.end());
 
@@ -2344,16 +2381,16 @@ std::string structure::create_non_poly(const std::string &entity_id, std::vector
 	auto &pdbx_nonpoly_scheme = m_db["pdbx_nonpoly_scheme"];
 	std::size_t ndb_nr = pdbx_nonpoly_scheme.find("asym_id"_key == asym_id and "entity_id"_key == entity_id).size() + 1;
 	pdbx_nonpoly_scheme.emplace({
-		{"asym_id", asym_id},
-		{"entity_id", entity_id},
-		{"mon_id", comp_id},
-		{"ndb_seq_num", ndb_nr},
-		{"pdb_seq_num", res.get_auth_seq_id()},
-		{"auth_seq_num", res.get_auth_seq_id()},
-		{"pdb_mon_id", comp_id},
-		{"auth_mon_id", comp_id},
-		{"pdb_strand_id", asym_id},
-		{"pdb_ins_code", "."},
+		{ "asym_id", asym_id },
+		{ "entity_id", entity_id },
+		{ "mon_id", comp_id },
+		{ "ndb_seq_num", ndb_nr },
+		{ "pdb_seq_num", res.get_auth_seq_id() },
+		{ "auth_seq_num", res.get_auth_seq_id() },
+		{ "pdb_mon_id", comp_id },
+		{ "auth_mon_id", comp_id },
+		{ "pdb_strand_id", asym_id },
+		{ "pdb_ins_code", "." },
 	});
 
 	return asym_id;
@@ -2375,15 +2412,13 @@ void structure::create_water(row_initializer atom)
 	{
 		asym_id = struct_asym.get_unique_id();
 
-		struct_asym.emplace({
-			{"id", asym_id},
-			{"pdbx_blank_PDB_chainid_flag", "N"},
-			{"pdbx_modified", "N"},
-			{"entity_id", entity_id},
-			{"details", "?"}
-		});
+		struct_asym.emplace({ { "id", asym_id },
+			{ "pdbx_blank_PDB_chainid_flag", "N" },
+			{ "pdbx_modified", "N" },
+			{ "entity_id", entity_id },
+			{ "details", "?" } });
 	}
-	
+
 	auto &atom_site = m_db["atom_site"];
 	auto auth_seq_id = atom_site.find_max<int>("auth_seq_id", "label_entity_id"_key == entity_id) + 1;
 	if (auth_seq_id < 0)
@@ -2397,13 +2432,13 @@ void structure::create_water(row_initializer atom)
 	atom.set_value("label_entity_id", entity_id);
 	atom.set_value("auth_seq_id", std::to_string(auth_seq_id));
 
-	atom.set_value_if_empty({"group_PDB", "HETATM"});
-	atom.set_value_if_empty({"label_comp_id", "HOH"});
-	atom.set_value_if_empty({"label_seq_id", "."});
-	atom.set_value_if_empty({"auth_comp_id", "HOH"});
-	atom.set_value_if_empty({"pdbx_PDB_model_num", 1});
-	atom.set_value_if_empty({"label_alt_id", ""});
-	atom.set_value_if_empty({"occupancy", 1.0, 2});
+	atom.set_value_if_empty({ "group_PDB", "HETATM" });
+	atom.set_value_if_empty({ "label_comp_id", "HOH" });
+	atom.set_value_if_empty({ "label_seq_id", "." });
+	atom.set_value_if_empty({ "auth_comp_id", "HOH" });
+	atom.set_value_if_empty({ "pdbx_PDB_model_num", 1 });
+	atom.set_value_if_empty({ "label_alt_id", "" });
+	atom.set_value_if_empty({ "occupancy", 1.0, 2 });
 
 	auto row = atom_site.emplace(atom.begin(), atom.end());
 
@@ -2412,16 +2447,16 @@ void structure::create_water(row_initializer atom)
 	auto &pdbx_nonpoly_scheme = m_db["pdbx_nonpoly_scheme"];
 	int ndb_nr = pdbx_nonpoly_scheme.find_max<int>("ndb_seq_num") + 1;
 	pdbx_nonpoly_scheme.emplace({
-		{"asym_id", asym_id},
-		{"entity_id", entity_id},
-		{"mon_id", "HOH"},
-		{"ndb_seq_num", ndb_nr},
-		{"pdb_seq_num", auth_seq_id},
-		{"auth_seq_num", auth_seq_id},
-		{"pdb_mon_id", "HOH"},
-		{"auth_mon_id", "HOH"},
-		{"pdb_strand_id", asym_id},
-		{"pdb_ins_code", "."},
+		{ "asym_id", asym_id },
+		{ "entity_id", entity_id },
+		{ "mon_id", "HOH" },
+		{ "ndb_seq_num", ndb_nr },
+		{ "pdb_seq_num", auth_seq_id },
+		{ "auth_seq_num", auth_seq_id },
+		{ "pdb_mon_id", "HOH" },
+		{ "auth_mon_id", "HOH" },
+		{ "pdb_strand_id", asym_id },
+		{ "pdb_ins_code", "." },
 	});
 }
 
@@ -2433,18 +2468,14 @@ branch &structure::create_branch()
 	auto entity_id = entity.get_unique_id("");
 	auto asym_id = struct_asym.get_unique_id();
 
-	entity.emplace({
-		{"id", entity_id},
-		{"type", "branched"}
-	});
+	entity.emplace({ { "id", entity_id },
+		{ "type", "branched" } });
 
-	struct_asym.emplace({
-		{"id", asym_id},
-		{"pdbx_blank_PDB_chainid_flag", "N"},
-		{"pdbx_modified", "N"},
-		{"entity_id", entity_id},
-		{"details", "?"}
-	});
+	struct_asym.emplace({ { "id", asym_id },
+		{ "pdbx_blank_PDB_chainid_flag", "N" },
+		{ "pdbx_modified", "N" },
+		{ "entity_id", entity_id },
+		{ "details", "?" } });
 
 	return m_branches.emplace_back(*this, asym_id, entity_id);
 }
@@ -2641,22 +2672,19 @@ std::string structure::create_entity_for_branch(branch &branch)
 		if (VERBOSE)
 			std::cout << "Creating new entity " << entityID << " for branched sugar " << entityName << '\n';
 
-		entity.emplace({
-			{"id", entityID},
-			{"type", "branched"},
-			{"src_method", "man"},
-			{"pdbx_description", entityName},
-			{"formula_weight", branch.weight()}});
+		entity.emplace({ { "id", entityID },
+			{ "type", "branched" },
+			{ "src_method", "man" },
+			{ "pdbx_description", entityName },
+			{ "formula_weight", branch.weight() } });
 
 		auto &pdbx_entity_branch_list = m_db["pdbx_entity_branch_list"];
 		for (auto &sugar : branch)
 		{
-			pdbx_entity_branch_list.emplace({
-				{"entity_id", entityID},
-				{"comp_id", sugar.get_compound_id()},
-				{"num", sugar.num()},
-				{"hetero", "n"}
-			});
+			pdbx_entity_branch_list.emplace({ { "entity_id", entityID },
+				{ "comp_id", sugar.get_compound_id() },
+				{ "num", sugar.num() },
+				{ "hetero", "n" } });
 		}
 
 		auto &pdbx_entity_branch_link = m_db["pdbx_entity_branch_link"];
@@ -2670,19 +2698,17 @@ std::string structure::create_entity_for_branch(branch &branch)
 			auto &s2 = branch.at(stoi(l2.get_auth_seq_id()) - 1);
 			auto l1 = s2.get_atom_by_atom_id("C1");
 
-			pdbx_entity_branch_link.emplace({
-				{"link_id", pdbx_entity_branch_link.get_unique_id("")},
-				{"entity_id", entityID},
-				{"entity_branch_list_num_1", s1.get_auth_seq_id()},
-				{"comp_id_1", s1.get_compound_id()},
-				{"atom_id_1", l1.get_label_atom_id()},
-				{"leaving_atom_id_1", "O1"},
-				{"entity_branch_list_num_2", s2.get_auth_seq_id()},
-				{"comp_id_2", s2.get_compound_id()},
-				{"atom_id_2", l2.get_label_atom_id()},
-				{"leaving_atom_id_2", "H" + l2.get_label_atom_id()},
-				{"value_order", "sing"}
-			});
+			pdbx_entity_branch_link.emplace({ { "link_id", pdbx_entity_branch_link.get_unique_id("") },
+				{ "entity_id", entityID },
+				{ "entity_branch_list_num_1", s1.get_auth_seq_id() },
+				{ "comp_id_1", s1.get_compound_id() },
+				{ "atom_id_1", l1.get_label_atom_id() },
+				{ "leaving_atom_id_1", "O1" },
+				{ "entity_branch_list_num_2", s2.get_auth_seq_id() },
+				{ "comp_id_2", s2.get_compound_id() },
+				{ "atom_id_2", l2.get_label_atom_id() },
+				{ "leaving_atom_id_2", "H" + l2.get_label_atom_id() },
+				{ "value_order", "sing" } });
 		}
 	}
 
@@ -2725,12 +2751,26 @@ void structure::cleanup_empty_categories()
 		obsoleteEntities.push_back(entity);
 	}
 
+	auto validator = m_db.get_validator();
+
 	for (auto entity : obsoleteEntities)
+	{
+		std::string entityID = entity["id"].as<std::string>();
+		if (validator)
+		{
+			for (auto linked : validator->get_links_for_parent("entity"))
+			{
+				if (auto cat = m_db.get(linked->m_child_category))
+					cat->erase(cif::key(linked->m_child_keys.front()) == entityID);
+			}
+		}
+
 		entities.erase(entity);
+	}
 
 	// the rest?
 
-	for (const char *cat : {"pdbx_entity_nonpoly"})
+	for (const char *cat : { "pdbx_entity_nonpoly" })
 	{
 		auto &category = m_db[cat];
 
@@ -2836,4 +2876,89 @@ void structure::validate_atoms() const
 	assert(atoms.empty());
 }
 
-} // namespace pdbx
+static int compare_numbers(std::string_view a, std::string_view b)
+{
+	int result = 0;
+	double da, db;
+
+	using namespace cif;
+	using namespace std;
+
+	std::from_chars_result ra, rb;
+
+	ra = selected_charconv<double>::from_chars(a.data(), a.data() + a.length(), da);
+	rb = selected_charconv<double>::from_chars(b.data(), b.data() + b.length(), db);
+
+	if (not(bool) ra.ec and not(bool) rb.ec)
+	{
+		auto d = da - db;
+		if (std::abs(d) > std::numeric_limits<double>::epsilon())
+		{
+			if (d > 0)
+				result = 1;
+			else if (d < 0)
+				result = -1;
+		}
+	}
+	else if ((bool)ra.ec)
+		result = 1;
+	else
+		result = -1;
+
+	return result;
+}
+
+void structure::reorder_atoms()
+{
+	auto &atom_site = m_db["atom_site"];
+
+	atom_site.sort([](row_handle a, row_handle b)
+		{
+			int d;
+
+			// First by model number
+			d = a.get<int>("pdbx_PDB_model_num") - b.get<int>("pdbx_PDB_model_num");
+			if (d == 0)
+				d = a.get<std::string>("label_asym_id").compare(b.get<std::string>("label_asym_id"));
+			if (d == 0)
+			{
+				auto na = a.get<std::optional<int>>("label_seq_id");
+				auto nb = b.get<std::optional<int>>("label_seq_id");
+
+				if (na.has_value() and nb.has_value())
+					d = *na - *nb;
+				else if (na.has_value())
+					d = 1;
+				else if (nb.has_value())
+					d = -1;
+			}
+
+			if (d == 0)
+			{
+				auto na = a.get<std::optional<int>>("auth_seq_id");
+				auto nb = b.get<std::optional<int>>("auth_seq_id");
+
+				if (na.has_value() and nb.has_value())
+					d = *na - *nb;
+				else if (na.has_value())
+					d = 1;
+				else if (nb.has_value())
+					d = -1;
+			}
+
+			if (d == 0)
+				d = compare_numbers(a.get<std::string>("id"), b.get<std::string>("id"));
+
+			return d;
+			//
+		});
+
+	// atom_site.set_validator(nullptr, m_db);
+
+	// for (int nr = 1; auto r : atom_site)
+	// 	r["id"] = nr++;
+
+	// atom_site.set_validator(m_db.get_validator(), m_db);
+}
+
+} // namespace cif::mm

src/parser.cpp

@@ -837,6 +837,9 @@ void parser::produce_datablock(std::string_view name)
 
 	const auto &[iter, ignore] = m_file.emplace(name);
 	m_datablock = &(*iter);
+
+	if (m_validator)
+		m_datablock->set_validator(m_validator);
 }
 
 void parser::produce_category(std::string_view name)

src/pdb/pdb2cif.cpp

@@ -5909,7 +5909,8 @@ void PDBFileParser::Parse(std::istream &is, cif::file &result)
 {
 	try
 	{
-		mDatablock.set_validator(result.get_validator());
+		if (mDatablock.get_validator() == nullptr)
+			mDatablock.load_dictionary();
 
 		PreParseInput(is);
 
@@ -6373,10 +6374,11 @@ void read_pdb_file(std::istream &pdbFile, cif::file &cifFile)
 {
 	PDBFileParser p;
 
-	cifFile.load_dictionary("mmcif_pdbx.dic");
-
 	p.Parse(pdbFile, cifFile);
 
+	if (not cifFile.empty() and cifFile.front().get_validator() == nullptr)
+		cifFile.front().load_dictionary("mmcif_pdbx.dic");
+
 	if (not cifFile.is_valid() and cif::VERBOSE >= 0)
 		std::cerr << "Resulting mmCIF file is not valid!\n";
 }
@@ -6421,8 +6423,8 @@ file read(std::istream &is)
 	}
 
 	// Must be a PDB like file, right?
-	if (result.get_validator() == nullptr)
-		result.load_dictionary("mmcif_pdbx.dic");
+	if (not result.empty() and result.front().get_validator() == nullptr)
+		result.front().load_dictionary("mmcif_pdbx.dic");
 
 	return result;
 }

src/pdb/pdb2cif_remark_3.cpp

@@ -1480,6 +1480,9 @@ bool Remark3Parser::parse(const std::string &expMethod, PDBRecord *r, cif::datab
 
 		for (auto &cat1 : best.parser->mDb)
 		{
+			if (cat1.empty())
+				continue;
+
 			auto &cat2 = db[cat1.name()];
 
 			// copy only the values in the first row for the following categories

src/pdb/reconstruct.cpp

@@ -100,10 +100,35 @@ void checkEntities(datablock &db)
 
 	for (auto entity : db["entity"].find("formula_weight"_key == null or "formula_weight"_key == 0))
 	{
-		const auto &[entity_id, type] = entity.get<std::string, std::string>("id", "type");
+		auto &&[entity_id, type] = entity.get<std::string, std::string>("id", "type");
 
 		float formula_weight = 0;
 
+		if (type.empty())	// yes, that happens
+		{
+			const auto comp_id = db["atom_site"].find_first<std::string>("label_entity_id"_key == entity_id, "label_comp_id");
+			auto compound = cf.create(comp_id);
+			if (compound != nullptr)
+			{
+				if (compound->is_base() or compound->is_peptide())
+					type = "polymer";
+				else if (compound->is_water())
+					type = "water";
+				else
+				{
+					if (db["pdbx_entity_branch_link"].contains("entity_id"_key == entity_id))
+						type = "branched";
+					else
+						type = "non-polymer";
+				}
+			}
+
+			if (type.empty())
+				throw std::runtime_error("Entity without type and cannot determine what it should be");
+			
+			entity["type"] = type;
+		}
+		
 		if (type == "polymer")
 		{
 			int n = 0;
@@ -144,9 +169,11 @@ void checkEntities(datablock &db)
 			if (comp_id.has_value())
 			{
 				auto compound = cf.create(*comp_id);
-				assert(compound);
 				if (not compound)
-					throw std::runtime_error("missing information for compound " + *comp_id);
+				{
+					std::cerr << "missing information for compound " << *comp_id << "\n";
+					continue;
+				}
 				formula_weight = compound->formula_weight();
 			}
 		}
@@ -416,6 +443,8 @@ void checkAtomRecords(datablock &db)
 	for (int id : db["entity"].find<int>("type"_key == "polymer", "id"))
 		polymer_entities.insert(id);
 
+	std::set<std::string> missingCompounds;
+
 	for (auto row : atom_site)
 	{
 		residue_key_type k = row.get<std::optional<std::string>,
@@ -446,11 +475,18 @@ void checkAtomRecords(datablock &db)
 		std::string asym_id = get_asym_id(k);
 		std::string comp_id = get_comp_id(k);
 
+		if (missingCompounds.contains(comp_id))
+			continue;
+
 		bool is_polymer = polymer_entities.contains(row["label_entity_id"].as<int>());
 		auto compound = cf.create(comp_id);
 
 		if (not compound)
-			throw std::runtime_error("Missing compound information for " + comp_id);
+		{
+			missingCompounds.insert(comp_id);
+			std::cerr << "Missing compound information for " << comp_id << "\n";
+			continue;
+		}
 
 		auto chem_comp_entry = chem_comp.find_first("id"_key == comp_id);
 
@@ -590,18 +626,18 @@ void checkAtomAnisotropRecords(datablock &db)
 			row["type_symbol"] = parent["type_symbol"].text();
 		}
 
-		if (row["pdbx_auth_alt_id"].empty())
+		if (row["pdbx_auth_alt_id"].empty() and not parent["pdbx_auth_alt_id"].empty())
 			row["pdbx_auth_alt_id"] = parent["pdbx_auth_alt_id"].text();
-		if (row["pdbx_label_seq_id"].empty())
+		if (row["pdbx_label_seq_id"].empty() and not parent["pdbx_label_seq_id"].empty())
 			row["pdbx_label_seq_id"] = parent["label_seq_id"].text();
-		if (row["pdbx_label_asym_id"].empty())
+		if (row["pdbx_label_asym_id"].empty() and not parent["pdbx_label_asym_id"].empty())
 			row["pdbx_label_asym_id"] = parent["label_asym_id"].text();
-		if (row["pdbx_label_atom_id"].empty())
+		if (row["pdbx_label_atom_id"].empty() and not parent["pdbx_label_atom_id"].empty())
 			row["pdbx_label_atom_id"] = parent["label_atom_id"].text();
-		if (row["pdbx_label_comp_id"].empty())
+		if (row["pdbx_label_comp_id"].empty() and not parent["pdbx_label_comp_id"].empty())
 			row["pdbx_label_comp_id"] = parent["label_comp_id"].text();
-		if (row["pdbx_PDB_model_num"].empty())
-			row["pdbx_PDB_model_num"] = parent["pdbx_PDB_model_num"].text();
+		// if (row["pdbx_PDB_model_num"].empty() and not parent["pdbx_PDB_model_num"].empty())
+		// 	row["pdbx_PDB_model_num"] = parent["pdbx_PDB_model_num"].text();
 	}
 
 	if (not to_be_deleted.empty())
@@ -811,6 +847,18 @@ void createEntityPoly(datablock &db)
 
 					non_std_monomer = true;
 				}
+				else
+				{
+					// c_type = "other";
+
+					letter_can = c->one_letter_code();
+					if (letter_can == 0)
+						letter_can = 'X';
+
+					letter = '(' + comp_id + ')';
+
+					non_std_monomer = true;
+				}
 
 				if (type.empty())
 					type = c_type;
@@ -877,7 +925,7 @@ void createEntityPoly(datablock &db)
 
 void createEntityPolySeq(datablock &db)
 {
-	if (db.get("entity_poly") == nullptr)
+	if (auto cat = db.get("entity_poly"); cat == nullptr or cat->empty())
 		createEntityPoly(db);
 
 	using namespace literals;
@@ -928,7 +976,10 @@ void createEntityPolySeq(datablock &db)
 
 void createPdbxPolySeqScheme(datablock &db)
 {
-	if (db.get("entity_poly_seq") == nullptr)
+	if (auto cat = db.get("entity_poly"); cat == nullptr or cat->empty())
+		createEntityPoly(db);
+
+	if (auto cat = db.get("entity_poly_seq"); cat == nullptr or cat->empty())
 		createEntityPolySeq(db);
 
 	using namespace literals;
@@ -986,6 +1037,10 @@ void comparePolySeqSchemes(datablock &db)
 	auto &ndb_poly_seq_scheme = db["ndb_poly_seq_scheme"];
 	auto &pdbx_poly_seq_scheme = db["pdbx_poly_seq_scheme"];
 
+	// Don't bother if ndb_poly_seq_scheme is empty
+	if (ndb_poly_seq_scheme.empty())
+		return;
+
 	// Since often ndb_poly_seq_scheme only contains an id and mon_id item
 	// we assume that it should match the accompanying pdbx_poly_seq
 
@@ -1065,7 +1120,7 @@ bool reconstruct_pdbx(file &file, std::string_view dictionary)
 	// ... and any additional datablock will contain compound information
 	cif::compound_source cs(file);
 
-	if (db.get("atom_site") == nullptr)
+	if (auto cat = db.get("atom_site"); cat == nullptr or cat->empty())
 		throw std::runtime_error("Cannot reconstruct PDBx file, atom data missing");
 
 	auto &validator = validator_factory::instance()[dictionary];
@@ -1073,7 +1128,7 @@ bool reconstruct_pdbx(file &file, std::string_view dictionary)
 	std::string entry_id;
 
 	// Phenix files do not have an entry record
-	if (db.get("entry") == nullptr)
+	if (auto cat = db.get("entry"); cat == nullptr or cat->empty())
 	{
 		entry_id = db.name();
 		category entry("entry");
@@ -1320,26 +1375,26 @@ bool reconstruct_pdbx(file &file, std::string_view dictionary)
 
 	db["chem_comp"].reorder_by_index();
 
-	file.load_dictionary(dictionary);
+	db.load_dictionary(dictionary);
 
 	if (db.get("atom_site_anisotrop"))
 		checkAtomAnisotropRecords(db);
 
 	// Now create any missing categories
 	// Next make sure we have struct_asym records
-	if (db.get("struct_asym") == nullptr)
+	if (auto cat = db.get("struct_asym"); cat == nullptr or cat->empty())
 		createStructAsym(db);
 
-	if (db.get("entity") == nullptr)
+	if (auto cat = db.get("entity"); cat == nullptr or cat->empty())
 		createEntity(db);
 
 	// fill in missing formula_weight, e.g.
 	checkEntities(db);
 
-	if (db.get("pdbx_poly_seq_scheme") == nullptr)
+	if (auto cat = db.get("pdbx_poly_seq_scheme"); cat == nullptr or cat->empty())
 		createPdbxPolySeqScheme(db);
 
-	if (db.get("ndb_poly_seq_scheme") != nullptr)
+	if (auto cat = db.get("ndb_poly_seq_scheme"); cat == nullptr or cat->empty())
 		comparePolySeqSchemes(db);
 
 	// skip unknown categories for now

src/pdb/validate-pdbx.cpp

@@ -29,7 +29,7 @@
 namespace cif::pdb
 {
 
-condition get_parents_condition(const validator &validator, row_handle rh, const category &parentCat)
+condition get_parents_condition(const validator_base &validator, row_handle rh, const category &parentCat)
 {
 	condition result;
 
@@ -189,6 +189,7 @@ bool is_valid_pdbx_file(const file &file, std::string_view dictionary, std::erro
 				for (auto asym_id : struct_asym.find<std::string>("entity_id"_key == entity_id, "id"))
 				{
 					if (pdbx_poly_seq_scheme.count(
+							"entity_id"_key == entity_id and
 							"asym_id"_key == asym_id and
 							"mon_id"_key == mon_id and
 							"seq_id"_key == num and
@@ -202,6 +203,7 @@ bool is_valid_pdbx_file(const file &file, std::string_view dictionary, std::erro
 			for (const auto &[seq_id, mon_id, hetero] : pdbx_poly_seq_scheme.find<int, std::string, bool>("entity_id"_key == entity_id, "seq_id", "mon_id", "hetero"))
 			{
 				if (entity_poly_seq.count(
+						"entity_id"_key == entity_id and
 						"mon_id"_key == mon_id and
 						"num"_key == seq_id and
 						"hetero"_key == hetero) != 1)
@@ -304,8 +306,8 @@ bool is_valid_pdbx_file(const file &file, std::string_view dictionary, std::erro
 
 			if (not seq_can.has_value())
 			{
-				if (cif::VERBOSE > 0)
-					std::clog << "Warning: entity_poly has no sequence for entity_id " << entity_id << '\n';
+				if (cif::VERBOSE > 1)
+					std::clog << "Warning: entity_poly has no canonical sequence for entity_id " << entity_id << '\n';
 			}
 			else
 			{

src/symmetry.cpp

@@ -32,7 +32,7 @@
 
 #include "symop_table_data.hpp"
 
-#include <Eigen/Eigenvalues>
+#include <Eigen/Eigen>
 
 namespace cif
 {
@@ -103,9 +103,9 @@ sym_op::sym_op(std::string_view s)
 	auto b = s.data();
 	auto e = b + s.length();
 
-	int rnri = 256;	// default to unexisting number
+	int rnri = 256; // default to unexisting number
 	auto r = std::from_chars(b, e, rnri);
-	
+
 	m_nr = static_cast<uint8_t>(rnri);
 	m_ta = r.ptr[1] - '0';
 	m_tb = r.ptr[2] - '0';
@@ -121,7 +121,7 @@ std::string sym_op::string() const
 	auto r = std::to_chars(b, b + sizeof(b), m_nr);
 	if ((bool)r.ec or r.ptr > b + 4)
 		throw std::runtime_error("Could not write out symmetry operation to string");
-	
+
 	*r.ptr++ = '_';
 	*r.ptr++ = '0' + m_ta;
 	*r.ptr++ = '0' + m_tb;
@@ -163,41 +163,16 @@ transformation::transformation(const matrix3x3<float> &r, const cif::point &t)
 
 void transformation::try_create_quaternion()
 {
-	float Qxx = m_rotation(0, 0);
-	float Qxy = m_rotation(0, 1);
-	float Qxz = m_rotation(0, 2);
-	float Qyx = m_rotation(1, 0);
-	float Qyy = m_rotation(1, 1);
-	float Qyz = m_rotation(1, 2);
-	float Qzx = m_rotation(2, 0);
-	float Qzy = m_rotation(2, 1);
-	float Qzz = m_rotation(2, 2);
+	Eigen::Matrix3f rot;
 
-	Eigen::Matrix4f em;
+	rot << m_rotation(0, 0), m_rotation(0, 1), m_rotation(0, 2),
+		m_rotation(1, 0), m_rotation(1, 1), m_rotation(1, 2),
+		m_rotation(2, 0), m_rotation(2, 1), m_rotation(2, 2);
 
-	em << Qxx - Qyy - Qzz, Qyx + Qxy, Qzx + Qxz, Qzy - Qyz,
-			Qyx + Qxy, Qyy - Qxx - Qzz, Qzy + Qyz, Qxz - Qzx,
-			Qzx + Qxz, Qzy + Qyz, Qzz - Qxx - Qyy, Qyx - Qxy,
-			Qzy - Qyz, Qxz - Qzx, Qyx - Qxy, Qxx + Qyy + Qzz;
-
-	Eigen::EigenSolver<Eigen::Matrix4f> es(em / 3);
-
-	auto ev = es.eigenvalues();
-
-	for (std::size_t j = 0; j < 4; ++j)
+	if (rot * rot.transpose() == Eigen::Matrix3f::Identity() and rot.determinant() == 1)
 	{
-		if (std::abs(ev[j].real() - 1) > 0.01)
-			continue;
-		
-		auto col = es.eigenvectors().col(j);
-
-		m_q = normalize(cif::quaternion{
-			static_cast<float>(col(3).real()),
-			static_cast<float>(col(0).real()),
-			static_cast<float>(col(1).real()),
-			static_cast<float>(col(2).real()) });
-
-		break;
+		Eigen::Quaternionf qe(rot);
+		m_q = normalize(cif::quaternion{ qe.w(), qe.x(), qe.y(), qe.z() });
 	}
 }
 
@@ -297,7 +272,7 @@ point spacegroup::operator()(const point &pt, const cell &c, sym_op symop) const
 {
 	if (symop.m_nr < 1 or symop.m_nr > size())
 		throw std::out_of_range("symmetry operator number out of range");
-	
+
 	transformation t = at(symop.m_nr - 1);
 
 	t.m_translation.m_x += symop.m_ta - 5;
@@ -316,7 +291,7 @@ point spacegroup::inverse(const point &pt, const cell &c, sym_op symop) const
 {
 	if (symop.m_nr < 1 or symop.m_nr > size())
 		throw std::out_of_range("symmetry operator number out of range");
-	
+
 	transformation t = at(symop.m_nr - 1);
 
 	t.m_translation.m_x += symop.m_ta - 5;
@@ -450,13 +425,13 @@ int get_space_group_number(const datablock &db)
 
 	if (_symmetry.size() != 1)
 		throw std::runtime_error("Could not find a unique symmetry in this mmCIF file");
-	
+
 	return _symmetry.front().get<int>("Int_Tables_number");
 }
 
 // --------------------------------------------------------------------
 
-std::tuple<float,point,sym_op> crystal::closest_symmetry_copy(point a, point b) const
+std::tuple<float, point, sym_op> crystal::closest_symmetry_copy(point a, point b) const
 {
 	if (m_cell.get_a() == 0 or m_cell.get_b() == 0 or m_cell.get_c() == 0)
 		throw std::runtime_error("Invalid cell, contains a dimension that is zero");
@@ -491,7 +466,7 @@ std::tuple<float,point,sym_op> crystal::closest_symmetry_copy(point a, point b)
 		while (fsb.m_x + 0.5f < fa.m_x)
 		{
 			fsb.m_x += 1;
-			s.m_ta += 1;			
+			s.m_ta += 1;
 		}
 
 		while (fsb.m_y - 0.5f > fa.m_y)
@@ -503,7 +478,7 @@ std::tuple<float,point,sym_op> crystal::closest_symmetry_copy(point a, point b)
 		while (fsb.m_y + 0.5f < fa.m_y)
 		{
 			fsb.m_y += 1;
-			s.m_tb += 1;			
+			s.m_tb += 1;
 		}
 
 		while (fsb.m_z - 0.5f > fa.m_z)
@@ -515,7 +490,7 @@ std::tuple<float,point,sym_op> crystal::closest_symmetry_copy(point a, point b)
 		while (fsb.m_z + 0.5f < fa.m_z)
 		{
 			fsb.m_z += 1;
-			s.m_tc += 1;			
+			s.m_tc += 1;
 		}
 
 		auto p = orthogonal(fsb, m_cell);

src/text.cpp

@@ -69,7 +69,7 @@ bool iequals(const char *a, const char *b)
 {
 	bool result = true;
 	for (; result and *a and *b; ++a, ++b)
-		result = tolower(*a) == tolower(*b);
+		result = kCharToLowerMap[uint8_t(*a)] == kCharToLowerMap[uint8_t(*b)];
 
 	return result and *a == *b;
 }
@@ -80,7 +80,7 @@ int icompare(std::string_view a, std::string_view b)
 	auto ai = a.begin(), bi = b.begin();
 
 	for (; d == 0 and ai != a.end() and bi != b.end(); ++ai, ++bi)
-		d = tolower(*ai) - tolower(*bi);
+		d = (int)kCharToLowerMap[uint8_t(*ai)] - (int)kCharToLowerMap[uint8_t(*bi)];
 
 	if (d == 0)
 	{
@@ -98,7 +98,7 @@ int icompare(const char *a, const char *b)
 	int d = 0;
 
 	for (; d == 0 and *a != 0 and *b != 0; ++a, ++b)
-		d = tolower(*a) - tolower(*b);
+		d = (int)kCharToLowerMap[uint8_t(*a)] - (int)kCharToLowerMap[uint8_t(*b)];
 
 	if (d == 0)
 	{

src/utilities.cpp

@@ -63,9 +63,9 @@ std::string get_version_nr()
 
 // --------------------------------------------------------------------
 
-#ifdef _WIN32
+#if defined(_WIN32) or defined(__MINGW32__)
 }
-#include <Windows.h>
+#include <windows.h>
 #include <libloaderapi.h>
 #include <wincon.h>
 
@@ -235,7 +235,7 @@ void progress_bar_impl::print_progress()
 	float progress = static_cast<float>(m_consumed) / m_max_value;
 	
 	if (width < kMinBarWidth)
-		std::cout << (100 * progress) << '%' << std::endl;
+		std::cout << (100 * progress) << "%\n";
 	else
 	{
 		uint32_t bar_width = 7 * width / 10;
@@ -329,7 +329,7 @@ void progress_bar_impl::print_done()
 	if (msg.length() < width)
 		msg += std::string(width - msg.length(), ' ');
 
-	std::cout << '\r' << msg << std::endl;
+	std::cout << '\r' << msg << '\n';
 }
 
 progress_bar::progress_bar(int64_t inMax, const std::string &inAction)

src/validate.cpp

@@ -25,6 +25,7 @@
  */
 
 #include "cif++/validate.hpp"
+#include "cif++/category.hpp"
 #include "cif++/dictionary_parser.hpp"
 #include "cif++/gzio.hpp"
 #include "cif++/utilities.hpp"
@@ -138,7 +139,7 @@ int type_validator::compare(std::string_view a, std::string_view b) const
 				ra = selected_charconv<double>::from_chars(a.data(), a.data() + a.length(), da);
 				rb = selected_charconv<double>::from_chars(b.data(), b.data() + b.length(), db);
 
-				if (not (bool)ra.ec and not (bool)rb.ec)
+				if (not(bool) ra.ec and not(bool) rb.ec)
 				{
 					auto d = da - db;
 					if (std::abs(d) > std::numeric_limits<double>::epsilon())
@@ -232,7 +233,7 @@ bool item_validator::validate_value(std::string_view value, std::error_code &ec)
 			ec = make_error_code(validation_error::value_is_not_in_enumeration_list);
 	}
 
-	return not (bool)ec;
+	return not(bool) ec;
 }
 
 // --------------------------------------------------------------------
@@ -284,6 +285,27 @@ const item_validator *category_validator::get_validator_for_aliased_item(std::st
 
 // --------------------------------------------------------------------
 
+void validator_base::report_error(std::error_code ec, bool fatal) const
+{
+	if (m_strict or fatal)
+		throw validation_exception(ec);
+	else if (VERBOSE > 0)
+		std::cerr << ec.message() << '\n';
+}
+
+void validator_base::report_error(std::error_code ec, std::string_view category,
+	std::string_view item, bool fatal) const
+{
+	auto ex = item.empty() ? validation_exception(ec, category) : validation_exception(ec, category, item);
+
+	if (m_strict or fatal)
+		throw ex;
+	else if (VERBOSE > 0)
+		std::cerr << ex.what() << '\n';
+}
+
+// --------------------------------------------------------------------
+
 void validator::add_type_validator(type_validator &&v)
 {
 	auto r = m_type_validators.insert(std::move(v));
@@ -397,25 +419,76 @@ std::vector<const link_validator *> validator::get_links_for_child(std::string_v
 	return result;
 }
 
-void validator::report_error(std::error_code ec, bool fatal) const
+// --------------------------------------------------------------------
+
+extended_validator::extended_validator(std::vector<const validator *> validators)
+	: m_validators(validators)
 {
-	if (m_strict or fatal)
-		throw validation_exception(ec);
-	else
-		std::cerr << ec.message() << '\n';
+	std::vector<std::string> names, versions;
+
+	for (auto v : m_validators)
+	{
+		names.emplace_back(v->name());
+		versions.emplace_back(v->version());
+		m_strict = m_strict or v->is_strict();
+	}
+
+	m_name = cif::join(names, "; ");
+	m_version = cif::join(versions, "; ");
 }
 
-void validator::report_error(std::error_code ec, std::string_view category,
-	std::string_view item, bool fatal) const
+const type_validator *extended_validator::get_validator_for_type(std::string_view type_code) const
 {
-	auto ex = item.empty() ?
-		validation_exception(ec, category) :
-		validation_exception(ec, category, item);
+	const type_validator *result = nullptr;
 
-	if (m_strict or fatal)
-		throw ex;
-	else
-		std::cerr << ex.what() << '\n';
+	for (auto v : m_validators)
+	{
+		result = v->get_validator_for_type(type_code);
+		if (result)
+			break;
+	}
+
+	return result;
+}
+
+const category_validator *extended_validator::get_validator_for_category(std::string_view category) const
+{
+	const category_validator *result = nullptr;
+
+	for (auto v : m_validators)
+	{
+		result = v->get_validator_for_category(category);
+		if (result)
+			break;
+	}
+
+	return result;
+}
+
+std::vector<const link_validator *> extended_validator::get_links_for_parent(std::string_view category) const
+{
+	std::vector<const link_validator *> result;
+
+	for (auto v : m_validators)
+	{
+		auto links = v->get_links_for_parent(category);
+		result.insert(result.end(), links.begin(), links.end());
+	}
+
+	return result;
+}
+
+std::vector<const link_validator *> extended_validator::get_links_for_child(std::string_view category) const
+{
+	std::vector<const link_validator *> result;
+
+	for (auto v : m_validators)
+	{
+		auto links = v->get_links_for_child(category);
+		result.insert(result.end(), links.begin(), links.end());
+	}
+
+	return result;
 }
 
 // --------------------------------------------------------------------
@@ -426,101 +499,172 @@ validator_factory &validator_factory::instance()
 	return s_instance;
 }
 
-const validator &validator_factory::operator[](std::string_view dictionary_name)
+const validator_base &validator_factory::operator[](std::string_view dictionary_name)
 {
-	try
-	{
-		std::lock_guard lock(m_mutex);
+	category audit_conform("audit_conform");
+	for (auto part : cif::split(dictionary_name, ";", true))
+		audit_conform.emplace({ { "dict_name", part } });
 
-		for (auto &validator : m_validators)
-		{
-			if (iequals(validator.name(), dictionary_name))
-				return validator;
-		}
-
-		// not found, try to see if it helps if we tweak the name a little
-
-		// too bad clang version 10 did not have a constructor for std::filesystem::path that accepts a std::string_view
-		std::filesystem::path dictionary(dictionary_name.data(), dictionary_name.data() + dictionary_name.length());
-
-		if (dictionary.extension() != ".dic")
-		{
-			auto dict_name = dictionary.filename().string() + ".dic";
-
-			for (auto &validator : m_validators)
-			{
-				if (iequals(validator.name(), dict_name))
-					return validator;
-			}
-		}
-
-		// not found, add it
-		auto data = load_resource(dictionary_name);
-
-		if (not data and dictionary.extension().string() != ".dic")
-			data = load_resource(dictionary.parent_path() / (dictionary.filename().string() + ".dic"));
-
-		if (data)
-			construct_validator(dictionary_name, *data);
-		else
-		{
-			std::error_code ec;
-
-			// might be a compressed dictionary on disk
-			std::filesystem::path p = dictionary;
-			if (p.extension() == ".dic")
-				p = p.parent_path() / (p.filename().string() + ".gz");
-			else
-				p = p.parent_path() / (p.filename().string() + ".dic.gz");
-
-#if defined(CACHE_DIR) or defined(DATA_DIR)
-			if (not std::filesystem::exists(p, ec) or ec)
-			{
-				for (const char *dir : {
-# if defined(CACHE_DIR)
-						 CACHE_DIR,
-# endif
-# if defined(DATA_DIR)
-							 DATA_DIR
-# endif
-					 })
-				{
-					auto p2 = std::filesystem::path(dir) / p;
-					if (std::filesystem::exists(p2, ec) and not ec)
-					{
-						swap(p, p2);
-						break;
-					}
-				}
-			}
-#endif
-
-			if (std::filesystem::exists(p, ec) and not ec)
-			{
-				gzio::ifstream in(p);
-
-				if (not in.is_open())
-					throw std::runtime_error("Could not open dictionary (" + p.string() + ")");
-
-				construct_validator(dictionary_name, in);
-			}
-			else
-				throw std::runtime_error("Dictionary not found or defined (" + dictionary.string() + ")");
-		}
-
-		return m_validators.back();
-	}
-	catch (const std::exception &ex)
-	{
-		std::string msg = "Error while loading dictionary ";
-		msg += dictionary_name;
-		std::throw_with_nested(std::runtime_error(msg));
-	}
+	return construct_validator(audit_conform);
 }
 
-const validator &validator_factory::construct_validator(std::string_view name, std::istream &is)
+const validator_base &validator_factory::construct_validator(const category &audit_conform)
 {
-	return m_validators.emplace_back(parse_dictionary(name, is));
+	if (audit_conform.empty())
+		throw std::runtime_error("Empty audit_conform category, cannot create a validator");
+
+	std::lock_guard lock(m_mutex);
+
+	std::vector<const validator *> validators;
+
+	for (const auto &[name, version] : audit_conform.rows<std::string, std::optional<std::string>>("dict_name", "dict_version"))
+	{
+		auto &v = construct_validator(name, version);
+		validators.emplace_back(&v);
+	}
+
+	if (validators.size() == 1)
+		return *validators.front();
+
+	// override mode, last dictionary is most important
+	std::reverse(validators.begin(), validators.end());
+		
+	for (auto &ev : m_extended_validators)
+	{
+		if (ev.m_validators == validators)
+			return ev;
+	}
+
+	return m_extended_validators.emplace_back(validators);
+}
+
+const validator &validator_factory::construct_validator(std::string_view name,
+	std::optional<std::string> version)
+{
+	for (auto &v : m_validators)
+	{
+		if (version.has_value())
+			check_version(name, *version, v.version());
+
+		if (v.name() == name)
+			return v;
+	}
+
+	std::filesystem::path dictionary(name);
+
+	auto data = load_resource(name);
+
+	if (not data and dictionary.extension().string() != ".dic")
+		data = load_resource(dictionary.parent_path() / (dictionary.filename().string() + ".dic"));
+
+	if (not data)
+	{
+		std::error_code ec;
+
+		// might be a compressed dictionary on disk
+		std::filesystem::path p = dictionary;
+		if (p.extension() == ".dic")
+			p = p.parent_path() / (p.filename().string() + ".gz");
+		else
+			p = p.parent_path() / (p.filename().string() + ".dic.gz");
+
+#if defined(CACHE_DIR) or defined(DATA_DIR)
+		if (not std::filesystem::exists(p, ec) or ec)
+		{
+			for (const char *dir : {
+# if defined(CACHE_DIR)
+					 CACHE_DIR,
+# endif
+# if defined(DATA_DIR)
+					 DATA_DIR
+# endif
+				 })
+			{
+				auto p2 = std::filesystem::path(dir) / p;
+				if (std::filesystem::exists(p2, ec) and not ec)
+				{
+					swap(p, p2);
+					break;
+				}
+			}
+		}
+#endif
+
+		if (std::filesystem::exists(p, ec) and not ec)
+		{
+			auto in = std::make_unique<gzio::ifstream>(p);
+
+			if (not in->is_open())
+				throw std::runtime_error("Could not open dictionary (" + p.string() + ")");
+
+			data.reset(in.release());
+		}
+		else
+			throw std::runtime_error("Dictionary not found or defined (" + dictionary.string() + ")");
+	}
+
+	return construct_validator(name, version, *data);
+}
+
+const validator &validator_factory::construct_validator(std::string_view name,
+	std::optional<std::string> version, std::istream &is)
+{
+	auto v = parse_dictionary(name, is);
+
+	if (version.has_value() and VERBOSE >= 0)
+	{
+		auto vv = v.version();
+
+		if (vv.empty())
+			std::clog << "Could not check version of dictionary " << name << " since this info is missing\n";
+		else
+			check_version(name, *version, vv);
+	}
+
+	return m_validators.emplace_back(std::move(v));
+}
+
+bool validator_factory::check_version(std::string_view name, std::string_view expected, std::string_view found)
+{
+	bool result = true;
+	auto el = cif::split(expected, ".");
+	auto fl = cif::split(found, ".");
+
+	auto eli = el.begin();
+	auto fli = fl.begin();
+
+	while (eli != el.end() and fli != fl.end())
+	{
+		int e_int, f_int;
+		if (auto [ptr, ec] = std::from_chars(eli->begin(), eli->end(), e_int); ec != std::errc{})
+		{
+			std::clog << "Could not parse requested version string for dictionary " << std::quoted(expected) << "\n";
+			result = false;
+			break;
+		}
+
+		if (auto [ptr, ec] = std::from_chars(fli->begin(), fli->end(), f_int); ec != std::errc{})
+		{
+			std::clog << "Could not parse version string in dictionary " << name << " " << std::quoted(found) << "\n";
+			result = false;
+			break;
+		}
+
+		if (f_int > e_int) // newer version, assume this is ok
+			break;
+
+		if (f_int < e_int)
+		{
+			std::clog << "The version in dictionary " << name << " is lower than requested, this may cause validation errors\n";
+			result = false;
+			break;
+		}
+
+		++eli;
+		++fli;
+	}
+
+	return result;
 }
 
 } // namespace cif

test/CMakeLists.txt

@@ -1,16 +1,20 @@
 # We're using the older version 2 of Catch2
 
-find_package(Catch2 QUIET)
+if(NOT(Catch2_FOUND OR TARGET Catch2))
+	find_package(Catch2 QUIET)
 
-if(NOT Catch2_FOUND)
-	FetchContent_Declare(
-		Catch2
-		GIT_REPOSITORY https://github.com/catchorg/Catch2.git
-		GIT_TAG v2.13.9)
+	if(NOT Catch2_FOUND)
+		include(FetchContent)
 
-	FetchContent_MakeAvailable(Catch2)
+		FetchContent_Declare(
+			Catch2
+			GIT_REPOSITORY https://github.com/catchorg/Catch2.git
+			GIT_TAG v2.13.9)
 
-	set(Catch2_VERSION "2.13.9")
+		FetchContent_MakeAvailable(Catch2)
+
+		set(Catch2_VERSION "2.13.9")
+	endif()
 endif()
 
 list(
@@ -30,10 +34,10 @@ add_library(test-main OBJECT "${CMAKE_CURRENT_SOURCE_DIR}/test-main.cpp")
 
 target_link_libraries(test-main cifpp::cifpp Catch2::Catch2)
 
-if(${Catch2_VERSION} VERSION_GREATER_EQUAL 3.0.0)
-	target_compile_definitions(test-main PUBLIC CATCH22=0)
-else()
+if("${Catch2_VERSION}" VERSION_LESS 3.0.0)
 	target_compile_definitions(test-main PUBLIC CATCH22=1)
+else()
+	target_compile_definitions(test-main PUBLIC CATCH22=0)
 endif()
 
 foreach(CIFPP_TEST IN LISTS CIFPP_tests)
@@ -49,8 +53,7 @@ foreach(CIFPP_TEST IN LISTS CIFPP_tests)
 		target_compile_definitions(${CIFPP_TEST} PUBLIC CATCH22=1)
 	endif()
 
-	target_link_libraries(${CIFPP_TEST} PRIVATE Threads::Threads cifpp::cifpp
-		Catch2::Catch2)
+	target_link_libraries(${CIFPP_TEST} PRIVATE cifpp::cifpp Catch2::Catch2)
 	target_include_directories(${CIFPP_TEST} PRIVATE "${EIGEN_INCLUDE_DIR}")
 
 	if(MSVC)

test/model-test.cpp

@@ -53,8 +53,8 @@ TEST_CASE("create_nonpoly_1")
 	cif::VERBOSE = 1;
 
 	cif::file file;
-	file.load_dictionary("mmcif_pdbx.dic");
-	file.emplace("TEST"); // create a datablock
+	auto &&[dbi, ignore] = file.emplace("TEST"); // create a datablock
+	dbi->load_dictionary("mmcif_pdbx.dic");
 
 	cif::mm::structure structure(file);
 
@@ -82,7 +82,7 @@ _atom_site.pdbx_formal_charge
 # that's enough to test with
 )"_cf;
 
-	atoms.load_dictionary("mmcif_pdbx.dic");
+	atoms.front().load_dictionary("mmcif_pdbx.dic");
 
 	auto &hem_data = atoms["HEM"];
 	auto &atom_site = hem_data["atom_site"];
@@ -159,14 +159,14 @@ _struct_asym.details                       ?
 _atom_type.symbol   C
 )"_cf;
 
-	expected.load_dictionary("mmcif_pdbx.dic");
+	expected.front().load_dictionary("mmcif_pdbx.dic");
 
 	if (not(expected.front() == structure.get_datablock()))
 	{
-		REQUIRE(false);
-		std::cout << expected.front() << '\n'
+		std::cerr << expected.front() << '\n'
 				  << '\n'
 				  << structure.get_datablock() << '\n';
+		REQUIRE(false);
 	}
 }
 
@@ -177,8 +177,8 @@ TEST_CASE("create_nonpoly_2")
 	cif::VERBOSE = 1;
 
 	cif::file file;
-	file.load_dictionary("mmcif_pdbx.dic");
-	file.emplace("TEST"); // create a datablock
+	auto &&[dbi, ignore] = file.emplace("TEST"); // create a datablock
+	dbi->load_dictionary("mmcif_pdbx.dic");
 
 	cif::mm::structure structure(file);
 
@@ -270,7 +270,7 @@ _struct_asym.details                       ?
 _atom_type.symbol   C
 )"_cf;
 
-	expected.load_dictionary("mmcif_pdbx.dic");
+	expected.front().load_dictionary("mmcif_pdbx.dic");
 
 	REQUIRE(expected.front() == structure.get_datablock());
 
@@ -354,7 +354,7 @@ _struct_asym.details                       ?
 #
 )"_cf;
 
-	data.load_dictionary("mmcif_pdbx.dic");
+	data.front().load_dictionary("mmcif_pdbx.dic");
 
 	cif::mm::structure s(data);
 

test/unit-3d-test.cpp

@@ -148,39 +148,37 @@ TEST_CASE("dh_q_0")
 	cif::point axis(1, 0, 0);
 
 	cif::point p(1, 1, 0);
-	
-	cif::point t[3] =
-	{
+
+	cif::point t[3] = {
 		{ 0, 1, 0 },
 		{ 0, 0, 0 },
 		{ 1, 0, 0 }
 	};
 
 	auto a = cif::dihedral_angle(t[0], t[1], t[2], p);
-	REQUIRE_THAT(a, Catch::Matchers::WithinRel(0, 0.01f));
+	REQUIRE_THAT(a, Catch::Matchers::WithinRel(0.f, 0.01f));
 
 	auto q = cif::construct_from_angle_axis(90, axis);
 
 	p.rotate(q);
 
-	REQUIRE_THAT(p.m_x, Catch::Matchers::WithinRel(1, 0.01f));
-	REQUIRE_THAT(p.m_y, Catch::Matchers::WithinRel(0, 0.01f));
-	REQUIRE_THAT(p.m_z, Catch::Matchers::WithinRel(1, 0.01f));
+	REQUIRE(std::abs(p.m_x - 1.f) < 0.01f);
+	REQUIRE(std::abs(p.m_y - 0.f) < 0.01f);
+	REQUIRE(std::abs(p.m_z - 1.f) < 0.01f);
 
 	a = cif::dihedral_angle(t[0], t[1], t[2], p);
-	REQUIRE_THAT(a, Catch::Matchers::WithinRel(90, 0.01f));
+	REQUIRE(std::abs(a - 90.f) < 0.01f);
 
 	q = cif::construct_from_angle_axis(-90, axis);
 
 	p.rotate(q);
 
-	REQUIRE_THAT(p.m_x, Catch::Matchers::WithinRel(1, 0.01f));
-	REQUIRE_THAT(p.m_y, Catch::Matchers::WithinRel(1, 0.01f));
-	REQUIRE_THAT(p.m_z, Catch::Matchers::WithinRel(0, 0.01f));
+	REQUIRE(std::abs(p.m_x - 1.f) < 0.01f);
+	REQUIRE(std::abs(p.m_y - 1.f) < 0.01f);
+	REQUIRE(std::abs(p.m_z - 0.f) < 0.01f);
 
 	a = cif::dihedral_angle(t[0], t[1], t[2], p);
-	REQUIRE_THAT(a, Catch::Matchers::WithinRel(0, 0.01f));
-
+	REQUIRE(std::abs(a - 0.f) < 0.01f);
 }
 
 TEST_CASE("dh_q_1")
@@ -228,62 +226,103 @@ TEST_CASE("dh_q_1")
 
 // --------------------------------------------------------------------
 
-TEST_CASE("m2q_0, *utf::tolerance(0.001f)")
+// TEST_CASE("m2q_0")
+// {
+// 	for (std::size_t i = 0; i < cif::kSymopNrTableSize; ++i)
+// 	{
+// 		auto d = cif::kSymopNrTable[i].symop().data();
+
+// 		cif::matrix3x3<float> rot;
+// 		float Qxx = rot(0, 0) = d[0];
+// 		float Qxy = rot(0, 1) = d[1];
+// 		float Qxz = rot(0, 2) = d[2];
+// 		float Qyx = rot(1, 0) = d[3];
+// 		float Qyy = rot(1, 1) = d[4];
+// 		float Qyz = rot(1, 2) = d[5];
+// 		float Qzx = rot(2, 0) = d[6];
+// 		float Qzy = rot(2, 1) = d[7];
+// 		float Qzz = rot(2, 2) = d[8];
+
+// 		Eigen::Matrix4f em;
+
+// 		em << Qxx - Qyy - Qzz, Qyx + Qxy, Qzx + Qxz, Qzy - Qyz,
+// 			Qyx + Qxy, Qyy - Qxx - Qzz, Qzy + Qyz, Qxz - Qzx,
+// 			Qzx + Qxz, Qzy + Qyz, Qzz - Qxx - Qyy, Qyx - Qxy,
+// 			Qzy - Qyz, Qxz - Qzx, Qyx - Qxy, Qxx + Qyy + Qzz;
+
+// 		Eigen::EigenSolver<Eigen::Matrix4f> es(em / 3);
+
+// 		auto ev = es.eigenvalues();
+
+// 		std::size_t bestJ = 0;
+// 		float bestEV = -1;
+
+// 		for (std::size_t j = 0; j < 4; ++j)
+// 		{
+// 			if (bestEV < ev[j].real())
+// 			{
+// 				bestEV = ev[j].real();
+// 				bestJ = j;
+// 			}
+// 		}
+
+// 		if (std::abs(bestEV - 1) > 0.01)
+// 			continue; // not a rotation matrix
+
+// 		auto col = es.eigenvectors().col(bestJ);
+
+// 		auto q = normalize(cif::quaternion{
+// 			static_cast<float>(col(3).real()),
+// 			static_cast<float>(col(0).real()),
+// 			static_cast<float>(col(1).real()),
+// 			static_cast<float>(col(2).real()) });
+
+// 		cif::point p1{ 1, 1, 1 };
+// 		cif::point p2 = p1;
+// 		p2.rotate(q);
+
+// 		cif::point p3 = rot * p1;
+
+// 		REQUIRE_THAT(p2.m_x, Catch::Matchers::WithinRel(p3.m_x, 0.01f));
+// 		REQUIRE_THAT(p2.m_y, Catch::Matchers::WithinRel(p3.m_y, 0.01f));
+// 		REQUIRE_THAT(p2.m_z, Catch::Matchers::WithinRel(p3.m_z, 0.01f));
+// 	}
+// }
+
+TEST_CASE("m2q_0a")
 {
 	for (std::size_t i = 0; i < cif::kSymopNrTableSize; ++i)
 	{
 		auto d = cif::kSymopNrTable[i].symop().data();
 
-		cif::matrix3x3<float> rot;
-		float Qxx = rot(0, 0) = d[0];
-		float Qxy = rot(0, 1) = d[1];
-		float Qxz = rot(0, 2) = d[2];
-		float Qyx = rot(1, 0) = d[3];
-		float Qyy = rot(1, 1) = d[4];
-		float Qyz = rot(1, 2) = d[5];
-		float Qzx = rot(2, 0) = d[6];
-		float Qzy = rot(2, 1) = d[7];
-		float Qzz = rot(2, 2) = d[8];
+		Eigen::Matrix3f rot;
+		rot << d[0], d[1], d[2], d[3], d[4], d[5], d[6], d[7], d[8];
 
-		Eigen::Matrix4f em;
+		// check to see if this matrix contains a true rotation
+		if (rot * rot.transpose() != Eigen::Matrix3f::Identity() or rot.determinant() != 1)
+			continue;
 
-		em << Qxx - Qyy - Qzz, Qyx + Qxy, Qzx + Qxz, Qzy - Qyz,
-		      Qyx + Qxy, Qyy - Qxx - Qzz, Qzy + Qyz, Qxz - Qzx,
-			  Qzx + Qxz, Qzy + Qyz, Qzz - Qxx - Qyy, Qyx - Qxy,
-			  Qzy - Qyz, Qxz - Qzx, Qyx - Qxy, Qxx + Qyy + Qzz;
+		Eigen::Quaternionf qe(rot);
 
-		Eigen::EigenSolver<Eigen::Matrix4f> es(em / 3);
+		auto q = normalize(cif::quaternion{ qe.w(), qe.x(), qe.y(), qe.z() });
 
-		auto ev = es.eigenvalues();
-
-		std::size_t bestJ = 0;
-		float bestEV = -1;
-
-		for (std::size_t j = 0; j < 4; ++j)
-		{
-			if (bestEV < ev[j].real())
-			{
-				bestEV = ev[j].real();
-				bestJ = j;
-			}
-		}
-
-		if (std::abs(bestEV - 1) > 0.01)
-			continue; // not a rotation matrix
-
-		auto col = es.eigenvectors().col(bestJ);
-
-		auto q = normalize(cif::quaternion{
-			static_cast<float>(col(3).real()),
-			static_cast<float>(col(0).real()),
-			static_cast<float>(col(1).real()),
-			static_cast<float>(col(2).real()) });
-		
 		cif::point p1{ 1, 1, 1 };
 		cif::point p2 = p1;
 		p2.rotate(q);
 
-		cif::point p3 = rot * p1;
+		cif::matrix3x3<float> rot_c({
+			rot_c(0, 0) = d[0],
+			rot_c(0, 1) = d[1],
+			rot_c(0, 2) = d[2],
+			rot_c(1, 0) = d[3],
+			rot_c(1, 1) = d[4],
+			rot_c(1, 2) = d[5],
+			rot_c(2, 0) = d[6],
+			rot_c(2, 1) = d[7],
+			rot_c(2, 2) = d[8]
+		});
+
+		cif::point p3 = rot_c * p1;
 
 		REQUIRE_THAT(p2.m_x, Catch::Matchers::WithinRel(p3.m_x, 0.01f));
 		REQUIRE_THAT(p2.m_y, Catch::Matchers::WithinRel(p3.m_y, 0.01f));
@@ -291,7 +330,7 @@ TEST_CASE("m2q_0, *utf::tolerance(0.001f)")
 	}
 }
 
-// "TEST_CASE(m2q_1, *utf::tolerance(0.001f)")
+// "TEST_CASE(m2q_1")
 // {
 // 	for (std::size_t i = 0; i < cif::kSymopNrTableSize; ++i)
 // 	{
@@ -337,7 +376,7 @@ TEST_CASE("m2q_0, *utf::tolerance(0.001f)")
 // 			static_cast<float>(em(bestJ, 0)),
 // 			static_cast<float>(em(bestJ, 1)),
 // 			static_cast<float>(em(bestJ, 2)) });
-		
+
 // 		cif::point p1{ 1, 1, 1 };
 // 		cif::point p2 = p1;
 // 		p2.rotate(q);
@@ -390,17 +429,17 @@ TEST_CASE("symm_3")
 	REQUIRE(sg.get_name() == "P 21 21 2");
 }
 
-TEST_CASE("symm_4, *utf::tolerance(0.1f)")
+TEST_CASE("symm_4")
 {
 	using namespace cif::literals;
 
 	// based on 2b8h
 	auto sg = cif::spacegroup(154); // p 32 2 1
 	auto c = cif::cell(107.516, 107.516, 338.487, 90.00, 90.00, 120.00);
-	
-	cif::point a{   -8.688,  79.351, 10.439 }; // O6 NAG A 500
-	cif::point b{  -35.356,  33.693, -3.236 }; // CG2 THR D 400
-	cif::point sb(  -6.916,   79.34,   3.236); // 4_565 copy of b
+
+	cif::point a{ -8.688, 79.351, 10.439 };  // O6 NAG A 500
+	cif::point b{ -35.356, 33.693, -3.236 }; // CG2 THR D 400
+	cif::point sb(-6.916, 79.34, 3.236);     // 4_565 copy of b
 
 	REQUIRE_THAT(distance(a, sg(a, c, "1_455"_symop)), Catch::Matchers::WithinRel(static_cast<float>(c.get_a()), 0.01f));
 	REQUIRE_THAT(distance(a, sg(a, c, "1_545"_symop)), Catch::Matchers::WithinRel(static_cast<float>(c.get_b()), 0.01f));
@@ -411,12 +450,12 @@ TEST_CASE("symm_4, *utf::tolerance(0.1f)")
 	REQUIRE_THAT(sb.m_y, Catch::Matchers::WithinRel(sb2.m_y, 0.01f));
 	REQUIRE_THAT(sb.m_z, Catch::Matchers::WithinRel(sb2.m_z, 0.01f));
 
-	REQUIRE_THAT(distance(a, sb2), Catch::Matchers::WithinRel(7.42f, 0.01f));	
+	REQUIRE_THAT(distance(a, sb2), Catch::Matchers::WithinRel(7.42f, 0.01f));
 }
 
 // --------------------------------------------------------------------
 
-TEST_CASE("symm_4wvp_1, *utf::tolerance(0.1f)")
+TEST_CASE("symm_4wvp_1")
 {
 	using namespace cif::literals;
 
@@ -427,7 +466,7 @@ TEST_CASE("symm_4wvp_1, *utf::tolerance(0.1f)")
 
 	cif::crystal c(db);
 
-	cif::point p{ -78.722, 98.528,  11.994 };
+	cif::point p{ -78.722, 98.528, 11.994 };
 	auto a = s.get_residue("A", 10, "").get_atom_by_atom_id("O");
 
 	auto sp1 = c.symmetry_copy(a.get_location(), "2_565"_symop);
@@ -442,10 +481,9 @@ TEST_CASE("symm_4wvp_1, *utf::tolerance(0.1f)")
 	REQUIRE_THAT(sp2.m_x, Catch::Matchers::WithinAbs(p.m_x, 0.5f));
 	REQUIRE_THAT(sp2.m_y, Catch::Matchers::WithinAbs(p.m_y, 0.5f));
 	REQUIRE_THAT(sp2.m_z, Catch::Matchers::WithinAbs(p.m_z, 0.5f));
-
 }
 
-TEST_CASE("symm_2bi3_1, *utf::tolerance(0.1f)")
+TEST_CASE("symm_2bi3_1")
 {
 	cif::file f(gTestDir / "2bi3.cif.gz");
 
@@ -455,18 +493,15 @@ TEST_CASE("symm_2bi3_1, *utf::tolerance(0.1f)")
 	cif::crystal c(db);
 
 	auto struct_conn = db["struct_conn"];
-	for (const auto &[
-			asym1, seqid1, authseqid1, atomid1, symm1,
-			asym2, seqid2, authseqid2, atomid2, symm2,
-			dist] : struct_conn.find<
-				std::string,int,std::string,std::string,std::string,
-				std::string,int,std::string,std::string,std::string,
-				float>(
-			cif::key("ptnr1_symmetry") != "1_555" or cif::key("ptnr2_symmetry") != "1_555",
-			"ptnr1_label_asym_id", "ptnr1_label_seq_id", "ptnr1_auth_seq_id", "ptnr1_label_atom_id", "ptnr1_symmetry", 
-			"ptnr2_label_asym_id", "ptnr2_label_seq_id", "ptnr2_auth_seq_id", "ptnr2_label_atom_id", "ptnr2_symmetry", 
-			"pdbx_dist_value"
-		))
+	for (const auto &[asym1, seqid1, authseqid1, atomid1, symm1,
+			 asym2, seqid2, authseqid2, atomid2, symm2,
+			 dist] : struct_conn.find<std::string, int, std::string, std::string, std::string,
+			 std::string, int, std::string, std::string, std::string,
+			 float>(
+			 cif::key("ptnr1_symmetry") != "1_555" or cif::key("ptnr2_symmetry") != "1_555",
+			 "ptnr1_label_asym_id", "ptnr1_label_seq_id", "ptnr1_auth_seq_id", "ptnr1_label_atom_id", "ptnr1_symmetry",
+			 "ptnr2_label_asym_id", "ptnr2_label_seq_id", "ptnr2_auth_seq_id", "ptnr2_label_atom_id", "ptnr2_symmetry",
+			 "pdbx_dist_value"))
 	{
 		auto &r1 = s.get_residue(asym1, seqid1, authseqid1);
 		auto &r2 = s.get_residue(asym2, seqid2, authseqid2);
@@ -492,7 +527,7 @@ TEST_CASE("symm_2bi3_1, *utf::tolerance(0.1f)")
 	}
 }
 
-TEST_CASE("symm_2bi3_1a, *utf::tolerance(0.1f)")
+TEST_CASE("symm_2bi3_1a")
 {
 	using namespace cif::literals;
 
@@ -504,23 +539,20 @@ TEST_CASE("symm_2bi3_1a, *utf::tolerance(0.1f)")
 	auto struct_conn = db["struct_conn"];
 	auto atom_site = db["atom_site"];
 
-	for (const auto &[
-			asym1, seqid1, authseqid1, atomid1, symm1,
-			asym2, seqid2, authseqid2, atomid2, symm2,
-			dist] : struct_conn.find<
-				std::string,std::optional<int>,std::string,std::string,std::string,
-				std::string,std::optional<int>,std::string,std::string,std::string,
-				float>(
-			cif::key("ptnr1_symmetry") != "1_555" or cif::key("ptnr2_symmetry") != "1_555",
-			"ptnr1_label_asym_id", "ptnr1_label_seq_id", "ptnr1_auth_seq_id", "ptnr1_label_atom_id", "ptnr1_symmetry", 
-			"ptnr2_label_asym_id", "ptnr2_label_seq_id", "ptnr2_auth_seq_id", "ptnr2_label_atom_id", "ptnr2_symmetry", 
-			"pdbx_dist_value"
-		))
+	for (const auto &[asym1, seqid1, authseqid1, atomid1, symm1,
+			 asym2, seqid2, authseqid2, atomid2, symm2,
+			 dist] : struct_conn.find<std::string, std::optional<int>, std::string, std::string, std::string,
+			 std::string, std::optional<int>, std::string, std::string, std::string,
+			 float>(
+			 cif::key("ptnr1_symmetry") != "1_555" or cif::key("ptnr2_symmetry") != "1_555",
+			 "ptnr1_label_asym_id", "ptnr1_label_seq_id", "ptnr1_auth_seq_id", "ptnr1_label_atom_id", "ptnr1_symmetry",
+			 "ptnr2_label_asym_id", "ptnr2_label_seq_id", "ptnr2_auth_seq_id", "ptnr2_label_atom_id", "ptnr2_symmetry",
+			 "pdbx_dist_value"))
 	{
-		cif::point p1 = atom_site.find1<float,float,float>(
+		cif::point p1 = atom_site.find1<float, float, float>(
 			"label_asym_id"_key == asym1 and "label_seq_id"_key == seqid1 and "auth_seq_id"_key == authseqid1 and "label_atom_id"_key == atomid1,
 			"cartn_x", "cartn_y", "cartn_z");
-		cif::point p2 = atom_site.find1<float,float,float>(
+		cif::point p2 = atom_site.find1<float, float, float>(
 			"label_asym_id"_key == asym2 and "label_seq_id"_key == seqid2 and "auth_seq_id"_key == authseqid2 and "label_atom_id"_key == atomid2,
 			"cartn_x", "cartn_y", "cartn_z");
 
@@ -540,7 +572,7 @@ TEST_CASE("symm_2bi3_1a, *utf::tolerance(0.1f)")
 	}
 }
 
-TEST_CASE("symm_3bwh_1, *utf::tolerance(0.1f)")
+TEST_CASE("symm_3bwh_1")
 {
 	cif::file f(gTestDir / "3bwh.cif.gz");
 
@@ -555,15 +587,15 @@ TEST_CASE("symm_3bwh_1, *utf::tolerance(0.1f)")
 		{
 			if (a1 == a2)
 				continue;
-			
-			const auto&[ d, p, so ] = c.closest_symmetry_copy(a1.get_location(), a2.get_location());
+
+			const auto &[d, p, so] = c.closest_symmetry_copy(a1.get_location(), a2.get_location());
 
 			REQUIRE_THAT(d, Catch::Matchers::WithinAbs(distance(a1.get_location(), p), 0.5f));
 		}
 	}
 }
 
-TEST_CASE("volume_3bwh_1, *utf::tolerance(0.1f)")
+TEST_CASE("volume_3bwh_1")
 {
 	cif::file f(gTestDir / "1juh.cif.gz");
 
@@ -573,4 +605,3 @@ TEST_CASE("volume_3bwh_1, *utf::tolerance(0.1f)")
 
 	REQUIRE_THAT(c.get_cell().get_volume(), Catch::Matchers::WithinRel(741009.625f, 0.01f));
 }
-

test/unit-v2-test.cpp

@@ -50,6 +50,47 @@ cif::file operator""_cf(const char *text, std::size_t length)
 
 // --------------------------------------------------------------------
 
+TEST_CASE("text_1")
+{
+	CHECK(cif::iequals("TEST", "test"));
+	CHECK(cif::iequals(std::string_view{"TEST"}, std::string_view{"test"}));
+
+	CHECK(cif::icompare("TEST", "test") == 0);
+	CHECK(cif::icompare(std::string_view{"TEST"}, std::string_view{"test"}) == 0);
+
+	CHECK(cif::icompare("TEST1", "test") > 0);
+	CHECK(cif::icompare(std::string_view{"TEST1"}, std::string_view{"test"}) > 0);
+
+	CHECK(cif::icompare("aap", "noot") < 0);
+	CHECK(cif::icompare(std::string_view{"aap"}, std::string_view{"noot"}) < 0);
+}
+
+// --------------------------------------------------------------------
+
+TEST_CASE("from_chars_1")
+{
+auto f = R"(data_TEST
+#
+loop_
+_test.v
+616.487
+616.487000
+    )"_cf;
+
+	auto &db = f.front();
+	auto &c = db.front();
+
+	auto r1 = c.front();
+	REQUIRE(r1.get<double>("v") == 616.487);
+	REQUIRE(r1["v"].compare(616.487) == 0);
+
+	auto r2 = c.back();
+	REQUIRE(r2.get<double>("v") == 616.487);
+	REQUIRE(r2["v"].compare(616.487) == 0);
+}
+
+// --------------------------------------------------------------------
+
 TEST_CASE("id_1")
 {
 	REQUIRE(cif::cif_id_for_number(0) == "A");
@@ -733,7 +774,6 @@ save__cat_2.desc
 	auto validator = cif::parse_dictionary("test", is_dict);
 
 	cif::file f;
-	f.set_validator(&validator);
 
 	// --------------------------------------------------------------------
 
@@ -764,7 +804,7 @@ _cat_2.desc
 	} data_buffer(const_cast<char *>(data), sizeof(data) - 1);
 
 	std::istream is_data(&data_buffer);
-	f.load(is_data);
+	f.load(is_data, validator);
 
 	SECTION("one")
 	{
@@ -886,7 +926,6 @@ save__cat_1.c
 	auto validator = cif::parse_dictionary("test", is_dict);
 
 	cif::file f;
-	f.set_validator(&validator);
 
 	// --------------------------------------------------------------------
 
@@ -909,7 +948,7 @@ mies Mies
 	} data_buffer(const_cast<char *>(data), sizeof(data) - 1);
 
 	std::istream is_data(&data_buffer);
-	f.load(is_data);
+	f.load(is_data, validator);
 
 	auto &cat1 = f.front()["cat_1"];
 
@@ -1048,7 +1087,6 @@ save__cat_2.desc
 	auto validator = cif::parse_dictionary("test", is_dict);
 
 	cif::file f;
-	f.set_validator(&validator);
 
 	// --------------------------------------------------------------------
 
@@ -1082,7 +1120,7 @@ _cat_2.desc
 	} data_buffer(const_cast<char *>(data), sizeof(data) - 1);
 
 	std::istream is_data(&data_buffer);
-	f.load(is_data);
+	f.load(is_data, validator);
 
 	auto &cat1 = f.front()["cat_1"];
 	auto &cat2 = f.front()["cat_2"];
@@ -1251,7 +1289,6 @@ save__cat_2.parent_id3
 	auto validator = cif::parse_dictionary("test", is_dict);
 
 	cif::file f;
-	f.set_validator(&validator);
 
 	// --------------------------------------------------------------------
 
@@ -1295,7 +1332,7 @@ _cat_2.parent_id3
 	} data_buffer(const_cast<char *>(data), sizeof(data) - 1);
 
 	std::istream is_data(&data_buffer);
-	f.load(is_data);
+	f.load(is_data, validator);
 
 	auto &cat1 = f.front()["cat_1"];
 	auto &cat2 = f.front()["cat_2"];
@@ -1472,7 +1509,6 @@ cat_2 3 cat_2:cat_1:3
 	auto validator = cif::parse_dictionary("test", is_dict);
 
 	cif::file f;
-	f.set_validator(&validator);
 
 	// --------------------------------------------------------------------
 
@@ -1509,7 +1545,7 @@ _cat_2.parent_id3
 	} data_buffer(const_cast<char *>(data), sizeof(data) - 1);
 
 	std::istream is_data(&data_buffer);
-	f.load(is_data);
+	f.load(is_data, validator);
 
 	auto &cat1 = f.front()["cat_1"];
 	auto &cat2 = f.front()["cat_2"];
@@ -1712,7 +1748,6 @@ cat_2 1 cat_2:cat_1:1
 	auto validator = cif::parse_dictionary("test", is_dict);
 
 	cif::file f;
-	f.set_validator(&validator);
 
 	// --------------------------------------------------------------------
 
@@ -1749,7 +1784,7 @@ _cat_2.parent_id_2
 	} data_buffer(const_cast<char *>(data), sizeof(data) - 1);
 
 	std::istream is_data(&data_buffer);
-	f.load(is_data);
+	f.load(is_data, validator);
 
 	// auto &cat1 = f.front()["cat_1"];
 	auto &cat2 = f.front()["cat_2"];
@@ -2108,7 +2143,6 @@ cat_2 1 '_cat_2.num'  '_cat_3.num'  cat_3
 	auto validator = cif::parse_dictionary("test", is_dict);
 
 	cif::file f;
-	f.set_validator(&validator);
 
 	// --------------------------------------------------------------------
 
@@ -2150,7 +2184,7 @@ _cat_3.num
 	} data_buffer(const_cast<char *>(data), sizeof(data) - 1);
 
 	std::istream is_data(&data_buffer);
-	f.load(is_data);
+	f.load(is_data, validator);
 
 	auto &cat1 = f.front()["cat_1"];
 	auto &cat2 = f.front()["cat_2"];
@@ -2393,7 +2427,6 @@ cat_2 1 '_cat_2.num'  '_cat_3.num'  cat_3
 	auto validator = cif::parse_dictionary("test", is_dict);
 
 	cif::file f;
-	f.set_validator(&validator);
 
 	// --------------------------------------------------------------------
 
@@ -2435,7 +2468,7 @@ _cat_3.num
 	} data_buffer(const_cast<char *>(data), sizeof(data) - 1);
 
 	std::istream is_data(&data_buffer);
-	f.load(is_data);
+	f.load(is_data, validator);
 
 	auto &cat1 = f.front()["cat_1"];
 	auto &cat2 = f.front()["cat_2"];
@@ -2990,7 +3023,6 @@ save__cat_1.name
 	auto validator = cif::parse_dictionary("test", is_dict);
 
 	cif::file f;
-	f.set_validator(&validator);
 
 	// --------------------------------------------------------------------
 
@@ -3013,7 +3045,7 @@ _cat_1.name
 	} data_buffer(const_cast<char *>(data), sizeof(data) - 1);
 
 	std::istream is_data(&data_buffer);
-	f.load(is_data);
+	f.load(is_data, validator);
 
 	REQUIRE(f.is_valid());
 
@@ -3185,7 +3217,6 @@ save__cat_1.name
 	auto &validator = cif::validator_factory::instance().construct_validator("test_dict.dic", is_dict);
 
 	cif::file f;
-	f.set_validator(&validator);
 
 	// --------------------------------------------------------------------
 
@@ -3212,19 +3243,21 @@ _cat_1.name
 	} data_buffer(const_cast<char *>(data), sizeof(data) - 1);
 
 	std::istream is_data(&data_buffer);
-	f.load(is_data);
+	f.load(is_data, validator);
 
-	REQUIRE(f.is_valid());
+	CHECK(f.is_valid());
 
 	std::stringstream ss;
 	ss << f;
 
 	cif::file f2(ss);
-	REQUIRE(f2.is_valid());
+	REQUIRE(f2.empty() == false);
+	f2.front().load_dictionary();
+	CHECK(f2.is_valid());
 
 	auto &audit_conform = f2.front()["audit_conform"];
-	REQUIRE(audit_conform.front()["dict_name"].as<std::string>() == "test_dict.dic");
-	REQUIRE(audit_conform.front()["dict_version"].as<float>() == 1.0);
+	CHECK(audit_conform.front()["dict_name"].as<std::string>() == "test_dict.dic");
+	CHECK(audit_conform.front()["dict_version"].as<float>() == 1.0);
 }
 
 // --------------------------------------------------------------------
@@ -3292,7 +3325,6 @@ save__cat_1.id_2
 	auto validator = cif::parse_dictionary("test", is_dict);
 
 	cif::file f;
-	f.set_validator(&validator);
 
 	// --------------------------------------------------------------------
 
@@ -3317,7 +3349,7 @@ _cat_1.id_2
 	} data_buffer(const_cast<char *>(data), sizeof(data) - 1);
 
 	std::istream is_data(&data_buffer);
-	f.load(is_data);
+	f.load(is_data, validator);
 
 	auto &cat1 = f.front()["cat_1"];
 
@@ -3456,7 +3488,7 @@ TEST_CASE("compound_test_1")
 	cif::compound_factory::instance().push_dictionary(gTestDir / "REA_v2.cif");
 	auto compound = cif::compound_factory::instance().create("REA_v2");
 	REQUIRE(compound != nullptr);
-	REQUIRE(compound->id() == "REA_v2");
+	REQUIRE(cif::iequals(compound->id(), "REA_v2"));
 }
 
 // --------------------------------------------------------------------